OAK RIDGE NATIONAL LABORATORY Gptrat.EB by ONION Carbidt CORPORATION • -F0r the U S ATOMIC ENERGY COMMISSION

QRNL-TM-4370

)

CHEMICAL DEVELOPMENT SECTION B SEMIANNUAL PROGRESS REPORT MARCH 1, 1973 TO AUGUST 31, 1973

: S V1 'i! ' I

«« Jb fS

' A .-V

W-t-'

. t " • •/

OAK RIDGE NATIONAL LABORATORY GPtRAT.EB BY ONION CARBIDt CORPORATION • -f0R THE U S ATOMIC ENERGY COMMISSION

\ H - ORNL-TM-U370

Contract No. W-7^05-eng-26

CHEMICAL DEVELOPMENT SECTION B SEMIANNUAL PROGRESS REPORT

March 1, 1973 to August 31, 1973

PART I

Compiled "by

c. D. Scott c. A. Burtis D. J. Crouse L. M. Ferris ——— NOTICE J. C. Mailen This report was prepared as an account of work Pitt sponsored by the United States Government. Neither W. W. the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights.

FEBRUARY 1974

NOTICE This document contains information of a preliminary nature and was prepared primarily for internal use at the Oak Ridge National Laboratory, It is subject to revision or correction and therefore does not represent a final report.

OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37^30 operated "by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION 'M & ^WTI t's iii

TABLE 0? CONTENTS Page

PUBLICATIONS ix

ORAL PRESENTATIONS xiii

SUMMARY xix

1. INTRODUCTION 1

2. FAST ANALYZER DEVELOPMENT Z 2.1 Miniature East Analyzer ?. 2.1.1 Instrumentation Development k 2.1.2 System Performance and Evaluation lo 2.1.3 On-Line Computer 3U 2.1.U Development of Chemical Assays 37 2.1.5 Dynamic Loading of Liquids 37 2.1.6 Fluorescence Monitor 5k 2.1.7 Light Scattering Monitor 61 2.1.8 Portable Data Printer 70 2.2 Portable Fast Analyzer 71 2.3 Fluorescence Fast Analyzer 72 2.3.1 Instrumentation 75 2.3.2 Fluorescence Sensitivity and Fluorescence Tracers 77 2.3.3 Fluorescence Polarization Measurements 75 2.3.U Fluorescence Referencing 62 2.3.5 Calculation of Enzyme Activity and Substrate Concentration 35 2.3.6 Effectiveness of Dynamic Referencing 5?" 2.3.7 Development of Enzyme Analyses yl 2.3.8 Immunological Assays . 10U 2.3.9 Fluorometric Determination of Uranium 112 2.3.10 Future Development 112 2.k Chemical Assay Methods llU 2.H.1 Triglycerides for Large 15-Place GeMSAEC llU 2.k.2 Serum Uric Acid with the Miniature Fast Analyzer 121 2.1+.3 Additional Serum Chemistry Assays for Miniature Fast Analyzer 127 2.5 Evaluation and Operation of Fast Analyzers in a Clinical Laboratory . 128 2.5.1 15-Place GeMSAEC Fast Analyzer 128 2.5.2 Miniature Fast Analyzers 129 2.5.3 Population Studies 130 2.6 References for Section 2 138 BLANK PAGE iv

Page

3. GENETIC MONITORING 1^2 3.1 Sample Preparation lU2 3.1.1 First Prototype Rotor IU3 3.1.2 Advanced Rotor Design lU5 3.2 Automated Elution Electrophoresis lU8 3.3 Isoenzyme Stripping by Use of the Fast Analyzer 150 3.U References for Section 3 157

U. HIGH-RESOLUTION ANALYTICAL SYSTEMS 158 U.l Prototype Systems 158 U. 2 Computer Systems l6l U.2.1 System Description l6l U.2.2 Computer Programs 162 U. 3 Systems Development - 170 U.3,1 Cation Exchange Separation of Ninhydrin-Positive Constituents 171 4.3.2 Liquid Chromatographic Analysis of Blood Serum . 177 U.3.3 Application of Fluorescamine to Monitoring of Urinary Polyamines l8o U. 3 • Analysis of Nucleosides and N-Bases 19U .3.5 Separations Systems ;.... * 196 U.U Identification of Body Fluid Constituents 203 U.U.I New Identifications 20U U.U. 2 Identification of a-Methoxyhomovanillic Acid 2 QU U.U. 3 Urinary Nucleoside Excretion by Patients vith Cancer 206 U.5 Bxoerimental ?,ecults and Applications 217 U.5.1 Effects of L-BOPA and a DOPA Decarboxylase Inhibitor 217 U.5.2 LC Profiles of Aromatic Acids for Children with Neurological Disorders 228 U.5.3 Body Fluid Constituents of Patients with Rare Pathologic Conditions 232 U.5.U Urinary Metabolites of U-Hydroxyacetanilide .... 235 U.6 References for Section U 239

5. HACKOI>'OLECuLAH SEPARATIONS Zkk 5.1 Purification of Erythropoietin 2UU 5.1.1 Preparation of Affinity Columns 2U5 5.1.2 Purification by Indirect Affinity Chromatography 2U7 cccxii

Page

5.2 Purification of Colony Stimulating Factor 250 5.3 Isolation of Purified Transfer Ribonucleic Acids 252 5.3.1 Selection of Factor Tu 255 5.3.2 Isolation of Factor Tu 256 5.3.3 Complex I and II Formation 259 5.U Distribution of Products 265 5.5 Separation and Comparison of Sequence of Three Formylmethionine tRNAs from E. coli 268 5.5.1 Separation of tRNA|Met and tRNA™e^ 269 5.5.2 Comparison of Ribonuclease Tx and Pancreatic Ribonuclease Digests 269 5.5.3 Relationships Among Three tRNAsfMet 270 5.6 Sequence Studies on Phenylalanine tRNA from Calf Liver 270 5.6.1 Base Composition of Calf Liver tRKAPhe 27k 5.6.2 pancreatic Ribonuclease Digestion of tRNAph© ... 276 5.7 References for Section 5 278 6. HYDROGEN PRODUCTION 28l 6.1 Enzymatic Hydrogen Production 28l 6.2 Thermal Hydrogen Production 285 6.3 References for Section 0 286

7. WATER POLLUTION STUDIES 288 7.1 Automated Analysis of Dissolved Organic Compounds in Polluted Waters 288 7.1.1 Effluents from Sewage Treatment Plants 290 7.1.2 Natural Waters 303 7.2 Environmental Effects of Antifoulants 311 7.3 References for Section 7 * 3i2

8. RADIOACTIVE WASTE DISPOSAL 315 8.1 Radiation Effects in Salt Mine Waste Repositories 31p 8.1.1 Radiation and Stored Energy of Salt Specimens .. 323 8.1.2 Results of Thermal Annealing Studies 325 8.1.3 Correlation of Energy Storage Data 328 8.1.U Evolution cf Hg upon Salt Dissolution 336 8.1.5 Loss of Water from Lyons Salt During Radiation Exposure 339 8.1.6 Gases in Unirradiated and Irradiated Lyons Salt 3^1 8.1.7 Future Work 3^ vi

Page

8t2 Waste Disposal "by Hydraulic Fracturing 8.2.1 Waste Injections 3^5 8.2.2 Environmental Impact Statement 3^-8 8.2.3 Sludge Characteristics 3^8 8.2.1* Leaching Studies 3^9 8.3 Wet Oxidation of Plutonium 355 8.U References for Section 8 356

9. ENVIRONMENTAL SURVEYS 359 9.1 The Nuclear Fuel Cycle - Milling of Uranium Ores 359 9.2 Nuclear Industry Wastes 3^1 9.2.1 Use of Evaporation for the Treatment of Liquids in the Nuclear Industry 361 9.2.2 Solid Waste Practices at Nuclear Power Plants 3^3 9.3 Mercury Studies 3^ 9.3.1 Survey of Mercury Reprocessors 3^5 9.3.2 Usage of Mercury by Agencies of the U.S. Government 3^5 9.U Biological Response to Proprietary Chemicals Used in Cooling Water <... * 3^6 9.5 References for Section 9 3^7 10. SEPARATIONS PROCESSES 368 10.1 Recovery of Uranium from Wet-Process Phosphoric Acid . 368 10.1.1 Process Flowsheet 3^9 10.1.2 Solvent Composition 370 10.1.3 Effects of Process Variables on Extraction ... 370 lO.l.U stripping of Uranium 371 10.1.5 Process Demonstration 372 10.2 Separation of Radium from Uranium Ore Tailings 378 10.3 Separation of Alpha Emitters from Reprocessing Wastes 381 10.k High-Pressure Ion Exchange Studies 386 10.5 References for Section 10 391 11. CHEMICAL APPLICATIONS OF NUCLEAR EXPLOSIVES 39^ 11.1 Contamination by Tritium ...... 392 11.2 Contamination by Fission Products 396 11.3 References for Section 11 397 vii

Page

12. CHEMISTRY OF RADIOIODINE .. 398 12.1 Mercuric Nitrate Scrubbing Process 398 12.1.1 Methyl Iodide Removal in Packed and Bubble-Cap Columns 399 12c1.2 Decomposition Rates of Various Organic Iodides ij-02 12.1.3 Process Demonstrations 12.2 Formation of Organic Iodides in Process Systems 1+10 12.3 References for Section 12 . Hll

13. CTR SUPPORTIVE RESEARCH ^12 13.1 Equilibria in the Hydrogen Isotope—Lithium Systems .. ^12 13.1.1 Total Pressure Method * U13 13.1.2 Partial Pressure Method h22 13.2 Phase Equilibria in the Li-K System I+23 13.3 Permeation of Deuterium Through Structural Metals at Low Pressures ^27 13.^ Permeation of Deuterium Through Metals Coated with Oxide Films k35 13.5 Chemisorption of Tritium on C-raphite ^3$ 13.6 EffectAnalogs osf Stronof Molteg Magnetin Saltcs Fields on Aqueous bk2 13.7 Thermodynamics and Mass Spectrometry of Vanadium Fluorides W? 13.8 The Solubilities of Hydrogen, Deuterium, and Helium in Molten Li2BeF4 kb-9 13.9 References for Section 13 ^+50

lU. THERMODYNAMICS OF MOLTEN-SALT SYSTEMS

lU.l Chemistry of Tellurium in Molten Li2BeF4 lk.2 Electrostatic Energies and Heats of Formation of 3d Metallic Difluorides b57 lU.3 Lattice Energies of Cubic Alkaline-Earth Oxides. Affinity of Oxygen for Two Electrons ^60 1I4.4 Relationship Bet-ween Sonic Velocity and Entropy in Molten UCfe k66 lU.5 References for Section lU U69

15. COAL CONVERSION STUDIES U72 15.1 References for Section 15 ^ viii

Page

APPENDIX I k75 APPENDIX II 1+93 PUBLICATIONS

Author(s) Title Publication

Arnold, W. D., R. Preliminary Evaluation of Methods 0RNL-TM-U02U. Salmon, K. H. for the Disposal of Tritiated Lin, and W. Water from Nucleaily Stimulated deLaguna Natural Gas Wells

Burtis, C. A., W. F. Development of an Analytical System Clin. Chem. Iff, 895 (1973). Johnson, J. C. Mailen, Based Around a Miniature Fast J. B. Overton, T. 0. Analyzer Tiffany, and M. B. Watsky

Burtis, C. A., et al. Gravity Zero (C0) Analytical Clini- 0RNL-TM-U027. cal Laboratory System Progress Report for the Period August "1 to October 31, 1972

Burtis, C. A. Gravity Zero (G0) Analytical Clini- 0RNL-TM-U225. cal Laboratory System Annual Progress Report for the Period April 15, 1972, to May 1, 1973

Cantor, S. Solubility of BF3 in Salts of J. Nucl, Mater, ffi, 177 (1973). Molten-Salt Reactor Interest

Chilcote, D* D., and Mathematical Analysis of Normaliza- Anal. Chem. k5, 721 (1973). C. D. Scott tion Techniques Used in Chromatog- raphy (Publications - continued)

Author(s) Title Publication

Chilcote, D. D., and Rapid Chromatographic Analysis Anal. Letters 6_, 531 (1973). J. E. Mrochek of Nucleosides and N-Bases in Physiologic Fluids

Chilcote, D. D. Computer Technique for Identifica- Clin. Chem. Ig, 826 (1973). tion of Chromatographic Peaks

Katz, S., W. W. Sensitive Fluorescence Monitoring Clin. Chem. 8, 817-20 (1973). Pitt, Jr., and of Aromatic Acids after Anion- G. Jones, Jr. Exchange Chromatography of Body Fluids

Lloyd, M. H., and A New Solvent Extraction Process Nucl. Technol. 18, 205 (1973)- 0. K. Tallent for Preparing PuO^ Sols

Mailen, J. C., J. B. Techniques for Fabrication and Anal. Letters 6, 2^5 (1973). Overton, C. A. Assembly of Rotors for CJse in Burtis, and C. D. a Miniature Fast Analyzer Scott

Mrochek, J. E., S. Monitoring Phenylalanine-Tyrosine Clin. Chem. lg, 927 (1973)• R. Dinsmore, and Metabolism by High-Resolution D. W. Ohrt Liquid Chromatography (Publications - continued)

Author(s) Title Publication

Neel, J. V., T. 0. Approaches to Monitoring Human Vol. 3, ed. by A. Hollaender, Tiffany, and N. G. Populations for Mutation Rates and pp. 105-50, Plenum Publishing Corp., Anderson Genetic Disease in Chemical Mutagens New York, N. Y., 1973-

Ross, R. G., C. E. The Oxide Chemistry of Protactinium J. Inorg. Nucl. Chem. 433 (1973) Bamberger, and in Molten Fluorides C. F. Baes, Jr.

Scott, C. D., and A Miniature Fast Analyzer System Anal. Chem. jg, 327-A (1973). C, A. Burtis

Scott, C. D. Health Care Delivery and Advanced Science 180, 1339 (1973). Technology

Scott, C. D., and Use of Sequential Columns of J. Chromatog. 83, 383 (1973). N. E. Lee Microreticular and Pellicular Ion- Exchange Resins in the High- Resolution Separation of Complex Biochemical Mixtures

Smith, F. J. The Solubilities of Thorium and J. Less-Common Metals 32, 297 (1973) Samarium in Liquid Lithium-Lead Solutions

Tiffany, T. 0., C. A. A Dynamic Multicuvette Fluorometer- Anal. Chem. kj?, 1716-23 (1973). Burtis, J. C. Mailen, Spectrophotometer Based on the and L. H. Thacker GeMSAEC Fast Analyzer Principle (Publications - continued)

Author(s) Title Publication

Tiffany, T. 0., M. B. Fluorometric Fast Analyzer: Some Clin. Chem. lg, 871-82 (1973)• Watsky, C. A. Burtis, Applications to Fluorescence and L. H. Thacker Measurements in Clinical Chemistry

Tiffany, T. 0., D. D. The Determination of Kinetic Clin. Chem. 19, 909-18 (1973). Chilcote, and C. A. Parameters by Use of a Small- Burtis Computer Interfaced Fast Analyzer An Addition to Automated Clinical Enzymology

Waalkes, T. P., S. R. Urinary Excretion by Cancer J. Nat. Cancer. Inst. £1, 271-7U, Dinsmore, and J. E. Patients of the Nucleosides July (1973). Mrochek -Dime thylguanos inc. 1-Methylinosine, and Pseudouridine ORAL PRESENTATIONS

Speaker and Coauthor(s) Title Place Presented

Burtis, C. A., W. F. Development of an Analytical System Fifth Annual Symposium on Advanced Johnson, J. C. Mailen, Based Around a Miniature Fast Analytical Concepts for the Clinical J. B. Overton, T. 0. Analyzer Laboratory, Oak Ridge, Tenn., Tiffany, and M. E. March 15-16, 1973. Watsky

Burtis, C. A. The Centrifugal Fast Analyzer Annual Convention of the Arkansas Society for Medical Technology and American Medical Technologists, Little Rock, Ark., April 12, 1973-

Burtis, C. A. Miniature Laboratory System for Eighth Annual Meeting of the Associa- Space Flight? and Health Care tion for the Advancement of Medical Delivery Instrumentation, Washington, D. C., March 23, 1973-

Burtis, C. A. Automated Clinical and Serological Annual Meeting of the Tennessee Analysis for Large-Scale Processing Public Health Association and Conference of Tennessee Public Health Workers, Gatlinburg, Tenn., March 28, 1973.

Burtis, C. A. Applications of a GeMSAEC Fast Department of Pharmacology, Univer- Analyzer sity of California at San Francisco, San Francisco, Calif., July 5, 1973- (Oral Presentations - continued)

Speaker and Coauthor(s) Title Place Presented

A Mathematical Analysis of Normali- Eighth International Symposium on Chilcote, D. D. zation Techniques Used in Chroma- Advances in Chromatography, Toronto, tography Canada, April 16-19, 1973-

An Identification Scheme for Fifth Annual Symposium on Advanced Chilcote, D. D. Chromatographically Separated Analytical Concepts for the Clinical Constituents that can be Used Laboratory, Oak Ridge, Tenn., with a Small Computer March 15-16, 1973-

The Indole Analyzer: A Simple Eighth Annual Meeting, Association Chilcote, D. D. Sensitive Liquid Chromatographic for the Advancement of Medical Analyzer for Naturally Occurring Instrumentation, Washington, D. C., Indoles in Body Fluids March 21-2U, 1973-

The Indole Analyzer: A Simple 165th ACS National Meeting, Dallas, Chilcote, D. D. Sensitive Liquid Chromatographic Tex., April 8-13, 1973- Analyzer for Naturally Occurring Indoles in Body Fluids

Computer Assisted Chromatographic National Meeting of AIChE: Sym- Chilcote, D. D. Analysis of Body Fluid Samples posium of Engineering Methods in Medical Diagnostic Techniques, Detroit, Mich., June 3-6, 1973- (Oral Presentations - continued)

Speaker and Title Place Presented Coauthor(s) • mil a<

Hurst, F. J. Recovery of Uranium from Wet- Annual ACS Meeting, Chicago, Process Phosphoric Acid by Ex- Illinois, August 27-31, 1973- traction with Octylphenyl Phosphoric Acid

Katz, S., W. W. Sensitive Fluorescence Monitoring Fifth Annual Symposium on Advanced Pitt, Jr., and of Aromatic Acids After Anion- Analytical Concepts for the Clinical G- Jones, Jr. Exchange Chromatography of Body Laboratory, Oak Ridge, Tenn., Fluids March 15-16, 1973-

Mailen, J. C. The Development of a Miniature 75th National Meeting of AIChE, Fast Analyzer Detroit, Mich., June 6, 1973-

Mrochek, J. E. Mon i tor ing Phenylala nine-Tyros ine Fifth Annual Symposium on Advanced Metabolism by High-Resolution Analytical Concepts for the Clinical Liquid Chromatography Laboratory, Oak Ridge, Tenn., March 15-16, 1973-

Mrochek, J. E. High-Resolution Liquid Chromatog- ACS Symposium, "Newer Analytical raphy - A Hew Analytical Tool for Techniques in Clinical Chemistry," the Clinical Chemist at Regional Meeting of ACS, Cleveland, Ohio, May Ik, 1973- (Oral Presentations - continued)

Speaker 8nd Coautnor(s) Title Place Presented

Liquid Chromatographic Analysis of ACS Symposium on Analytical Advance Urinary Nucleosides in Normal ana in Clinical and Medicinal Chemistry Mrochek, J. E. Malignant States at National ACS Meeting, Chicago, # 111., Aug. 30, 1973. Introduction to the Fifth Annual Oak Ridge National Laboratory, Oak Symposium on Advanced Analytical Ridge, Tenn., March 15, 1973- Scott, C. D. Concepts for the Clinical Labora- tory

High Resolution Liquid Chromatog- Iowa State University, Ame:;, Iowa, raphy as Applied to the Separation April 9, 1973. Scott, C. D. of Complex Biochemical Mixtures Scott, C. D. High Resolution Liquid Chromatog- 3-M Company, St. Paul, Minn., raphy as Applied to the Separation April 11, 1973. of Complex Biochemical Mixtures

Scott, C. D., and Use of Sequential Columns for Micro- International Symposium on Column N. E. Lee reticular and Pellicular Ion- Liquid Chromatography, Interlaken, Exchange Resins in the High- Switzerland, May if, 1973• Resolution Separation of Complex Biochemical Mixtures (Oral Presentations - continued)

Speaker and Coauthor(s) Title Place Presented

Scott, C. D. Recent Developments in Fast University of Birmingham, Birmingham, Analyzer Systems England, May 7, 1973-

Scott, C. D. High Resolution Liquid Chroma- University of Salford, Salford, tography England, May 8, 1973-

Scott, C. D. High Resolution Liquid Chromatog- Philadelphia Section of AACC, raphy as Applied to the Clinical Hershey, Pa., May 16, 1973- Laboratory

Scott, C. D., and Advances in the Development of High 75th National Meeting of AIChE, W. W, Pitt, Jr. Resolution Liquid Chromatography as Detroit, Mich., June 6, 1973- Applied to Analyses of Physiologic Fluids

Scott, C. D. Fast Analyzer Systems American Society for Pharmacology and Experimental Therapeutics, East Lansing, Mich., Aug. 23, 1973-

Tiffany, T. 0., Potential Use of the Fluorometric Fifth Annual Symposium on Advanced M. B. Watsky, and Fast Analyzer for Enzymatic and Analytical Concepts for the Clinical C. A. Burtis Immunological Assays in the Laboratory, Oak Ridge, Tenn., Clinical Laboratory March 15-16, 1973- (Oral Presentations - continued)

Speaker and Coauthor(s) Title Place Presented

Tiffany, T. 0., Determination of Kinetic Enzyme Fifth Annual Symposium on Advanced D. D. Chilcote, and Parameters for General Clinical Analytical Concepts for the Clinical C. A. Burtis Use, Genetic Screening, and Laboratory, Oak Ridge, Term., Biomedical Research Through the March 15-16, 1973- Use of a Small Computer-Interfaced Fast Analyzer

Tiffany, T. 0. The Fluorescent Fast Analyzer Spring Meeting, Southeast Section of the American Association of Clinical Chemists, Gatlinburg, Tenn., April 5-6, 1973-

Tiffany, T. 0. Recent Developments in Fast Joint Spring Meeting CSCC and AACC Analyzers (Upstate N. Y. Section), Hamilton, Ontario, Canada, May 3-U, 1973*

Tiffany, T. 0., The Use of the Fluorometric Fast 25th National Meeting of the C. A. Burtis, and Analyzer for Measuring Enzyme American Association of Clinical M. B. Watsky Activity (A Discussion of Direct Chemists, New York City, July 15-20, Referencing of Fluorescence 1973- Assays) xix

SUMMARY

Fast Analyzer Development Two miniature analytical systems, each consisting of a miniature Centrifugal Fast Analyzer (CFA), several rotors, discrete loading station, a rotor cleaning station, and either a computer or portable data printer, have been developed and fabricated and are currently undergoing evaluation in the clinical laboratories of the NASA Manned Space Center, Houston, Texas, and the Health Division of the Oak Ridge National Laboratory. In addition, the technique of dynamic liquid introduction has been extensively investigated and parametrically optimized. Other developments include a light-scattering and fluorescent monitor and a digital data printer. The first prototype of a portable "battery-operated analyzer has been fabricated and is currently being evaluated. New applications for the fluorometric CFA include: (1) a technique for dynamic calibration in order to equate the fluorescent signal with the concentration of the analyte; (2) a kinetic method for determination of acid phosphatase; (3) a competitive binding assay for insulin using fluorescein-tagged insulin; and (U) a method for flucrometrically measuring uranium.

Genetic Monitoring

A system for preparing plasma, wa.shed red blood cells, and red blood cell lysates is under development. Preli- minary work has begun on an automated elution electrophoresis system to improve the resolution and speed of electrophoresis, and the adaptation of the computerized Centrifugal Fast Analyzer to evaluate kinetic enzyme parameters as another means of genetic screening is being studied.

High-Resolution Analytical Systems High-resolution separation and analyses of complex biochemical mixtures have been achieved using liquid chromatographic analyzers which employ colorimetric-, ultraviolet-, or fluorescence-detecting flow monitors. A small computer system has been developed and built which can simultaneously monitor four channels of infor mation from two chromatographs in real-time while analyzing the data from one channel either on-line or from a previous run stored on a cassette tape. A xxi

Hydrogen Production

Programs -were initiated to study two methods of hydrogen production. The first method involves a cylic process in which a soluble substrate (Nag^gQj.) is oxidized while reducing water to form Hg in the presence of the two enzymes ferredoxin- and hydrogenase. The substrate is then to be recycled bacii to the reduced form by a thermal process that releases 02. The net result is the production of hydrogen and oxygen from water. Methods and conditions were determined for growing several grains of Clostridium pasteurianum bacteria as a source of crude enzymes for use in studying the basic kinetics of the enzyme-catalyzed reaction. In the second approach, data are being used to evaluate several chemical reactions that might be combined in a cycle where the net result is the decomposition of water. Solvents ranging from organic compounds to molten salts are being evaluated.

Water Pollution Studies

Automated, high-resolution liquid chromatographs with ultraviolet absorbance and cerate oxidation monitors are being used to analyze various polluted waters. Samples of sanitary sewa.ge treatment plant effluents and polluted natural waters from several different sites ha.ve been concentrated up to 10,000-fold and chromatographed on a high- pressure anion exchange column. The presence of numerous organic contaminants at concentrations less than 1 fig/ liter has been revealed in each of these samples. Thirty- seven of the contaminants in sewage plant effluents have been identified; in addition, 17 chlorine-containing compounds have been tentatively identified as being present in chlorinated sewage. Approximately 1% of the chlorine used in disinfecting sewage results in the formation of stable chlorine-containing organic compounds. This study of chlorination effects is being extended to the use of chlorine as an antifoulant in condenser cooling waters.

Radioactive Waste Disposal

Supportive experimental investigations were conducted relative to three different methods of radioactive waste disposal: (l) salt-mine repositories for high-level solid wastes, (2) shale fracturing process for intermediate-level waste solutions, and ^3) Pressurized Aqueous Combustion (PAC) for ashing combustible wastes. Controlled experiments designed to provide information on the effects of different exposure variables on the amount BLANK PAGE xxii of stored energy were conducted. Samples of salt ("both synthetic salt crystals and bedded salt from a mine) were exposed to gamma rays at various combinations of dose rates, doses, and salt temperatures, and subsequently analyzed for accumulated stored energy. Stored energies ranging from 0 to 18.3 cal/g were observed for total gamma doses between 1.2 and 1?.^ x rads. The salt samples were irradiated at average dose rates between 0.l6 and 7.7 x 107 rads/hr. The amounts of Hg evolved upon aqueous dissolution of irradiated salt is proportional to the amount of stored energy in the NaCl. Detailed information on the performance of the existing shale fracturing facility was obtained during a series of four injections of intermediate-level waste solutions. An Environmental Impact Statement covering the modifications to the ORNL liquid waste handling system was written. Leach rates of 9°Sr, 137 Cs, 239Pu, and 24:4 Cm from grouts cured 0, 7, and 28 days were obtained to provide a better assessment of the environmental impact of shale fracturing for disposal of ILW. A study of the effect of pH on the completeness of oxidation of an aqueous slurry of simulated glove-box waste during PAC showed that the completeness of combustion decreased from 88% at pH 5 to 83% at pH 11. This program was cancelled as of June 1, 1973 > a^d the final report has been published.

Environmental Surveys

Comprehensive surveys to assist in the assessment of environmental effects associated with milling of uranium, production of power at nuclear power plants, and usage of toxic metals and compounds are in progress or approaching completion. Present practices of effluent control and waste management in the uranium milling industry have been surveyed. Most of the literature (i.e., over 200 publications) pertinent to uranium mill processes has been abstracted and catalogued. The calculation of radioactive source terms for the airborne dusts from the mill stacks and for liquid effluents from a solvent extraction mill is now in progress. Operating and design data on evaporators at nuclear installations were collected by direct contacts with 30 • organizations, and a report was prepared. The principal emphasis was placed upon data concerning the system decontamination factors achieved for different classes of radionuclides. All of the available information in AEC dockets concerning practices with solid radioactive waste at nuclear power plants has been collected and is being summarized in in a generic report. Surveys on the use and. disposition of mercury in the United States, as well as on the "biological response to proprietary chemicals used in cooling water, are in progress. A report covering mercury reprocessors has "been issued; a report concerning the usage of mercury by U.S. government agencies is in preparation.

Separations Processes

Investigations in separations chemistry include work in four areas: (l) by-product recovery of uranium from wet-process phosphoric acid; (2) separation^of radium from uranium ore tailings; (3) separation of alpha emitters from reprocessing wastes; and CU) study of the applicability of high-pressure ion exchange techniques to certain hydrometallurgical separations. An alternative first-cycle process for recovering uranium from wet-process phosphoric acid was developed on a bench scale. This includes extraction with octylphenylphosphoric acid in an aliphatic diluent, followed by stripping by oxidation to the hexavalent state with sodium chlorate contained in ~10 M P04. During this step the uranium in the strip solution is concentrated to 15 to 20 g of uranium per liter. The abundance of radium in various fractions of a typical sandstone-type uranium ore was determined. The leachability of radium from ore and HgS04-leached slime solids was studied using various acids and salt solutions. Greater than 90$ of the radium in slime solids was recovered in a cascade leach with 1 M sodium chloride solution. The presence of calcium appeared to have only a minor effect on the precipitation of radium in the presence of BaS04. Several organic extractants were screened for their relative utility in extracting cerium and europium (stand- ins for americium and curium) from nitric acid and nitrate salt solutions. Only octylphenylphosphoric acid and a "pyro" phosphoric acid gave significant Ce(lll) extraction from nitric acid solutions above 1 M in concentration. Several compounds showed efficient extraction of Ce(lll) from LiNC^ —0.2 M HNC^ solutions. In tests with 0.02 M di(2-ethylhexyl)phosphoric acid, Zr-Nb was extracted; however, Ce(lll), Eu, and Ru were essentially rejected. Studies were initiated to determine whether pressurized ion exchange techniques might be useful for resolving difficult metal separation problems. The separation of nickel from cobalt in ammonium hydroxide—ammonium carbonate liquors produced by the Caron process uging 15- to 25-n Dowex 50-X8 cation exchange resin is being investigated. Chemical Applications of Nuclear Explosives

Possible tritium "behavior during the recovery of oil from oil shale broken with nuclear explosives was studied by exposing shale samples to tritiated water vapor and recovering oil from the contaminated shale in a bench- scale retort designed to generate a moving retorting front. The oil recovered by moving the furnace down a; 30-in. shale bed at the rate of 6 in./hr contained about 30% of the tritium found in the retort products, in a concentration range of 0.13 to 0.1*1 n.Ci/ml. Decreasing the furnace travel rate to 2.3 in./hr had essentially no effect on the tritium concentration of the oil; however, in a second test at the lower retorting rate, in which water vapor was added to the gas stream entering the retort, the tritium concentration of the oil was reduced to 0.06 |j.Ci/ml.

Chemistry of Radioiodine

The most recent work in this program has been concerned primarily with development of a mercuric nitrate scrubbing process for removing radioiodine from off-gases. The most important recent finding is that the mercuric nitrate process is relatively ineffective for removing aromatic iodides from gas streams and that the presence of aromatic compounds interferes with the removal of Is and methyl iodide. Aromatics were found to be readily iodinated by iodine in 7 to 12 M nitric acid; hence aromatic organics can be converted to aromatic iodides in the mercuric nitrate scrub solution.

CTR Supportive Research

Studies related to tritium management in fusion reactor systems and to the recovery of tritium from liquid lithium and molten Li2BeF4 coolant-blankets have been initiated. Rates of permeation of hydrogen isotopes through oxide-free metals (e.g., Ni) were found to vary with the half-power of the permeating gas pressure in the range of 1CT3 to 750 torr at 600-700°C. The permeation rates were decreased markedly in the presence of oxide films formed by Dig -IfeO mixtures, and the pressure dependence of the rates was increased to somewhere between half- and first-power. Equilibria in the hydrogen isotope-lithium systems are being measured to provide data necessary for evaluation of methods for recovering tritium from liquid lithium blankets. Initial studies have been involved with determining the pressure of hydrogen in equilibrium at high XXV temperatures vith LiH-Li solutions that have low LiH concentrations. Data obtained at 888 ± 2°C obeyed Sieverts' law, as expected, when the LiH concentration in the lithium phase was very low. Several aqueous solutions of relatively low electrical conductivity were passed through strong magnetic fields at high velocity to simulate molten salts. Faraday's Law was obeyed in each case; that is, an electric field was induced in the fluid perpendicular to the magnetic field. Similar behavior is expected with molten Li2BeF4, which has a much higher electrical conductivity than the aqueous analogs used. The solubilities of Hg and D2 in molten Li2 BeF4 were determined throughout the temperature range 773 to 973°K and were found to be practically identical. Thus, reasonable estimates of the solubil.ity of T2 in Li2BeF4 can be made.

Thermodynamics of Molten-Salt Systems

Studies of the chemistry of tellurium in molten Li2BeF4 were initiated to acquire a more fundamental knowledge of the role that tellurium plays in the corrosion of metals in reactor systems, particularly molten-salt reactors. Estimates of the heats of formation of selected metal difluorides were made to help predict potentially corrosive reactions which these metals might undergo in fission or fusion reactors utilizing molten fluoride salts. Theoreti- cal work related to the compressibility of molten oxide fuels was also done.

Coal Conversion Studies

Studies of the conversion of coal to liquid and gaseous fuels by catalyzed and pressurized hydrogenation have been initiated. Preliminary studies have been made on the use of HF in the gaseous mixture to accelerate the hydrogenation process. HF concentrations up to 20 mole did not have a favorable effect. -1-

1. INTRODUCTION

This report represents the first of a series of semiannual reports in which progress for the period March 1, 1973 to August 31, 1973 is reported for the programs in Chemical Development Section B of the

Chemical Technology Division. It is intended that these progress reports will document "basic experimental results much of which will be preliminary in nature, and progress in the various nonexperimental efforts in the

Section will be summarized. Although these progress reports will represent a permanent repository of the results from the various programs in the Section, it is expected that topical reports and open literature publication will ultimately be used to correlate experimental results and to present conclusions based on these results.

Obviously, the subject matter in these progress reports will change with time as the Section becomes involved in other areas, but insofar as possible, there will be a continuity of reporting. This first report will be given in two parts. Part I will include progress on all programs except those dealing with the LMFBR. Programs associated with the LMFBR will be discussed in Part II. BLANK PAGE -2-

2. FAST ANALYZER DEVELOPMENT

Centrifugal fast analytical systems are under development to provide

automated, fast analyses of molecular components in physiologic fluids

and other liquids "by use of colorimetric, fluorometric, and sedimentation

assays. These systems are primarily based on the GeMSAEC principle"*"

and include a multicuvet rotor which rotates through a stationary monitor

•whose output is processed in real-time by means of an on-line computer

or other data processing system.

A miniature analytical system is being developed in an effort to

decrease the overall size and cost. Integral components of such a system

include a miniature analyzer, several plastic rotors of various designs,

an automated sample and reagent loader, and a modified version of a previously developed data printer. Two such systems have been fabricated

and are undergoing testing and evaluation in the clinical laboratories of

the Manned Spacecraft Center of NASA, and the Health Division of Oak

Ridge National Laboratory. Many routine clinical assays have been

adapted for use with the system, and special emphasis has been placed in

operation in the multichemistry analytical mode. A portable micro system

is also under development, and alternate monitoring systems including those measuring fluorescence and light scatter are being studied.

2.1 Miniature Fast Analyzer C. A. Burtis, W. F. Johnson, J. C. Mailen, T. 0. Tiffany, J. B. Overton, R. C. Lovelace, and M. B. Wat sky

The successful development of centrifugal Fast Analyzers for use in

the clinical laboratory has led to consideration of the same principle -3- for a miniature clinical analyzer that would be capable of operating in zero gravity conditions of space flight as well as ground-based operation.

The unique centrifugal mode of sample and reagent handling, coupled with rapid automated processing of the data, makes this analytical concept particularly adaptable for space applications. Further, the possibilities of miniaturizing the analyzer and using extremely small quantities of

sample and prepackaged reagents are compatible with the weight and space

limitations for space flight application; in addition, these features are very attractive for ground-based operation, especially in the small

clinical laboratory, the emergency laboratory, the pediatric laboratory,

and medical research laboratories.

Previous studies indicated that a miniature analyzer based on the 2-4

GeMSAEC concept was feasible, and that it was possible to design and

fabricate operable, miniaturized systems that will be useful not only in

space-flight applications but also in ground-based laboratories. Recent

objectives have been to design and fabricate two miniature analyzer

systems for evaluation as ground-based clinical systems. These analyzers

will undergo testing in the clinical laboratories of the NASA Manned

Spacecraft Center, Houston, Texas, and the Health Division, Oak Ridge

National Laboratory, Oak Ridge, Tennessee. A complete system includes

the miniaturized analyzer, several plastic rotors, a portable data printer,

an automated sample and reagent loader, and a rotor washing station. Each

system will also have the capability of on-line computer applications.

Accompanying these systems will be several analytical assay protocols. 2.1.1 Instrumentation Development

Two miniature analytical systems have "been designed, fabricated, and assembled. Each of these systems, as shown in Fig. 2.1, consists of a miniature Fast Analyzer, several rotors, an automated sample-reagent loader, a data system, and a rotor cleaning station. The heart of the analytical system is an improved model of earlier miniature Fast Analyzer prototypes. As with the earlier prototypes, this model (Fig. 2.2) is compact (it occupies only a square foot of bench space) and weighs only

30 lb. Its primary function is to spin a 17-cuvet rotor at speeds up to

5000 rpm (normal operating speed is 1000 rpm) through a stationary optical system. The optical system consists of a quartz-iodine-tungsten light source located in a movable housing mounted above the rotor, interference filters, and a miniature photomulitplier (PM) tube. The six interference filters (3^0, 1+00, kl5, *+85, 550, and 620 nm; Ditric Company, Marlboro,

Mass.) are contained in a movable filter wheel which is mounted just under the rotor housing. These filters are positioned manually in the optical light path by means of the filter selector switch. As with all

Fast Analyzers, the basic raw data generated by spinning cuvets through the optical system consist of a synchronized series of voltage signals which can be displayed by an oscilloscope and acquired and processed by means of either an on-line computer or other data systems.

Aside from these similarities, several improvements have been made in the present model, including the addition of a temperature monitoring and control system and a control circuit to automatically maintain the reference output of the PM tube at a preset level. PHOTO NO. 0248 - 73 A

Fig. 2.1. Analytical System Based on the Use of a Miniature

Centrifugal Fast Analyzer. -6-

PHOTO NO. 0244-73 A

OPTICAL HEAT LAMP LAMP

ROTOR AND CUVET SYNCHRONIZATION CIRCUITS

ROTOR

CONTROLS

AUTOMATED PM CONTROL SYNCHRONIZATION CIRCUIT

Fig. 2.2. Miniature Centrifugal Fast Analyzer. Automated Temperature Monitoring and Control. — The importance of monitoring and controlling the temperature of the; reaction mixture in a cuvet, especially in kinetic analyses, has led to the development of a system to perform these functions automatically in the miniature Fast

Analyzer. To measure the temperature of the rotor, a calibrated thermistor is embedded in a copper-tipped, stainless steel pin. This pin is mounted on the rotor holder of the analyzer and also serves as one of the two indexing pins that mate with matching holes in the rotor to ensure a single orientation of the rotor in its holder. Once a rotor is in place, the measuring section of the pin contacts the center portion of the rotor midway between two cuvets. The signal from the thermistor is transmitted from the rotor holder through a pair of slip rings and brushes to an operational amplifier circuit, and the rotor temperature is displayed on a meter.

The thermistor output is also used with an additional operational amplifier circuit as a proportional controller to maintain the cuvet temperature a.t a preset level. The output from the controller powers the output of an additional lamp, mounted above the rotor, which acts as an infrared source to heat the rotor. The rotor temperature can be maintained from somewhat above ambient up to 37-5°C; typically it is controlled at

30.0° ± 0.2°C in an ambient temperature of 25° ± 1°C.

Automated Photomultiplier Output. — The second improvement in the analyzer has been the addition of a control circuit that automatically maintains the transmission pulse obtained from the solution in the reference cuvet (cuvet l) at a constant, preset level, usually from 9 to

9.5 V, by varying the high voltage supply for the photomultiplier tube. -8-

One of the three photometric detectors (shown in Fig. 2.2) is used to provide a signal indicating that cuvet 1 is in the optical light path. The automated control circuit then obtains the photomultiplier voltage signal produced by the light transmission peak of the solution contained in cuvet 1, and compares this value to a reference voltage signal which has been previously preset by the operator via the reference voltage potentiometer located on top of the instrument cabinet. After the voltage of the transmission signal obtained from cuvet 1 has been compared with the preset voltage, the automated circuit varies the high voltage supplied to the photomultiplier until the two voltages are equal. This modification eliminates the necessity for manually adjusting the photomultiplier voltage supply each time a different filter is placed in the optical light path. Such a feature proves to be quite advantageous when performing multichemistry analyses that require sequential and repetitive filter changes. An additional benefit is that it also provides for maximum digital resolution and performance of the analyzer's data interface.

Rotors. — The design, fabrication, and assembly of multifunctional rotors for use with a miniature Fast Analyzer have been previously 2-5 reported. An unassembled rotor consists of three separate parts

(i.e., the body and the two windows). Initially the rotor components were fabricated from acrylic plastic using a manual machining method.

This method is useful when only a few rotors of various designs are required. However, the resulting components tend to be relatively,

expensive; also, rotor-to-rotor reproducibility becomes a problem when

a large number of rotors are produced. Consequently, a computer- -9- controlled machining technique was employed to fabricate the components for 1+0 rotors which are supplied with each of the analytical systems.

The rotors, an example of which is shown in Fig. 2.3, were assembled using a silicone adhesive (Dow Corning, 31^5 RTV; Dow Corning Corporation,

Midland, Mich.).

Automated Sample-Reagent Loader. — The sample-reagent loader, which is shown in Fig. 2.1+, will sequentially and automatically obtain and dispense aliquots of reagent(s) and sample(s) into their respective cavities in the rotor. A unique turntable and carousel assembly was designed to allow the loader to operate in any one of four loading modes.

As shown in Fig. 2.5, the first is a testing mode in which repetitive aliquots of the same sample and reagent are loaded and which is useful in evaluating the performance of the loader and/or the analyzer. The second and third are multiple-sample:single-chemistry and single-sample: multiple-chemistry modes, respectively. The fourth mode is a combination of t'.ie second and third modes and allows for multiple-sample rmultiple- chemistry analysis.

The heart of the loader consists of two Micromedic Automatic Pipettes

(Model 25001+, Micromedic Systems, Inc., Philadelphia, Penn.) which, deliver reagent and sample, respectively. Both of these pipettes are operated in the sample-diluting mode, which entails aspirating an accurately measured volume of sample or reagent into a small-diameter, flexible delivery tube and then dispensing and following it with a preset quantity of diluent (usually distilled water). With these pipettes, the volumes aspirated and dispensed are set by adjusting index counters on the control panel of the pipette. This adjustment allows the different PHOTO NO. 3737-72

r .

- i 3!

0 < 2 | OAK RIDGE NATIONAL LABORATORY |

Fig- 2.3* Seventeen-Cuvet Plastic Rotor Used in the Miniature Fast

Analyzer. Discrete loading side on right, dynamic loading side on left.

(Scale in inches.) -J.JL-

PHOTO NO. 0236- 73A

REAGENT SAMPLE PiPETTE PIPETTE

SAMPLE a REAGENT PROBES

SAMPLE AND REAGENT CAROUSELS

CONTROLS

* m^

Fig. 2.b. Automated Sample-Reagent Loader Developed for Use with

the Miniature Centrifugal Fast Analyzer. PHOTO NO. 0453-73

Multi-Sample •• Single-Chemistry 1 H CO F

Single-Sample: Multi-Chemistry Multi-Sample: Multi-Chemistry

Fig. 2.5. Four Loading Modes of an Automated Sample-Reagent Loader. volumes to "be varied from 5 to 100$ of full pump capacity. Each of the pipettes used in the sample-reagent loader contains 20-|_l1 a,nd 50-p.l sample and diluent pumps. Typically, the two pipettes are set to obtain and deliver 10 |il of sample followed "by 50 [j.1 of diluent, and 20 \il of concentrated reagent followed "by 50 JJLX of diluent, respectively.

Data Printer. — To circumvent the cost and immobility of a computer- based data system, a small, portable data device has been developed. An improved version of it is shown in Fig. 2.6. Briefly, this unit can be described as a device for acquiring, processing, and printing the results of analyses carried out in a miniature Fast Analyzer. The analyzer transmission pulses are converted electronically to absorbance values, which are recorded as printed dots on a data card. Vertical height on the card is proportional to absorbance, with the maximum absorbance and the absorbance span to be covered on the card being set by means of calibrated potentiometers.

Tne original design was modified and improved by the addition of three additional sets of absorbance maximum and span potentiometers. Thus, the photometric outputs from cuvets 2-5, 6-9, 10-13, and 14-17 may be processed at the same or different absorbance maxima and spans. This modification has greatly increased the versatility of the printer. A typical data card obtained from the modified data printer is shown in

Fig. 2.7. Another modification has been the addition of an electronic module to perform repetitive scans automatically at timed intervals during an analytical run. This new feature has been found to be quite useful in processing data from enzyme assays. -11*-

PHOTO NO. 1366-73A

Fig. 2.6. Modified Data Printer Developed for Use with the Miniature Centrifugal Fast Analyzer. -15-

ORNL DWG. 72-13505 Rl

LDH-L S60T CPK SGPT .55I 46 .49 .48

O * IO UJ O

CD J < .35- .36J .16- *.38

CUVE** T 2345 6789 10 12 13 14 15 16 17

SAMPLE 1234 1234123 4 1234

Fig. 2.7. Typical Data Card Obtained from Data Printer When System

Is Operated in the Multiple-Chemistry:Multiple-Sample Analytical Mode. -i6-

Rotor Cleaning Station. — When the rotor fabrication and assembly- techniques discussed earlier are used, the individual cost per rotor is about $70 each. This cost has made it imperative to reuse the individual rotors, -which in turn has led to the development of a module (Fig. 2.8) to automatically clean and dry a rotor after an analytical run.

The operation of this module, or cleaning station, involves an initial emptying of the cuvets of their reaction mixtures, followed successively by a 4o-ml flush of the rotor cavities with a 2:1 methanol- water mixture and by air-drying of the rotor. This cleaning operation requires 10 min per rotor, but recent developments indicate that the time can probably be decreased to 5 min or less. With the development of this cleaning station, rotors have been used as many as 100 times without encountering problems.

2.1.2 System Performance and Evaluation

The miniature analyzer was extensively tested and evaluated in order to ensure that it is capable of performing routine clinical analyses on very small sample volumes (l to 10 |al) with a high degree of precision and accuracy.

Optical Performance. — Since a miniature Fast Analyzer is basically a multicuvet photometer, it must provide accurate and precise photometric measurements. To determine if it meets this specification, reference solutions of the reduced form of nicotinamide adenine dinucleotide (NADH) in Tris buffer (0.1 mole/liter, pH 7.*0 and £-nitrophenol (PNP) in Tris buffer (0.1 mole/liter, pH 10.5) were prepared, and their absorbances at 3^0 and i+00 nm were determined using both the miniature analyzer system PHOTO. NO. 0240-73A

RESERVOIR FOR CLEANING SOLUTION

ROTOR COMPARTMENT AIR DRYER

ACTUATING SWITCHES

ROTORS

Fig. 2.8- Rotor Cleaning Station Used for Cleaning and Drying the

Rotors Used in the Miniature Centrifugal Fast Analyzer. -18- and a Cary lU spectrophotometer (Cary Instruments, Monrovia, Calif.).

The data, as listed in Table 2.1, indicate that the values obtained from the miniature analyzer agree to within 1 to 2% with those obtained by the Cary 1^. In addition, when the absorbances of the individual solutions are plotted vs their NADH concentrations (Fig. 2.9), the photometric output of the miniature analyzer is shown to be linear up to an absorbance of 2.0 at 3^0 nm.

In regard to photometric precision, the absorbances of the NADH solutions were repetitively determined. The results obtained were statistically processed and yielded the data listed in Table 2.2. As can be seen, the photometric precision of the analyzer is quite satisfactory, as the absorbances of the individual solutions were determined with a variation of only ± 0.06 to 0.20$, which corresponds to actual absorbance values of 0.0003 to 0.00^-0.

These results indicate that the miniature Fast Analyzer system provides the optical linearity, accuracy, and precision required in order for chemical procedures to be performed with the system.

Temperature Control and Monitoring. — One of the unique features of the miniature Fast Analyzer is the use of a thermistor in th-: rotor housing to measure the temperature of the rotor. The thermistor, together with its associated circuitry, is an integral part of the circuit that controls the rotor temperature.

To compare the temperature measured by the monitoring thermistor with the actual temperature of a liquid contained within a cuvet, a calibrated thermistor was inserted into an individual cuvet of a special test rotor. This thermistor was then connected to its external monitoring -19-

Table 2.1. Absorption Measurements from the Miniature Centrifugal Fast Analyzer and a Cary Spectrophotometer

Absorbance (0.5-cm cell) Miniature Wavelength Concentration Fast " Cary (nm) Solution (|j.mole/ml) Analyzer Spectrophotometer

3U0 NADHa 0.085 0.261+2 0.268

0.128 0.3931 0.398

0.256 0.80': 2 0.800

0.320 O.9897 O.983

0.^22 1.31^9 1.320

0.533 1.6662 I.683

0.6U0 1.9826 2.026

kOO PNP13 0.01 0.09^7 0.092

0.025 O.2365 0.23?

0.05 0.V711 O.kSO

0.075 0.701+6 0.687

0.10 0.9358 0.917

^ADH (nicotinamide adenine dinucleotide, reduced) prepared in Tris buffer (0.1 mole/liter, pH 7.^).

•u PUP (p-nitrophenol) prepared in Tris buffer (0.1 mole/liter, pH 10.5). -20-

ORN L DWG 72- M7IIRI

O COMPUTER DATA A OSCILLOSCOPE DATA

O.I 0.2 0.3 0.4 0.5 NADH CONC.(millimole/ liter)

Fig. 2.9. Optical Linearity of the Miniature Centrifugal Fast

Analyzer at 3^0 nm» -21-

Table 2.2. Photometric Precision Obtained with the Miniature Fast Analyzer

Mean Absorbancea Standard "1 Cone, of NADH (0.5-cm cell, Deviation RSD (l_imole/ml) 3I+0 nm) (Absorbance Units) (#)

0.085 0.261+2 0.0003 0.11

0.128 0.3931 0.000k 0.10

0.256 0.80^2 0.0006 0.07

0.320 0.9897 0.0006 0.06

0.1+22 1.31^9 0.0010 0.08

0.533 1.6662 0.0020 0.12

0.61+0 1.9826 0.00^0 0.20

81 10 data points averaged per concentration.

Relative standard deviation or coefficient of variation. -22- circuitry by means of an extra set of slip rings mounted above the rotor. Subsequently, 100-M.I aliquots of water were placed into each of the reagent cavities of the rotor. The rotor and its contents were then brought to a preset temperature by slowly rotating the rotor at 100 rpm with a concurrent high heat influx from the heating lamp of the analyzer.

After reaching the preset temperature, as indicated by the display meter, the contents of the reagent cavities were transferred into their respective cuvets, and the temperatures of the rotor and cuvet were simultaneously monitored for a period of time. A typical trace from such an experiment is shown in Fig. 2.10. As anticipated, the rotor and cuvet temperatures are slightly offset. However, due to the parallel nature of this offset and the precision of the controlling circuit (±0.2°C), it is possible to control the cuvet temperature at a constant value.

Consequently, the output of the rotor thermistor and its display is set and calibrated to maintain the cuvet temperature at 30.0 ± 0.2°C.

To evaluate the temperature monitoring and control system under routine analytical conditions, the miniature Fast Analyzer was used to analyze a serum sample for various enzyme activities at 30°C. As an independent check, the identical analyses were also performed using a

Gilford 300 N Spectrophotometer (Gilford Instrument Laboratories, Inc.,

Oberlin, Ohio), whose cuvet temperature was regulated a.t 30°C. To ensure that data handling, data manipulations, and calculations of final enzyme activity would be performed in an identical manner, both instruments were interfaced to an on-line computer. As the results from this comparative study show, equivalent results were obtained from each instrument (Table 2.3). In addition, good analytical precision was ORNL DWG 72-8794

I DO JO I

Fig. 2.10. Temperatures of the Rotor and Cuvet as Recorded During a Test of the Temperature Monitoring and Control System of the Miniature

Fast Analyzer. -2b-

Table 2.3. Comparison of the Analytical Results Obtained from Enzyme Analysis Using a Miniature Fast Analyzer and a Gilford 300 N Spectrophotometer

Enzyme Activity, I.U. liter""1" min"1 at 30°C Q Miniature Fast Analyzer Enzyme Gilford 300 Nb M C BSD11 Mean (%)

ALP bb.6 1.8 b5.3

ALT 12.3 2.9 12.1 AST 17.1 3.2 17.2 CK 87.7 o.b 86.5

LD-L 66.7 1.6 6b. 0

cL Reaction conditions: sample volume = 0.010 ml; total reaction volume = 0.130 ml; total run time = bOO sec; n = 16.

Reaction conditions: sample volume = 0.20 ml; total reaction volume = 2.60 ml; total run time = HOO sec.

°Mean activity of l6 observations.

^Relative standard deviation (coefficient of variation). -25- obtained. The relative standard deviation (RSD) ranged from only OA to 3'2"Jo, which includes not only temperature variation but pipetting and instrumental factors as well.

Rotor Reproducibility. — Due to their central role, each of the assembled rotors was tested for its cuvet variability with regard to light transmission properties and pathlength. The results of these tests are shown in Table 2.b. It is evident that the techniques for fabricating and assembling the rotors resulted in quite uniform and reproducible rotors since the RSD's for the rotor-to-rotor variation for cuvet light transmission and pathlength were only ±1.02 and 0.55%> respectively.

Performance of Automated Sample-Reagent Loader. — In a series of repetitive experiments, various volumes of a concentrated blue dextran

(Pharmacia Fine Chemicals, Inc., Piscataway, N.J.) dye solution (7.*+ mg/ml) were diluted to a final volume of 0.130 ml in each case, and the absorbances of the resulting diluted solutions were measured at 620 nm using the miniature Fast Analyzer. Results of these experiments indicated that the delivered volumes of the sample and reagent pumps were linear over the 1- to 20-[.il volumetric range of the pumps (Fig. 2.11), and that precision ranged from ±3*5% to 1 p.1 to ±0.25% at 20 jil (Table

2.5) • Gravimetric analysis indicate.! accuracy of the Micromedic pipettes to be within the 1 to 2% specification listed by the manufacturer. Addi- tional experiments indicated a carryover of 1$ or less between sample

compartments.

Evaluation of Data Printer. — When interfaced with an on-line

computer the miniature Fast Analyzer has been shown to be capable of providing linear, accurate, and precise photometric results. Consequently, -26-

Table 2.if. Variation in Transmission and Pathlength of Thirty Rotors Fabricated for Use with the Miniature Fast Analyzer

Average Intrarotor Interrotor Variation Variation RSD RSD Mean Mean Ho) m

Cuvet " transmission 9-531 0.72 9.531 1.02

Pathlength (cm) 0.50^ 0.1+7 0.50U 0.55

3# Cuvets were filled with water, measurements were photomultiplier output.

Cuvet 1 was filled with water; cuvets 2-17 were filled with a blue dextran dye solution of known absorbance. ORNL DWG. 73-793RI

Volume (pi)

Fig. 2.11. Volumetric Linearity of the Sample and Reagent Pumps

Used in the Automated Sample-Reagent Loader. -28-

Table 2.5. Precision of the Automated Sample-Reagent Loader at Various Sample Volumes

Resuitsa Relative Sample Standard Standard Volume Mean Deviation Deviatio{%) n (ND Absorbance (Absorbance Units)

1 0.032*4- 0.0027 3.1+8

2 0.0636 0.0021+ 2.67

5 0.1611 0.0018 0.50 10 0.3282 0.0025 0.58

15 0.1+8U-9 0.0016 0.1+1

20 O.6I+33 0.0023 0.25

9* Sixteen replicate aliquots of a concentrated "blue dextran solution (7.^ mg/ml) were obtained at each of the indicated sample volumes and diluted to a final volume of 0.130 ml. The absorbances of the diluted solutions were then measured at 620 nm using the miniature Fast Analyzer. -29- to evaluate the performance of the mechanical Data Printer, "both it and

an on-line computer were interfaced (in parallel) with the miniature

Fast Analyzer. With this experimental arrangement, the absorbances of various dye solutions and neutral density filters were determined at various wavelengths. A plot of the linear regression analysis of the

resulting data is shown in Fig. 2.12. It is evident from the close

correlation of the data that nearly identical absorbance values were

obtained from the two data systems. In addition, the absorbances of the

solutions were measured in a Cary lb spectrophotometer and were found to be essentially the same as those obtained by either the Data Printer or

the computer.

To determine the reproducibility of the Data Printer, replicate ali-

quots of a p-nitrophencl solution having an absorbance of 0.190 at b05

nm were placed in the cuvets of the miniature Fast Analyzer. With the maximum and span potentiometers of the Data Printer set to give an absor-

bance span of 0.15 to 0.25 on the vertical axis of the data card, ten

replicate scans were made and the absorbances of the solutions in the

individual cuvets during each scan were measured. By statistically

averaging and evaluating the resulting data, reproducibility of the Data

Printer was found to correspond to an absorbance of ±0.0015. Although

this value is higher than the ±0.0005 obtained with the computer, it is

probably acceptable under most analytical conditions.

To determine performance under routine analytical conditions, 16

replicate aliquots of Hyland control sera I or II (Hyland Laboratories,

Costa Mesa, Calif.) were analyzed for their lactate dehydrogenase-

lactate substrate (LDH-L) activity and the results were processed -30-

ORNL-DWG 72-5245

1.002* + 0.0(D 3 I 0.999 / /i s * / •

0 0.2 0.4 0.6 0.8 1.0 ABSORBANCE(GeMSAEC COMPUTER)

Fig. 2.12. Correlation of Absorbance Values Obtained Using an On-

Line Computer and the Data Printer. -31-

simultaneously "by the Data Printer and the computer. The enzyme activities obtained "by each data system agreed to -within 0.8 to 2.0$ (Table 2.6).

The computer-derived data were observed to be slightly more precise; however, when one considers that the computer performs its calculations on

1+00 data points per cuvet as compared to only 2 per cuvet for the Data

Printer, the performance of either system is quite satisfactory.

To evaluate the capability of the four groups of absorbance maximum

and span potentiometers for giving equivalent results, each group was set

to give an absorbance maximum of 0.5 and an absorbance span of 0.05 to

0.5. Sixteen replicate aliquots of a serum sample were assayed for their

alkaline phosphatase activity. Their resultant changes in absorbance

at 1+00 nm were monitored at 1-min intervals using the Data Printer. The

average activity was found to be 36.70 I.U., with a coefficient of varia-

tion of only ±1.7$. This indicates that equivalent results may be

obtained from any of the four groups of cuvets.

To evaluate the Data Printer's capability for processing data in a

multiple-chemistry:single-sample analytical mode, a fresh serum sample

was assayed in quadruplicate for its LDH-L, serum glutamate oxaloac.etate

transaminase (SG0T), creatine phosphokinase (CPK), and serum glutamic

pyruvic transaminase (SGPT) activities. The results obtained are listed

in Table 2.7. These results were quite encouraging since they indicate

that the Data Printer performs quite satisfactorily in the processing of

data from multiple chemistry analyses.

Total System Performance. — To evaluate the overall performance of

the analytical system, a single serum sample was repetitively analyzed

for its alkaline phosphatase and AST activities. The resulting data -32-

Table 2.6. Comparison of Enzyme Activities Obtained by Means of the Miniature Fast Analyzer, Interfaced with Either a Computer or the Mark I Data Printer

Enzyme Coefficient Activity of (I.U. liter"1 Variation3, Serum Sample Data Device min"1 ao 30°C) {%)

Hyland I Computer13 63.3 I.71

Data Printer0 6k.5 Z.kl

Hyland II Computer 139-8 1.01

Data Printer0 1^0.9 1.6l aBased on 16 replicates.

"^A PDP 8/l computer was used with the following program constants: delay interval = 20 sec; observation interval = 15 sec; number of data readings averaged per observation = 10; analysis time = 10 min; activity = absorbance change per minute multiplied by 3860. cData Printer conditions: scan every min for 10 min; absorbance change between 1 and 10 min used to calculate enzyme activity. Enzyme activity = absorbance change per 9 min multiplied by 926. -33-

Table 2.7. Multiple Chemistry Analysis Using the Data Printer as a Data Acquisition Device

Absorbance Activity Enzyme Cuvet Nos. Maximum Span (I.U./liter)

LDH-L 2-5 0.6 0.5 73-3 ± 0.8$

SGOT 6-9 O.k 0.2 12.7 ± 5.9$

CPK 10-13 0.35 0.2 97.7 ± 2.14

SGPT llf-17 0.35 0.2 13.8 ± 2.8$

A fresh serum sample was assayed in quadruplicate for each of the four enzym.es.

^At a temperature of 30.0°C. -367-

•were acquired, processed, and statistically evaluated by means of the on-line computer. As shown in Table 2.8, the analytical precision of the system was quite good. The within-run variation for the alkaline phosphatase analyses averaged ±0.75% and rose to only ±1.5% for the run-to- run variation. At a lower enzyme activity, as indicated by the SGOT data, the within-run and run-to-run variations were ±5.30 and ±6.88%, respectively. However, considering that an SGOT activity of 10.93 I.U. liter 1 min"1 at 30°C corresponds to an absorbance change of only

0.0025 min"1, the observed precision is adequate and compares favorably 7 with that obtained with the larger Fast Analyzer.

2.I.3 On-Line Computer

The miniature Fast Analyzer has been successfully interfaced with a previously developed Fast Analyzer data system. 3 As shown in Fig. 2.I3, it has also been interfaced with a compact digital computer system which has been recently developed for use with Fast Analyzers. This data

system is composed of a PDP-8/E computer (Digital Equipment Corporation, Maynard, Mass.) with 8192 words of core memory, a 1200-Hz line-frequency- based clock, an ARS 33 Teletype, a Sykes cassette-tape unit (Sykes Datatronics Inc., Rochester, N. Y.), and the necessary analog and digital interfaces for coupling the analyzer to the computer hardware.

The high-level language, FOCAL , was modified and extended with appropriate software to accept data under real-time control of the line- frequency clock. Modified subprograms have been added to permit program

*F0CAL (FOrmula CALculation) is a conversational language and is a registered trade name of the Digital Equipment Corporation, Maynard, Massachusetts. -35-

Table 2.8. Analytical Precision of the Miniature Fast Analyzer System in Performing Enzyme Assays

Average Within- Hun-to-Run Run Variation3, Variation^3 Relative Relative Standard Standard Mean Deviation Mean Deviation Enzyme Activity0 Ho) Activity0 a)

Alkaline ^ 0.75 1.51 phosphatase

e SG0T 10.93 5.30 10.93 6.88

£1 Sixteen replicate aliauots were assayed during a single run, and the resulting data were statistically analyzed.

One hundred twelve replicate aliquots were assayed over a seven- run period, and the resulting data were statistically analyzed. c -1 -1 Enzyme activity is expressed as I.U. liter min at 30°C.

^Reaction conditions: wavelength, ^00 nm; sample volume, 0.010 ml; total volume, 0.130 ml; total run time, 200 sec. eReaction conditions: wavelength, 3^0 nm; sample volume, 0.010 ml; total volume, 0.130 ml; total run time, 1+00 sec. PHOTO NO. 0254-73 A

Pig. 2.13. A Compact Computer Data System Developed for Use with

Fast Analyzers. (Shown here interfaced with a miniature Fast Analyzer.) -37- and data storage on and retrieval from the cassette tape unit. Direct access features of the tape transport permit storage of data from many analyses for later correlation and study, as -well as reprocessing of the original data with new data processing techniques.

As with the earlier system, all of the analytical computer programs are written in FOCAL and are stored and retrieved from the cassette unit upon operator command. The routine operating programs include calibration, concentration measurements (both end-point and kinetic), and enzyme rate analysis routines.

2.1.U Development of Chemical Assays

Chemical procedures for acid and alkaline phosphatase, creatinine phosphokinase (CPK), lactic dehydrogenase-lactate substrate (LDH-L), serum glutamate oxaloacetate transaminase (SGOT), serum glutamate pyruvate transaminase (SGPT), glucose, blood urea nitrogen, serum triglyceride, calcium, total bilirubin, and for multienzyme:sample analyses have been developed or adapted for use with the miniature Centrifugal

Fast Analyzer.

2.1.5 Dynamic Loading of Liquids

Liquids may be introduced into the miniature Fast Analyzer in either a discrete or dynamic mode. In the discrete mode, discrete aliquots of liquids are obtained by means of a pipetting device and transferred and dispensed into their respective cavities in a stationary rotor.

A device to automatically load a rotor In the discrete mode has been developed and previously discussed in Sect. 2.1.1 of this report. The main disadvantages of discrete loading are the relatively high cost of -38- instrumentation to automate it and the decreased portability of such a loading system.

To circumvent these two disadvantages, the technique of dynamic loading of liquids into a spinning rotor is being developed. Basically the technique consists of dynamically introducing a volume of liquid

(reagent or sample) into the centrifugal field generated by the spinning rotor of the Fast Analyzer (Fig. 2.1^). The rotor can be designed to apportion the volume of liquid into discrete measured volumes that are automatically transferred into their respective cuvets for reaction initiation followed by photometric monitoring of the individual reactions. Q Initial studies indicated the feasibility of using dynamic loading with a miniature Fast Analyzer. In the interim we have developed a dye dilution technique for evaluating the precision and accuracy of dynamic loading and have identified and investigated several parameters that influence its operation. The Measurement of Dynamically Loaded Volumes. — To measure the volume per cuvet obtained by dynamically loading a known volume of liquid into a spinning rotor, a dye dilution technique was developed. In this technique precise and accurate aliquots of a dextran blue dye solution of known absorbance are discretely dispensed into cuvets 2-17 of a rotor. (Note: cuvet 1 serves as a reference cuvet and is filled with water.) These aliquots are then transferred into their respective cuvets. A known volume of water is then dynamically introduced into the spinning rotor. After mixing, the absorbances of the diluted dye solutions are measured. The volume dynamically introduced into each cuvet can be calculated using the following equation: -39-

ORNL DWG 72-3204

KNIFE-EDGE SEPARATORS

ROTOR

CUVET LIQUID STREAM

Fig* 2.1^. Dynamic Apportionment of a Liquid Stream in a

Centrifugal Field by a Centrifugal Fast Analyzer Rotor. -to-

ys , a2> , (1) where

V2 = volume of dynamically loaded liquid,

VI = volume of discretely loaded liquid,

A2 = absorbance of diluted dye solution,

A1 = absorbance of undiluted dye solution.

This equation has been incorporated into a rewly developed computer program for use -with the on-line computer -which automatically calculates the dynamically loaded volume for each cuvet. To evaluate the splitting performance of a rotor, a statistical analysis is performed on the calculated volumes and absorbances of the diluted dye solution in cuvets

2-17. A correction for cuvet-to-cuvet variability inherent in the rotor is also made. To evaluate run-to-run -variability within a cuvet, the data from several runs can be stored for subsequent recall and statistical processing.

Using this technique, several of the parameters important in dynamic loading were identified and investigated. They included rotor design and geometry, rotor speed, injection rate, and probe diameter.

Rotor Design and Geometry. — A 17-place rotor (shown previously in

Fig. 2.3) has been designed and fabricated for use with the miniature

Fast Analyzer. This rotor is a multifunctional device in that sample(s) or reagent(s) can be loaded either discretely or dynamically. In the dynamic mode, solutions can be injected into the spinning rotor and diverted into the rotary path of its splitting vanes. The resulting apportioned aliquots are simultaneously transferred into their respective -1+1-

'j.wets. In testing the original design several modifications were found to "be necessary (Fig. 2.15). The computer-controlled machining method used to fabricate the rotor components tended to produce small, thin curls of plastic on the face of the splitting vanes with a concurrent loss in the splitting performance. However, these curls were easily removed by a polishing procedure which resulted in sharp, smooth surfaces on the face and walls of the splitting vanes.

A second modification concerned the volume capacity of the receiving chambers of the rotor. In the original design (Fig. 2.l6), each had a capacity of approximately 20 |al. In testing this design, reasonable splitting performance was obtained when the injected volume per cuvet did not exceed 30 \±1 per cuvet. However, when the volume per cuvet was increased to either 1+0 or 50 p.1, splitting performance decreased. This was attributed to an overfilling of the individual receiving chambers during the dynamic loading operation. On increasing the chamber capacity to 50 \il} equivalent performance was obtained independent of the volume loaded per cuvet (Fig. 2.l6). However, a;t a volume of 60 \il per cuvet, equivalent performance was obtained irrespective of the volume capacity of the receiving chamber. A possible explanation for this anomaly is that the total injected volume required to load 60 y.1 Per cuvet is 1.020 ml. This volume is relatively large considering the internal geometry of the rotor, and when injected was observed to flood the center portion of the rotor. Thus, under the centrifugal field of the analyzer, this flooding results in the formation of a symmetrical, volumetric donut near the center of the rotor. Due to the centrifugal field this volumetric donut is forced towards the rotor periphery and drains at an equal rate ORNL DWG. 73-8642

RECEIVING CHAMBERS (20 fi\) RECEIVING CHAMBER (50 I )

SPLITTING VANES - ORIGINAL MODIFIED

Fig. 2.15. Modification of the Dynamic Loading Side of a 17-Place

Rotor Used in a Miniature Centrifugal Fast Analyzer. -1+3-

ORNL DWG 73-8639

LOADED VOLUME (pi)

Fig. 2.16. Precision of Dynamic Loading as a Function of Volume

Capacity of Rotor Receiving Chambers. -1+1+- through the receiving chambers into the cuvets of the rotor. Evidently, at lower volumes this flooding phenomenon does not occur and unequal apportionment results.

A third modification involved the internal geometry of the rotor immediately adjacent to the splitting vanes. If the injected liquid was directed onto the walls of the splitting vanes, an aerosol was often formed, resulting in poor splitting performance. This problem was eliminated "by machining a hemispherical cavity into the center of the rotor. In this present modification the injected liquid stream now strikes the smooth face of the hemispherical cavity, which improves the splitting performance of the rotor.

In addition to rotor modifications, a syringe and probe guide was added to the analyzer. The syringe and probe can now be accurately and reproducibly positioned for each analysis, and this has greatly improved the splitting performance of the system.

Rotor Speed. — Before the above modifications were made to the rotor, it was observed that the optimum rotational speed of the rotor was dependent on the volume injected per cuvet. For example, when the volume injected was 10 p.1 per cuvet, optimum performance was obtained at 1+000 rpm. When the volume loaded per cuvet was increased to 50 p.1, optimum performance was achieved at 2000 rpm. However, this inverse relationship was not observed when the modified rotor was tested. As shown in Table

2.9, essentially equivalent performance was obtained for each of the four rotational speeds tested. Consequently, a rotational speed of either

2000 or 3000 rpm is now routinely used to dynamically load the rotors. -1+5-

Table 2.9. Effect of Rotor Speed on Splitting Performance During Dynamic Loading3,

Rotation Measured Volume (ul) Volume Injected (|il) Speed Standard Total Per Cuvet (rpm) Mean Deviation

170 10 1000 8.1+ 0.6 2000 7.0 1.1 3000 7.h 1.1 1+000 9.7 0.9

3lf0 20 1000 19.3 0.7 2000 20.1 0.7 3000 21.3 0.8 1+000 21.6 1.2

510 30 1000 30.9 1.0 2000 28.9 0.7 3000 29.8 1.3 1+000 28.7 1.5

850 50 1000 50.1+ 1.6 2000 50.1 1.8 3000 50.0 1.6 1+000 50.0 1.6

'Conditions: Total volume injected in 1.5 sec using a Fast Micromedic Automatic Pipette; injection probe, 0.01 in. ID; wavelength = 620 nm. g Probe Diameter. — An earlier report indicated that if a liquid is dynamically introduced into a rotor in the form of large drops rather than as a continuous stream, unequal apportionment could result, since the individual drops might be directed to only a few cuvets. This effect can be minimized by introducing the liq_uid through a small-diameter delivery tip. To experimentally test this hypothesis, various volumes of liquids were introduced into the spinning rotor through delivery probes having different tip diameters. A Fast Micromedic Automatic Pipette was used to inject the liquid through probes having an internal tip diameter of either 0.010, 0.020, or 0.030 of an inch. For each of the three probes, the standard deviations obtained from the four injected volames (i.e., 10,

20, 30, and 50 per cuvet) were averaged and this value was plotted against the internal tip area of the delivery probe.

The data in Fig. 2.17 clearly indicate the advantages of using a delivery probe having a tip of very small internal diameter. However, a problem of a practical nature must be considered in selecting a delivery probe. For operational ease, mechanical stability, cost, and availability, it would be desirable to use a hypodermic needle as a delivery probe.

However, a needle having an internal diameter of 0.010 in. is equivalent to a 33-gauge needle. This type of needle is often difficult to obtain and it is easily obstructed during operation. This practical problem can be circumvented by using Teflon delivery probes which are tapered down to an internal tip diameter of 0.010 In. The tips of these probes can be cut 1 in. from their furthermost end, and then slipped over the end of a 26-gauge needle. The result is an easily obtainable delivery probe having a small-diameter, flexible tip. Since only the tip and a ;thort -U7-

ORNL DWG 73-8640 r

h<-

U> J

UJ o < cr UJ > <

TIP AREA OF LOADING PROBE (cmxIO*) Fig. 2.17- Precision of Dynamic Loading as a Function of Tip Area of Loading Probe. -US- section of the probe is of the small diameter, this reduces the possibility of plugging. In addition, the flexible Teflon tip will not scratch the rotor -when, as frequently happens, the delivery probe is inadvertently inserted too far into a spinning rotor. Injection Rate. — When dynamically loading a liquid into a rotor, the injection rate depends on the total volume injected and the time required to inject this volume. To investigate the effect of injection rate on splitting performance, a Fast Micromedic Automatic Pipette was used as a dispensing device. With this pipette, the volume dispensed can be easily varied, but is always delivered in 1.5 sec. In addition, the delivery probe of the pipette can be easily changed, which allows one to vary the linear velocity of the dispensed liquid. The experimental procedure included placing the Micromedic Pipette above and behind the miniature analyzer. A 1-ml pump was placed into the pipette and a delivery probe attached to it. The delivery probe was placed into the probe guide of the analyzer at a 60° angle relative to the horizontal plane of the rotor. The tip of the delivery probe was inserted into and positioned within the hemispherical cavity of the rotor such that its injected stream would be delivered onto the smooth wall of the cavity. Experiments were conducted in which the injected volume and tip diameter were varied and their effect on splitting performance determined. In general, for a given probe diameter, equivalent splitting performance was obtained when the volumetric flow rate was varied from 0.113 to O.567 ml/sec. However, when the data obtained from the three probes were plotted as a function of the linear velocity of the injected liquid (Fig. 2.18), splitting performance was found to improve, with an increase ORNL DWG. 73-8641 R!

T T

0.010 - in. PROBE 0.020-in. PROBE 0.030 - in. PROBE

i VIO

JL ± ± J. 200 300 400 500 600 700 800 900 1000 LINEAR VELOCITY (cm/sec)

Fig. 2.18. Precision of Dynamic Loading as a Function of the Linear

Flow Velocity of the Injected Liquid. -50- in linear velocity. This improvement is quite marked when the linear velocity is increased from 25 to 200 cm/sec. Above 200 cm/sec, the performance curve flattens and equivalent performance is obtained in the linear velocity range of 200 to 1000 cm/sec.

This decrease in effects at the higher linear velocities is of practical significance since it indicates that equivalent performance can be obtained within a range of injection times. Since linear velocity is a function of the volumetric flow rate and area of the flow vessel, given a probe of known diameter (i.e., 0.010 in.), the time required to inject a given volume of liquid at a given linear velocity can be easily calculated. For example, equivalent results can be obtained, when a volume of 170 nl (10 ul/cuvet) is injected into a rotor within a time range of from 0.3 to 1.7 sec. At 850 nl this range increases from 1.7 to 8.3 sec.

The practical advantage achieved from this range of injection times is that a simple device can be manually used to load liquids instead of an expensive, automated one. A manual device, such as a hypodermic syringe, requires an operator and thus the time of injection would be expected to vary due to operator individuality (Fig. 2.19). However, almost equivalent results can be obtained even though the injection time varies slightly. Equivalent results were obtained in an experiment in which performance of an automatic dispensing device was compared with that of manual operation (Table 2.10).

Precision of Dynamic Introduction. — Utilizing the previously mentioned improvements, successive experiments were performed to determine the run-to-run precision of dynamic loading. As shown in Table 2.11, -51-

PHOTO NO 0820-73

A *AUfZER

Fig. 2.19. dynamic Introduction of a Liquid into the Rotor of

Miniature Fast Analyzer Using a Kypoderiuic Syringe and Needle. -52-

Table 2.10. Comparison of the Splitting Performance of a Manual Versus Automatic Dynamic Loading Device

Conditions: Rotor speed at time of injection = 3000 rpm, •wavelength = 620 nm.

Volume Standard Deviation (j-il) Injected Calculated Within Cuvet-to- Device (nD (nD Cuvet Cuvet a Manual 50 50.5 0.8 1.5

Automatic 50 50.1+ 1.3 1.8 al-ml Hamilton syringe, 26-gauge needle having a flexible 0.010 in.-ID tip.

Fast Micromedic Automatic Pipette, 1-ml pump. -53-

Table 2.11. Run-to-Run Precision of Dynamic Loading9,

st andard Run Volume (Ml) Deviation No. Injected Mean (MD

1 50 1+9.2 1.6

2 50 51.1 1.2

3 50 50.8 1.3 k 50 1*9.7 0.1+

5 50 51.0 0.7

6 50 50.5 1.0 1-6, Avg. 50 50.li- 1.0

SI Rotor speed of 3000 rpm and introduction by 1-ml Hamilton Syringe with 26-gauge needle having a flexible 0.010-in.- ID tip. b Sixteen values per run. -54- excellent run-to-run precision was obtained as the average cuvet volume over a six-run test (n = 96) was 50.4 ± 1.0 nl. A series of experiments was also performed in which precision vs volume injected was determined.

These data are summarized in Table 2.12.

Analytical Comparison of Discrete vs Dynamic Loading of Reagents. —

A series of samples was analyzed for their various enzyme activities using the miniature Fast Analyzer. In one set of experiments, both the samples and reagents were loaded discretely using the automated sample- reagent loader (Sect. 2.1.1). In a second set of experiments the samples were loaded discretely with the automated sample-re agent loader, and then the rotors were placed in the analyzer and the reagents were dynamically introduced. As seen in Table 2.13, equivalent results were obtained using either discrete or dynamic loading of the reagents. These data indicate that reagents may be dynamically loaded into a rotor with no loss in analytical performance.

2.1.6 Fluorescence Monitor

A fluorescence monitor for Fast Analyzers has been previously developed for use with a 15-place GeMSAEC Fast Analyzer.9A0 Due mechanical and physical constraints of the analyzer's rotor assembly, this optical system utilizes a top frontal excitation and emission optical configuration instead of the more widely used right-angle excitation/emission configuration. As discussed in Sect. 2.3., this has proved to be successful in its initial evaluations with respect to both sensitivity and versatility in methodology development.

When transferring the fluorescence monitoring technology to the miniature Fast Analyzer, the mechanical and physical constraints of the -55-

Qi Table 2.12. Precision of Dynamic Loading Versus Volume Injected

Conditions: Miniature Fast Analyzer, 620 nm; rotor speed, 30°0 rpm; injection device, 1-ml Hamilton Syringe with a 26-gauge needle having a flexible 0.010-in.-ID tip; for each volume mean, N = 16.

Standard Volume Measured Volume Deviation Injected (|il) Mean (|j.l) (Ml)

10 10. k 0.6

20 20.0 0.5

30 33.2 1.0

50 50.5 1.1 cL Rotor speed of 3000 rpm and introduction by 1-ml Hamilton Syringe with 26-gauge needle having a flexible 0.010-in,-ID tip. -389-

Table 2.13. Comparison of Analytical Results Obtained with Either Discrete8, or Dynamic*5 Loading of Reagents

Sample Enzyme Activity (i.U. liter"1 min"1, 30°C Code ALP° LDH-Ld SGOT6 Number Discrete Dynamic Discrete Dynamic Discrete Itynamic

1 30.6 31.9 6 2.8 59.9 11.4 10.8 2 210.2 201.5 217.6 213.2 41.9 3 37.6 38.9 70.8 67.8 5.7 5.5 4 38.9 39.8 67.9 63.3 6.8 5.2 5 32.5 33.1 62.2 58.9 10.1 8.0 6 58.7 57-7 70.1 76.5 12.9 12.0 7 27.7 27.0 100.0 86.8 11.0 9.6 8 ^3-9 ^3-7 97.2 95.9 10.2 8.3 9 ^2.3 41.6 87.8 87.3 25.0 23.5 10 52.3 52.0 84.4 83.0 14.2 12.6 11 58.7 58.5 93.^ 96.8 12.8 8.3 12 56.0 56.9 56.8 53.9 6.7 5.6 13 38.7 38.4 66.6 66.1 6.7 4.0 i4 37.2 38.3 68.0 67.5 4.6 3.8 15 30.0 30.5 61.3 57.7 10.3 8.9 16 210.1 202.9 216.2 220.7 42.2 b2.3 Average 62.8 62.1 92.7 91.0 14.6 13.1 a Volumes loaded: Discrete 1.0 p.1 sample + 50 p.1 diluent Discrete 20 [il reagent + 50 diluent

Volumes loaded: Discrete 10 |j.l sample + 50 [il diluent Dynamic 60 \xl reagent (diluted 1:3)

CConditions: 400 nm; 30°C; dynamic loading speed, 3000 rpm; delay interval = 10 sec; observation interval = 10 sec; number of observations = 20. cL Conditions: 3^0 nm; 30°C; other conditions identical to those in footnote c.

Q Conditions: 34-0 nm; 30°C; other conditions identical to those in footnote c except observation interval = 15 sec. -57-

15-place GeMSAEC Fast Analyzer can "be circumvented. Since the rotor of the miniature Fast Analyzer is small, portable, and self-contained, at least two approaches can be made: (l) adaptation of right-angle excitation/emission using specially designed rotors with end-windows in each cuvet normal to the standard top rotor; and (2) frontal excitation/ emission as previously developed for the 15-place GeMSAEC Fast Analyzer.

The former can also be used for monitoring light scattering. Two such systems have been built for the miniature Fast Analyzer. Each is currently undergoing evaluation for sensitivity, versatility, and practicality as it relates to operation of the system in both fluorometric and photometric modes and as it relates to the maintenance of the compactness and mobility of the miniature analyzer.

Right-Angle Excitation Rotor. - A right-angle rotor and detection system developed for the miniature analyzer is shown in Fig. 2.20. This rotor can be used for both fluorescence and light-scattering measurements.

The rotor was designed specifically for lo.v-level fluorescence measurements where a right-angle rotor would have the advantage of reduced reflection and light scatter over the frontal system. Also, the top window and the emission detector can be positioned to minimize problems associated with the inner filter effect. The emission volume element (as seen by the detector) was placed at the outer (excitation window side) edge of the cuvet in contrast to centering it in the middle of the cuvet window. The rotor, as seen in Fig. 2.21, has solid cylinders of ultraviolet transmitting (UVT) acrylic plastic windows glued into the end of each cuvet, allowing excitation to occur directly through the end window and normal to the overhead detection system. The emission signal (90° PHOTO NO 1968-738

PHOTOMULTIPLIER TUBE

FOCUSING LENS ROTOR (NOTE WINDOWS LIGHT PIPE ON OUTER WALL OF IN POSITION ROTOR ) FOR RIGHT-ANGLE EXCITATION

Fig. 2.20. Optical Configuration of a Monitoring System to Measure Light Scattering with a Miniature Fast Analyzer. -59-

PHOTO NO. 1969-73 A

Fig. 2.21. Rotor Used to Measure Light Scattering with a Miniature

Fast Analyzer. (Note that plastic windows have been sealed into the external wall of each cuvet.) «6o- to the excitation source) is monitored by a photomultiplier tube and filter system positioned above the rotor. This arrangement "was developed to use an existing 1P28 photomultiplier tube which would allow evaluation of the performance with respect to the frontal system on the 15-place analyzer.

A semiquantitative evaluation of the rotor and detection sensitivity was made using solutions of sodium fluorescein in 0.01 N NaOH. Sodium fluorescein solutions concentrations of 0.2, 0.4, 0.6, 0.8, 1.0, 2.0, 4, —10 8 6, 8, and 10 ng/ml (5.0 x 10" mole/liter to 2.5 x 10" mole/liter) were prepared and placed in the rotor, with the first five solutions being pipetted in duplicate. The excitation source was a 150-W xenon arc lamp coupled to a Bausch and Lomb high-intensity gradient monochromator. The analog emission signals of these samples are shown in Fig. 2.22. The blank emission for 0.01 N NaOH can be seen (cuvets l6 and 17) to be very small as compared to cuvets 1 and 2, which contain 0.2 ng/ml (5.0 x lo"1^ mole/liter) sodium fluorescein. This is indicative of a detection limit in the range of 1.0 x 10"10 mole/liter. At this level, adsorption of the fluorescence material onto walls of the dilution vials becomes a limiting factor (along with blank emission) due to many factors, including Raman emission and filter leakage of scattered light.

Surface Fluorescence. — A second approach to fluorescence measurements on the miniature Fast Analyzer has been to adapt a surface fluorescence optical system to the analyzer. The advantage of such an approach is twofold. First, no redesign of rotors is required, and the same rotors can be used for fluorescence as are currently used for photometric analyses.

Second, the system would use the existing photomultiplier tube of the -6l-

ORNL DWG. 73-8635RI

0 V

c/) l-

-10 V-

2 4 6 8 10 12 14 16 II I I I II I M 11 I I I II '—i—' V W S-1 0.1 0.2 0.4 Ofi 0.8 I 2 4 6 8 BLANK

SODIUM FLUORESCEIN CONCENTRATION (ng /ml in 0.01 N NaOH)

Fig. 2.22. Emission Signals Obtained from the Miniature Fast Analyzer Using a Right-Angle Fluorescence Monitor. analyzer and. with the successful adaptation of a miniature high-intensity lamp source, the compactness and mobility of the present analyzer would be maintained. A surface fluorescence optical system has been developed for the analyzer. The primary component of the system is a mirror assembly that fits in a slot cut into the end of the filter wedge. This unit is shown in Fig. 2.23. The mirror is set at an angle of 30° to the filter holder plane, which results in the direction of excitation light at an angle of incidence of 6o° into the bottom window of the rotating rotor. The mirror extends into the first cylindrical filter holder to permit direct! on of the exciting light to the maximum possible depth in the cuvet as the cuvet is centered above the aperture of the photomultiplier tube. An objective lens is used to focus the emission signal, and 0.5-in.- diam barrier filters of the desired wavelength are used to isolate the emission signal. A quartz fiber optical bundle has been employed for direction of excitation light to the spinning rotor irom a xenon lamp source. The -8 current sensitivity of the system is about 10 ng/ml or 2.5 x 10 mole/ liter of sodium fluorescein in 0.01 K NaOH. The sensitivity of the analyzer can be improved by removing the current l/4-in. aperture in the photomultiplier holder and replacing it with a variable aperture, which would allow most of the emission signal to fall on the photomultiplier tube window in the fluorescence mode of operation, but would provide a l/8-in. aperture for the photometric mode.

A suitable small light source that has sufficient intensity in the uv and visible to permit fluorescence measurements and photometric measurements is under consideration. The mirror assembly has made it PHOTO NO. 1974-73 A

CUVET APERTURES ENLARGED FOR SURFACE EXCITATION

1 ON CO I MIRROR ASSEMBLY

LIGHT PIPE FOCUSING LENS

Fig. 2.23- Surface Fluorescence Optical System for Use with the

Miniature Fast Analyzer. -6k- possible to eliminate the fiber optical bundle completely and mount the

lamp source close to the mirror system.

2.1.7 Light Scattering Monitor

The right-angle rotor was developed specifically for low-level

fluorescence, particularly as such measurements relate to competitive

binains fluoroimmunoassays. However, It can also be used as a monitor

for light scattering. Consequently, one of the important and practical

uses of the rotor will be for specific protein analyses. The monitoring

of the intensity of light scattering as a function of antigen concentra-

tion after aggregate equilibration has become of interest recently for

so-called specific protein analyses. Serum IgG, IgA, IgM, a-1-trypsin

inhibitor, transferin, and C'3 complement factor are among some ko

specific serum proteins that can be assayed. These analyses are based

on the aggregation of immunoglobulins with their respective antibodies.

The aggregation can be measured by determining the change in light

scattering either during the aggregate formation (kinetic) or after

aggregate formation (equilibrium).

Currently, flowing stream analyzers are used to perform such speci-

fic protein analyses. This technique is relatively rapid and sensitive,

and brings to the clinician and clinical chemist rapid techniques for

looking at specific serum proteins which were formerly not possible to ana

lyze routinely due to their low concentration in serum ((j.g/ml to mg/ml).

One of the advantages of the flowing-stream approach is its ability

to determine antigen excess. This is accomplished simply by noting

those peaks on the strip-chart recorder that are inverted. If one looks -65- at scatter intensity as a function of antigen concentration at a constant antibody titer (concentration), the intensity increases in a nonlinear ma.nner to a plateau, and then is seen to again decrease and to return to near zero intensity at high antigen concentration. The problem is to determine, for any given sample intensity, which side of the intensity maximum is being measured. This is determined in manual procedures by running two samples, one more diluted than the other. If the resulting scatter intensity of the more dilute sample is less than the other sample, the antigen concentration must be to the left of (less than) that concentration resulting in the intensity maximum. In flowing-stream systems, concentration gradients of antigen occur in a flowing stream of constant antibody concentration. If the antigen is in excess, the scatter intensity along the gradient will increase until antigen excess occurs; the intensity will then decrease, giving the appearance of an inverted neak-

A disadvantage of the flowing-stream analyzer is its relatively large reagent consumption. As an example, flow rates of 0.3 ml/min of diluted antibody (Ab) are common. For a 1:25 dilution of Ab, this results in approximately 0.75 to 1.0 ml of Ab consumption per hour of operation for the analysis of 60 to 100 samples. Thus, a flowing-stream analyzer has a relatively large consumption of antibody, which is a limited and expensive biological reagent.

Because of its small volume requirements and light-scattering monitoring capabilities, the miniature Fast Analyzer with the right- angle rotor is being developed to perform specific protein analyses. In an initial study, 100 \il of antihuman IgG diluted 1:25 with 0.9$ NaCl solution was pipetted in reagent cavities 1 through 17 of the right-angle rotor. Fifty microliters each of diluent was added to sample cavities 1 and 2; 50 |il of human IgG (2.73 mg/ml) diluted 1:1000 with 0.9% NaCl solution was added to sample cavities 3 through 6; 50 \il of human IgG (6.1 mg/ml) diluted 1:1000 was added to sample cavities 7 through 10; 50 p.1 of human IgG (11.5 mg/ml) diluted 1:1000 was added to sample cavities 11 through lU; and samples containing 23.0 mg/ml IgG diluted 1:1000 were added to sample cavities 15 through 17. Excitation was set at UlO nm, and the rotor containing the IgG samples and antibody was accelerated to 500 rpm. A photograph of the analog signal was taken immediately

(Tq), and a. similar photograph was taken at 16 min. The results shown in Fig, 2.2k indicate that increasing intensity was obtained with increasing concentrations of IgG at 16 min. The system is apparently very sensitive in that it can detect microgram-per-milliliter quantities of IgG and it requires the use of only h (j.1 of antibody per test (100 vi 1 of a 1:25 dilution). This volume of antibody can be reduced further by both decreasing the reaction volume and using more specific Ab.

The quantitation of the IgG by light scattering is a somewhat unique approach in which the rate of IgG/anti-IgG aggregation is monitored (Fig. 2.25). The change in light scattering intensity (I °) as a function of IgG concentration is shown in Fig. 2.26. These data indicate that an assay based on a nonlinear fixed-time approach could provide direct kinetic determination of IgG concentration in less than 3 rain from the initiation of the reaction. The time required for equilibrium light-scatter procedures is approximately 20 min for IgG and longer for other assays, while a time of 18 hr is generally used for double immuno- diffusion procedures which are commonly employed. In regard to the -b7-

ORNL DWG. 73-8636

0- (a) T= 0 min

10 V- 0 - (b) T = 16 min

-10 V—

INCREASING CONCENTRATION OF IgG IN CONSTANT ANTI IgG

Fig. 2.2b. Light Scattering from IgG—Anti-IgG as Determined on the Miniature Fast Analyzer. -68-

0RNL-DW6 73-4321 1600

1400

1200

>- h- cn 1000 Ui

£ 800 I-

400

200

4 6 8 TIME (min)

Fig. 2.25. The Interaction of Human IgG—Anti-IgG vs Time as a

Function of IgG Concentration. -69-

ORNL-DWG 73-2056

Fig. 2.26. Fixed-Time Rate of Change of IgG-Anti-IgG Interaction as a Function of Time. -70- question of antigen excess, there is some literature evidence -which suggests that the initial rate of formation of antigen/antibody complex as monitored by light scattering tends to reach a constant value at higher antigen concentration, and the rate does not decrease at antigen excess. Therefore, this type of data could be used to establish a cut- off value for the acceptable rate, and any samples that had values appearing at this limiting rate would be diluted and analyzed again.

2.1.8 Portable Data Printer

A portable digital Data Printer is currently being developed for use with the miniature and portable Centrifugal Fast Analyzers. It will print instantaneous absorbance values of each of 17 cuvets on thermal-sensitive paper in response to an external print command signal. The electronic package, which has been fabricated and is currently being tested, utilizes a logarithmic amplifier to convert the transmission pulses to absorbance data pulses. These absorbance data pulses are then digitized by a 12-bit A/D converter, and the resultant digital data are stored in a 38^-bit random-access memory arranged as 32 words of 12 bits each.

After the data are accumulated in memory, printing begins. An 8- digit printer utilizing a 7-segment thermal print-head and thermal- sensitive paper is being used as the recording device. It prints the

2-digit cuvet number, followed by a blank space (or an 8 if the reference cuvet 1 is dirty\ a decimal point, then a U-digit absorbance value at the rate of 2 lines per second.

This first model has been designed to record the absorbance in the range of from 0 to O.8192 utilizing the total count capacity of U096 -71- counts of the 12-bit A./D converter. This is somewhat less capacity than

is desirable, but it will allow the techniques for storage and data retrieval to be tested. This will lead to the use of a more sophisticated

digital calculator integrated circuit as the central data processor in the next model.

2.2 Portable Fast Analyzer

C. D. Scott and L. H. Thacker

A portable fast analyzer system has been developed, and a working

prototype has been fabricated and tested. A U-in. cubical case contains

the system power supply (a Ni-Cd battery pack), rotor drive motor, a

miniature photomultiplier (PM) tube, a converter to develop the required

high voltage for the PM tube, and optics to direct the light from trans-

mission samples to the PM tube. Controls provide rotor speed and

PM voltage adjustments, and allow the PM tube to be rotated to positions

for sensing either transmitted light or emitted fluorescence. The 12-W

quartz-halogen source lamp is powered by the Ki-Cd battery pack, but is

externally mounted for improved heat dissipation.

On top of the ^-in. cubical case are the 2-l/U—in.-diam, 8-cuvet

rotor and a mount for the distal end of a bifurcated, half-quartz, half-

glass fiber optic illuminator and fluorescence detection system. The

proximal end of the quartz-fiber arm is inserted in the source lamp

holder, which also contains an interference filter for selecting the

incident wavelength for each type of analysis. The proximal end of the

glass fiber arm is returned to an adapter, on top of the instrument case,

which contains an interference filter for isolating emitted fluorescent -72- wavelengths and provides an aperture to the photomultiplier tube. Output coaxial connectors transmit light measurement signals and rotor and cuvet pulses.

A series of tests was performed to evaluate the capabilities of this prototype to make both absorbance and fluorescence measurements. Absor- bance tests were made with solutions of blue dextran in water (l/2$ formaldehyde as preservative) with known absorbances at 620 nm. The

corresponding range of absorbances for the solutions in the rotor cuvets provided a comparison (Fig. 2.27). The results indicate reasonable linearity up to an absorbance of about 0.6.

Fluorescence measurements were made using solutions of fluorescein in 0.01 N NaOH. The excitation light was defined by a U80-nm cut-off interference filter, and a 520-nm cut-on barrier filter was used to pass emitted light to the PM tube. Because of the limited PM tube voltage available in the battery-operated prototype, concentrations below 10

Hg/ml were not detectable; however, the results indicate essentially

linear response in the concentration range of 10-100 (ig/ml (Fig. 2.28).

2.3 Fluorescence Fast Analyzer

C. A. Burtis, T. 0. Tiffany, and W. F. Johnson

During the past several months, development has progressed on a fluorescence monitoring system (see Sect. 2.1.6) for the Fast Analyzer10

The goal has been to obtain an instrument that could be used as either a multicuvet spectrophotometer or a fluorometer through a single intercon- version of the optical system. This goal has been achieved. The

following sections demonstrate some of the ideas concerning the development and use of such a multicuvet fluorometer. In addition, the same -73-

ORNL DWG 73-8638 " 1— 1 f 1— I - T • I -J—' " O 1.0 -

0.9 - o / O CM <0 0.8 - mm

0.7 - / 0 - uJ o 0.6 - - < 0 0.5 — - CD cr /r o 0.4 - - CO CD / < 0.3 - / - QC LJ 0.2 - - >Nl /o 0.1 < 0 1 1 1 I I 1 1 1 0 0.1 0.2 03 0.4 0.5 0.6 0.7 0.8 0.9 1.0 ABSORBANCE OF TEST SOLUTIONS AT 620nm

Fig. 2.27. Absorbance Measurements Using the Portable Fast Analyzer. -7U-

ORNL DWG -73- 8637 T T T T

L EXCITATION FILTER: 480 nm CUT-OFF BARRIER FILTER'. 520 nm CUT-ON

± _L 0 10 20 50 100 FLUORESCEIN (/ig/ml)

Fig. 2.28- Response of the Portable Fast Analyzer to Solutions of

Fluorescein. -75- instrument is "being used for rapid spectrophotometry kinetic studies of the conversion of iodine to various products in concentrated nitric acid, further demonstrating the attainment of the stated goal and the versatility of such a Fast Analyzer.

2.3.I Instrumentation

A basic objective in the development of a fluorometric Fast Analyzer is measurement of solute concentration over a broad concentration range -5 -11 (e.g., from 10 to 10 M). Two major problems in spanning this concentration range are: (l) obtaining sufficient sensitivity of the optical system at the lower range of concentrations through minimization of scattered light and nonemitted stray light; and (2) minimization of the inner filter effect. The basic criterion for the development of the fluorometric Fast Analyzer to be used for concentration measurements was established from considering the problem of the Inner filter effect.

Minimization of this effect by use of frontal or surfac12e fluorescence measurements has been discussed by Brand and Wilholt.

Several decisions concerning instrumental parameters for fluorescence concentration measurements were incorporated into an existing Fast

Analyzer to give the instrument shown in Fig. 2.29. Excitation is achieved with a 150-W xenon-mercury source coupled with a small monochromator and a stabilized dc power supply (Schoeffel Instrument

Company, Westwood, N. J.). A quartz optical fiber bundle U mm in diameter and 10 cm long (Schott Optical Company, Duryea, Penn.) provides the optical path between the monochromator and the cuvet surface. Exci- tation is at an angle of incidence of 60°, and the emitted fluorescence ORNL PHOTO NO. 0496A-7

LIGHT SOURCE Si flj'lisysSH& i S mk I'-,.- PHOTOMULTIPLIER

LIGHT PIPE *

ft FOCUSING LENS

4 ROTOR tA'N

Fig. 2.29. Centrifugal Fast Analyzer for Surface Fluorescence Monitoring. -77- signal is detected normal to the rotating cuvet surface. Barrier filters (Baird Atomic, Inc., Cambridge, Mass.), cut-on/cut-off filters

(Ditric Company, Marlboro, Mass.), or sharp-cut filters are used in place of an emission monochromator to obtain greater sensitivity. This optical configuration permits the use of the analyzer for fluorescence or absorbance measurements after shifting the optical fiber to a mirror assembly which directs the source intensity through the rotating cuvets.

The emission filters have to be removed for photometric measurements.

An operational amplifier with a gain of 10 or 100 provides emitted signal amplification, which results in a reduction in photomultiplier voltage, and hence instrumental noise. The amplified signal is fed directly into a DEC PDP-8E computer, which permits the same type of data reduction that is available for spectrophotometric measurements.

2.3.2 Fluorescence Sensitivity and Fluorescence Tracers

The sensitivity of the fluorometric Fast Analyzer has been investigated using solutions of sodium fluorescein in the nanogram-per-milliliter concentration range. The excitation monochrornator was set at U80 nm. A.

500-nm cut-off filter purchased from the Ditric Filter Company (Marlboro,

Mass.) was placed at the entrance of the quartz fiber optic to provide a. more effective wavelength cut-off and to eliminate second-order excitation energy from the monochromator. A barrier filter, 500 nm to 600 nm, was used on the excitation side in conjunction with a 530-fl-m sharp-cut filter to further isolate the emission signal. Solutions of sodium fluorescein in the concentration range of 0.2 to 8 ng/ml were analyzed.

The detection limit, defined as the concentration of sodium fluorescein -78- producing a fluorescence intensity signal twice the background signal, was about Uo pg/ml, with a useful analytical range above 0.2 ng/ml for

sodium fluorescein. This detection limit is indicative of, but not

directly equivalent to, the detection limit of fluorescein-labeled

compounds because of the slight loss of intensity of fluorescein upon 13 conjugation with protein. However, it does indicate that the presence

of a fluorescence label at a concentration of approximately 100 pg/ml

can be detected.

Fluorescein-labeled insulin was prepared to demonstrate the sensi-

tivity of the analyzer. Extensive information is available concerning

the preparation, biological activity, and immunoactivity of fluorescein-

labeled insulin.An estimation of fluorescein isothiocyanate

(FITC) insulin concentration was made from spectrophotometric data, and

a FITC-insulin solution of the concentration U.O vig/ml was prepared in

0.1 M Tris buffer, pH 7.U, containing 0.1$> bovine albumin. The amount of

fluorescein contained in a 1-ng/ml preparation of FITC-insulin, assuming 1

mole of label per mole of insulin, Is approximately 70 pg/ml. From these

considerations, it is evident that one should be able to detect FITC-

insulin in the concentration range 1 to 10 ng/ml, and Fig. 2.30 shows

this to be the case. The level of insulin in plasma at the upper limit 11 of normal is on the order of 25 M- units/ml or approximately 1 ng/ml.

The data indicate that we are approaching a detection limit for FITC-

labeled insulin that will make it suitable for use as a tracer In a

fluoroimmuno competitive binding assay for Insulin.

2.3.3 Fluorescence Polarization Measurements

Fluorescence polarization measurements can be used in studying -79-

0RNL-DW6 73-2068

FITC-INSULIN (ng/ml)

Fig. 2.30. Relative Fluorescence Intensity vs Concentration for

FITC-Insulin. -uu- antigen-antibody reactions. The interesting aspect of the use of this

type of measurement is the potential it offers for the direct measurement of antigen or antibody concentration without resorting to a variety of procedures for separating free antigen from bound antigen. It is felt that fluorescence polarization measurements are possible by using a

fluorometric Fast Analyzer.

To demonstrate fluorescence polarization measurements with the Fast

Analyzer, a. series of solutions of sodium fluorescein (0.5 p.g/ml) was prepared to contain increasing concentrations of ethylene glycol. Film polarizers were obtained and adapted so that the polarizer on the exci-

tation side was fixed but the polarizer on the emission side of the

emission analyzer could be rotated through an angle of 90°. Fluorescence polarization measurements were made on all solutions in the spinning

rotor by setting the emission analyzer (polarized) in a perpendicular position, taking a series of readings for digital averaging purposes, rotating the polarizer 90°, and then taking a second set of intensity

readings. These data were then used to calculate polarization, p, as

follows:

I - I J L (?\ II x where

I|| = light intensity parallel to light source, Ij^ = light intensity perpendicular to light source.

A plot of polarization vs ethylene glycol concentration for sodium

fluorescein in 0.01 N NaOH is shown in Fig. 2.31. A plot of reciprocal -81-

ORNL-DWG 73-2053Rl

0 20 40 60 80 100 PERCENT ETHYLENE GLYCOL

Fig. 2.31. Fluorescence Polarization Measurements of Sodium

Fluorescein in Increasing Concentrations of Ethylene Glycol. -82- polarization (l/p) vs reciprocal viscosity (T/T]) for solutions of sodium

-6 fluorescein (0.5 x 10 g/ml) in increasing concentrations of glycerol,

•where

7] = viscosity

T = absolute temperature, is shown in Fig. 2.32. These data show that fluorescence polarization measurements can be made using the fluorometric Fast Analyzer. Duplicate samples produced replicate polarization values close to ±0.001. Efforts are being made to upgrade the quality of polarizers and to establish correction for artifacts in polarization measurements so that the value of p for fluorescein in glycerol will8 l more closely approach the calculated value of O.Ml reported by Perrin. The placement of the analyzer polarizers at an angle other than the customary 90° should not affect the measurement of fluorescence polarization.1 9

2.3.U Fluorescence Referencing

Fluorescence measurements, like some absorption measurements, will be more precise if internal referencing can be used. Due to variations in excitation intensity, to detector sensitivity, and to excitation volume elements, it is difficult to obtain the relationship of absolute emission intensity to molar absorptivity for any given fluorescence compound on any particular fluorescence instrument. Furthermore, since fluorescence measurements involve molecules in their excited state, these measurements become more dependent on buffers, pH, ionic strength, temperature, and other related solvent properties. Therefore, the

simplest approach to referencing fluorescence measurement is to relate -83-

ORNL-DWG 73-2309RI 22 / 20

16 / • 14

0 8 f

6 / •4/ 4 / •/ 2

0 3 4 5 6 8 10 T 1 4 /fl (deg. poise" xlO" )

Fig- 2.32. Reciprocal Polarization vs Reciprocal Viscosity for

Solutions of Sodium Fluorescein in Increasing Concentrations of Glycerol as Determined Simultaneously on the Fast Analyzer. -Qk- the relative intensity of fluorescence of the unknown sample to the relative fluorescence intensity of a known concentration of a fluorescent compound. This fluorescent compound can "be the monitored component of a particular reaction measured under identical reaction conditions, or it can be a recognized fluorescent reference compound whose intensity is known relative to that of the compounds of interest.

The multicuvet feature of the Fast Analyzer makes it possible to reference dynamically by using one cuvet as a nonfluorescence blank, a second cuvet for the fluorescent reference compound, and the remaining cuvets for determining enzyme activity or compound concentration by the referenced fluorescence measurement. During each rotation of the cuvet rotor, the relative intensities of the blank, the reference cuvet, and the sample cuvets are read and then averaged over successive rotations.

The data for each cuvet for each reading inteval are stored, and the normalized intensity for each cuvet at each reading interval is calculated by obtaining the ratio of sample intensity to the corrected reference intensity:

"^sample I - I reference blank

In this manner, the sample intensity is obtained in units of relative intensity of sample per unit of relative intensity of reference compound; and since the concentration of the reference compound is known, the

sample concentration or enzyme activity can be readily obtained. Use of

the dynamic referencing technique not only provides a means of rapidly

calculating concentration but also serves as a double-beam measurement by continuously reading a fluorescent standard and normalizing over -85- long-term instrumental drift and intensity changes.

Two forms of dynamic referencing have been adapted for use with the fluorometric Fast Analyzer. The first, called direct substrate referencing, involves the use of a known concentration of the substrate or product being measured during the analysis. The second, called dynamic ratio referencing, involves the use of a reference compound, such as quinine bisulfate in 0.1 N HgSC^, with determination of the ratio of the relative intensity of this compound to that of the monitored species of the reaction.

The latter method is used when the reaction product is not suitable for direct substrate referencing. An example of this case would be the use of NADH solutions, which are not stable over long periods of time.

2.3.5 Calculation of Enzyme Activity and Substrate Concentration

The use of dynamic referencing techniques makes it possible to readily calculate enzyme activity from the normalized intensity data. A. computer program has "been written in FOCAL language to enable the analyst to obtain enzyme activities in international units using either direct or ratio referencing. This program uses a linear regression analysis of the data for each cuvet in order to obtain a normalized change in intensity per minute. Enzyme activity, in international units, is calculated from the following expression for direct referencing:

(C )(1000)(V ) RR RR Enzyme Activity (I.U.) = — where

C = reference substrate or product concentration ((ag/ml)3 Rn MW = molecular weight of the reference compound, -86-

V = total reaction volume, .h. V = volume of sample used in the reaction, b Upon initiation of the computer program, the analyst is queried as to -which referencing method is to "be employed. If ratio referencing is selected, a value for the ratio is entered into the program and the enzyme activity is calculated from the following expression:

(CR)(1000)(V) , Enzyme Activity (x.u.) - W^Xy (l^J, - W) < where A^ is the ratio of the relative intensity per micromole per liter of substrate.

A similar expression can be obtained for a fluorescence substrate factor in which the intensity of the reference compound is substituted for the molar absorptivity:

(m )(c )(v ) (S.F.) = r-^ , (6) 4 (l0 )(MWr)(Vg)(AR) where S.F. = substrate multiplication factor, MW = the molecular weight of the substrate (e.g., glucose), s MW^ = the molecular weight of the reference compound,

A^ = the micromolar intensity ratio of NADH to quinine bisulfate.

Use of the factor 10^ results in units of mg/lOO when the substrate factor is multiplied by the normalized intensity change, and the substrate concentration is then given by:

Substrate Concentration = (S.F.) C — ——— \ (7) reference "" blank -87-

2.3.6 Effectiveness of Dynamic Referencing

The effectiveness of dynamic referencing (ratio or substrate) can

"be evaluated in several ways. Of these, we selected the following: (l) assay of enzyme of known activity at two different levels in at least two runs made at each level; (2) comparison of the activity of an enzyme, obtained by ratio referencing to that obtained by direct substrate referencing; and (3) calculation of substrate concentration through multiplication of the relative intensity data by a substrate factor.

LDH Activity Using Ratio Referencing. — An LDH assay kit using the conversion of lactate to pyruvate (Calbiochem, Los Angeles, Calif.) was prepared according to the manufacturer's recommendations. Control Sera

I and II with measured LDH activities of 25 I.U. and 79 I.U. respectively, at 25°C were prepared and diluted 1:10 using 0.9$ NaCl. Quinine sulfate

(0.2 ng/ml in 0.1 N HgSO^) was used as the fluorescence reference, and the blank solution was 0.1 N HgSO^. The control serum and reagents were pipetted, added to each of the remaining cuvet cells, and the analyses were made. Under the conditions of the assay, 5 M-l of sample was

analyzed in a total reaction volume of 550 [il. The results of four

analyses (two levels of control sera in duplicate) are shown in Table 2.1^-.

The analytical precision was represented by a coefficient of variation

(C.V.) of 3 "to 5$. Analysis time was of the order of 3 to 7 min, and

the temperature was maintained in the range of 2^ to 25°C«

Alkaline Phosphatase Activity by Ratio and Direct Referencing

Techniques.— A second means of checking the referencing method was to

compare the activity measured for a compound using the ratio procedure

with that obtained by direct referencing with the substrate prepared -88-

Ta"ble 2.lU. Fluorescence LDH Activity Assay Using Quinine Sulfate as a Reference Standard

Range of LDH Standardized Activity Temperature Result 3 Run Sample ' (I.U.) (°c) (I.U.)

1 Control II 79 ± 5 25 78

2 Control II 79 ± 5 25 81

3 Control I 25 ± k 2k.5 23

k Control I 25 ± h 25 25

Reaction conditions: 5 i~il of sample in a total assay volume of 550 lil; ratio - 0.032; reference standard concentration = 0.2 p.g of quinine sulfate per milliliter in 0.1 N HgSO^. under conditions of the reaction. Since the ratio of the intensity of

^-methylumbelliferone to the intensity of quinine sulfate has "been determined, U-methylumbelliferone can "be purified and used. Because b-methylumbelliferone itself represents the product of an enzyme hydrolysis, it can "be included in the assay "buffer and used as the referencing material. A serum sample from a patient suffering from acute hepatitis was used in the assay, and various dilutions of substrate were used, both to establish a comparison between the two techniques of referencing and also to determine an optimum substrate for the assay. Table 2.15 lists the enzyme activity results as a function of U-methylumbelliferone concentration. The correlation is good and shows that similar results can be obtained using either approach.

Glucose Concentration Using Ratio Referencing. — It has been previously demonstrated that spectrophotometric enzymic end-point substrate analysis can be used to obtain reliable glucose concentrations by multi-

* plication of the absorbance change of the reaction by a substrate factor.'

Similar assays can be performed by using the fluorometric Fast Analyzer and ratio referencing. Serum glucose standards representing concentrations of 6b, 80, 107, and 160 mg/dl, at a 1:20 dilution, were further diluted 1:5 with 0.9$ NaCl. Quinine sulfate in 0.1 N HgSO^ was again used as the fluorescence reference and 0.1 N HgSO^ "was the blank reference. The remaining reagent wells (b through 15) received 500-ial aliquots of hexokinase glucose reagent (Calbiochem). Four serum glucose

standards were introduced in triplicate into the remaining cuvet wells.

The reaction was initiated, and Intensity readings were obtained starting

at 5 sec, and at 20-sec intervals thereafter, for a total reaction time -90-

a Table 2.15. Comparison of Alkaline Phosphatase Activities Obtained by Direct and Ratio Referencing

Activity (I.U. at 25°C) Concentration Direct ^ Ratio c (m M) Referencing Referencing

0.1 7.2 7.2

0.2 13.6 13.7

O.U 23.9 25.5

0.6 31.6 3lA

0.8 39-3 39-9 1.0 b3.b ^5.8

3.0 67.1 67.6

SL Each sample consisted of 5 M-l of enzyme (alkaline phosphatase- hepatic) in 550 i±l of 0.3 M MAP buffer, pH 10.5.

Against 0.2 ng/ml U-methylumbelliferone in 0.3 M MAP buffer, pH 10.5.

°Ratio = 9'9;' reference standard -was quinine sulfate, U- (ig/ml in 0.1 N H2S0^. -Ql- of 22p sec. A substrate factor of 555 "was calculated (see Table 2.16)

Again, the results were satisfactory and further demonstrated the potential of direct, dynamic, automatic fluorescence referencing. (The actual reaction sample volume for each of these glucose assays repre-

sents 0.5 ial of sample, a concentration that is fivefold more dilute than was used in the spectrophotometry kinetic glucose determinations.)

2.3.7 Development of Enzyme Analyses

The main advantages of fluorometric substrate and enzyme activity assays are sensitivity (e.g., they are 100 to 1000 times more sensitive than photometric measurements) and the availability of a number of fluorogenic compounds that permit direct equilibrium, or kinetic assay of substances which could not be previously determined (or at least not readily determined) by photometric methods. One example of the latter

is the use of the fluorogenic substrate h-methylumbelliferyl phosphate

for the rapid, direct, kinetic assay of acid phosphatase in acidic buffer. Acid phosphatase assays have been carried out photometrically by first allowing the reaction to proceed in acidic buffer and then

changing the pH rapidly to generate the chromophore, or by coupling a

product with a diazo reagent to produce a colored compound that could be related in some manner to the activity of enzyme present. The former

procedure is not a directly monitored analysis, while the latter suffers

from the presence of a lag phase and the side reactions of other com-

.runds in the serum reacting with the coupling diazo compound. To

circumvent both of these problems, a fluorometric acid phosphatase and

also an alkaline phosphatase assay have been developed for use with the

fluorometric Fast Analyzer. -92-

Table 2.l6. Tabulation of Equilibrium Hexokinase Data Obtained Using Fluorescence Ratio Referencing

Glucose Calculated Concentration^ Glucose Concentration Sample3, (mg/dl) (mg/dl)

1 6k 62

2 80 81

3 107 109

k 160 158 diluted serum control.

Glucose concentration determined spectrophotometrically (see ref. 20). -93-

p-Nitrophenyl phosphate is commonly used in the spectrophotometric assay of acid and alkaline phosphatases. The reaction in basic solution results in the formation of the p-nitrophenylate ion, "which can be detected by colorimetry. However, at pH values significantly below the pK of p-nitrophenol, the product is colorless. Thus, when using the

substrate p-nitrophenylphosphate for acid phosphatase assays in the pH range of k to 6, it is necessary to continue the reaction for several minutes at that pH and then adjust the pH to 9 °r greater to obtain the activity. This is the usual manner in which acid phosphatase assays have been carried out spectrophotometrically. Obviously, the procedure is not direct and is subject to problems and errors associated with an inability to follow the reaction directly. Fortunately, the fluorescence intensity of the hydrolysis product of the substrate, ^-methylumbellif eryl phosphate in the pH range from k to 6 is 300 to 1000 times greater than the sub-

strate; therefore, it can be used as the basis of a direct kinetic fluorometric enzyme assay for acid phosphatase.

Another raaction of interest is represented by monitoring the change * • in normalized fluorescence intensity of methylumbelliferyl phosphate

\ from pH 2.0 to pH 7.0 (Fig. £-33). What is observed is a fluorometric

titration curve representing the dissociation of a proton from the

second hydroxyl group of the phosphate moiety. The pK^ of the ionization

can be calculated from these data. The intensity of the product l+-methylumbelliferone does not change throughout the pH range to 6.0, as is

shown in Table 2.17.

Alkaline Phosphatase. — The use of k-methylumbelliferyl phosphate

as a substrate for alkaline phosphatase was first reported by Fernley ORNL-DWG 73-2551 1.1

1.0

0.9

0.8

< / 0.7 o 7 0.6 oL_ at (J0> XI «c in3 ® 0.5

pKg= 5.2 AT 25°C

0.4

0.3 I /

0.2

0.1

6 7 8 10 11 pH

Fig. 2.33. Fluorometric Titration Curve of ^-Methylumbellif eryl Phosphate as a Function of pH. -95-

Table 2.17. Normalized Intensity of 1|—Methylumbelliferyl Phosphate and b-Met hylumbe lliferone in Acidic Buffer

Relative Relative Change Intensity of Intensity of in 3 a PH Substrate - Product Intensity R°

4.00 0.0021+7 5.060 5.058 2048

4.25 0.00269 5.060 5-057 1881

4.50 O.OO365 5.060 5.056 138.6

4.75 0.00485 5.060 5.055 10U3

5.00 0.00705 5.060 5.053 717

5.25 0.00982 5.060 5.050 515

5.50 0.01303 5.060 5.0V7 388

5.75 0.0149 5.060 5.045 339

6.00 O.OI63 5.060 5.044 310 aRatio of intensity of micromole/liter sample to micromole/liter quinine sulfate. bRati o of product and substrate intensities. -9b-

21 et al. Its use In a clinical procedure was first introduced "by 22 Cornish, Neale, and Posen. Contrary to the results obtained by the latter workers, we find the substrate reasonably stable in AMP and MAP buffers, and have chosen to look at the assay under similar conditions of the current state-of-the-art assay for alkaline phosphate, using

p-nitrophenyl phosphate in MAP buffer that is 0.01 M in MgClCi . Figure 2.34 is a plot of reaction velocity vs substrate concentration obtained from human intestinal alkaline phosphatase and liver alkaline phosphatase,, The activity-vs-substrate curves are similar in substrate optima to those obtained by spectrophotometric procedures using p-nitrophenyl phosphate in the same assay buffer. The linearity of enzyme activity as a function of ei.zyme dilutions is shown in Fig. 2.35. The data were obtained by diluting a human serum alkaline phosphate sample of elevated concentration (2^0 I.U./liter at 30°C) with 0.9% saline. Our initial investigations suggest that a fluorescence alkaline phosphatase assay can be set up for human serum alkaline phosphatase using 4-methylumbelliferyl phosphate at an assay concentration of 5 x 10 mole/liter, and 2- amino-2-methyl-l-propanol buffer (0.2 to 0.6 M), pH 10.5* A reaction volume of 550 p.1 was used for these assays in each case, with the sample volume equivalent to 5 m-1 . The assay time was in the range 3 to

5 min. The activity of normal human serum, using the fluorescence assay procedure at 25°C, is about 50$ of that obtained with the spectrophotometric procedure at 30°C. The analytical variation is represented by a

C.V. of 2 to 5°!o using the current prototype instrument.

Acid Phosphatase. — The development of a kinetic acid phosphatase assay for hiuiian prostatic acid phosphatase using 4-methylumbelliferyl -97-

ORNL-DWG 73-2063R2 140

120

| 100

o> 80 o E =1 LJ

|= 40

0 0 1 2 3 4 5 6 SUBSTRATE CONCENTRATION (moles/liter x 103)

Fig. 2.34. Alkaline Phosphatase Activity vs 4-Methylumbelliferyl

Phosphate Concentration from Two Different Enzyme Sources. -98-

ORNL-DWG 73-2056

ENZYME DILUTION (7o)

Fig. 2-35' Linearity of Alkaline Phosphatase Activity. -99- phosphate in an acidic buffer has not been previously reported. 23 Guilbault has assayed for potato acid phosphatase activity using umbelliferyl phosphate as substrate. This assay can compete most favorably with spectrophotometric methods that either require analysis at pH

5.0 and subsequent pH adjustment, or the use of a coupled dye assay procedure.

The substrate 4-methylumbelliferyl phosphate is only slightly fluorescent when compared with the fluorescence of 4-methylumbelliferone in the pK range 3.0 to 7-0. It is apparent from these initial observations concerning the relationship between the fluorescence intensities of the product and the substrate that a procedure for assaying for acid phosphate can be developed if the enzyme is active toward this substrate.

We have chosen an assay based on the use of serum controls containing human prostatic acid phosphatase.

In a review of the literature concerning human prostatic acid phos- 25 phatase, Tsuboi and Hudson reported that human prostatic acid phosphatase is inhibited by sulfhydryl reagents such as p-chloromercuribenzoate and by metal ions such as copper. Therefore, we have chosen to specify an assay buffer containing 0.05 M sodium acetate, 10 M EDTA, and 10

M (3-mercaptoethanol to stabilize the enzyme. Enzyme properties, including pH optimum, substrate optimum, enzyme stability, and inhibition of acid phosphatase by L-tartrate have been investigated.

Substrate Optimum. — Enzyme activity as a function of substrate concentration is shown in Fig. 2.36. From this figure, it can be seen that maximum reaction rate is attained at approximately b x 10 mole/ liter. The K of the enzyme for 4-methylumbelliferyl phosphate at pH ORNL-DWG 73-2054 R2

c E

4 CD {tz

1— 1 o < LxJ tr 0 Ot 23456789

SUBSTRATE CONCENTRATION (moles/liter x 105)

Fig. 2.36. Human Prostatic Acid Phosphatase Activity as a Function

of 4-Methylimibelliferyl Phosphate Concentration. -101-

5.0 was calculated to be 0.9 x 10~^ mole/liter. Substrate inhibition was _tr noted at substrate concentrations above 8 x 10 mole/liter. The relatively low K^ value for 4-methylumbelliferyl phosphate indicates that the assay can be performed at very low substrate concentrations, thereby representing a savings with regard to reagent. L-Tartrate as an Inhibitor of Activity. — The inhibition of acid phosphatase (prostatic) activity as a function of L-tartrate concentration was investigated. At the concentration of L-tartrate in the assay, >10 M, essentially 95$ inhibition is seen. Figure 2.37 is a plot of the reciprocal of the reaction velocity vs the reciprocal of the substrate concentration for increasing concentrations of L-tartrs.te. -The plot takes on a form that cart> be described by noncompetitive inhibition, which would be expected, due to the dissimilar nature of the substrate 26 and inhibitor.

Optimum pH. — The maximum activity was obtained at pH 4.0, with an activity near zero at pH 7.0 and essentially half the maximum value at pH 2.7. The hydrolysis rate was determined throughout the pH range 2.7 to 7.0 by measuring the activity in the absence of enzyme under the same conditions as were used to obtain the pH-vs-activity curves. Figure 2.38 shows the enzyme activity vs pH.

Enzyme Stability. — The control serum containing human prostatic acid phosphatase was stable over a period of at least one week when prepared in an acetate buffer. We have noted that normal serum acid phosphatase removed from the clot immediately after centrifugation is stable when diluted in acid buffer but loses approximately 50$ activity over a

24-hour period upon refrigeration without prior buffer dilution. The -102-

ORNL-DWG 73-2230RI

5 x icr5 M L-TAfRTRATE i

c E • 1 x 10"5I \ L-TAR TRATE ^ Q)

Q)U1 O / £ •/ /

o o UJ > / • / / w /, / NO L-T£RTRAT E

0 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 /SUBSTRATE (moles/liter X106)"1

Fig. 2.37* Reciprocal Velocity vs Reciprocal 4-Methyluiribelliferyl

Phosphate as a Function of L-Tartrate Concentration. -103-

1600 ORNL-DWG 73-2230RI

> v

6 \ < > <1>

3 • \\ •

4 / \

1 1 - 6 pH

Fig. 2.38. Enzyme Activity of Human Prostatic Acid Phosphatase vs pH. -104- normal acid phosphatase activity in serum samples under the assay conditions described was 1.0 I.U. ±0.5 I.U. at 25°C; the abnormal control serum had an activity of 4.1 I.U. The precision of replicate analysis was represented by a C.V. of 5

The acid phosphatase procedure provides a direct kinetic measurement of acid phosphatase and, as such, should be useful for rapid monitoring of human serum acid phosphatase, particularly since measurements of this type are related to abnormally high levels of prostatic acid phosphatase. The reaction is immediate upon initiation, that is, without la.g phases or necessity of pH adjustment for end-point monitoring. In addition, it does not involve the problems inherent in the diazo coupling (to produce colored products) method that is currently employed in direct kinetic 23 spectrophotometric assays.

2.3.8 Immunological Assays The development of immunological assays and their adaptation to the Fast Analyzer (miniature and 15-place) are being approached by: (a) the development of fluorescence polarization techniques for quantitation of interacting antigen-antibody or hapten-antibody systems, and (b) the development of competitive binding fluoroimmuno assays.

Fluorescence Polarization. — Fluorescence polarization, although somewhat limited in sensitivity, offers one method for the direct quantitation of equilibrated solutions of antigen-antibody or hapten-antibody without first laboriously separating the bound antigen-antibody complex from free antigen. This is particularly promising with regard to quantitation of circulating antibodies formed in autoimmuno disease or as a consequence of infection of the body by foreign material. It has been -105- demonstrated that fluorescence polarization measurements can he made by using the 15-place fluorometric Fast Analyzer, and this appears to be an open area for development to produce an automated fluorescence polarization immunological assay for circulating antibody (see Sect. 2.3.3).

Competitive Binding Assays. — The equilibration of an antigen with its specific antibody, or the equilibration of a substrate with a specific binding protein and the subsequent equilibration of the same system with a tracer-labeled antigen or solute produces a means of determining the original concentration of the antigen or substrate by measuring the amount of tracer bound, or the amount of free tracer remaining in solu-

tion after removing the bound complex. This can be represented by:

Antigen + Antibody Antigen-Antibody Complex ,

* *

Labeled Antigen + Antibody Antigen - Antibody Complex.

The amount of labeled antigen bound to the antibody is an inverse function of the amount of antigen present in the system. Tlie same consideration

applies to hapten-antibody and specific protein-bindiag systems.

The tracers employed so far have consisted of radioactively labeled

compounds which have provided the sensitivity required for measuring substances in serum in the concentration range 10-1 1 to 10 -12 ^/liter. The

procedures developed have been tedious, requiring several separation f steps, and involving the sequential counting of samples for as long as

10 min per sample. In addition, these radioimmunoassays commonly employ 125 I as a label. The short half-life of: this isotope, 60 days, makes

it difficult to maintain reagents on the shelf and thus reduces the

availability of immunoassays to small-volume hospital laboratories due -io6- to the maintenance cost of reagents "between assays. The final criticism involves the problems associated with handling radioactive tracers in hospital laboratories. If a reasonably stable fluorescence tracer- labeled compound can be obtained, if the sensitivity required for these assays can be achieved, and if much of the tedium can be eliminated from these assays by automation of separation steps and rapid measurement of remaining fluorescence tracer, then the fluoroimmunoassay will find a place in both biochemical and immunochemical research, as well as in clinical diagnostics. Toward this end, an effort has been made to develop one or two model fluoroimmunoassay systems to evaluate and demonstrate the feasibility, and perhaps the utility, of such an approach.

Considerable information about fluoresceinthiocarbonyl insulin is available. It is known that reaction of insulin with fluoroescein- isothiocyanate (FITC) can result in the formation of three labeled compounds, mono-, di-, and tri-labeled FITC-insulin. It is also known that mono-labeled FITC-insulin maintains most of its biological activity and all of its immunological activity; however, with further labeling, the immunological activity decreases. It is apparent that careful preparation of mono- and di- FITC-insulin can result in compounds very useful for developing a fluoroimmunoassay of insulin. Therefore, an attempt was made to prepare predominantly mono-labeled FITC-insulin and to measure its immunoactivity as compared with human insulin, through the use of a radioimmunoassay procedure. The objectives of the second effort, which is currently in progress, are to produce purified preparations of mono- and di-labeled FITC-insulin and to demonstrate their immunoactivity using the radioimmunoassay procedure. -107-

lli- Preparation of FITC-insulin. — It has been shown "by Bromer et al.

' hat the most reactive group in insulin toward FITC is the n-terminal amino group of the B chain, which is phenylalanine. Further reaction of FITC with insulin results in a reaction with the n-terminal glycine of the A chain, and finally, an e-amino group of a lysine residue found on the B chain. If an FITC-.insulin molar ratio of 1:1 is used in the reaction, and a reaction time of 5 to 10 min is employed, about 50$ of the insulin will have reacted tc form about 95$ mono- and 5$ di-labeled insulin. Higher FITC:insulin molar ratios and longer reaction times

(e.g., 70 min) will produce all three substituted compounds. In our first preparation of FITC-insulin, a 1:1 molar ratio was chosen and the reaction was carried out as described below.

Crystalline beef insulin (57 mg) was dissolved in 6 ml of 0.2 M sodium carbonate-bicarbonate buffer, pH 9.1. To this solution, 4 mg of powdered fluorescein isothiocyanate (isomer I, Sigma Chemical Company,

St. Louis, Mo.) was quickly added. During the addition of FITC and throughout the reaction, vigorous stirring was maintained. At the end of 15 to 30 min, the pH of the reaction mixture was rapidly lowered to 4.5; by the addition of 6 M HC1, and FITC insulin was precipated from solution. The precipitate was dissolved in carbonate buffer, and either passed over a Sephadex G-25 column or again precipitated by adjusting the pH to k.5. The spectrophotometric scan of this material

is shown in Fig. 2.39. Measurement of the immuno activity of this

insulin derivative as compared with human insulin was performed using a

double-antibody radioimmunoassay with reagents from Schwartz/Mann

Company (Orangeburg, N. Y.). Figure 2.40 shows a plot of the percentage ORNL-DWG 73-2062

X,WAVELENGTH (nm)

Fig. 2.39. Spectrophotometry Scan of FITC-insulin in 0.1 M Tris

Buffer, pH 7-5- -109-

ORNL DWG 73-9967

Fig- 2.40. Percentage of Insulin Bound to Antibody as a Function of Insulin Concentratir-a. -110- of FITC-insulin and human insulin "bound to antibody as a function of insulin concentration.

It is Important to "be able to look at both monosubstituted and disubstituted FITC-insulin. Therefore, in a second preparation, an FITC: insulin molar ratio of 1.5:1 "was used. In this procedure, 29.4 mg of

FITC was added to 286 mg of insulin in 30 ml of carbonate buffer, 0.2 M, pH 9.1. The reaction was allowed to proceed for 45 min, and then the

FITC-insulin was precipitated by rapidly lowering the pH to 4.5. After precipitating the FITC-insulin derivatives three times at pH 4.5, 6o mg of the FITC-insulin was dissolved in 3 ml of buffer consisting of 7 M urea, 0.01 M Tris, and 0.15 M NaCl. This was applied to a DEAE A-25

Sephadex column, and the material was eluted in 0.01 M Tris buffer using a linear NaCl gradient. The NaCl gradient increased in concentration from 0.15 M/liter to 1.0 M/liter. A uv-chromatograph indicated that the

FITC-insulin derivatives were separated into three fractions. The spectrofluorometric characterization of eluate samples from the three fractions, carried out in 1 N NaOH, confirmed that fraction 1 had a molar ratio of 1:1, fraction 2 had a molar ratio of 2:1, and fraction 3 bad a molar ratio of 3-1- However, when attempting to measure immunoreactivity of fractions 1 and 2 (mono- and di-labeled FITC-insulin), little or no reactivity was found. An effort is being made to find the source of the loss of activity. The successful preparation of active FITC derivatives

(mono- and di-) is a necessary prerequisite for the completion of the development of the fluoroimmunoassay for insulin.

In regard to future development, the fluoroimmunoassay for insulin will be continued to produce an assay that is sensitive to within one •Ill' order of magnitude of current RIA assays of insulin. In addition, we have "begun the investigation of the possible development of a fluoroimmunoassay for L-thyroxine through the use of fluorescamine-labeled

L-thyrcxine as a tracer. The problem here is that L-thyroxine is a hapten and it does not in itself have antigenic activity; however, when covalently bound to albumin, etc., it becomes an antigenic site on the larger molecule, and antibody specific for the L-thyroxine antigenic site can be obtained. The mode of covalently linking thyroxine to the carrier protein determines how sensitive the resulting antibody is to the active site of thyroxine as opposed to the remaining functional groups of this amino acid. The structure of thyroxine is shown below:

I I

Either the carboxyl group or the amino group can be activated for substitution onto the carrier protein. Apparently, this makes considerable difference in the specificity of the antibody produced as evidenced by the relative activity of desamino L-thyroxine to antibody. The activity of the antiserum from one manufacturer is 7$ that of L-thyroxine, and this suggests that any substitution of the amino group would be detrimental as far as activity of a tracer is concerned. However, a second producer of antithyroxine antibody has an antiserum that is at least 75$ active and has a certain potential of being active toward fluorescamine- substituted L-thyroxine. The preparation of this derivative, the -112- measurement of its fluorescence, and its activity toward L-thyroxine antibody will "be considered in the very near future.

2.3.9 Fluorometric Determination of Uranium There is an anticipated need for analytical techniques in nuclear fuel processing plants for monitoring the concentrations of uranium and

Plutonium. The development of a sensitive fluorometric Fast Analyzer has made it possible to monitor the fluorescence intensity of several different solutions simultaneously and rapidly. This has led to consideration of the Fast Analyzer for measuring the fluorescence emission of dissolved uranium at different concentrations. Six solutions of uranium -3 -3 in 3 M were prepared with concentrations of 3 x 10 M, 1.5 x 10 M, 6 x lQ-k M, 3 x 10"14" M, 1.5 x 10~k M, and 3 x 10-5 M. Plots of log relative fluorescence intensity vs log concentration when excited at 260, 280, and 300 nm are shown in Fig. 2.1+1. The emission signal was monitored from 500 to 600 nm by the use of appropriate interference filters placed between the objective lens and the photomultiplier tube. The detection of uranium samples at concentrations of 1 x 10~^ M can be obtained, and this sensitivity can be increased through careful optimization of the optical system for this particular problem. 2.3.10 Future Development An additional modification of the fluorescence detection system is under consideration. It is known that some molecules have fluorescence -9 -6 decay times of the order of 10 to 10 sec, with many systems falling -9 -8 in the range of 10 to 10 sec. Thus, we are presented with a small, but finite, luminescence period following excitation. A second fact is ORNL DWG 73-9968 10*1 T T • -EXCITATION AT 305 nm A- EXCITATION AT 260 nm

u K Z

U> H -no2 UJ

1 1 10"! 10-4 10-3 10-5 2+ 10- U05 CONCENTRATION (W>

2+ Fig. 2.41. Relative Emission Intensity at 500 to 600 nm of U02

in 5 M H2S04 as a Function of Concentration and Excitation Wavelength. -114- that fluorescence detection limits are a function of, among other things, "background noise due to light scatter and Raman emission from the solvent. It seems reasonable, then, that if one were able to pulse the excitation source in such a manner that several pulses occurred during the dwell time of any given cuvet in the monitoring system, and if the emitted light -9 -9 could be monitored from 1 x 10 to 5 x 10 sec after each lamp pulse, only the emission signal due to the fluorescence species and possible contaminants would be detected. A system based on these considerations would have a significant potential in lowering detection limits, but it would require a rapid-pulsed, high-intensity source. A delayed fluorescence detection system is now under consideration.

2.4 Chemical Assay Methods

C. A. Burtis, T. 0. Tiffany, J. B. Overton, and M. B. Watsky

Work is continuing on the development of chemical assay methods using photometric quantitation.

2.4.1 Triglycerides for Large 15-Place GeMSAEC Several methods are used for the analysis of serum triglycerides:

p rj (a) a colorimetric procedure using the Eegriwe reaction, (b) a 29 colorimetric procedure using Schryver's reaction, (c) a colorimetric OQ procedure using the Hantsch reaction, ' (d) a fluorometric procedure using the Hantsch reaction, ' D and (e) an enzymic procedure using a 34 35 coupled dehydrogenase assay. Each of these procedures requires prior hydrolysis of serum triglycerides to glycerol and free fatty acids and then subsequent analysis of the released glycerol to quantitate the initial serum triglyceride level. The colorimetric and fluorometric triglyceride methods require initial extraction of serum with isopropanol -115- and aluminum oxide, or other lipid adsorbents, to remove interfering lipids "before saponification of triglycerides. The enzymic triglyceride methods use either KOH as a saponifying agent, or more recently, lipase to release glycerol, followed "by a glycerol kinase-pyruvate kinase- lactate dehydrogenase coupled assay. The latter procedure uses the stoichiometric oxidation of reduced nicotinamide adenine dinucleotide

(NADH) to quantitate the amount of released glycerol. In addition it utilized enzymic hydrolysis, which eliminates the requirement of a previous serum extraction and thus simplifies the analysis. However, one disadvantage exists in the use of the enzymic hydrolysis method due to phosphatase contamination of the lipase in the commercially prepared reagent." During photometric determination of glycerol, the contaminant introduces a competing reaction which produces an elevated final value. Of course, this can be corrected by using a "reagent blank" under the same reaction conditions as the samples and subtracting this blank value from the total value obtained. Alkaline phosphatase in serum does not ordinarily interfere, except at markedly elevated concentrations, because of the small sample volumes required for the analysis.

The analysis of serum triglycerides using enzymatic hydrolysis coupled with an enzyme assay is being investigated for use on the Fast

Analyzer. Both an enzymic integral and an enzymic fixed-time substrate

19 analyses are being studied. The advantage of using an end-point procedure is that the blank-corrected absorbance change can be related directly to triglyceride concentration without the necessity of using calibration standards; however, 6 to 8 min is required for completion of the reaction and the reagent blank must be determined during each -116- run. On the other hand, the fixed-time procedure requires only 2 or 3 min of analysis time but demands the use of calibration standards.

Integral (End-Point) Analysis. — This analytical approach was adapted for the larger 15-place GeMSAEC Fast Analyzer as follows: first,

10 |il of HgO along with 500 |al of triglyceride reagent was metered into inner reagent well 2 of the transfer disk to serve as a reagent blank.; then 10-|il volumes of the samples and reference solutions along with 500 pi of reagent, were individually dispensed into inner reagent wells 3 through 15. All pipetting and dispensing were performed semiautomatically with a Micromedics Sample and Reagent dilutor and dispensor (Micromedics

Systems, Inc., a division of Rohm and HP-as, Philadelphia, Pa.). The transfer disk, containing the sample and reagent minus glycerol kinase, was agitated carefully as an additional mixing step and allowed to stand covered with a Lucite cover for 10 min to ensure complete enzymic hydrolysis of the serum triglycerides. As a final step in the procedure,

500 pi of distilled water was added to reagent well 1 as a reference solution, and 20 |il of glycerol kinase was pipetted manually into each of the outer sample cavities 2 through 15. The transfer disk was then placed into the Fast Analyzer, and the reaction was initiated by accele- ration of the analyzer rotor. It can be seen from the reaction progression curve shown in Fig. 2.1+2 that the reaction rate does not go to zero as glycerol is exhausted from the reaction mixture; instead, a rate which is due to the reaction blank is exhibited.

Two questions can be raised concerning the reagent blank: with what precision can it be obtained, and what contribution does the serum alkaline phosphatase make to the reagent blank? The precision of the blank value -117-

ORNL DWG 72- II746RI

pLQT OF T Tfl c FiQr' U- J'O.

^fwmtl f it . . 0:0 . S^PUNG /filE^tiHL \fr\t Gmh = 3b Minutes = o ^nPLts/cuuhiE •= 20 •moNo^. = o

mwmnce Tint 1,500-'

• 1.2feCL

J)i 1,22(0 J

o.m- • R

O.IMO • • • # • • • *

0100 & * • • vi "( :i • -r r i" < i \ '>ft i 'i "•«' i fi i r i it » r c »• * i fi't't'r r r . - v ' I WE IH HC0HD$< . .

Fig. 2.42. Reaction Progression Curve for the En r.ymat ic BeJ ermina* tion of Serum Triglyceride. -118- in one disk of lU reagent blanks over a 3-min period is 0.0064 ± 0.0003 absorbance unit. However, since the temperature varies from run to run, it is necessary to run a blank with each analysis because of the temperature-dependent enzymic nature of the blank. Several samples were run to answer the second question pertaining to the conbribution of the serum alkaline phosphatase to the blank. In each case the reaction was monitored for 15 min, after which all the glycerol had been exhausted. The concentrations of the samples ranged from 8l mg/dl to 278 mg/dl (expressed as triolein); the data for the overall reaction show that for this group of samples, the contribution to the overall reagent blank due to serum alkaline phosphatase was very small.

In the analysis for triglyceride samples in the normal and slightly elevated range the coefficient of variation was 2 to 2$. To resolve the question of how the results of the enzymic integral (end-point) method relate to those obtained with a more conventional clinical method using extraction, saponification, and a fluorometric determination of the released glycerol, several samples were analyzed using both procedures. The correlation of the enzymic procedure on the Fast Analyser with the standard Kessler Fluorometric procedure^ yielded, for Zk samples, a regression line of y » U.5 •$• 0.9UX and a correlation coefficient of 0.?75. The mean value for the enzymic procedure was 134.7 mg/dl while the mean value for the fluorometric method was 138.4 mg/dl.

Non-linear Fixed-Time Analysis. — The fixed-time analysis of serum triglycerides was made in the same way as described in the previous section, except that the reagent blank was eliminated from cuvet 2 and aliquots of glycerol standards equivalent to triolein concentrations of -119-

100 mg/dl, 200 mg/dl, and 300 mg/dl were placed in sample cavities 2,

3, and 4 respectively. Accuracy of the fixed-time procedure depends upon the standards; however, with proper quality control, an accuracy greater than 2% can "be achieved by careful weighing of known amounts of glycerol.

This can be checked by using the integral (end-point) enzymic substrate program.

An alternate approach to the enzymic determination of serum triglyceride consists of determining the absorbance change between two

fixed-time absorbance readings and subsequently relating this absorbance

change to concentration via the use of glycerol standards. If all of

the series-coupled reactions in the assay follow pseudo-first-order kinetics, then the absorbance change during any selected fixed-time

interval will be a linear function of initial glycerol concentration.

Since the contribution of the blank is the same for each standard, it

will be eliminated; thus, it will not be necessary to run a separate

reagent blank. Figure 2.k3 shows the relationship between absorbance

change and glycerol concentration, expressed as triolein concentration

in units of mg/dl, for several time periods. The function does not become linear until the reaction has proceeded for 2 min. However, the

consideration concerning elimination of the blank still holds; and with

the use of the nonlinear fixed-time program, the fixed-time approach

serves as a useful procedure for the enzymic determination of serum

triglycerides. The precision of the fixed-time analysis was quite good

with a mean value of 190 mg/dl and coefficient of variation of 1.1^.

The correlation between the results of the end-point analysis and the

fixed-time analysis is reasonable; these data for several samples are -120-

ORNL DWG. 72-II750RI

Glycerol as mg/IOOml Triolein

Fig. 2.^5. Effect of Time on the Enzymatic Assay of Serum Tri- glyceride. -121- shown in Table 2.18. The end-point analysis was performed both by us and by medical technologists in the ORNL Health Division, and the results are in excellent agreement.

Fluorometric Triglyceride Analysis. — Comparison of the results of the enzymic spectrophotometric triglyceride analysis were made by for- warding aliquots of several serum samples to Bio-Science Laboratories

(Van Nuys, Calif.) for analyses.

2. b.2 Serum Uric Acid with the Miniature Fast Analyzer

The typical assay for uric acid in the clinical laboratory involves

the chemical oxidation of uric acid with phosphotungstic acid. The

procedure, although effective for this purpose, is not specific and the

oxidation of a variety of serum metabolites also occurs. Uric acid can

be chemically and enzymatically oxidized to allantoin and carbon dioxide.

Kalchor introduced the use of purified porcine liver uricase for the

highly specific assay of uric acid in biological fluids. Uricase has

been demonstrated to be almost totally specific for uric acid and has

been shown to exhibit little or no activity toward uric acid derivatives 37 or compound analogs. Praetorious and Paulsen modified Kalchor's

procedure to develop the current clinical method for uric acid using

uricase (porcine) at 295 nm. Their procedure required a 1:30 serum

dilutor with a blank reading of 0.7 to 1.0 absorbance unit at 293 nm.

The porcine uricase preparations used for routine enzymic uric acid

determinations are relatively crude, exhibiting low activity and high

reagent blanks at 293 nm. Mohler's recena t introduction of a highly purified and active bacterial uricase with a low reagent blank (e.g., -122-

Table 2.18. Comparison of Replicate Analyses for Serum Triglycerides Using Fixed-Time and End-Point Methods

Type of Analysis Sample* Fixed-Time End-PointC (mg/ml) (mg/ml)

Control 238 2^0 A 2k3 2b0 B 169 175 C 152 lb6 D 230 234 E 93 90 F 161 16b

G 104 101

H 168 166

J 188 182 K 162

L 169 M 73 N 223 0 79

P 212 • •—•

SLSample s represent 10- to 16-hr fasting serum samples. Fixed-time analysis. Delay time, 13 sec; = 3 min. Q End-point determination with a reaction time of 8 min. -123- less than 0.05 absorbance units) has renewed interest in the enzymic serum uric acid determination at 292 nm,

Based on the uricase method, a uric acid procedure is "being developed for use with the miniature Fast Analyzer. The problem of adapting uric acid procedures to the miniature analyzer is twofold and involves both instrumental changes and chemistry development. First, a suitable light source must be obtained for operation at 292 nm, and a rotor capable of uv transmission must be available. Second, the chemistry has to be suitably adapted to the existing automated sampling and diluting sysr.em.

Light Source and Rotor. — Three light sources were available for testing with the miniature Fast Analyzer: (1) a. 22-W quartz-iodine lamp; (2) a low-pressure mercury lamp coupled to a 300-nm emitting phosphor: and (3) a deuterium lamp source coupled to a monochromator and quartz optical fiber bundle. Each source was evaluated with respect to linearity of absorbance at 290 to 293 nm by using calibrated uric acid solutions whose absorbances ranged from 0 to 2.0 absorbance units. A rotor with quartz windows was fabricated by sandwiching a black acrylic body between a bottom quartz disk and top quartz annular ring. A plastic insert, which was cut to fit in the top center portion of the rotor, contains the sample reagent and dynamic loading ports. The procedure for assembling the rotor is identical to that used for assembling the 5 all-plastic rotors.

Figure 2.bk shows the linearity of absorbance of the uric acid solutions at 292 nm as a function of window material and lamp used. The linearity of absorbance at 292 nm for the uv/phosphor source, and the deuterium source with the quartz rotor, is shown in Fig. 2.1+5. As -124-

Fig. 2.44. Linearity of Absorbance of Uric Acid Solution at 292 nm as a Function of Concentration as Measured by the Miniature Fast Analyzer with Different Light Sources and Rotors. -125-

ORNL DWG 73-99 73R1

Fig. 2.45. Linearity of Absorbance of Uric Acid Solutions at

292 nm as a Function of Concentration as Measured by the Miniature Fast

Analyzer with a Deuterium Lamp Source and a uv Phosphor Source. -126-

shown in these figures, the most nearly linear response was obtained with the quartz rotor. In addition, the uv/phosphor source has reasonable

linearity and potential application. The main disadvantage of this lamp

source is its extremely low intensity, which results in a higher noise

level (e.g., ±1.5 x 10 absorbance unit uncertainty as contrasted to

±0.7 x 10 absorbance unit with the deuterium source). The performance

of the quartz iodine lamp was not acceptable. However, it has recently been reported that a higher-wattage GE halogen lamp operated at 12 V for

the analyzer period provided adequate uv intensity at 292 nm and reason- 39 able lamp life. However, due to the immediate availability of the

deuterium source, the uric acid procedure was developed using the quartz

rotor in conjunction with the deuterium source and a 200- to 400-nm mono-

chromator as the optical system.

Chemistry Development. — Due to its high specific activity, the

procedure was developed using bacterial uricase obtained from Nova

Industry (New York, N. Y.). This uricase has been found to be stable

for a year in 0.1 M borate buffer, pH 9.1. The development of the method was straightforward and primarily involved developing the procedure

within the operating conditions of the automated sample-reagent loader

as discussed earlier (Sect. 2.1.1). This device aspirates small aliquots

of samples and reagents into their probes, and then dispenses and follows

them with a preset quantity of diluent into their respective rotor cavi-

ties. Thus, a concentrated preparation of the reagent must be used. To

meet this condition, the uricase reagent was prepared in a 1.0 M borate

buffer (pH 9-1)• Consequently, during the loading procedure, 15 |il

of reagent is aspirated and diluted approximately 1:10 in the final -127- reaction volume. The analysis is thus performed at a buffer concentration of 0.1 M and a pH of 9-1* which is the optimal condition for this assay. The preparation of the enzyme in the concentrated form did not affect the activity of the enzyme. The detailed analysis procedure is presented in Appendix 1A.

This procedure was evaluated for several weeks to determine its precision and to correlate the results obtained from it with parallel data obtained from an existing procedure using a continuous flow analyzer.

In the correlation, both analytical methods used the same standards. The main difference between the two systems is the reducing procedure that is used with the continuous flow analyzer. A phosphotungstic acid procedure has been designed to minimize nonuric acid interference. More than

200 samples have been analyzed with each system, and the results have been correlated by linear regression methods. Typical correlation coefficients of 0.91 to O.96 have been obtained, with mean uric acid values of 5*85 mg/100 ml and 5.95 mg/100 ml for the continuous flow analyzer and the enzymic uric acid procedure respectively. The C. V. of the analysis Is 3 to b% in the normal range (e.g., 4 to 6 mg per 100 ml), and 2Pj0 in the abnormal range of 8 to 10 mg per 100 ml. Because of these encouraging results, the uric acid procedure on the miniature Fast

Analyzer will replace the continuous flow analyzer method and will be routinely used -by the ORNL Health Division.

2.4.3 Additional Serum Chemistry Assays for Miniature Fast Analyzer

Twelve chemical assay procedures already developed have been adapted for use with the miniature Fast Analyzer and five of the procedures are now routinely used in the ORNL clinical laboratory. Detailed procedures -128- for the following are presented in Appendix 1 (Section B through M):

B, alkaline phosphatase; C, acid phosphatase; D, creatinine phosphokinase

(CPK); E, lactic dehydrogenase—lactate substrate (LDH-L); F, glutamate oxaloacetate transaminase (SGOT); G, serum glutamate pyruvate transaminase

(SGPT); H, serum glucose; I, blood urea nitrogen; J, serum triglyceride;

K, calcium, L, total bilirubin; M, multi-enzyme:multi-sample.

2.5 Evaluation and Operation of Fast Analyzers in a Clinical Laboratory

C. A. Burtis, T. 0. Tiffany, W. F. Johnson, and J. B. Overton

A 15-place GeMSAEC Fast Analyzer has been routinely operated in the clinical laboratory of the Health Division of the Oak Ridge National

Laboratory, thereby allowing an evaluation under routine conditions. It has been extremely useful, not only in providing data relevant to the operation of the analyzer but also in providing useful medical information for the ORNL employees. As this evaluation is successfully

concluded, the older analyzer will be replaced with the newer miniature

Fast Analyzer.

2.5.1 15-Place GeMSAEC Fast Analyzer

It has been demonstrated that the 15-place GeMSAEC Fast Analyzer will provide accurate and precise data with a minimum of instrumental

downtime. It has been used in the clinical chemistry laboratory of the

ORNL Health Division for the past 2 years. The operation began with the

routine determination of glucose, albumin, total protein, SGOT, LDH, and

alkaline phosphatase in serum. The reliability and performance capability of this analyzer was demonstrated, and its operation began to be directed -129- to more selective tests to better serve the health needs of the L&x>ra-

cory. In the routine mode of operation, Laboratory personnel undergoing periodic l8-month health examinations were asked to present themselves at the time of the examination in a fasting state (12 to 16 hr without food). The Fast Analyzer was then used to perform the following tests on the blood sample drawn at that time: glucose (fasting and 2-hr post-

100-g glucose load), triglycerides (fasting), SGOT, LDH, and alkaline phosphatase. The laboratory has also been participating in two outside quality control programs, one set up by the Center for Disease Control

(CDC), and the other directed by the College of American Pathologists

(CAP). The program results obtained by the laboratory technologists using the Fast Analyzer have generally been very good. Such quality control programs are important in maintaining useful results in the clinical laboratory.

Miniature Fast Analyzers

To evaluate the miniature Fast Analyzer under routine conditions, a

small analytical system based around a miniature Fast Analyzer has been placed in the clinical laboratory of the Health Division of the Oak

Ridge National Laboratory. After an initial evaluation and comparison,

the older 15-place GeMSAEC Fast Analyzer will be phased out and replaced with the miniature one.

The miniature fast analytical system consists of a miniature Fast

Analyzer, several rotors, an automated sample-reagent loader, a rotor

cleaning station, and a computerized data system. Elements of this

system are shown in Figs. 2.1 and 2.13. -130-

Using the previously described on-line computer (Sect. 2.1.3), an operating program has been developed for routine operation of the system.

The program has an executive program which initiates the daily operation, and by a series of operator questionrresponse routines, will automatically load the required program into the computer core.

2.5.3^ Populatiomm n Studiein. msm One interesting result of the use of the Fast Analyzer in the ORNL clinical laboratory comes from the statistical evaluation of the laboratory test results over a period of time to establish normal ranges of blood chemicals for the ORNL population according to sex and age.

Serum Glucose. — For example, great care in all laboratory techniques and use of proper quality control have produced results that should be very close to the so-called true serum glucose value. The precision of this assay using the more accurate enzyme procedures has been shown previously to be in the range of 2 to 3%. Table 2.19 gives the Fast

Analyzer results for glucose from the ORKL population by sex and age, and

Table 2.20 shows the standard deviation of the test results, again as a function of sex and age, for both the fasting sample and the sample taken 2-hr after ingestion of 100 g of glucose. Such statistical results, when the test is known to be reliable, provide several benefits. First, the normal range as a function of age and sex is established for both the fasting and the 2-hr glucose tests. Second, the data show a definite age depenciw^e which is particularly evident after ingesting the 100-g of glucose. Third, it is noted from the results in these two tables that the population has a constant mean fasting glucose value and -131-

Table 2.19. Statistical Evaluation of Serum Glucose Data from the ORHL Population

Fasting Samples Two-Hour Sample Population Number After Glucose Loading Group of 8, Mean Range Mean Rangea Subjects (mg/100 ml) (mg/100 ml) (mg/100 ml) (mg/100 ml)

Total population 605 87-5 68 - 107 39 - 152

Male total 481 89 70 - 106 (87) 40 155

Male (20-30) 18 88 74 - 101 (96) 59 — 122

(30-40) 102 96 69 - 103 (86) 42 mm 131

(40-50) 205 88 72 - 105 (86) 46 mm 148

(50-60) 135 90 71 - 110 (104) 37 - 171

(>60) 21 93 65 - 121 (120) 42 198

Female total 124 83 62 - 105 (88) 40 mm 137

Female (20-30) 25 77 60 - 95 (85) 51 - 125

(30-40) 20 80 67 - 93 (89) 33 - 137

(40-50) 47 86 62 - 110 (94) 43 - 136

(50-60) 24 88 68 - 108 (85) 35 - 153

(>60) 5 85 77 _ 93 (84) 61 — 108

^ean value ± two standard deviations. -132-

Table 2.20. Standard Deviation of Serum Glucose Levels in the ORNL Population as a Function of Age

Number Standard Deviation (mg/dl) of Sample Two Hours Group Age Subjects Fasting Sample After Glucose Loading

Male 1 (20-30) 18 ±7 ±13

2 (30-40) 102 ±9 ±16

3 (40-50) 205 ±8 ±21

4 (50-60) 135 ±10 ±30

5 (> 60) 21 ±14 ±38

Female 1 (20-30) 28 ±9 ±18

2 (30-40) 20 ±6 ±26

3 (40-50) 47 ±12 ±23

4 (50-60) 25 ±10 ±29

5 (> 60) 5 +4 ±12 -133- standa,rd deviation as a function of age. However, when a glucose load of 100 g is given, the population reflects a diverse response to this

stress and the standard deviations for the glucose loading tests "become

strongly age dependent. Such results demonstrate a reduction in the ability to control glucose with increased age, and can "be viewed as metabolic data reflecting the loss of glucose control as one facet of

human aging. Similar data are being obtained for serum triglycerides.

The above discussion provides an example of the function of the

Fast Analyzer in the ORNL Health Division chemistry laboratory. In

addition, this type of information obtained from a large static popu-

lation representing a variety of occupations can be used to furnish

information regarding preclinical disease not available from hospital populations. This is not to suggest the use of the ORNL population for

experimentation, but to suggest the effective use of existing population

da.ta to obtain more extensive information concerning the relationship between test results and subclinical disease. This is only possible when reliable, accurate, and precise analytical instrumentation is

available.

Serum Triglycerides. — Quality control is maintained during an

analytical run in two ways. A glycerol standard and an elevated human

serum triglyceride control (Tri-El, A. R. Smith Laboratories, Los Angeles,

Calif.) are determined during each run to provide a check on the perfor-

mance of the coupled assay for glycerol (glycerol standard) and on the

effectiveness of the preanalysis lipase digestion of triglyceride (human

serum control). Quality control samples analyzed over a two-week period

using human serum resulted in a mean value of 276 mg/100 ml with a range -134- of 266 to 285 mg/100 ml and a run-to-run coefficient of variation of 3.5%.

The commercially established concentration of triglycerides in this serum control (established by a similar enzymic procedure) was 275 mg/lOO ml with an acceptable range of 266 to 285 mg/100 ml. Values reported here are based on triolein as the reference triglyceride.

Participation in outside quality control programs has resulted in values within the acceptable laboratory performance range, but higher than the mean value obtained by the use of extraction methods. However, in reviewing the results for all enzymic triglyceride values for reference and volunteer laboratories participating in the CDC proficiency testing

(March 9> 1973)> it is interesting to observe that these values are consistently higher than those obtained by using extraction procedures.

This suggests that a certain amount of free glycerol is present in such samples and is not being extracted along with the triglycerides prior to hydrolysis and quantitation. The source of this free glycerol is most likely artifactual contamination of the sample vessels (with glycerol), but it could also arise from endogenous breakdown of sample triglycerides with lipase contained in the sample.

As previously indicated, correlation of the Fast Analyzer results with those from an outside evaluation of the samples using a fluorometric extraction procedure was excellent; mean sample values for both procedures agreed within 2 to 3% ahd the correlation coefficient was 0. 98. This discussion was presented here to provide some idea of the quality of the results as compared to other methods, outside evaluation, and internal quality control. -135-

During the past several months, fasting serum triglyceride values have "been obtained as a routine health maintenance examination for the

ORNL employees. These triglyceride analyses were performed using the enzymic procedures described in this work. The statistical data for 5^-2 analyses are shown in Table 2.21. No attempt was made to eliminate the several samples that gave abnormally high results (e.g., above 4-00 mg/100 ml as triolein) as has been done in similar studies. The largest number in this population study were in the age group from 35 to 50 years of age. Table 2.21 shows a trend toward an increase in both mean values and range with Increasing age. The lower values for females are demonstrated in this study.

Figure 2.46 is a histogram of the male population distribution of values from ages 30 to 60 broken into three groups. Age group (a) includes those values for individuals 30 to 39 years of age (123 samples) and represents the pattern of normals expected, with 86$ of the population below 185 mg/100 ml and 95$ below 230 mg/100 ml. Above 39 (group b, number of subjects = 171), the normal population spread greatly increases, with several values above 400 mg/100 ml. Again, 8l$ of the population is below I85 mg/100 mlj but the range has increased, and 95$ of the population is below 280 mg/100 ml. The last age group, 50 to 59, again shows a much broader distribution from group (a) with 80$ or better below 185 mg/100 ml, but 95$ below 280 mg/100 ml, showing again the broadening of the range with age.

Although the data from this 4-month study are preliminary, the study

indicates that the normal triglyceride values for a predominantly male

population with a mean age value of 45 are going to be considerably -136-

Table 2.21. Triglyceride Values Obtained for ORNL Population

Mean Value Number of (mg/100 ml) Range Samples

Male

All ages (20-65) 142 64 - 220 460

20-29 138 71 - 205 22

50-39 124 65 - 185 115

40-49 152 66 - 238 171

50-59 143 68 - 218 133

60-65 178 75 - 220 19

.Female

All ages (20-59) 105 55 - 160 82

20-29 95 72 - 118 25

30-39 96 61 - 127 17

4 0-49 94 50 - 138 22

50-59 l4l 67 - 215 18

Total Population

All ages 137 60 - 213 542

^ean value ± two standard deviations. No attempt was made to eliminate high values from this study or to correct the samples for the 4 to 10 mg/ 100 ml elevation of values due to endogenous free glycerol. ORNL-DWG 73-7821 20

(a) A GE GROUP 2 (1VI/U-E ) 30 -39 10 1 in ^ 0 M hu LILi LAJLJ 1 —•— 3 20 Q > Q (b) A GE GROlJ P 3 (M/M.E ) 40 -49 - 10 1 1 u_ o Q: If UJ JU III JrA | 0 Jm mytBm • 3 20

0 50 100 150 200 250 300 350 400 >400 TRIGLYCERIDE VALUES (mg/100 ml, AS TRIOLEIN)

Fig. 2.46. Histogram of Triglyceride Values for Three Age Groups. 1 higher than typical ranges quoted previously, "based on few samples and 30 weighted toward a lower age group and higher female population.

Summary and Conclusion. — Two approaches to the enzymic analysis of

serum triglycerides using the Fast Analyzer were presented, and the advantages and disadvantages of each approach were discussed. Each procedure was shown to give good analytical results; however, from a pragmatic viewpoint, the fixed-time procedure can "be performed more rapidly with

"better precision (1 to 2% as compared with 2 to 4%) and is more rapidly adaptable to routine analysis. The end-point procedure serves as an excellent method for checking the accuracy of the standards used for the fixed-time analysis. Used in this context, both procedures serve a valuable and useful function.

A preliminary normal study was presented based on the ORNL population, which is predominantly male with a mean age of 45. The values obtained show age and sex trends and demonstrate higher values for this

type of population. We are currently in the process of standardizing

this procedure with Dr. Gerry Cooper's lipid research laboratory at CDC.

We are beginning to investigate the problem of phospholipids and endo-

genous glycerol in an effort to accurately define the test and to be able

to better evaluate the laboratory statistical data in light of this information. 2.6 References for Section 2

1. N. G. Anderson, "Basic Principles of Fast Analyzers," Amer. J. Clin. Pathol. 53, 778 (1970).

2. N. G. Anderson, C. A. Burtis, J. C. Mailen, C. D. Scott, and D. D. Willis, "Feasibility of Miniaturization of a Fast Analyzer," Anal. Lett. 5, (1972). -139-

3. C. A. Burtis, J. C. Mailen, W. F. Johnson, C. D. Scott, T. 0. Tiffany, and N. G„ Anderson, "Development of a Miniature Fast Analyzer," Clin. Chem. 18, 753 (1972). h. C. D. Scott and C. A. Burtis, "Miniature Fast Analyzer Systems," Anal. Chem. 1+5, 327 (1973).

5. J. C. Mailen, J. B. Overton, C. A. Burtis, and C. D. Scott, "Techniques for Fabrication and Assembly of Rotors for Use in a Miniature Fast Analyzer," Anal. Lett. 6(3), 2*4-5 (1973). 6. W. F. Johnson, J. C. Mailen, C. A. Burtis, T. 0. Tiffany, and C. D. Scott, "Development of a Portable Data Processor with Mechanical Data Output for Use with a Miniature Fast Analyzer," Clin. Chem. 18, 767 (1972).

7. C. A. Burtis, W. F. Johnson, J. E. Attrill, C. D. Scott, N. Cho, and N. G. Anderson, "Increased Rate of Analysis by Use of a 42- Cuvet GeMSAEC Fast Analyzer," Clin. Chem. 17, 686 (1971).

8. C. D. Scott and J. C. Mailen, "Dynamic Introduction of Whole-Blood Samples into Fast Analyzers," Clin. Chem. 18, 7^9 (1972).

9. T. 0. Tiffany, C. A. Burtis, J. C. Mailen, and L. H. Thacker, "Dynamic Multicuvette Fluorometer-Spectrophotometer Based on the GeMSAEC Fast Analyzer Principle," Anal. Chem. 1+5, 1716 (1973).

10. T. 0. Tiffany, M. B. Watsky, C. A. Burtis, and L. H. Thacker, "Fluorometric Fast Analyzer: Some Applications to Fluorescence Measurements in Clinical Chemistry," Clin. Chem. (in press).

11. Chem. Tech. Div. Ann. Progr. Rept. for Period Ending March 31, 1973; ORNL-4883, p. 83.

12. L. Brand and B. Wilholt, "Fluorescence Measurements," in Methods in Enzymology, Vol. XI, ed. by C. H. W. Hirs, Academic Press, New York, 1967, P. 776.

13. R. C. Nairn, Fluorescent Protein Tracing, Williams and Wilkins Co., Baltimore, 1964, p. 9. W. W. Bromer, S. K. Sheehan, A. W. Be ins, and E. R. Arquilla., "Preparation and Properties of Fluoresceinthiocarbamoyl Insulins," Biochem. 6, 2378 (1967).

15. E. R. Arquilla, W. W. Brower, and D. Mercola., "Immunological Confor- mation and Biological Activity of Insulin," Diabetes 18, 193 '1969). 16. K. Federlin, Immunopathology of Insulin, Springer-Verlag, New York, 1971, P. 33. -iko-

17. C. N. Hales and P. J. Randle, "immunoassay of Insulin/' Biochem. J. 88, 137 (1963). 18. F. Perrin, "Polarization of Fluorescence Light by Means of Molecules in the Excited State," J. Phys. Radium 7, 390 (1926).

19. A. J. Pesce, C.-Rosen, and T. L. Pasby, "Polarized Fluorescence," in Fluorescence Spectroscopy, Marcek Dekker, Inc., New York, 1971, p.87.

20. T. 0. Tiffany, J. M. Jansen, C. A. Burtis, J. B. Overton, and C. D. Scott, "Enzymatic Kinetic Rate and End-Point Analyses of Substrate, by Use of a GeMSAEC Fast Analyzer," Clin. Chem. 18, 829 (1972). 21. H. N. Fernley and P. G. Walker, "Kinetic Behavior of Calf-Intestinal Alkaline Phosphatase with Methylumbelliferyl Phosphate," Biochem. J. 97, 95 (1965). 22. C. J. Cornish, F. C. Neale, and S. Posen, "An Automated Fluorometric Alkaline Phosphatase Microassay with U-Methylumbelliferyl Phosphate as Substrate," Amer. J. Clin. Pathol. 53, 68 (1970).

23. G. G. Guilbault, S. H. Sadar, R. Glazer, and J. Haynes, "Umbelli- ferone Phosphate as a Substrate for Acid and Alkaline Phosphatase," Anal. Lett. 1(5), 333 (1968).

2k. D. L. Fabiny-Byrd and G. Ertinghausen, "Kinetic Method for Deter- mining Acid Phosphatase Activity in Serum with Use of the 'Centri- fichemV' Clin. Chem. 18, 8kl (1972).

25. K. K. Tsuboi and P. B. Hudson, "Acid Phosphatase III Specific Kinetic Properties of Highly Purified Human Prostatic Phosphomon- esterase," Arch. Biochem..Biophys. 55, 191 (1955). 26. J. Westley, Enzymic Catalysis, Harper and Row, New York, 19^9, p. k2.

27. M. Kawade, "Microdetermination of Lipids in Serum," Mie Med. J. 11, 399 (1962). 28. E. W. Rice, "Triglycerides ("Neutral Fats") in Serum," in Standard Methods of Clin. Chem. 6, ed. by Roderick P. MacDonald, Academic Press, New York, 215-22 (1970).

29. F. Galletti, "An Improved Colorimetric micromethod for the Deter- mination of Serum Glycerides," Clin. Chim. Acta 15, iQk (1967).

30. M. J. Fletcher, "A Colorimetric Method for Estimating Serum Triglycerides," Clin. Chim. Acta 22, 393 (1968).

31. L. B. Foster and R. T. Dunn, "Stable Reagents for Determination of Serum Triglycerides by a Colorimetric Hantzsch Condensation Method," Clin. Chem. 19, 338 (1973). -141-

32. G. Kessler and H. Lederer, "Fluorometric Measurement of Trigly- cerides in Automation," in Analytical Chemistry Technicon Symposium 1965; I-. T. Skeggs, Jr., et al., eds,, Mediad, New York, 1966, p. 341.

33- b. Edwards, C. Falkov/ski, and M. E. Chilcote, "Semiautomated Fluorometric Measurement of Triglycerides," Stand. Meth. Clin. Chem. 7, 69 (1972). 34. 0. Wieland, "Glycerol," in Methods of Enzymatic Analysis, H. U. Bergmeyer, ed.,, New York, Academic Press/1963, pp. 211-14.

35. G. Bucolo and H. David, "Quantitative Determination of Serum Tri- glycerides "by Use of Enzymes," Clin. Chem. 19, 476 (1973). 36. H. M. Kalchor, "Differential Spectrophotometry of Purine Compounds by Means of Enzymes. I. Determination of Hydroxypurine Compounds," J. Biol. Chem. 167, 429 (1947).

37. E. Praetorius and H. Paulsen, "Enzymatic Determination of Uric. Acid," Scand. J. Lab. Invest. 5, 273 (1953). 38. J. L. Mohler, "A New Bacterial Uricase for Uric Acid Determination," Anal. Biochem. 38, 65 (1970).

39' 0. Ertinghausen, private communication. -142-

3- GENETIC MONITORING

Mutations in man may be caused by environmental insults such as ionizing radiation, chemical pollutants, etc. as well as natural causes. It is desirable to monitor the mutation rate so that dangerous increases can be detected. Since the mutation rate is postulated to be quite low, 1 x 10- x5 mutation per locus per generation, 1 a proper monitoring program would require measurements in a large population, preferably with techniques that evaluate the effects on many loci.

It has been proposed that high-resolution gel electrophoresis of red blood cell enzymes be used to monitor for 10 to 20 biochemical variants 2 associated with known genetic differences. We have initiated a program in this area to give technologic support to such a genetic screening program. Areas of interest will be: (1) automating the sample preparation step; (2) automating steps in gel electrophoresis methods; and (3) developing a completely automated, rrrultisample, high-resolution, electrophoresis system. Biochemical analytical systems such as the Fast Analyzer will also be evaluated as tools for monitoring biochemical indicators of genetic difference.

3.1 Sample Preparation

T. 0. Tiffany, C. D. Scott, w. F. Johnson, and N. E. Lee

If red blood cells are to be analyzed in the genetic screening program, a rapid means of separating stabilized blood into plasma and washed cells is needed. This separation and cell washing now require a tedious series of manual operations including centrifugations, mixings, and pipettings. Development of a semiautomatic device for processing a -143- number of blood samples in parallel is desired. It is proposed that this operation be performed in a processing centrifuge with an appropriately designed rotor.

3.1.1 First Prototype Rotor

The first prototype rotor (Fig. 3-1) separated stabilized blood into plasma and cell fractions, washed the cell fractions, lysed the blood cells, and allowed sampling of both the plasma and the lysed cells. The rotor was held in a fixture which could be rotated at speeds up to 2000 rpm and required several manual operations. Provisions were made for application of vacuum both at the edge of the rotor and at the large center hole of the top cover of the rotor. The latter vacuum application was to permit washing solution (0.9% NaCl) and red blood cells (RBC) to be mixed.

For operation, blood samples (1.5 nil each) were loaded into the blood chambers through the central port with the rotor at rest. The rotor was then brought to 2500 rpm and held at this speed until the cells and plasma were well separated. At this point, the rotor was slowed to about 1000 rpm and vacuum was applied around the edge of the rotor. This sequence of operations induced a partial vacuum in the plasma chambers, which drew the plasma into the plasma chambers through the transfer tubes. The rotor was then stopped, and the plasma was drained to the bottom of each plasma

chamber. The sloped chambers caused the plasma to drain to a tube inlet

at the outer edge of the chamber.

An automatic pipette was then inserted into the tapered sampling

ports to remove the plasma. Subsequently, the rotor was rotated at ORNL PHOTO NO. I046A-73

ROTOR

H -p4=-

Fig. 3.1. Top View of First Prototype of the Blood Separation Rotor. -145- about 2000 rpm, and a volume of physiological saline was injected into the port at the top center of the rotor to wash the separated RBC. This volume (about 0-5 to 1 ml per sample) was essentially equally divided by the splitting vanes and directed into the blood chambers. A Teflon ball with a hole drilled through the center was used to provide a seal at the center hole, and vacuum was applied at this point. The application of vacuum caused air to be drawn through the sampling ports, which resulted

in agitation and suspension of the blood cells. The rotating seal was

then removed and the cells were re sediment ed, after which the wash liquid was drawn off into the plasma chamber. The cell washing procedure was

repeated any desired number of times.

To lyse the cells, a last injection of distilled water or other

lysing liquid was made, and air was drawn through the blood chambers to mix this liquid with the cells. When mixing was complete, the rotor was

stopped and the lysed cells were removed via an automatic pipette inserted

in the sampling ports. Finally, the rotor was cleaned and dried by

drawing cleaning solutions (methanol and air) through all chambers and

tubes, using the vacuum applied to the edge of the rotor.

3.1.2 Advanced Rotor Design

A second rotor design incorporated additional automated features.

One problem with the first prototype rotor was that the total operation

was contained in a single rotor. It was decided that the processes of

plasma separation and sampling, etc., could be accomplished more readily

if the plasma chamber, blood cell chambers, and final lysate collecting

chambers were separate from the working centrifuge-syphon portion of the

rotor. Therefore, a new rotor system was designed to incorporate this -146- feature. The new system has provisions for "blood separation, RBC washing, and an BBC lysing rotor in which the initial whole blood samples are individually pipetted; however, an outer, separate, annular ring containing chambers (for each sample sector) for collecting plasma, RBC, and lysate in three successive operations is also used (Fig. 3.2). In the final concept, the operation of adding whole "blood will "be automated.

The new rotor system allows one to add whole blood samples, centrifuge them, and transfer the plasma by vacuum, speed control, and syphoning to the outer plasma collection ring. At this point the vacuum is shut off and the system is immediately braked to a stopped position. The plasma ring is then rapidly removed, a wash collection ring is inserted, and the RBC are dynamically washed several times with 0.9$ saline. Subsequently, the system is stopped, a RBC collection ring is inserted, and portions of the RBC are removed by application of a vacuum syphon to the collection ring. Next, the collection ring is replaced with the final lysate ring. Finally, a lysing solution is dynamically added to the inner rotor, the RBC, and lysed, the pH is adjusted to pack the stroma, and the lysate is transferred to its collection ring in a manner similar to the previous sample collection steps. The separation rotor and sampling ring are keyed so they will fit in only one position, ensuring sample identification. It should be noted that the new system has 16 sample sectors with blood sample volumes of 5 ml> in contrast to 8 sample sectors with sample volumes of 1 to 1.5 ml in the first prototype.

The important feature of the new system is that collected plasma, RBC, and lysate can themselves be sampled at the end of the process, or -147-

ORNL DWG 73-6913

WASH SOLUTION PORT

WASH SOLUTION SAMPLE RECEIVING ROTOR BODY INTRODUCTION CUP QHANNEL

Fig. 3«2. Design of Advanced Rotor and Sample Ring for Whole Blood Processing. -148- they can be sampled while other separation processes are in progress, Thus, sampling in this system does not limit farther separation and preparation steps as it did in the single rotor design. The final step involves washing the rotor and rings. The rotor can be dynamically washed with water and dried with an alcohol wash solution. This latter process requires approximately 10 min.

The system is now undergoing initial evaluation. Further modifications will probably be effected to make it a completed working system, but it will perform the separation steps as outlined.

3.2 Automated Elution Electrophoresis C. D. Scott and N. E. Lee Although elution electrophoresis of one type or another has tradi- tionally been used in a preparative mode of operation, it also possesses great advantages over the conventional manual operations necessary in high-resolution gel electrophoresis. Not only can the elution electrophoretic techniques be automated, but the detection of separated proteins is greatly simplified. Furthermore, the photometric response of the eluted species is amenable to precise quantitation.

Preliminary investigations in the area of elution electrophoresis have been initiated. A single column system has been designed and built in which a serum sample can be injected into a stationary bed of gel beads and eluted with a buffer immediately after the electrophoresis is complete (Fig. 3'3)* Simultaneous electrophoresis and elution are also possible with this system. A highly sensitive flow-uv-photometer operating at 254 and 280 nm is used to detect the separated eluted species. The glass column is 0.4-cm-ID x 20 cm long with a center entrance for the ORNL DWG 73-6906RI

Fig. 3.3. Elution Electrophoresis System. sample. Electrode ports are located at either end of the glass column, and a maximum electrical potential of 600 V is possible. Since a large amount of heat is generated during electrophoresis due to the reasonably high input of energy, the column is jacketed, and cooling water is circulated during the analysis. A positive displacement pump is used to deliver the buffer to the column and the effluent from the column passes directly to the miniature photometer. Preliminary results indicate that over ten electrophoretic peaks can be monitored from 5 M*1 of serum in 60 min when acrylamide gel beads (100 to 200 mesh) and a phosphate buffer are used.

The emphasis of the initial work has primarily been centered on column design. Since this system is not isolated from the pump and photometer during electrophoresis, special design problems are introduced. A low dead-volume sample injection port is necessary for maximum resolution of the separated peaks. Bubble formation at the electrodes is a challenging design problem, especially under the constraint of minimal backmixing of the separated species as they are eluted. Several variations in the design of the electrode ports have been tested. The current column design is quite usable, although more testing will continue in this area.

3.3 Isoenzyme Stripping by Use of the Fast Analyzer

T. 0. Tiffany, C. A. Burtis, and D. D. Chilcote

Determinations of kinetic enzyme parameters, such as the Michaelis-

v an

laboratories and have had little clinical application. However, due to -151- the presence of isozymes in human serum, and enzyme variants in red blood cells, white blood cells, and human serum, it is of interest to develop computer programs that can be used on-line with the Fast Analyzer to rapidly determine kinetic parameters, including K^ and

Enzyme kinetics have been used primarily for the investigation of reaction mechanism. It would be desirable to expand the use of enzyme kinetics for analytical purposes to tell what, as well as how much, is present and to provide the qualitative, as well as the quantitative, approach needed, for example, in determining genetic variants or defining circulating levels of elevated serum isozymes. These objectives can be achieved through the use of different substrate concentrations, inhibitors, different buffer pH, denaturants, or different thermal properties of enzymes.

A computer program has been written which enables the rapid determination of K and V for a single enzyme acting on one or two m max ° ° ° substrates in a single run through the determination of initial velocities at several different substrate dilutions, with all reactions progressing in parallel. In this manner, time-dependent denaturation of a partially purified enzyme, or an unstable serum enzyme, is eliminated. The data are plotted on the computer scope for rapid visualization. After visual inspectio* n of the data,' K m and V max', estimates of the error in these two parameters are obtained using an iterative best-fit algorithm. The

total parameter analysis requires approximately 10 min. Work is now. progressing on the use of similar programs to determine K and V"max for two enzymes acting on the same substrate. The results of this work have potential application in diagnostic differentiation of serum isozymes, -152-

V, V 1 S + 2 S /-. v K1 + S 2 where V^, V^ and K^ and K^ represent the maximal velocity and the Michaelis constants of the two enzymes. An attempt to linearize Eq. (l) by forming the reciprocal of v and S yields the following equation, which is obviously not linear:

L a [ 1 + Mi) - c v r [i + d(i)]

Therefore, in examining a reciprocal plot that was generated from data obtained from a heterogeneous enzyme mixture of two or more enzymes acting upon the same substrate, nonlinearity in the plot will be seen (Fig. provided that the K 1s of the two isozymes differ more than

twofold. The data in Fig. 3*4 were generated using Eq. (l) and representative K and V data. The solution of Eq. (l) for K_, K^, V_,

m max ^ JL 2 1

and V2 can be obtained, and the procedure for doing so has been written for the PDP-8 computer interfaced to the Fast Analyzer. In an evaluation of this program and the total Fast Analyzer data- taking process, alkaline phosphatase tissue extracts were used to check the isozyme stripping routine. Human intestinal alkaline phosphatase and ORNL DWG 73-9970RI 4 ISOENZYME I, Km = 1.0x I0" M, Vmax = 5 micromoles mM min-

.-3 ISOENZYME 2, Km = 1.0 xlO * M, Vmax= 50 micromoles mM min~' 1/ VELOCITY VS l/SUBSTRATE SUBSTRATE/VELOCITY VS SUBSTRATE 550 200

440 £ 160 r-

330 -

220

l/CONCENTRATION SUBSTRATE CONCENTRATION (millimole-1) (millimoles)

Fig. 3'b. Effect of Isozymes on the Linearity of Reciprocal Plots. -154- beef placental alkaline phosphatase were used to simulate isozymes. The

Km and Vmax values for both enzymes were determined using the single substrate routine. A mixture of the two enzymes was made to contain activities of approximately 150 I.U. per liter of placental alkaline phosphatase. Substrate concentrations in the range 0.055 to 11.0 millimoles liter were chosen, and ten p-nitrophenyl phosphate dilutions were used. The reciprocal plot and parametric evaluation are given in Fig. 3.5' The results indicate that the data can be fitted to the model and a solution can be obtained successfully. However, an effort was made to secure more accurate data, using similar substrate dilutions both for the single substrate analysis and the isozyme stripping routine. The results, shown in Table 3.1, represent a definite improvement in the accuracy with 'which the. isozymes can be determined. The enzyme extracts used were unpurified enzyme preparations and probably contain small quantities of other alkaline phosphatase isozymes which would affect the two enzyme models chosen for this work. The results indicate a promising start in the use of heterogeneous enzyme kinetics to obtain both qualitative and quantitative information concerning more than one enzyme present in the same reaction mixture, and provide incentive for further, more definitive studies involving electrophoretic comparisons on systems such as glutamate oxaloacetate transaminase serum isozymes.

The Michaelis-Menton constant, K , is a qualitative property of the enzyme itself, whereas V is representative of the concentration of the enzyme in solution. Thus, one has both a qualitative and a quantitative means of determining the identity and concentration of each isozyme present. Solutions have been presented for measuring Vn and V after ORNL- DWG 73-620RI 160 LINEAR REGRESSION ANALYSIS 1 /rmi = 6.04 mmoles liter" 1 140 /fm2= 0.44 mmole liter" = 0.195 mmole liter"1 min"1 =0.067 mmole liter"1 min"1 120

| 100 u. a>

80 o E E

5 60

0 8 12 16 20 (mmole"1 liter) 5]

Fig. Effect of Substrate Concentration on a Mixture of

Placental and Intestinal Alkaline Phosphatase Isozymes. -156-

Table 3.1. Evaluation of a Known Mixture of Placental Alkaline Phosphatase and Intestinal Alkaline Phosphatase for K and V m max

Si Parameter Value Expected Value Obtained

3 3 K (placenta), M 0.43 x io~ (ref. b) O.36 X 10" ml

3 3 Km (intestine), M 3.04 X io~ (ref. b) 3.19 x 10~

Total activity (V + V ), 0.2k3a 0.239°

-1 -1 millimoles ml min

V (placenta), 0.099k 0.071d 1 millimoles ml min"^*

V (intestine), 0.144* 0.107

millimoles ml min""*"

Percentage placenta 36 30 Percentage intestine 64 70

Sum of V obtained from ratio of placenta and intestine, max

Values obtained using the single substrate computer program.

Sum of values obtained from isoenzyme stripping program.

^Values obtained from computer program for isoenzyme stripping program. -157-

K^ and K^ have "been determined separately. We .are proposing the solution of V , V2, K^, and K^ without any prior knowledge of the two enzymes except that they exist, and that the difference between K^ and K^ under the selected conditions is a factor of 2 or greater. The former method has the advantage that V1 and V2 can be determined using only two different substrate dilutions instead of ten or more; however, it also has the disadvantage that Km must first be determined using purified preparations.

3.1+ References for Section 3 1. J. V. Neel, "The Detection of Increased Mutation Rates in Human Populations," Perspective Biol. Med. 14, 522 (1971). 2. J. V. Neel, T. 0. Tiffany, and N. G. Anderson, in Approaches to Monitoring Human Populations for Mutation Rates and Genetic Disease in Chemical Mutagens, Vol. 3, A. Hollaender, ed., Plenum Publishing Co., New Yor, (1973). -158-

4. HIGH-RESOLUTION ANALYTICAL SYSTEMS

A number of disease states may give early warning of their onset by subtle changes in excretion levels of molecular constituents in physiological body fluids. Prompt medical treatment based upon these early symptoms may enable disease prevention or, at least, minimization of its debilitating effects. Detection and monitoring of these early symptoms will require that research and clinical laboratories have automated, high-resolution analytical systems with which to identify and measure large numbers of the molecular constituents. To attempt to fill this need, several liquid chromatographic systems are being developed. These high-resolution systems, having the advantage of minimal sample preparation, enable a large number of important, nonvolatile metabolites to be monitored with minimal use of skilled technical labor by means of automated data processing methods.

Two analytical systems developed through the prototype phase are currently being evaluated at other laboratories, and the development of several other systems is now under way. Many interesting experimental results have been obtained with these advanced analytical systems.

4.1 Prototype Systems

D. D. Chilcote, W. W. Pitt, and G. Jones Eighteen prototype high-resolution analytical systems based on ORNL developments are being used in other laboratories (Table 4.1). All of these systems, with the exception of the UV-Analyzer at Duke University Medical Center, have exhibited reliable, trouble-free operation during the current report period. The Duke machine has been plagued by repeated -159-

Table 4.1. High-Re solution Analytical Prototype Systems in Other Laboratories

Principal Type of Prototype Laboratory Investigator System

Clinical Center, NIH Dr. Donald S. Young UV-Analyzer, Mark II UV-Analyzer, Mark II-A Carbohydrate Analyzer, Mark II Carbohydrate Analyzer, Mark III University of Texas Dr. R. Rodney Howell UV-Analyzer, Mark II Medical School, Carbohydrate Analyzer, Houston Mark III Duke University Medical Dr. William Kelley Center UV-Analyzer, Mark II-A Eli Lilly Clinical Dr. Robert Wolen Laboratory UV-Analyzer, Mark II-A q University of Salford Dr. R. W. Oliver UV-Analyzer, Mark II-A Chemical Technology Mr. P. B. Orr UV-Analyzer, Mark II Division, ORNL Biology Division, ORNL Dr. James Epler UV-Analyzer, Mark II-A UV-Analyzer, Mark II-A (Modified)a Advanced Waste Treatment Mr. Charles Mashni UV-Analyzer, Mark II Laboratory of ORM-EPA (Modified) University of Oregon Dr. Adam Lis UV-Analyzer, Mark II-Aa Medical School Q University of Cincinnati Dr. I. W. Chen UV-Analyzer, Mark II-A Medical School Erie County Laboratories Dr. Max Chilcote Carbohydrate Analyzer, Mark II (Modified) NCI Dr. T. P. Waalkes UV-Analyzer, Mark II-B13 Southwest Water Lab., Dr. Wayne Garrison UV-Analyzer, Mark II-B EPA fabricated by the other facilities using ORNL construction plans.

•u Being fabricated by ORNL for other agencies. -i6o- pump failures, and a replacement pump was scheduled to be installed during August. The three new computer systems have been constructed and extensively tested. The first system was delivered to the National Institutes of Health during the latter part of June. The use of synthetic urine has proven very useful in the initial shakedown of the computer system after delivery. At present, hardware reliability appears good.

Fabrication of two prototype UV-Analyzers for delivery to outside laboratories was initiated during this report period. A Mark II-A UV- Analyzer is being fabricated for the Environmental Protection Agency, for use by Dr. A. W. Garrison at the Southeast Environmental Research Laboratory in Athens, Georgia; and a Mark II-A UV-Analyzer, modified for rapid analysis for nucleosides is being fabricated for the National Cancer Institute, for use by Dr. T. P. Waalkes. The components for both systems have been ordered, and all of the recorders except one have been received. Fabrication of the systems proceeded in parallel through completion of the electrical wiring. At this point, the major effort was concentrated on the NCI analyzer, which will be delivered about September 1. The other analyzer is to be delivered during October.

In addition to the modifications for nucleoside analysis, three significant component changes are being made. The previously used high- * pressure pumps are being replaced with high-pressure Minipumps . These pumps are considerably lighter and smaller than the previously used pumps *

Registered trademark of Milton-Roy Company, St. Petersburg, Florida. and are rated "by the manufacturer for continuous service at 3000 psi.

Instead of the multipoint recorder that was used in the earlier proto-

types, a recorder with two drag pens is "being installed in these systems.

The third change is the use of an equipment rack fabricated at ORUL from

aluminum angle and sheeting, as opposed to previously used racks with

doors that were purchased from an outside vendor.

4.2 Computer Systems

D. D. Chilcote

Three miniaturized, computer-based data acquisition and analysis

systems have been developed and installed for on-line, real-time analysis

of the data from high-resolution chromatographic systems.

4.2.1 System Description

The computer systems, which are identical, are equipped with 12K words

of fast memory, a dual cassette tape transport, a real-time clock, an

automatic program loader, a digital input/output interface, and an analog

input subsystem (ADC). These interfaces enable the computer to simulta-

neously collect data from two chromatographs while analyzing stored data

from one chromatographic analysis.

Computer programming for the system is facilitated by the high-level

language FOCAL. This language has been expanded to include commands to

specify data acquisition rate, initiate or terminate d?»ta collection

from either analyzer, acc.vs data recorded for either analyzer, write/

read processed results onto/from cassette tape, and access the real-time

clock for program sequencing and timing. Acquisition of data from the

liquid chromatographic columns is performed by an automated, foreground -162-

subprogram which is scheduled to operate periodically by the real-time clock. Double buffering of the input data (two analog signals from photometers and associated digital control inputs) into a block of the memory for each active analyzer is performed by the foreground subprograms at the specified data rate. As the input data blocks are filled, they are written into a ring buffer on the data storage cassette for processing by a FOCAL language program written for an earlier computer system"1" which was modified to operate the new systems. The first computer system was delivered to the NIH on June 26, 1973* This system is presently interfaced to two UV-Analyzers. Although data acquisition from the two systems is often simultaneous, the analysis of the data is always sequential. The second computer system was delivered to Dr. Kelley's laboratory at Duke University on August 28, 1973. Late October is the tentative date set for delivery of the third computer system to Dr. Howell's laboratory in Houston. Each of the last two systems will only be interfaced to a single UV-Analyzer.

4.2.2 Computer Programs Listings of the computer analysis program, the stripping program, and the identification program that are currently being used in these systems are given in Appendix 2. A specific description of the major modifications that have been introduced into these programs is given in the discussion that follows.

Analysis Computer Program. — Numerous changes have been made in the

computer program for the analysis of chromatographic data. This program,

which was designed for the previous computer system, was described in an 1 2 earlier report. ' Some of these modifications were necessary for program operability since certain FOCAL functions used in the new computer system were not present in the former system. Also, some of the functions, such

as the FSUt function, had been modified, necessitating minor format

changes. The new system was designed to access data from a cassette tape

in blocks, rather than individually, from the analog-to-digital converter.

This mode of operation required some extensive program modifications.

9 Certain changes occurred in an attempt to improve the performance of the

program itself. Many of these changes were made possible because of the

greatly expanded capability and versatility of the new computer systems.

In the discussion to follow, the major emphasis will be placed on those

modifications which improved performance of the computer program, and

therefore the computer system as a whole.

The tape cassette has the capability of storing approximately 135

hr of data for a 20-sec sampling interval. Two chromatograms can therefore

easily be stored on one cassette. However, once the data-taking routine

is terminated, it cannot be reinitiated without losing all previous data

stored on the tape. The capability for storage of two 1+2-hr chromatograms

on one tape does exist, but the system is programmed in such a way that

the data-taking scheme cannot be started and stopped repetitively to

stack up runs on a cassette.

Two chromatographic analyses can be placed on a single tape by

starting the data-taking routine and allowing data to be collected con-

I lniiMii;: l.y from the start of the first run until the second run is complete, tn it IM nune, sonv> useless data will also be stored (the data between the

end or IVIIR firot. run and the beginning of the second). However, this

jjoaes no problnm RM long as there is ^ means of exactly locating the start of the second chromatogram. In our case, this can be done quite easily via the key switches for units 1 and 2 on the front panel.

At the start of the second analysis, one of the key switches is turned on, depending upon which unit is in operation. This activates a contact closure in the computer. The switch is left on for at least

5 min. The status of the contact closures is stored on tape, along with the ADC readings at the end of each sampling interval. Therefore, the start of the second chromatogram can easily be determined by searching the data tape for the exact point at which the appropriate contact closure was activated. Since one of the key switches (unit l) is used for an additional purpose, the search routine must verify that the contact closure remained activated for the required length of time.

During initialization of the analysis routine, the operator instructs the computer to analyze either the first or second chromatogram on the tape. If the second chromatogram is to be analyzed, a subroutine is called which searches for the beginning of that chromatogram. This routine locates the data block and the position within that block of the first ADC reading from the second run. Once this reading is located, the computer calculates the (absolute) time at which this reading occurs and outputs the time the chromatographic analysis began. The routine analysis program is then entered and executed.

When the operator initiates the dialysis program, the computer immediately queries the operator for certain input information. Once this information is obtained, the analysis of the chromatogram proceeds until all the data have been examined. No further interaction with the operator is required. The stripping routine, which is stored in a -165- separate file, is called in as it is needed. A rather complete output of information for each peak is generated during operation of this routine.

For each peak, the peak number, elution time, elution volume, flow rate, absorbance ratio, area, time the peak began, time the peak ended, and the baseline are recorded. Also, for each peak, the elution volume, absorbance ratio, flow rate, and area are stored as data for the identification program.

It is often desirable to operate the peak-strip routine on an envelope

of peaks that exhibits minima between the peaks, even though the individual peaks are well overlapped. The current analysis program treats each minimum as an indication of the end of the peak (or envelope, as the case may be). To increase the usefulness of the peak-strip routine, a modified version of the analysis program was constructed. This program allows the

operator to deconvolute any given group of peaks (up to three) in the

chromatogram. The program instructs the computer to request the time the

envelope began, the time it ended, the baseline, the flow rate, and the

elution time and volume of one of the peaks in the envelope. Thus, up to

three peaks of varying degrees of overlap can be analyzed. The initial

input information is obtained from the regular analysis program.

The total number of data points to be examined for each analysis is

determined by the operator. This eliminates analyzing hours of useless

data collected after the chromatographic analysis was completed. The

foreground program has no provision for automatically terminating the

acquisition of data. Since using a short FOCAL program to determine when

to end the data-taking routine would not provide economical use of computer

time, the automatic acquisition of data is normally terminated by the -166- operator via the keyboard. This often results in the collection of much data with no utility.

In general, once the data-taking scheme is initiated, the data analysis routine is also started. Thus the computer analyzes one block of data, then waits (cycling in a loop) until the next block becomes available. This operation, although providing essentially a real-time computer analysis of the data, ties the computer up for the duration of the chromatographic analysis. Some means of temporarily halting the analysis program so that the computer can be used for another application is highly desirable. This was accomplished by means of one key switch on the front panel. The analysis program was modified to repetitively determine the status of the contact closure associated with the key switch for unit 1. The time interval between each determination of contact closure status is such that activation of the contact closure is recognized immediately after the key is turned. Operating the key switch causes the program to execute a library swap upon itself, then halt. The computer is then ready for general use. A library return command entered by the operator via the keyboard causes the analysis program to resume execution at the exact point of termination. In this fashion, the computer is always available for miscellaneous computational work even though it is in the process of analyzing chromatographic data.

During a chromatographic analysis, problems are sometimes encountered

•with the volume dump mechanism. The two most common problems are double dumps and missed dumps. The double dump results in an abnormally high value for the flow rate, whereas the missed dump leads to an abnormally

low value for the flow rate. The program has been modified to check the -167- value of the currently determined flow rate against the value of the previous flow rate. If these values are not within suitable limits

ml/hr), the previously determined flow rate is used until the next flow rate determination occurs. When a double dump is encountered (one extraneous dump), the volume of the siphon device must be subtracted from the total accumulated volume. On the other hand, when a missed dump is encountered, the volume of the measuring device must be added to the total accumulated volume. The computer program is set up to perform this task automatically.

Peak-Strip Computer Program. — Effective use of the stripping routine, 3 which has been described in an earlier progress report, necessitated some programming changes. Since this program was too large to be combined with the analysis program and stored in a single file, it was placed in a separate file and called into fast memory when necessary. When this stripping program is required, a library swap command is executed. This command stores the data analysis program in its appropriate file and calls in the stripping routine. Since all variables associated with the analysis program are erased when the stripping routine is brought into the text buffer, those variables necessary for the operation of the stripping routine must be stored on tape before the swap and then retrieved after the swap. With this particular FOCAL package, the variable field is large enough to allow a three-peak envelope to be deconvoluted without placing the coefficients of the matrix in bulk storage.

The major problem involved ready access to all the smoothed absorbance data for a given envelope. Although the raw data from the ADC were stored on tape, the smoothed absorbance data were not. The smoothed data were -168- V calculated, used, and immediately discarded. Since the stripping routine was set up to analyze the absorbance data, some mechanism for temporarily storing these data was desired. In an on-line, essentially real-time analysis scheme, the program has no way of recognizing whether a peak is an envelope until a second pair of inflection points is located. Thus the smoothed data from each peak must be stored as they are determined until the end of the peak is located. If the peak is merely a single peak, these smoothed data are discarded. However, if an envelope is found, then these data are used in the deconvolution routine before being discarded.

In the current system, word retrieval from bulk storage via an FSTR instruction results in the transfer of a block of data from the tape to the FSTR buffer in memory. After a block is loaded into the FSTR buffer, additional deposits and retrievals of information to and from that particular block require no further (tme-consuming) tape transfers. A tape transfer does not occur until a different FSTR block is requested. Thus, for repetitive FSTR operations, much processing time is saved by utilizing

only one FSTR block, if possible. Each FSTR block contains 128 storage

locations. The stripping routine uses between 30 and 60 smoothed absorbance

readings for deconvolution. All the information necessary for proper

operation of the stripping routine can therefore be placed into one FSTR

block. This is the approach we have taken.

Once a peak is located, the successively determined, smoothed

absorbance readings are stored until the end of the peak is found. If

the peak is an envelope, the appropriate absorbance readings are trans-

ferred to another FSTR block for subsequent use in the stripping routine.

If the envelope contains less than 60 data points, all are used. For 60 -169- to 89 data points, every other reading is used; for 90 to 119 data points, every third reading is used, etc. Thus the time required for the numerous retrieval operations that occur during operation of the peak- strip routine is not unnecessarily lengthened by the abysmally slow tape transfer operation.

Once the peak-strip analysis is complete, a library return instruction retrieves the analysis program from bulk storage. The swap command is designed such that the analysis program, once it is called back in, resumes execution automatically. The peak information that is recorded from the peak-strip routine includes the peak number, elution volume, area, least- squares sum, and total envelope area, which is determined numerically during execution of the analysis program.

The operator has the option of using or not using the peak-strip routine. When the peak-strip routine is not used, the peaks are numbered consecutively. Thus, it is not apparent from the printout how many enve- lopes were located. This simplifies storage, retrieval, and matching of information for the peak identification routine.

Peak Identification Computer Program. — The identification program has also been modified, and now allows new identifications to be inserted k with ease. In the old program, the list of known peaks had to be in the order that they were eluted from the column. In the present case the known peaks can be in any order. The chromatogram is divided into four sections,

and the program works on a section-by-section basis. 3 The sectioning of the chromatogram was discussed in a previous report. For any given normalized elution volume, the program instructs the computer to scan all

known normalized elution volumes in that section for a possible match. -170-

Each storage location (on tape) that contains a known normalized elution volume corresponds to a numbered program line which contains the appropriate identification information for that compound. When a match is noted, the line containing the identification information for that peak is executed. Line numbers 3*01 to 3-^9 and storage locations 513 to 576 are for section 1, lines 3-51 to 3*99 and locations 577 to 64-0 are for section 2, lines 4.01 to 4.49 and locations 64l to 704 are for section 3, and lines 4.51 to 4.99 and locations 705 to 768 are for section 4. Thus^ two FSTR blocks are required for the identification program. However, all known normalized elution volumes for any given section are always stored on a single block, no crossover is permitted, and tape transfer operations are minimized. In our scheme, storage location 513 corresponds to line number 3.01, 514 to 3-02, 577 to 3*51, etc. Although 64 storage locations are specified for each section, only the first 50 can be used since only

50 program lines are allocated to each section.

4.3 Systems Development

S. Katz, H. Veening, D. D. Chilcote, and G. Jones

We are continuing our efforts relative to the development of high- resolution separation systems and more sensitive detection devices for application in biomedical research and clinical chemistry. As new developments occur in these areas, their applications to practical problems in the separation and detection of constituents present in physiologic body fluids are evaluated. -171-

4.3*1 Cation Exchange Separation of Ninhydrin-Positive Constituents

A compact ninhydrin-positive compound analyzer has "been developed for use as a high-resolution system for separation and detection of amino acids found in urine and blood sera.

A 0.45- by 200-cm column packed with Aminex A-7 (cation exchange resin with a particle siz.e of 7 to 11 Bio-Rad Laboratories, Richmond,

Calif.) is used for urine analyses. A 0.22- by 300-cm column packed with

Aminex A-6 (cation exchange resin with a particle size of 15 to 19 n) is used for the blood sera analyses. The remainder of the analyzer (Fig.

4.1) consists of: (l) a nine-chamber vessel for generating the eluent buffer concentration gradient; (2) a high-pressure (4000-psi) pump;

(3) a high-pressure, six-port, sample injection valve mounted on top of the column to minimize peak broadening; (4) a mixer-reactor where the column effluent is mixed with the ninhydrin reagent in a jet mixer and then passed through a coil of Teflon tubing immersed in boiling water;

(5) a reagent metering system; (6) a colorimeter which monitors the reaction mixture at 440 and 570 nm; and (7) a strip-chart recorder.

Using this equipment, a series of experiments was made during which the eluent concentration, pH gradients, temperature profile, flow rates, and sample volume were optimized to obtain the greatest number of peaks from samples of either a composite of "normal" urines or "pool" sera. 5 Columns. — 13ie columns were slurry packed with the Aminex resins which were supported on the stainless steel frit (0.5-H pore size) held on a reducing union. The resin was converted to the citrate form by mixing with the pH 5*25 buffer and filtering before packing the column.

Jackets welded around the column allowed the column temperature to be ORNL DWG 73-9966

2b. X X X X X X xlx 6-PORT SAMPLE NINE-CHAMBER INJECTION VALVE GRADIENT BOX PRESSURE O GAGE

HIGH - PRESSURE CHROMATOGRAPHIC HIGH COLUMN PRESSURE PUMP y -C3

I JET MIXER STRIP V CHART RECORDER IO-m COIL OF 0.04-cm-l D TEFLON TUBING

WASTE

L REACTION C. VESSEL ORNL COLORIMETER

Fig- 4.1. System for Separation and Detection of Ninhydrin-Positive

Compounds in Physiologic Fluids. -173- maintained at the desired 53°C with a constant-temperature circulator.

Several different column geometries were used in the tests, but the best results were obtained with the 0.4-5- by 200-cm column for urine and the

0.22- by 300-cm column for the blood sera samples. The flow rates, sample volumes, and reactor volume were all appropriately scaled to the column geometry.

The strongly basic cation exchange resins, Aminex A-5, -6, and -7 were investigated as column packing. The A-7 resin (7 to 11 p) was found to give the best resolution and hence was used for the urine samples.

However, with the smaller-cross-section, longer column used for blood sera samples, the pressure required to obtain the desired flow rate exceeded the capability of the pump when A-7 was used. Consequently,

A-6, which has greater permeability, was found to be more advantageous m for the sera samples.

Eluent Gradient. — An eluent buffer that increases in pH and in sodium citrate concentration is produced with a nine-chamber gradient generator. The various buffers needed for urine and sera sample analyses 6 7 were prepared by accepted procedures. 3 Tables 4.2 and 4.3 show how the individual chambers of the gradient generator box are filled for the two types of analyses. 3 The Mixer-Reactor. — The annular jet mixer previously described was used to mix the column eluate and ninhydrin reagent. This mixture immediately flows through the reactor, which consists of a coil of 18- to

20-gage Tefon tubing immersed in boiling water. The volume of the Teflon tubing was varied, depending on the volumetric flow rate, but was chosen so that the residence time for the eluate-reagent mixture would be approximately 20 to 27 min. -174-

Table 4.2. Individual Chamber Loading of Gradient Generator Box for Urine Analysis on 0.45- by 200-cm Column

Sodium Citrate Buffer (ml) 0.2 M, 0.2 M, 1.08 M, Chamber pH 2.95 pH 3-25 pH 5-25

1 100

2 100

3 100

4 78 22

5 68 32 f cr\ o 50

7 20 80

8 100

9 100 -175-

Table 4.3. Individual Chamber Loading of Gradient Generator Box for Sera Analysis on 0.22- by 300-cm Folded Column

Sodium Citrate Buffer (ml) 0.2 M, 0.2 M, 1.08 M, Chamber pH 2.95 pH 3-25 pH 5-25

1 36

2 36

3 36

k 30 6

5 26 10

6 18 18

7 6 30

8 36

9 36 -176-

Reagent Metering System. — The ninhydrin dissolved in dimethyl sulfoxide, with stannous chloride added as an oxidizing agent,^ is contained in a 5-liter glass vessel which is painted "black and has a 1-in. sight strip for viewing the liquid level. A nitrogen overpressure of approximately 8 psi and a 10-m length of 0.0k-cm ID Teflon tubing were used to obtain a steady flow of reagent at 6 ml/hr. Increased metering reliability and an exceedingly smooth flow of reagent were achieved, thus improving the resolution, as compared to that obtained with a reagent pump.

The ninhydrin reagent was prepared in the following manner:

(1) Place 3 liters of dimethyl sulfoxide in a 5-liter glass

vessel.

(2) Add 1 liter of lithium acetate--acetic acid buffer. (The

lithium acetate—acetic acid buffer contains 168 g of

lithium hydroxide monohydrate and 295 ml of glacial acetic

acid per liter.)

(3) Deaerate the solution with nitrogen.

(4) Add 80 g of ninhydrin and stir until dissolved. (Ninhydrin

should be weighed in hood.)

(5) Add 1.6 g of SnCl2.2H20 and stir until dissolved.

Colorimeter. — A dual-beam, dual-wavelength flow colorimeter based 8 9 on the same design as those previously developed at ORNL ' was used.

Interference filters were used to achieve the desired wavelengths of 440 and 570 nm. Since the citrate buffer does not contribute to the detected signal and thus does not cause a baseline shift, the reference flow cell was replaced with a quartz rod. The electronic and flow noise of the colorimeter was such that a full-scale absorbance of 0.05 could be used. -177-

Resuits. — Figure 4.2 shows typical examples of a chromatogram for a composite urine sample from eight normal male subjects (top chromatogram) and a chromatogram obtained from a pooled blood serum sample (bottom chromatogram). As many as 143 chromatographic peaks have been resolved from a single urine sample, while over 50 chromatographic peaks have been resolved from serum. A number of amino acids have been identified based on cochromatographic studies with reference standards.

4.3*2 Liquid Chromatographic Analysis of Blood Serum

When compared with ordinary specific methods, multicomponent determinations in a single analysis of blood serum appear to offer distinct

advantages with regard to decreased sample volume, cost, and time required.

Such multicomponent determinations have been demonstrated to be useful

in the cases of steroids"1"0 and indoles,and offer an additional

advantage of providing internal comparisons within a sample of co-measured

constituents.

The separation of urinary constituents such as indoles, aliphatic

acids, and aromatic acids by high-resolution ion-exchange chromatography 13 is well established, ^ and many of these are known to be indicators of

the state of health. However, application of liquid chromatography (LC) 14 to the analysis of blood serum has not been aggressively pursued. The 15 application of the cerate oxidative detector in series with the uv

detector provides additional, sensitive detection capability to an LC

system for analysis of blood serum. The utility of this system as a

blood serum analyzer is being evaluated.

The Chromatographic System. — The chromatographic system and its operation are similar to that described earlier for urine analysis. 15

-178-

ORNL DWG 73-9974RI

i T 1 1 1 n SAMPLE: 0.76 ml POOLED SERUM COLUMN: 0.22 x 300 cm WITH AMINEX A-6 (15-19/i) ELUENT RATE: 14 ml/hr I NINHYDRIN REAGENT: 6 ml/hr TEM PERATURE:55°C INITIAL PRESSURE 3 500 psi 570 nm

J i J 1 12 13 14 15 16 t >rmal Urine and Pooled Serum for Ninhydrin-Positive

I -179-

The volume of sample in the sample loop has been increased from about

0.3 nil for urine samples to 0.6 ml for ultrafiltered blood serum; larger volumes appear to introduce peak broadening, especially in the case of early-eluting compounds. Several interrelated parameters were changed to permit the total elution to be completed in less than 20 hr with minimal effect upon the separation. These changes are: (l) the column length was decreased from 150 to 100 cm; (2) the eluent buffer concentration gradient was reduced in total volume from 360 to 225 nil, and prepared in a nine-chamber gradient box with 25 g of acetate buffer solution (pH U) in each chamber (0.015 M in 1 and 2, 4 M in 3 and k, and

6.0 M in the remaining chambers); and (3) the column temperature x?as increased to 6o°c at 1.3 hr and for the remainder of the elution time.

Preparation of Serum Samples. — The colloids and very large molecules are separated prior to analysis by ultrafiltration through dialysis tubing with an inert-gas overpressure of 15 psi. About 80$, of the volume of the sample passes through the tubing during a l6-to 20-hr period and is collected in a refrigerated receiver. Samples are frozen until use.

Additional concentration is possible through freeze-drying but is limited to about a factor of 5 by the large volume of insoluble solids.

Greater concentration can be achieved by passing the ultrafiltered serum through a weak cation exchanger (Rexyn 102 H, Fisher Scientific, Fairlawn,

N. J.) prior to freeze-drying; however, some alteration of the subsequent chromatograms has been noted, indicating that sample changes may be occurring in the concentration process.

Results. — Chromatograms obtained from serum samples of a single individual after overnight fasting, 30 min after ingesting 100 g of glucose, and 1, 2, and 3 hr after ingesting the glucose (Fig. ^-.3)^ demonstrate the unique ability of this analysis to measure the effect of a single parameter upon many of the blood constituents. In excess of 30 uv peaks and 30 COD (cerate oxidative detection) peaks are found in each of the chromatograms with more than 50 constituents detected. Hippuric acid and several other aromatic acids were tentatively identified by their elution positions; these have not heretofore been detected in blood. The levels of detection, of the order of 0.01 to 0.1 |Jbg per cc of blood, should be useful in defining the absence of such important indicating compounds as 5-hydroxyindoleacetic acidr The larger peaks that appear have been verified as pseudouridine, tyrosine, hypoxanthine, uric acid, and pyruvic acid by collection of the peak-associated eluate in each case and subsequent rechromatography on a cation exchange column. Smaller peaks coincide, in a number of cases, with the elution positions and relative responses of known aliphatic and aromatic acids. A large sample of 500 cc of pooled blood serum has been collected and is being processed for concentration on the preparative UV-Analyzer prior to identification studies.

J+.3•3 Application of Fluorescamine to Monitoring of Urinary Polyamines

Studies have continued in an effort to adapt the reagent known as fluorescamine (Fluram ) to a sensitive analysis for polyamines in physiologic body fluids. A useful ion exchange separation of these compounds was previously developed, and efforts have continued to

Registered trademark of Roche Diagnostics, Nutley, U. J. -181-

ORNL DWG. 73-8626

Fig. 4.3- Analyses of Several Blood Serum Samples "by the W-

Analyzer with a Cerate Oxidative Detector. -182- improve this separation and adapt it for use with Fluram in an automated, complete system.

The ^olyamines generally include the following compounds: 1,3-diaminopropane, CH^WH^s putrescine, CHg^NHgS cadaverine, H2N( CH2)^m^; spermidine, H2N( CHg^NHC CH2)NH2; and spermine, H2N( GH2)CHg)CH2)^NK^•

Fluorescamine has "been reported to form a highly fluorescent product with 16 primary amines, and has "been demonstrated to he two orders of magnitude more sensitive for the detection of picomole quantities of amino acids in 17 amino acid analysis than the standard ninhydrin procedure. A manual fluorometric assay for nanogram quantities of proteins utilizing Fluram has 18 also "been reported. This reagent (which itself is nonfluorescent) reacts directly with primary amines in aqueous medium (pH 9 to 10) at room temperature, with a half-life of a fraction of a second. Excess reagent is destroyed "by water, with a half-lif1R e of several seconds, to form nonfluorescent hydrolysin products.

Chromatographic System. — Figure b.k is a schematic flow diagram of the ion exchange chromatograph which was constructed. Three Milton-Roy

Minipumps were used to pump the column eluent, the buffering agent (0.1 M

H-^BO^), and the fluorescamine solution through the system. The first four chambers of a nine-chamber gradient box supplied the eluent, which was pumped through a six-port sample injection valve supplied with a 0.^5~ml sample loop. A jacketed, nickel column was used in this investigation; its dimensions were 0.^5 cm (ZD) by 15 cm. A nickel column was used in order to minimize the possible corrosive effects of sodium chloride solutions. The column was slurry packed with Aminex A-5 resin (13- to 20-JJ. particles) which was supported by a stainless steel frit. The column was operated at 70°C. -183-

ORNL-DWG 73-6901A

Fig. b.k. Ion Exchange Chromatographic System for the Separation and Fluorescence Detection of Polyamines. -519-

In order to form Lhe fluorescamine derivatives of the polyamines at the optimum pH for fluorescence measurement, it was found necessary to readjust the pH of the column effluent by mixing it with 0.1 M boric

acid solution in an annular mixer as shown in Fig. k.k. The buffered

stream was then fed to a T-mixer supplied with a small magnetic stirring bar, where the fluorescamine reagent stream was introduced. A fluorescence

flow monitor was used to detect components eluting from the column.

The chromatographic system was provided with a back-pressure device at

a point beyond the detector in order to provide sufficient back-pressure

(4o to 50 psi) to prevent gas bubble formation when acetone was mixed with the aqueous stream at the T-mixer. This device consisted of a 45- by 0.45-cm stainless steel column packed with finely divided glass beads.

Fluorescence Properties of Fluorescamine-Polyamine Derivatives. —

Several experiments were carried out in order to determine the optimum

conditions for chromatographic operation. Activation and fluorescence

spectra were recorded for each of the fluorescamine derivatives of the

polyamines. In all cases, the maximum excitation and emission wavelengths were 395 nm and 475 nm, respectively, at a pH of 10.5.

The effect of pH on the fluorescent emission of the five polyamine-

fluorescamine complexes was investigated; the results are given in Fig.

4.5- It is apparent that fluorescence intensities for spermine and

spermidine are considerably lower than for the other three compounds.

The maximum fluorescence yield for putrescine, cadaverine, and 1,3-

diaminopropane is developed within the pH range 9 to 10, whereas the

optimum pH for spermine and spermidine is 10.5- These data suggest that

optimum sensitivity, using this reagent for the detection of all polyamines,

would be attained at a pH of about 10. -185-

ORNL DWG 73-6899 9.0 t 1 1 r—* 1 1 > EXC. WAVELENGTH: 395 nm EM. WAVELENGTH: 475 nm CONC. POLYAMINES: I ug/ml leach) ^ L PUTRESCINE tCONC. FLURAM: 100 ftq/ml

Fig. Relative Fluorescence Intensity us a Function of pH for

Several Polyaraines. (Measurements made manually on an Arainco-Bowman spectrophotofluorometer.) -186-

The effect of excess reagent on fluorescence intensity of the polyamine fluorophors was also studied. This was done by preparing a series of solutions, each containing 1.0 yxg/ml of polyamine, and by varying the weight ratio of fluorescamine to polyamine. The pH was held constant at 10.5' The results of this study illustrate that the curves for putrescine and cadaverine coincide (Fig. 4.6). Maximum fluorescence intensities in all cases occur at rather high fluorescamine/polyamine weight ratios (UCO/l or 500/1). Several chromatographic runs were carried out using a fluorescamine concentration of I4OO p-g/ml; however, the gain in sensitivity did not warrant the high cost associated with increased use of the reagent. Accordingly, the fluorescamine concentration utilized in the chromatographic system was 150 t*g/ml.

pH Gradient and Buffer Stream. — The pH gradient used to separate the polyamines was determined experimentally by monitoring aliquot fractions of eluent at the column exit (Fig. 4.7)' Since the polyamines have retention volumes ranging from 27 to 65 ml, it is apparent from the data in Fig. 4 .7 that the pH levels in this region of the gradient lie well outside the optimum pK region necessary for achieving maximum fluorescence intensities for the flaorophors resulting from the reaction between polysmines end fluorescamine. To improve sensitivity, the column effluent was buffered with 0.1 M boric acid. At a boric acid flow rate of 5»0 ml/ hr, the pH of the column effluent was lowered sufficiently so that maximum fluorescence intensity for all components could be achieved. The pH curve resulting from periodic measurements of the combined streams during gradient operation is also shown in Fig. 4.7' -187-

Flg. h.6> Effect of Fluorescamine Concentration on the Fluorescence

Intensity of Several Polyamines. (Measurements made manually on an

Aminco-Bo-wman spectrophotofluorometer.) -188-

ORNL DWG 73-6933R3

I .1 i I I 1 I I 1 I I 1.1 1 1 0 12 24 36 48 60 72 84 ELUTION VOLUME (ml)

Fig. h.7. curves Showing pH Gradient Used for the Separation of

Folyamines and pH of the Column Effluent After Suffering with 0.1 !J

H3B03 at a ?lov; pate of 5-0 ml/hr. -189-

Chromatographic Separations and Quantitative Response. — A typical separation of a standard mixture containing known amounts of the five polyamines is shown in Fig. 4.8. In this particular experiment, the boric acid stream was not employed. The resolution was satisfactory except for spermidine and cadaverine; however, even with the disappointing results for these two components, it is still possible to make quantitative measurements. The entire analysis can be completed in less than 2 hr under the conditions cited.

It was found that the fluorescamine method was sensitive for 1,3- diaminopropane, cadaverine, and putrescine. The use of the boric acid stream to lower the pH of the effluent stream markedly increased the sensitivity for these three compounds, as shown in the quantitative standardization plots in Fig. 4.9* The fluorescence response for each compound was first measured without the boric acid flow, and then a second time using a boric acid flow of 5»0 ml/hr. In all cases, peak heights were used. Lowering the pH of the column effluent effectively increases the response approximately twofold for each of the compounds in accordance with the results obtained previously in Fig. 4.5* The minimum detectable quantity for each of these compounds was 100 ng for 1,3-diaminopropane and cadaverine, and 50 ng for putrescine.

The responses for spermidine and spermine were considerably less.

The minimum detectable quantities for these two compounds were found to be 1.0 and 3»0 ng, respectively. Hence the effect of lowering the pH of the column effuent on the fluorescence intensity for these two compounds was only slight. -190-

ORNL DWG 73-6933R3

UJ z 100 Q ^ S tt 2 UJ 0C UJ Q_ 90 UJ Z CO UJ ?£§

10 1 30 60 90 120 TIME (min ) ± J I i L_ 18 36 54 72 ELUTION VOLUME (ml)

Fig. ^-.8. Chromatogram of a Standard Mixture of 12.0 /ag of

1,3-Diaminopropane, 3-6 fag of Putrescine, 15-0 \±g of Spermidine,

7*0 tag of Cadaverine, and 35 «0 ng of Spermine. -191-

ORNL DWG 73-6934 T T T WITH H3BO3 FLOW PUTRESCINE WITHOUT H3BO3 FLOW

400h

300

x CD UJ X <

UJ Q. 200

100

,3- DIAMINOPROPANE

—A- ____ .—DIAMINOPROPANE

1.0 2.0 3.0 MICROGRAMS INJECTED

Fig. 4-9' Standardization Plots for Putrescine, Cadaverine, and

1,3-Diaminopropane With and Without the Effect of Stream Buffering. -192-

Several "urine samples were analyzed for polyamines. Typical examples of urine from a clinically normal subject and urine from a cancer patient are shown in Fig. 4.10. In the chromatogram of a hydrolyzed, composite normal urine sample obtained from eight normal male persons

[Fig. 4.10(a)], peaks 3, 5, 6, and 7 appear in the normal retention positions for 1,3-diaminopropane, putrescine, spermidine, and cadaverine, respectively. A smaller unknown peak (No. 4) also appears. The chromatogram in Fig. 4.10(b) shows a hydrolyzed urine sample from a cancer patient. Peaks 5 ana 6 represent components located in the normal retention positions for putrescine and spermidine. Based on the previously established calibrations, the normal sample contains 0.7 M*g of putrescine per milliliter, 1.5 Pg of spermidine per milliliter, and a trace of cadaverine while urine from a cancer patient contains more than 2.0 Hg of putrescine per milliliter and 4.0 p,g of spermidine. It can be seen in these urine chromatograms that the amino acids and other less basic components elute early using these chromatographic conditions, leaving a reasonably clear area for the elution of the polyamines.

The separation procedure described for the polyamines is rapid and satisfactory. The sensitivity of fluorescamine toward 1,3-diaminopropane, putrescine, and cadaverine was acceptable for use in the analysis of physiological fluids; however, the detection capability for spermidine and spermine was disappointing. It is interesting to note that the latter two compounds are the only ones with secondary amine groups (spermidine has one, spermine has two). Despite the presence of the two terminal primary amine groups in each of these molecules, the secondary amine groups apparently either interfere with the fluorescamine reaction or have a ORNL DWG 73-9971

Fig. 4.10. Chromatograms of Polyamines in Urine from

(a) Clinically Normal Subjects and (h) a Cancer Patient. -529- quenching effect on the activation of fluorescence in the polyamine- fluorescamine derivatives.

4.3*4 Analysis of Nucleosides and N-Bases

The system for analyzing nucleosides and N-bases, as previously 4 2 developed was not capable of separating 7-methylguanine and N - dimethylguanosine. Both compounds are present in human urine, and the level of dimethylguanosine is suspected to be elevated in the urine 19 of cancer patients. ^ Therefore, the quantitation of this compound is of interest. In an .effort directed toward attaining a separation of these two compounds, the effect of pH on the separation was studied first. This was done by adjusting the pH of the initial eluent (dilute acetate buffer) with either acetic acid or ammonium hydroxide. The 2result s of chromato- graphing reference samples of 7-methylguanine and N -dimethylguanosine with eluents of various pH values are shown in Fig. 4.11. Decreasing the pH from a value of 4.2 gave a better separation than did increasing the pH from this value. Thus, lowering the pH of the eluent appeared to be the best approach. However, this has to be considered in the context of the optimum separation among all sample constituents eluting in this

chromatographic region.

After several tests, an eluent with a pH of 3*7 was prepared by

diluting (1:400) a concentrated ammonium acetate—acetic acid solution

(50 ml of NH^OH + 345 g of acetic acid per liter). When used in the

system, this eluent gave an acceptable separation. The operating conditionk s or parameters that differed from those described in an earlier report

included isocratic elution at pH 3-7 and a column temperature profile

in which temperature was increased to 6o°C over 1.5 or 2.5 hr. -195-

ELUTION VOLUME (ml)

Fig. lull. Effect of Eluent pH on Elution Volume of N3-Dimethyl- guanos ine and 7-Methylguanine. -196-

The variation of the elution volumes of many nucleosides and K-bases as a function of pH and temperature profile is given in Table k.h. The eluents used in this study were prepared by diluting stock buffer solutions which were 6 M in total acetate and had variable ammonia concentrations. k.3*5 Separations Systems

The need for rapid analysis of individual constituents in physiologic fluids is an obvious and continuing requirement in the clinical laboratory.

In addition, there is an overwhelming need for the development of high- resolution, quantifiable separations for high-molecular-weight compounds

such as proteins. Speed and analytical specificity are major requirements for acceptance of new clinical methods, and they furnish the impetus for development studies in the field of new separations systems.

Multicomponent Separation by Affinity Chromatography. — A preliminary

study was performed to determine the feasibility of utilizing affinity chromatography to achieve multicomponent separation in a single analysis.

A system was desired wherein several components of a sample could be removed in a common moving phase (the washing buffer) by a support medium to which specific antibodies had been attached, and from which

the components might subsequently be detached or "eluted" sequentially.

Our studies were carried out with an agarose support to which the

antibodies of either albumin or IgG had been attached. For purposes of

this study, the eluent buffer was selected by a solenoid valve, forced

through a sample injection valve and the affinity column by a peristaltic

pump, and then monitored at 280 nm by a uv photometer. The column was

15 mm ID and cooled to ice-water temperature. -197-

Table 4.4. Elution Volume (ml) of Nucleosides and H-Bases as a Function of pH and Temperature Profile

3.79 4. 20 4. U0 Rate of Temp. Change, °c/min 0.39 0.23 0.39 0.23 0.39 0.23

Cytidine 18.6 18.4 20.5 19.4 22.7 21.3

Uracil 30.7 31.0 31.4 30.8 31.8 31.7

Uridine 36.3 36.8 37-1 36.6 37.5 37.6

1-Methylinosine 42.5 ^3.6 43.2 43.7 43.9 45.7

7 -Me thy lxa n t h ine 47.1 48.0 49.I 48.7 50.3 50.6

Hypoxanthini; 51.7 53-0 52.6 53-4 53.3 55.0

Inosine 54. 4 57,5 54.9 57.9 55.8 60.1

1-Methylguanosine 59-3 62.5 59-7 62.7 60.1 64.6

7 -Me thy 1 guan i ne 63.0 66.5 67-0 72.0 67.7 76.1 2 N -Dimethylguanosine 66.2 71.7 67-0 72.0 67.7 75

3-Me thylxanthine 73-2 76.8 75.5 78.5 73.2 83.4 2 N ,7- Dime thylguanine 80.7 85.6 89.O 9^.2 —

Xanthine _ _ 82.7 __ —_ 8l. 1 97.0 -198-

A phosphate buffer (pH ?.C) was selected as the washing buffer, and a sodium perchlorate solution (2 as the eluting buffer, since these were found to be compatible with the two protein-antibody pairs selected for testing. Sodium aside (0.1c£) was added to prevent growth of organisms in the buffer. The tests showed that an analytical peak could be obtained with either component in as little as 45 min, with the analytical peak e.'.uting immediately after the "washing buffer was replaced with the eluting buffer. Tests with albumin showed that dilution of the eluting buffer by -i factor of 10 made very little difference in elution volume, and further dilution caused only a broadening of the analytical peak but did not effectively delay the elution. The results indicated that sequential elution of these two proteins from columns in sequence could not be achieved using this washing buffer.

A new design configuration, whereby multiple analyses could be performed, was tested. This design involved close coupling of affinity and centrifugal chromatography. The proposed system design includeJ a rotor that contains a number of columns, each with a different antibody coupled to its support, and utilizes a single flow system and detector to r-erforrn multiple protein separations from a single sample. As described If previously, centrifugal force would be used to move the eluent through

the columns with the eluate from each column being monitored by a single,

stationary photometer. In a test of a single tube mockup of this design, adequate flow through a column of sufficient capacity for detection was

achieved. Thus, feasibility of the design was demonstrated.

Preparation of Pellicular Gel Filtration Media. — Gel filtration presents an attractive approach to the separation of large macromolecules -199- that vary in size. Unfortunately, current techniques which utilize gel filtration for separation of important biochemical macromolecules have relatively poor resolution (e.g., as compared to high-resolution ion exchange chromatography); the eluted bands are quite broad. This is largely due to the extremely low diffusivity of the large molecules into the gel particles, which results in poor mass transfer between phases and leads to noticeable peak dispersion. The use of small particles per se is not an acceptable alternative since the gel spheres are too compressible to withstand the high pressure drop across the column that is necessary to maintain a reasonable flow rate. For these reasons, a small gel particle with rigid structural properties is desired.

The notable success of pellicular resins in ion exchange and adsorption chromatography suggests that this approach should be considered for gel filtration as well. A solid core would provide mechanical stability, and a thin film of gel would minimize adverse mass transfer effects. Finding a method for applying a uniform gel layer to a solid core is the problem that needs solution.

One possible solution would be to form a slurry of the liquefied gel and the solid core particles, and then nebulize this slurry to disperse the particulates (now coated with a thin film of gel) into an environment conducive to rapid gelation. For example, when using agarose gel, the hot slurry would be sprayed into a cold room to quickly set the gel layer, allowed to settle into a cold-water bath, and collected. The feasibility of nebulizing a slurry of solid particles approximately 0.1 to 1.0 in diameter has already been demonstrated in an entirely different appli- 20' cation. -200-

Preliminary investigations have resulted in the successful nebulization of slurries of hot agarose gel and smoll glass spheres (44 to 57 The atomised slurry was collected in an ice-water bath to promote rapid gelation. A cold room would be preferred since it would eliminate any possibility of dissolution of the gel film in the water. Unfortunately, the nebulizer used during the initial phase of this investigation resulted in an excessively high volumetric flow rate of the slurry. Consequently, large gel plaques formed on the surface of the cold-water bath. Attempts to reduce the air pressure, which would, in turn, decrease the flow rate, were unsuccessful. Lowering the flow rate caused the linear velocity of the hot gel in the capillary tube leading to the orifice to fall below the limit at which the glass spheres would be carried in this vertically moving stream. The nebulizer must therefore be redesigned to incorporate capillary tubes with smaller inside diameters. This may necessitate the use of smaller glass spheres to avoid plugging the orifice, a change which will allow a sufficiently high linear velocity of the slurry in the capillary tube but will keep the volumetric flow rate fairly low. Work is continuing along these lines.

A neutral resin, XAD-2 (20 to 50 M<), has also been coated by this method. The lower density of this material allowed the use of sufficiently low flow rates with the current nebulizer to avoid the formation of gel plaques on the surface of the water bath. However, the problem of plugging the nozzle was quite severe with these larger particles. In addition,

separating the gel spheres (formed during nebulization of the slurry) from

the coated neutral resin by gravity was difficult. Use of glass spheres

as the solid core eliminated this problem entirely. -201-

Chromatographic Analysis of Creatinine and 1-Methylnicotinamide. —

Numerous modifications of the basic Jaffe reaction ha.ve been used for the routine clinical analysis of creatinine in serum and urine. This approach involves measuring the color produced when picric acid and sodium hydroxide solutions are combined with urine or a protein-free filtrate from serum.

Unfortunately, many substances other than creatinine also react with these reagents, introducing a positive error into the analysis. This interference is normally much greater in serum than in urine.

Two major modifications have been introduced because of the high probability of interference. One has been to replace the end-point 21 determination with a kinetic analysis. Alternatively, a cation exchange resin has been used to isolate the creatinine in the sample prior to the 2 end-point determination. The first approach assumes that the initial rate of reaction is due almost solely to creatinine (i.e., the interfering substances have significantly lower reactions than creatinine). This approach demands a highly accurate, automatic kinetic measuring device.

The second approach assumes that the interfering compounds either do not adhere to the cation exchange resin or do not elute with creatinine.

However, since creatinine elutes quite late in the normal amino acid

analysis using a strongly acid cation exchange resin, it is reasonable to

assume that an eluent strong enough to remove creatinine will also probably

remove most of the other biochemical constituents adhering to the resin.

Since a cation exchange resin is used to prepare the sample for

subsequent analysis, it is quite logical to investigate the feasibility

of determining creatinine photometrically in the effluent stream during

the elution step. This would greatly simplify the overall procedure by -202- entirely eliminating the Jaffe reaction step. As mentioned above, a strong eluting buffer is essential for the rapid removal of creatinine.

The use of such an eluent has additional advantages with respect to developing a rapid chromatographic separation scheme for creatinine. The constituents that are less tightly held would be quickly removed by this buffer, which would have a tendency to minimize the chromatographic interference of other compounds with creatinine. Use of this buffer would also elute essentially all the compounds adhering to the column, thus obviating the need for a regeneration step.

To develop a successful chromatographic analysis scheme for creatinine,

two strict requirements must be met: (l) the detection limits must be reasonable, and (2) the eluted creatinine pet-k must be free from coeluting, uv-absorbing compounds. Preliminary investigations indicate that both of

these criteria can be met. In addition, 1-methylnicotinamide, eluting

immediately after creatinine, can be determined quantitatively. Tryptophan, which elutes just prior to creatinine, can also be determined by this method; however, because its molar absorptivity at 254 nm is low, the peak

in normal urine is quite small.

The chromatographic system set up to analyze urine samples for

creatinine consisted of a 0.22- by 25-cm jacketed stainless steel column

packed with Aminex A-6 resin, a high-pressure pump, a pressure gage, a

sample injection valve, and a uv photometer. A 0.5 M ammonium acetate—

acetic acid buffer solution, pH 4.4, was used as the eluent. Gradient

elution was not required. Operation of the system at 50°C and 1800 psig

produced a volumetric flow rate of 47 ml/hr. Under these conditions,

creatinine eluted at 22 min, tryptophan at 11 min, and 1-methylnicotinamide

at 36 min. The eluate stream was normally monitored at 254 nm. -203-

Use of a more concentrated eluent buffer (l M in acetate) decreased the elution volumes of the compounds of interest by a factor of two but

also caused some noticeable overlap of peaks. The use of the 0.5 M buffer gave baseline separation between the peak for creatinine and adjoining peaks. The maximum peak height of creatinine from 100 nl of normal urine was 80% of full scale deflection (0.05 absorbance units) when monitored

at 254 nm.

When the effluent stream was monitored at 280 nm, only a small base-

line undulation occurred at the elution position of creatinine. Since the molar absorptivity of creatinine at 280 nm is quite low, this observation would be expected if no interfering species were present. However, this

condition is neither necessary nor sufficient to prove noninterference.

A coeluting compound that absorbs strongly at 280 rim but very weakly at

254 nm could be tolerated. This obviously was not the case here.

Furthermore, our experience with the UV-Analyzer suggested that compounds which absorb strongly at 254 nr> but only weakly at 280 nm are not common;

hence our case for the lack of interferences is strengthened.

The described chromatographic procedure is suitable for clinical

laboratory use. Use of data processing, automated sample injection, and

multiple columns would provide rapid analysis for creatinine and N"1"-

methylnicotinamide, freeing skilled technical labor for other tasks.

4.4 Identification of Body Fluid Constituents

J. E. Mrochek and S. R. Dinsmore

Qualitative identification of the biochemicals separated by high-

resolution liquid chromatography is a necessary component of the Body

Fluids Analysis Program. Several new uv-absorbing compounds have been identified, -which brings the total number of identified constituents to

date to 127 (if carbohydrates are included, the total is 139)-

4.4.1 New Identifications

Eight new uv-absorbing compounds have been identified. Table 4-5

lists the new identifications, together with origins of the urine samples

from which they were isolated and identified. A ninth compound, listed

in Table 4.5 as a metabolite of 6-thioguanine, was previously tentatively 3 identified as having the methyl substituted on the free amino group.

Using mass spectra for a number of methyl-substituted thioguanines as a

basis, we have been able to conclusively show that the metabolite isolated

from leukemic patients receiving 6-thioguanine is the riboside, methylated

on the sulfur group. The substitution of this methyl group is in agreement 22 23 with findings reported by LePage and others. However, our work has

definitely shown that the product is a nucleoside, whereas the previously

reported identifications were of the free base. It is most probable that

the isolation techniques used in the earlier work resulted in hydrolysis

of the actual nucleoside metabolite, yielding the free base 6-methylthio-

guanine.

4.4.2 Identification of a-Methoxyhomovanillic Acid

Data from the UV-Analyzer indicates a relatively large chromatographic

peak, having maximum uv absorbance at 270 nm and eluting at about 25*5 hr

(just prior to 4-hydroxyhippuric acid), for urinary samples representing

seven different pathologic conditions: Lesch-Nyhan syndrome, chronic

leukemia, synovial sarcoma, embryonic neoplasia, malignant carcinoid,

multiple myeloma, and mental retardation. Table New Urinary Identifications and Their Sample Origins

Compound Sample Origin p-(3-Methoxy-4-hydroxyphenyl)hydracrylic acid Normal and pathologic samples a-Methoxy-a-(3-methoxy-4-hydroxyphenyl)acetic acid Lesch-Nyhan syndrome, mental retardation, a number of different types of cancer

4-Hydroxyacetanilide (sulfate conjugate) Drug metabolite

3-Methoxy-4-hydroxyacetanilide (sulfate conjugate) Drug metabolite

1,3-Dimethyluric acid Normal and pathologic samples, dietary- source

3,7-Dimethyluric acid Normal and pathologic samples, dietary source

1-Methyluric acid Normal and pathologic samples, dietary source

6 -Methylthioguano s inea Metabolite of 6-thioguanine in leukemia

3-Methoxy-4-hydroxycinnamic acid (glycine conjugate) Pathologic samples

Previously identified (tentatively) as N-methyl-6-thioguanosine. -206-

Sufficient material to perform extensive identification studies was

* separated from the concentrated urine sample of a patient with chronic leukemia. High-resolution mass ^spectrometry revealed that the compound had an empirical formula of C-^H-.JD.-J-U Id 2, and empirical formulas were obtained for a number of electron-impact-produced fragments which had associated metastable transitions (Fig. 4.12). The mass spectrum of the trimethylsilyl derivative of the compound showed that it contained two derivatizable hydrogens (Fig. 4.13). Metastable transitions, calculated from data in the low-resolution mass spectrum (Fig. 4.14) and analyzed in conjunction with the high-resolution data of Fig. 4.12, were consistent with interpretation of the fragmentation pattern as shown in Fig. 4.15.

The compound, a-methoxy-a-(3-methoxy-4-hydroxyphenyl)acetic acid, has been given the trivial name of a-methoxyhomovanillic acid. This interesting

compound not previously reported in urine, appears to be homologous with vanillmandelic acid (YMA), a well-known metabolite of the biogenic amines,

epinephrine and nonepinephrine; however, its metabolic origin and

pathological significance are unknown.

4.4.3 Urinary Nucleoside Excretion by Patients with Cancer

For some time the Body Fluids Analysis Group has been involved in

the liquid chromatographic (LC) study of urine samples from patients

with various neoplastic diseases. The primary purpose of this effort is to

A 100-ml volume of urine was passed through a 100-cm bed of XAD-4 resin (Rohm and Haas, Philadelphia, Penn.) and washed with two bed volumes of 0.015 M ammonium acetate—acetic acid buffer; the adsorbed compounds were eluted with methanol.

Samples and funding provided by the National Cancer Institute. ORNL DWG 73-1963

Difference from Measured Mass Empirical Formula Observed Metastable Losses Theoretical, mmu 212.0682 C10H12°5 -0.3 C H 167.0685 • 2 3°3 C9H11°3 -2.3

137.0601 c H -°3V3 -0.2 8 9°S --OH. 124.0527 +0.3 122.0378 -CH. Wa « +1.0 109.0300 -CHO C6H5°2 +1.0 95.0478 C H O-, 6 7 -CO -1.9 -H20 92.0255 W -0.7 81.0297 c5H5O -4.3 77.0367 C h -2.4 6 5 '

Fig. 4.12. High-Resolution Mass Spectrometric Data for Unknown

Body Fluid Constituent of Molecular Weight 212» 100 ORNL DWG 73-2536RI 149

>- f— 80- I—I en u 60- 181 356 LU An 166 > 40-- (—1 196 i CO 0 00 d20 + 341 1

JJ 1, - 0 I " I lilI" "lI IJL L 1 4 i—"1.T . I1—n h*"— r L l1i .:r 1—n 1 1 1 1 1 1 r1—r 70 90 110 130 150 170 130 E10 230 250 270 290 310 330 350 M/E

Fig. 4.13. GC-MS Spectrum (23*5 eV) for the TMS Derivative of the

New Metabolite, a-Methoxyhomovanillic Acid (mol. wt = 212), Isolated

from Urine. The molecular ion at m/e 356 shows that the compound has

two derivatizable protons. -209-

A1ISN31NI 3AIlV~l3d -210-

ORNL DWG. 73- I949RI

1OCH , + —19 7 »rl5 HC-COOH CH,

167 -15 (108.7*) ^0CH3 0 OH OH 137 OH 122 -88(72.5*)

* +

-29(72.8*L 18 (62.4*) H H 95 OH -26 (33.8 ) 1

Indicates Observed 28(60.2*) Metastables

- 28 (34.7 )• 53

Fig. 4.15^ Interpretation for the Mass Spectral Fragmentation of the New Urinary Metabolite Having the Empirical Formula ClOH1205 and Two

Derivatizable Hydrogens. -211- separate and identify urinary excretion products which might serve as biochemical indicators of the presence of malignant cell populations.

Current studies are concentrated on the measurement of three urinary nucleosides, pseudouridine (Y), 1-methylinosine (M^l), and N - 2 dimethylguanosine (M^ G), in urine samples from normal subjects and patients with cancer.

The literature contains little information concerning the quantitative excretion of urinary methylated nucleosides by normal subjects or patients with cancer. Several workers have reported data on the excretion of 24-26 urinary N-bases, but they took no precautions in their analytical procedures t27o preven28 t hydrolysis of the glycosidic bond of nucleosides. Fink et al. ' first identified the urinary ribonucleosides 1-methylinosine and N 2-dimethylguanosine ; however, the first analytical studies of the 29 excretion of these methylated nucleosides were reported by Chheda, who only gave data for three normal subjects. Pseudouridine, the third nucleoside of interest to us currently, was first identified by Davis and 30 Allen. Several workers have reported excretion levels of pseudouridine for normal subjects, and have noted elevated levels for patients with 31-33 leukemia, Hodgkin's disease, gout, and psoriasis.

Our studies of physiologic fluids3 9 4-} 9wit h the UV-Analyzer have shown 2 the methylated N-bases, 1-mechylhypoxanthine and N -dimethylguanine, to 2 be excreted in the form of their nucleosides, M^I and M^ G. The only methylated free N-base that we have found to be excreted in substantial

amounts is 7-methylguanine.

Previously2 , we reported the urinary excretion levels of Y, MI, and M^ G for a number of normal adults and patients with various -212- manifestations of malignant disease. We have now analyzed a total of

76 urine samples from cancer patients for y, and ^9 of these samples for

1-r.ethylinosine and K -dimethylguanosine. Normal excretion data for adults

(ages, 15 to 53) averaged 67-4 with a standard deviation of l4.3 mg/24 hr for y, 4.0 mg/24 hr (S.D. = 1.0) for M-Î, and 3-9 mg/24 hr (S.D. = 1.3) for M^ G. Similar data for children (ages, 3 to 13) averaged 35-4 mg/24 hr (S.D. = 80) for y, 2.5 mg/24 hr (S.D. = 0.9) for MÎ, and 2.5 mg/2U hr 2 (S.D. = 1.0) for M2 G. The data, summarized in Figs. 4.l6 and 4.17, indicate that 55$ of the results for f, and 57% of the results for MÎ 2 and Mg G, were greater than 1 S.D. above the results for normal subjects.

Using a 2 S.D. upper bound, the respective values were 51°!o for Y but only p 27$ for M,1I and ^ G.

Urine samples from three patients were analyzed for the three nucleosides before and after an unknown chemotherapeutic treatment regimen. The results, shown in Table 4.6, indicate a reduction in the excretion level of the three nucleosides for the patients with lung and breast cancer; however, an increased excretion was observed for the patient with stomach cancer in the post-treatment sample.

Two urine samples from a 20-year-old male patient from Africa were analyzed for the three urinary nucleosides. The patient was afflicted with Burkitt's lymphoma, and the malignancy was reported to be in an active state when the first sample was obtained. Several months later a sample was obtained when the disease was in remission. The patient was receiving no drugs when the first sample was taken, but was on BCG therapy for the second sample. The results (Table 4.7) indicate a substantial reduction 2 in the excretion of Y and M0 G when the disease was in remission. -213-

ORNL DWG 73-6933R3 ADUL T BREAS T OVARIA N LUN G AL L NORMA L MELANOM A LEUKEMI A CHILDRE N LEUKEMI A NORMA L CHILDRE N ADULT S CANCE R CARCINOM A TYPE S CARCINOM A 250 A 'OTHE R

X X

- -

- X

- - X

cvj - - X E 150 X X X X

X X UJ

Fig. 4.l6. Urinary Excretion of Pseudouridine (y) by Apparently

Normal Adults and Children Compared with That for Cancer Patients. -214-

ORNL DWG 73-7773

< < < z 2 z«t z 1-2 —i CO -Juj o U>U(J£ tno z z >(E OQ gi

CVI N.

z O K Ul 10 c 8 X— «* X, z X £C X_ 3 6 X - x X •Hfl- % XX f- NORMAL ADULT AVG. «X ^-lo- * V NORMAL CHILD AVG. a* y

M|1 M|G Mjl M|G M,I M|G M,1 M|G M,I M|G M,I M|GM,I M|G M,1 M|G M,I M|G

Fig. 4.17. Urinary Excretion of 1-Methylinosine (Mil) and

2 3 N -Dimethylguanosine (M2 G) by Apparently Normal Adults and Children

Compared with That for Cancer Patients. -215-

Table 4.6. Effect of an Unknown Chemotherapeutic Treatment Regimen on Urinary Nucleoside Excretion

Excretion, (mg/24 hr)

2 Sample Diagnosis Age Sex Treatment Y M-JI M2 G

167(010) Stomach 51 F No 105.2 6.4 6.8 170(013) cancer Yes 122.6 8.3 9.2

127 Lung 68 M No 113.7 9.6 8.0 157 cancer Yes 59.0 5.6 7.4

145(003) Breast 47 F No 66.9 5.4 5.6 190(026) cancer Yes 56.6 1.3 3.6 -216-

Table 4.7. Effect of Eemission on Urinary Nucleoside Excretion

Excretion, (mg/24 hr) Status of Drug 3 2 Age Sex Disease Treatment Y M1I M2 G

20 M Active None 210 5.1 9-6 In remission BCG 77 6.3 3.4 g Burkitt's lymphoma. -217-

The data obtained thus far are interesting, but the possible use of 2 these three urinary nucleosides, M-^I, and M^ G, as biochemical indicators of malignant cell populations remains to be tested. The incidence of elevated excretion in patients with malignant disease seems to be only about 50f0, which is not very encouraging. Additional studies relating physical and other signs of the status of the disease to nucleoside excretion are necessary during the course of treatment.

4.5 Experimental Results and Applications

J. E. Mrochek, S. Katz, and S. R. Linsmore

The high-resolution liquid chromatographic systems have been used to

analyze body fluids obtained from patients having various metabolic and pathologic disorders.

4.5.1 Effects of L-DOPA and a DOPA Decarboxylase Inhibitor

Parkinson's disease is a chronic neurological disorder characterized by tremors, ridigity of the limbs, hindrance of movement, and abnormal

facial expression. Its symptoms develop gradually in both males and

females between the ages of 50 and 65, with an incidence of about one

person in a thousand. Subnormal concentrations of dopamine and norepineph- 34

rine are found in the brains of patients with this disease. Treatment

is based upon the hypothesis that the defect in catecholamine synthesis

in the brain may be at least partially overcome by ingesting L-DOPA.

This cstecholamino acid, which is the precursor of dopamine,3 5i s able to penetrate the blood-brain barrier whereas the amine cannot. Significant

amounts of 3-methoxytyrosine, found in blood and cerebrospinal fluid,

apparently result from the action of catechol-0-methyl transferase (COMT) ft-

on the exogeneous drug, and undoubtedly, are accompanied by acidic urinary

metabolites associated with this pathway, as illustrated in Fig. 4.18.

Metabolites found in the blood after ingestion of L-(_ C3-3-methoxytyrosine

included 3-methoxy-4-hydroxyphenyllactic (vanillactic) acid, homovanillic

acid (HVA), and 3,4-dihydroxyphenylacetic acid (DOPAC)."^

The usual therapeutic treatment of Parkinson's disease with L-DOPA

involves gradual increases in the dosage of the exogenous drug to 37 40 relatively large amounts, ' limited by the degree of side effects and

based on minimizing the physical symptoms of the disease. Decarboxylase

inhibitors have been investigated as a means of decreasing decarboxylation

of the exogenous drug in the peripheral tissues, thus enabling increased 4l 4-2 amounts of DOPA to reach the brain. ' One of these decarboxylase

inhibitors, a-methyldopahydrazine (MK-485), is a potent inhibitor of dopa 43 decarboxylase in the peripheral tissues, but exhibits little effect on

cerebral enzymes since it does not significantl41 y penetrate the blood-brain barrier, as indicated by studies with rats. Its use in conjunction with

L-DOPA resulte44 d in a tenfold increase in the maximum plasma concentration of DOPA. With the combination drug treatment, reduced amounts of

exogenous L-DOPA were required to reach therapeutically useful drug dosages

for depressed patients45 ; in addition, the incidence of side effects was considerably lower.

It might be anticipated that metabolism of DOPA to 3-methoxytyrosine

via COMT (see Fig. 4.18) would be the major pathway (in peripheral tissues)

in the presence of a dopa decarboxylase inhibitor and, consequently,14 vanillactic acid excretion should be increased. Rats given C-labeled

3-methoxytyrosine excreted small amounts of labeled DOPA, dopamine, arid ORNL DWG. 73-2003 R-l

8 ?H CHt-C-COOH CH2-CH-C00H Homovanillic Acid 0—0- OH ON Vanilpyruvic Acid Vanillactic Acid

1" I NHt ch2-ch-cooh 'HJ-CM-COOH ch-ch2nh,

Tyrosine COWL izSSilSUfl (?) OCHB

OH OH / OH DOPA 3-Methoxytyrosine 3-Methoxytyramine DD C02 I"1 CHJ-CHJNHJ fc^AO i COMT = Catechol-O-Methyl Transferase DO H MAO = Monamine Oxidase vo DOPAMINE Homovanillic I ^/9-Hydroxylasa Acid AT = Aminotransferase PE-N-MT= Phenethanolamine -N-Methyl Transferase OH H CH-CHzNHt H-COOH DD - DOPA Decarboxylase Normetanephrine

VMA OH OH O/ Norepinephrine 3,4-DihydroxymandeliDihydrox^ c Vanilliillic r. Acid ) jpE-NJPE-N-M- T

Metanephrine COMT Epinephrine

Fig. 4.18. Detailed Metabolism of L-DOPA in Man. Pathways shown

with question marks have not been proved to exist in man. Aldehyde

intermediates are not shown for reactions involving MAO. -220-

DOPAC, indicating that some demethyiation occurred in the animal.

However, the bulk of the radioactivity (74$) was excreted as homovanillic 46 hi acid, vanillactic acid, and 3-methoxytyramine. Bryson and others have

speculated on the existence of another pathway to homovanillic acid, based on the lack of correlation between urinary dopamine and homovanillic acid. excretion. As shown in Fig. 4.18, at least three other pathways to homovanillic acid are possible: (a) monoamine oxidase (MAO) conversion

of 3-methoxytyramine, and (b) decarboxylation of vanilpyruvic acid. Both

of these, of course, are accomplishe46 d via 3-methoxytyrosine. Experimental studies by Bartholini et_ al. showed that cerebral 3-methoxytyramine in

the rat was probably not due to a direct decarboxylation of 3-methoxytyrosine but, instead, to 3-0-methylation of dopamine that had been formed by

demethyiation and subsequent decarboxylation of 3-methoxytyrosine. This represents a third possible pathway to HVA.

LC Data for Parkinsonian Patients Receiving Oral L-DOPA. — Urine

samples from two patients afflicted with Parkinson's disease and receiving

4 and 4.9 g of L-DOPA during each 24 hr period were chromatographed on the

UV-Analyzer. Unfortunately, chromatographic data for an untreated

Parkinsonian patient were not obtained. It might be expected that urinary

excretion of 3-^ethoxy-4-hydroxymandelic acid (VMA), HVA, and DOPAC would

be subnormal; however, the first two compounds are barely detectable and

the third is undetectable by uv monitoring of samples from normal subjects.

Thus, few chromatographic differences between normal subjects and untreated

Parkinsonian patients are expected using this detector to monitor the

metabolites of interest here. The chromatographic pattern obtained for

patient 1, receiving 4 g of L-DOPA, is shown in Fig. 4.19, where the ORNL OWG 73-I946RZ

Fig. 4.19. Elution Profile Observed on the UV-Analyzer for a

Parkinsonian Patient Receiving k g of L-DOPA Every 2b- hr. New and intensified peaks apparently due to the exogenous drug treatment are shaded. -222- shaded peaks represent new constituents compared to normal subjects, or normal constituents which were excreted in larger quantities than expected. lj.

Identified metabolites include sulfates of dopamine, 3,4-dihydroxyphenylacetic acid, and 3}4-dihydroxymandelic acid; free homovanillic acid and its glucuronide, free 3>4-dihydroxyphenylacetic acid; vanillactic acid; and a conjugate of 3-iuethoxy-4-hydroxyphenylacetamide, tentatively identified as a sulfate. Estimated 24 hr excretions for some of the identified compounds are compared in Table 4.8. The normetanephrine metabolite,

VMA, was identified; however, we were unable to determine its excretion from uv absorbance data since the quantity excreted was relatively small. 15 Use of the cerate oxidimetric detector may enable quantification of this important metabolite.

Although patient 1 received approximately 20$ less L-DOPA than patient 2, his excretion of all metabolites except DOPAC and vanillactic acid was greater than that of the latter subject (Table 4.8). Excretion of vanillactic acid was not detectable in the sample from patient 1, while patient 2 excreted about 50 mg/24 hr. This probably indicates greater conversion of the exogenous drug to 3-methoxytyrosine via COMT by patient 2, and its subsequent transamination to vanilpyruvic acid and reduction to vanillactic acid, as shown in Fig. 4.18. Higher excretion of dopamine sulfate, HVA, and 3>4-dihydroxymandelic acid and the absence of detectabio quantities of vanillactic acid probably indicate that the primary metabolic pathway for patient 1 is through dopamine and not through 3-methoxytyrosine. In addition to the chromatographic peaks shorn in Fig. 4.19, small amounts of a second sulfate conjugate were identified for dopamine and 3>4-dihydroxymandelic acid. Although we have -223-

Table 4.8. Estimated Urinary Excretion of Metabolites by Parkinsonian Patients Treated with L-DOPA

Estimated Excretion (mg/24 hr) Metabolite Patient la Patient 2

c Dopamine sulfate 151 112

3 -Me thoxy - 4- glucur ono s idopheny 1 - 72 56 acetic acidc

Vanillactic acid^ N. D.e 50

Homovanillic acid 1748 802

3,4-Dihydroxymandelic acid 158 66 (sulfate conj.)c

3,4-Dihydroxyphenylacetic acid 177 424

3,4-Dihydroxyphenylacetic acid 147 N.D.e (sulfate conj.)c

aSubject ingesting 4 g/24 hr of L-DOPA.

Subject ingesting 4.9 g/24 hr of L-DOPA.

Conjugate quantified using molar absorptivity of the unconjugated

compound. Quantified using the molar absorptivity of 4-hydroxyphenyllactic

-1 -1

acid at 280 nm, 1226 liter mole cm . eRot detectable. -224- indicated the primary metabolite to be the 4-sulfate in the figure, this is only a tentative assignment since we could not distinguish between 3- and 4-substitution of the sulfate group. LC Data for Parkinsonian Patients Eeceiving L-DOPA plus a DOPA * Decarboxylase Inhibitor. — Urine samples obtained from three patients with Parkinson's disease and being treated with L-DOPA and the dopa decarboxylase inhibitor a-methyldopahydrazine [jVK-4853 were chroma- 43 tographed. The chromatographic profile for acidic metabolites of L-DOPA is shown in Fig. 4.20 for one of the subjects. New constituents (compared to normal subjects not receiving the drug) and those normal constituents whose excretion was apparently increased by the combination drug treatment included dopamine sulfate, p-(3-methoxy-4-hydroxyphenyl)hydracrylic acid, homovanillic acid and its glucuronide, vanillactic acid, and the glycine conjugate of 3-niethoxy-4-hydroxycinnamic acid. The excretion of the compound £-(3-hydroxyphenyl)hydracrylic acid, normally present in all urinary chromatograms, also appeared to be elevated; however, in the absence of a control urine without drugs, this could not be verified. Quantification data for a number of these compoimds are shown in Table 4.9. As indicated in the discussion of DOPA metabolism, inhibition of dopa decarboxylase would probably emphasize the pathway through 3- methoxytyrosine, at least in the peripheral tissues. However, only subject 3 excreted a significantly large amount of one of the expected metabolites of this pathway, vanillactic acid. The other two subjects _

Received through the courtesy of Dr. D. S. Young, Clinical Center, National Institutes of Health, Bethesda, Md. -225-

2 g of L-DOPA Plus 0.3 g of a-Methyldopahydrazine, a Dopa Decarboxylase Inhibitor and intensified peaks apparently due to the exogenous drug treatment are shaded. ORNL DWG. 73- I950RI i i 1 i r 1 1 1— L-DOPA and MK-485, DOPA Decarboxylase Inhibitor 260nm 280nm

§ Z7 2B 29 i 3i i iV

Time Chr) brved on the UV-Analyzer for a Parkinsonian Patient Receiving iopahydrazine, a Dopa Decarboxylase Inhibitor, Every 2h hr. New

to the exogenous drug treatment are shaded. BLANK PAGE -2^6-

Table 4-9. Estimated Rates for Excretion of Dopa Metabolites by Patients with Parkinson's Disease and Receiving L-DOPA plus a-Methyldopahydra zinea

Estimated Excretion Rate (mg/24 hr) Patient 3^ Pat-ent 4b Patient 5° Q Dopamine sulfate 53 N.D.e 3-Me thoxy-4-glucurono s idophenyl- f f f acetic acid a 213 38 Vanillactic acid 194 177 69 Homovanillic acid N.D. N.D. N.D. 3,4-Dibydroxyphenylacetic acid 3,4-Dihydroxymr.ndelic acid N.D. 13 N.D. (sulfate conjugate)

3-Methoxy-4-hydroxycinnamic acid 4.5 N.D. 8.3 (glycine conjugate)*1

Each subject received 300 mg.

Subject received L-DOPA at a dosage level of 2 g/24 hr.

°Subject received L-DOPA at a dosage level of 1.4 g/24 hr.

Compound quantified using the molar absorptivity of the free amine.

eNot detectable by LC. f Compound present but not quantified because of peak overlap.

^Quantified using the molar absorptivity for 4-hydroxyphenyllactic acid at 280 nm, 1226 liter mole"1 cm"1.

^Quantified using the molar absorptivity for 3-methoxy-4-hydroxycinnamic -1 -I acid at 290 nm, 13,700 liter mole cm . -227- also excreted this aromatic acid but in smaller quantities (see Table 4.9).

Apparently, the metabolic pathways through DOPAC and 3>4-dihydroxymandelic acid (see Fig. 4.l8) were inhibited in the presence of the decarboxylase inhibitor, as shown by the virtual absence of these two metabolites from the profile (Table 4.9).

The excretion of small, but apparently elevated, amounts of £-(3- methoxy-4-hydroxyphenyl)hydracrylic acid (barely detectable by uv monitoring for normal subjects) in connection with this combination drug treatment is interesting since it may furnish supportive evidence for a pathway connecting 3-methoxy-4-hydroxycinnamic acid to this metabolite, as 48 postulated by Hicks, Young, and Wootton. These workers suggested that both of the substituted phenylhydracrylic acids were endogenous, based on their identification of these compounds in concentrated dialysates from two subjects who had renal problems and had been placed on a strict protein-free diet.

Interesting differences were noted among the three subjects on the combination drug treatment, and also between these subjects and those receiving only L-DOPA. The unknown chromatographic peak eluting at 34 hr

(Fig. 4.20) was significant, and common to the three subjects receiving the combination drug therapy. It has not been observed in samples from normals or from the two subjects receiving only L-DOPA. Identification of this compound and others may yield new information concerning the in vivo metabolism of L-DOPA in the presence of a decarboxylase inhibitor. -228-

4.5.2 LC Profiles of Arcmatic Acids for Children with Neurological Disorders Urine samples from a number of children with neurological seizures, infantile autism, extreme hyperactivity, and mental retardation have "been chromatographed on the UV-Analyzer. Qualitatively, we have noted several similarities in elevated excretion of certain aromatic acids for a number of these children in comparison to normal children of approximately the same age. Figure 4.21 allows a comparison, for three of these children, of the chromatographic region in which we have identified various acid metabolites of phenylalanine and tyrosine. Apparent elevations in the excretion of p-(3-hydroxyphenyl)hydracrylic, 4-hydroxyphenyllactic, 4-hydroxyhippuric, and 4-hydroxyphenylacetic acids were observed for a number of these children. Elevated excretion of vanillic acid and its glycine and glucuronide conjugates was also observed in several chromatograms; however, since the diet of these children was .unrestricted and vanillin is a common dietary constituent, this observation may not be significant.

Two of the chromatograms in Fig. 4.21 have a relatively strong chromatographic peak, eluting at about 2.5. hr, with maximum uv absorbance at 270 nm (see the upper two chromatograms of Fig. 4.21. This new chromatographic peak was of some interest to us because we had observed the same peak in chromatograms of patients afflicted with chronic leukemia, synovial sarcoma, embryonic neoplasia, malignant carcinoid, multiple myeloma, and Lesch-Nyhan syndrome. Identification studies characterized the compound as a homologue of VMA, a-methoxy-a-(3-methoxy-4-hydroxyp!'^M; 1.)

Received through the courtesy of Dr. M. Coleman, Washington, D.C. ORNL DWG. 73-1951

A040" Subject - Male,9 yrs. Diagnosis - Neurological Seizures, Hyperactivity,! Mental Retardation 260 nm 280 nm 2 70 nm —•—•—

CHfCOOH CHf-COOH Co0"'"'

—~ ^—

10 to

2.0 1.5 r- 1 1 1 1.0- Subject - Male, B yrs. .9 - Diagnosis - Infantile Autism, Mild .8- .7- 260 nm Hippuric Acid 280 nm 270 nm — .3-

CHf&COOl a .3f E Q HOOC^^COOH i.l a i.i vJiMrulv^ J I I I I I I I I I 23 24 25 26 27 26 29 30 31 32 33 Time (hr)

Fig. 4.21. Comparison of the Elution Patterns of Acid Metabolites for Three Children with Neurological Disorders. Elevated excretion of the tyramine metabolite, 4-hydroxyphenylacetic acid, and 4-hydroxyhippuric acid along with a new metabolite, a-methoxyhomovanillic acid, were observed for the two more severely affected children. BLANK PAGE -230- acetic acid (trivially named a-methoxyhomovanillic acid). Its pathologic significance or metabolic origin is unknown.

The estimated excretion rates (absolute for compounds that are available as reference standards, and relative for the others) for a number of these acidic metabolites are listed in Table 4.10 for the three neurologically affected children whose urinary chromatograms are shown in Fig. 4.21. Also listed, for comparison purposes, are the data obtained for three normal children of approximately the same age. It is apparent from the table that the levels of 4-hydroxyphenyllactic acid excreted by the children with neurological disorders are not abnormally high compared to the normal values that were obtained. With the exception of the normal excretion of 4-i.iydroxyphenylacetic acid by the child with mild infantile autism, excretion levels for the remainder of the aromatic acids were factors of 2 to 5 higher for the afflicted children than for the normal children.

The elevated urinary excretion of 4-hydroxyphenylacetic acid observed for several children with neurological disorders appears to qualitatively coincide with elevated excretioh of 4-hydroxyhippuric acid. In addition to the results noted here for children with neurological disorders, we have also observed simultaneous elevation of both 4-hydroxyphenylacetic acid and 4-hydroxyhippuric acid for patients having fibrous displasia, where elevation of the former is diagnostic. The literature, contains evidence from in vitro studies for the existence of a metabolic pathway connecting the precursor of 4-hydroxyhippuric acid, 4-hydroxybenzoic acid, with the 49 50 metabolism of tyramine 9 via an intermediate product, 4-hydroxy- benzaldehyde. The aromatic acid 4-hydroxyphenylacetic acid is, of course, a well-known metabolite of tyramine in humans. Table 4.10. Estimated Rates for Excretion of Some Acidic Metabolites by Three Children with Neurological Disorders and Three Normal Children

Estimated Excretion Rate (mg/24 hr) Neurological Infantile Infantile Normal Autism, Seizures, Autism, Children, Compound Age 9a Severe, Age 8b Mild, Age 8° Ages 12, 9, and 8 p-(3-Hydroxyphenyl)hydracrylic acid 42 26 35 12.5 (5-7)e f 42 g N.D. a-Methoxyhomovanillic acid 70 N.D. 54 d 44 30.1 (14.4) 4-Hydroxyphenyllactic acid 12 7.6 (2.2)e 4-Hydroxyhippuric acid''1 38 14 4o 10.0 (2.2)e 4-Hydroxyphenylacetic acid 31 9 aChild on a low-purine diet, which was also low in phenylalanine and tyrosine (l to 2 g/24 hr). Drugs, 950 mg of allopurinol and 7-l/2 mg of phenobarbital.

Drug, 5-hydroxytryptophan. c Drug, allopurinol. d —1 —1 Quantified using an estimated molar absorptivity of 1000 liter mole cm at 270 nm. e Standard deviation in parentheses. f -1 -1 Quantified using an estimated molar absorptivity of 2700 liter mole-' cm at 270 nm (VMA model).

^Not detectable. h i -1-1 Quantified using molar absorptivity of 4-hydroxybenzoic acid at 260 nm, 11,000 liter mole cm . The elevated urinary excretion of 4-hydroxyphenylacetic acid and 4-hydroxyhippuric acid, which we observed for several children with neurological disorders, may "be due to abnormal metabolism of tyrosine, perhaps resulting in increased formation of tyramine, the amine precursor of these aromatic acid metabolites. Some of the children excreted as much as 40 to 50 mg/24 hr. Our measured values for three normal children averaged 10 mg/24 hr, which was probably somewhat high because of the contribution of an adjacent, overlapping peak believed to be ascorbic acid sulfate in the three normal samples; this interfering peak was not present in samples from the afflicted children. Tocci, Phillips, and 51 Sager showed that diet had no effect on the excretion of 4-hydroxyphenylacetic acid, and they reported that its excretion averaged 3-9 mg/24 hr for 4l normal children ranging in age from 1.5 to 15 years and receiving unrestricted diets. If elevated excretion of this aromatic acid is related to increased production of tyramine for these children, it may be of interest, since the characteristic "pink spot" in paper chromatographic studies of urinary samples from individuals suffering from schizophrenia and Parkinson's disease has been identified as 52 tyramine. However, there is some controversy concerning this identi-

ficatio(3,4-dimethoxyphenyln since, more )recently ethylamine, th, esupposedl "pink spoty o"f hadietars beeyn originidentifie.5 3 d as

4.5*3 Body Fluid Constituents of Patients with Rare Pathologic Conditions * Samples from children with relatively rare diseases were analyzed using the UV-Analyzer coupled with the cerate oxidimetric detector. _

Samples obtained through the courtesy of Dr. P. M. Tocci, University of Miami. -233-

Observations for each of the three diseases, eosinophilic granuloma, cystic fibrosis (and infectious hepatitis), and Henoch-Schonlein disease, are listed below. <

Eosinophilic Granuloma. — The child, a 3-year-old male, was reported to be excreting abnormal amounts of 4-hydroxytryptamine (serotonin) and its metabolite, 5-hydroxyindoleacetic acid. After undergoing treatment, a second sample was received from this patient which indicated a reduction of his urinary excretion of serotonin and 5-hydroxyindoleacetic acid to normal.

A large amount of phenacetin, or Tylenol, was given to the child prior to his chemotherapeutic treatment. This is evidenced in the uv analysis by the large chromatographic peaks representing k-hydroxyacetanilide,

3-methoxy-4-hydroxyacetanilide, their glucuronide and sulfate conjugates, and an unidentified metabolite associated with this medication eluting just prior to hippuric acid.

Elevated excretion of the following compounds was observed:

glucuronosidotoluene, homovanillic acid, vanillylmandelic or a-ketoglutaric acid (compounds coelute), 4-hydroxyphenylacetic acid (21.6 mg/24 hr), and

5-hydroxyindoleacetic acid. Relatively high excretion of a-methoxyhomovanillic acid, the recently identified metabolite, was observed along with detectable amounts of kynurenine and 3-hydroxykynurenine. All of these are abnormal when compared to normal children.

Changes observed in the post-treatment sample included:

(a) Inosine-elevated excretion, usually barely detectable in

normal subjects and pretreatment sample.

(b) Kynurenine - not detectable. -234-

(c) 3-Hydroxykynurenine - still detectable. (d) Homovanillic acid and 4-glucuronosidotoluene - excretions were reduced considerably and are now about normal. (e) 4-Hydroxyphenyllactic acid - excretion very low in the pretreatment sample; increased to approximately a normal level after treatment.

(f) a-Methoxyhomovanillic acid - excretion is considerably decreased after treatment; may still be present but only at a trace level.

(g) 4-Hydroxyphenylacetic acid - 15»4 mg/24 hr; still elevated but decreased from pre-treatment level. (h) 4-Hydroxyhippuric acid - increased from a normal 5.8 mg/24 hr before treatment to 26.4 mg/24 hr after treatment. (i) 5-Hydroxyindoleacetic acid - not detectable. A number of large new chromatographic peaks in the post-treatment sample are probably related to the medication regimen. It is difficult to determine what changes were induced in the excretion of some important metabolites of tyrosine because of the large amount of phenacetin received in the pre-treatment regimen. One of the major metabolites of this analgesic drug is 3-methoxy-4-hydroxyacetanilide. The effect of its formation, and whether its hydroxylation in the 3-position is competitive with that of tyrosine forming DOPA, is an unknown factor. Cystic Fibrosis with Infectious Hepatitis. — This patient was an 11- year-old female, Patients with cystic fibrosis typically excrete elevated 1 5b amounts of q-hydroxyphenylacetic acid, and this patient was excreting 58.5 mg/24 hr, or 5 to 10 times the normal level.Other aromatic acids which were excreted at abnormally high levels were (3-(3-methoxy-4- -235- hydroxyphenyl)hydracrylic acid, 4-hydroxymandelic acid, and 4-glucurono-

sidotoluene. The glycine conjugate of 4-hydroxybenzoic acid, hydroxyhippuric acid, was excreted at an estimated 86 mg/24 hr, which is 10 to 12 times the level that we have found for normal children. Urinary k5 tyrosine excretion was abnormally high, and Van der Heiden et al. have

suggested that, in some cases, the source of abnormal excretion of phenolic and phenyl acids may be bacterial degradation of nonabsorbed

tyrosine and phenylalanine, based upon their observation of a patient with a severely impaired amino acid absorption.

Henoch-Schonlein Disease. — This patient was a 12-year-old male who was excreting massive amounts of b-hydroxyphenylpyruvic and

hydroxyphenyllactic acids. In addition to these two tyrosine metal: elites,

elevated excretion of tyrosine, 4-glucuronosido toluene, and the

octopamine metabolite, 4-hydroxymandelic acid, was observed. Three very

large fluorescent (cerate detector) peaks eluted late in the chromatogram,

two of which have tentatively been identified as substituted cinnamic

acids (perhaps 3-methoxy-4-hydroxy and 4-hydroxy substituted).

b.5*4 Urinary Metabolites of 4-Hydroxyacetanilide

Urinary excretions of metabolites of drugs are of great potential

use in defining the mechanisms by which the drug operates, the side

reactions, dosage parameters, and the interrelationship with the patho-

logical condition of the subject. Recognition of those patterns is also

important in interpreting the resulting chromatograms. The patterns

associated with ingestion of b-ethoxyacetanilide and 4-hydroxy-

acetanilide are of special interest because of their extensive use and 55-57 the concern about possible toxicological effects of some of the

metabolites. -236-

58 59 Since the work of Brodie and Axelrod, and Smith and Williams, phenacetin (4-ethoxyacetanilide) is thought to be converted to k- hydroxyacetanilide by oxidative O-dealkylation. This metabolite is detoxified by conjugation with sulfuric or glucuronic acid. Less than IFjo of 4-ethoxy&cetanilide has been found to be dsacetylated to P- 58 59 phenetidine, ' but this (4-ethoxyaniline) metabolite is reported to be 55 56 60 the origin of some toxic metabolites. ' Burtis, Butts, and Rainey identified another metabolite of phenacetin, 3-methoxy-4-hydroxyacetanilide and its glucuronide conjugate, in a patient with neuroblastoma. Our chromatographic studies of urine samples from several different patients receiving Tylenol (4-hydroxyacetanilide) have shown a marked similarity in the patterns of metabolites observed. The dual-detector 15 (uv and cerate oxidimetric detectors) system has been primarily used to investigate these metabolites because it is very sensitive to the metabolites of interest, and because the elution conditions enable observation of the late-eluting sulfate conjugates. Four urine samples from patients having leukemia, osteogenic sarcoma, colon carcinoma, and malignant carcinoid receiving Tylenol were analyzed on the UV/COD Analyzer. A sample from a fifth patient, a child with eosinophilic granuloma, was also analyzed; however, we do not know whether the medication was Tylenol or phenacetin in this case.

A representative chromatogram illustrating the separation of 8 metabolites of a single drug is shown in Fig. 4.22. The metabolites that have been identified include 3-methoxy-4-hydroxyacetanilide, its glucuronide and sulfate conjugates, k-hydroxyacetanilide, and its glucuronide and sulfate conjugates. Two additional metabolites have also been associated ORNL DWG 73-8607

I UO -<] I

3 4 5 6 7 6 9 10 li 12 13 14 15 16 17 16 20 21 22 23 24 25 26 27 29 30 32 33 34 35 39 40 41 43 44 TIME (hr) I 3-METH0XY-4-HYDR0XYACETANILIDE 2 UNKNOWN METABOLITE OF H ANION EXCHANGE CHROMATOGRAPH RUN CONDITIONS n 4-HYDROXYACETANILIDE 21 SULFATE OF I 02 cm « 150 cm STAINLESS STEEL WITH 8-12/1 DJAM AMINEX A-27 RESIN, HI GLUCURONIDE OF I an SULFATE OF n TEMPERATURE, AMBIENT TO 60"C AT 3 nr , ELUENT GRADIENT, 0015 M HE GLUCURONIDE OF n am UNKNOWN METABOLITE OF E TO 60 M AMMONIUM ACETATE, pH 4 4 , ELUENT FLOW RATE, 7 ml/hi , COLUMfTPRESSURE ,2700 psig

Fig. 4.22. Urinary Chromatograms from the W-Analyzer with Cerate

Oxidimetric Detector from a Child with Eosinophilic Granuloma Before

and After Treatment with 4-Hydroxyacetanilide. with this medication (labeled V and VIII, Fig. 4.22); however, they have not yet been identified.

The identification of sulfate conjugates has been materially aided by the discovery that the trimethylsilyl (TMS) derivative of the urinary conjugate apparently causes a deconjugat.ion. Gas chromatograms of the derivatized urinary conjugates yield peaks corresponding to the free compound and a molecular entity which seems by GC-MS to be a tri-TMS sulfite. This peak with methylene unit retention times of 14.98/13*38

(0V-17/0V-1) is characteristic of urinary sulfates and identifies the conjugate as a sulfate.

Our observations of patients receiving Tylenol indicate Lhst the major metabolite is the sulfate conjugate of U-hydroxyacet^nilide folicwod by ilo glucuronide. It is apparent that significant amounts of 3- methoxy-4-hydroxyacetanilide are produced, with the relative order being 6o sulfate > glucuronide > free compound. Burtis et al. suggested that this metabolite waa not normally observed after phenacetin ingestion; hov;ever, our data indicate that it is a normal metabolite of 4-hydroxy-

5»ci:.tbriilide.

The central analgesic action of Tylenol is apparently similar to that of the salicylates, but the mechanism of its pain-relieving action hax not been elucidated.^1 The formation of a catecholamine-like structure,

3,4-clihydroxyacetonilide, can be postulated as preceding the formation of the observed metaboJ.ite, 3-niethoxy-4-hydroxyacetanilide. This suggests involvement of a hydroxylating enzyme such as tyrosine hydroxylase, which catalyzes the conversion of tyrosine to 3,U-dihydroxyphenylalanine (DOPA).

One might speculate that enzymes involved in the formation of this particular -239- metabolite may be u.he connecting link to the analgesic action of the drug.

Further studies of this mechanism might be made by determining the effect of the drug on excretion of metabolites of the catecholamine pathway, such as homovanillic and vanillylmandelic acid.

4.6 References for Section 4

1. C. D. Scott and S. Katz, "Body Fluids Analysis Program Progress Report for the Period September 1, 1968, to February 28. 1969, ORNL-TM-2551 (1969). 2. C. D. Scott et al., Biochemical Separation Systems Development Section of the Molecular Anatomy Program Progress Report for the Period September 1, 1970, to February 28, 1971, ORNL-TM-3434 (1971).

3. Molecular Anatomy (MAN) Program 1st Semiannual Progress Report March 1 to August 31, 1972, 0RNL-4815.

4. Molecular Anatomy (KAN*' Program 1st Semiannual Progress Report September 1. 1972 ~:c February Lo, 19:7j, L-RHL067.

. 0. L». Sccti and II. E. Lee, "Dynamic Packing of Ion Exchange Chromato- graphic Columns," J. Chromatog. 42, 261 ',1969).

6. Beckman. Handbook 120-B, "Preparation and Storage of Reagents," Part 5.

7. B. C. Starcher, L. \r. Wenger, and L. D. Johnson, "A Modified Gradient Elution Procedure for Single Column Amino Acid Analysis," J. Chromatog. 54, 425 (1971). 8. L. H. Thacker, C. D. Scott, W. W. Pitt, Jr., "A Miniaturized Ultra- violet Flow Photometer for Use in Liquid Chromatographic Systems," J. Chromatog. 51, 175 (1970).

9. L. H. Thacker, W. W. Pitt, Jr., S. Katz, and C. D. Scott, " Minia- ture Photometers for Chromatography,: Clin. Chem. 16, 626 (1970). 10. K. B. Eik-Nes and E. C. Horning, Gas Phase Chromatography of Steroids, Springer-Verlag, New York, 1968.

11. D. D. Chilcote and J. E. Mrochek, "Chromatographic Analysis of Naturally Fluorescing Compounds: 1. Rapid Analysis of Nanogram Amounts of Indoles in Physiologic Fluids," Clin. Chem. 18, 778-82 (1972). -240-

12. D. D. Chilcote, "Column-Chromatographic Analysis of Naturally Fluorescing Compounds: II. Rapid Analysis of Indoleacetic Acid and 5-Hydroxyindoleacetic Acid in Biological Samples," Clin. Chem. 18, 1376-78 (1972).

13. S. Katz and C. A. Burtis, "The Relationship Between Chemical Structure and Elution Position in an Anion Exchange System," J. Chromatog. 40, 270-82 (1969). 14. C. D. Scott. R. L. Joiley, W. W. Pitt, Jr., and M. S. Johnson, "Prototype Systems for the Automated, High-Resolution Analysis of UV-Absorbing Constituents and Carbohydrates in Body Fluids," Amer. J. Clin. Path. 53, 701-12 (1970).

15. S. Katz, W. W. Pitt, Jr., and G. Jones, Jr., "Sensitive Fluorescence Monitoring of Aromatic Acids After Anion Exchange Chromatography of Body Fluids," Clin. Chem. 19, 817-20 (1973). 16. S. Udenfriend, S. Stein, P. Bohlen, ¥. Dairman, W. Leimgruber, and M. Weigele, "Fluorescamine: A Reagent for Assay of Amino Acids, Proteins, and Primary Amines in the Picomole Range," Science 178, 871 (1972).

17. S. Stein, P, Bohlen, J. Stone, W. Dairman, and S. Udenfriend, "Amino Acid Analysis with Fluorescamine at the Picomole Level," Arch. Biochem. Biophys. 155, 203 (1973). 18. P. Bohlen, S..Stein, W. Dairman, and S. Udenfriend, "Fluorometric Assay of Proteins in the Nanogram Range," Arch. Biochem. Biophys. 155, 213 (1973).

19. T. P. Waalkes, S. R. Dinsmore, and J. E. Mrpchek, "Increased Urinary Excretion of the Nucleosides, N2, N^-Dimethylguanosine, 1-Methylinosine, and Pseudouridine by Cancer Patients," J. Nat, Cancer Inst. 51, 271 (1973). 20. G. W. Israel and S. K. Friedlander, "High Speed Beams of Small Particles," J. Colloid. Interfacial Sci. 24, 330 (1967). 21. R. J. Mitchell, "Improved Method for Specific Determination of Creatinine in Serum"and Urine," Clin. Chem. 19, 408 (1973).

22. G. A. LePage and J. P. Whitecar, Jr., "Pharmacology of 6-Thiogranium in Man," Cancer Res. 31, 1627 (1971). 23. G. B. Elion, et al., "The Metabolism of 2-Amino-6-[(l-methyl-4- nitro-5-imidozolyl)thio] Purine in Man," Cancer Chemotherapy Rept. 8, 47 (1960).

24. R. W. Park et al., "Urinary Purines in Leukemia," Cancer Res. 22, 469 (1962). -241-

25. B. Weissmann, et al., "The Purine Bases of Human Urine. I. Sepa- ration and Identification," J. Biol. Chem. 224, 407 (1957). 26. 3. S. Mirvish, et al., "7-Methylguanine and Other Minor Urinary Purines: Values for Normal Subjects from Israel, Gaza, and Kenya, and for Patients -with Cancer of Various Organs or Cirrhosis of the Liver," Cancer 27, 736 (1971).

27. K. Fink, et al., "The Identification of 2-Dimethylamino-6-Hydroxy- purine and Its Ribonucleotide in Urine of Normal and Leukemia Subjects," Cancer Res. 23, 1824 (1963).

28. K. Fink and W. S. Adams, "Urinary Purines and Pyrimidines in Normal and Leukemia Subjects," Arch. Biochem. Biophys. 126, 27 (1968).

29. G. B. Chheda, et al., "Nucleosides in Human Urine. I. Isolation and Identification of N2-Dimethylguanosine, 1-Methylinosine, and N2-Methylguanosine from Normal Urine," J. Pharm. Sci. 58(1), 75 (1969). 30. F. F. Davis and F. W. Allen, "Ribonucleic Acids from Yeast Which Contain a Fifth Nucleotide," J. Biol. Chem. 227, 907 (1957).

31. W. S. Adams, F. Davis, and M. Nacatoni, "Purine and Pyrimidine Excretion in Normal and Leukemia Subjects," Am. J. Med. 28, 726 (1960).

32. K. J. Pinkard, et al., "Purine and Pyrimidine Excretion in Hodgkin's Disease," J. Nat. Cancer Inst. 49, 27 (1972).

33. S. Weissman, A. Z. Eisen, and M. Karon, "Pseudouridine Metabolism. II. Urinary Excretion in Gout, Psoriasis, Leukemia, and Heter- ozygous Oroticaciduria,," J. Lab. Clin. Med. 59, 852 (1962).

34. H. Ehringer and 0. Hornykiewicz, "Verteilung von Noradrenalin und Dopamin in Gehior des Menschen und Ihr Verhalten bei Erkrahkungen des Extrapyramidalen Systems," Klin. Wochschr. 1236 (1960).

35. H. Blaschko, "The Action of L-DOPA Decarboxylase," J. Phsiol. 101, (1942).

36. J. H. Fellman, J. R. Joyce, and J. J. Strandholm, "Analysis in Human Plasma of 3-Meth.oxytyrosine: A Metabolite of DOPA," Clin. Chim. Acta 32, 313 (1971). 37. N. S. Sharpiess and D. S. McCann, "DOPA and 3-0-Methyldopa in Cerebrospinal Fluid of Parkinsonism Patients During Treatment with Oral L-DOPA," Clin. Chim. Acta 31, 155 (1971). -242-

38. A. Pletscher, G. Bartholin!, and R. Tissot, "Metabolic Fate of L- [l4c]D0PA in Cerebrospinal Fluid and Blood Plasma of Humans," Brain Res. 4, 106 (1967).

39. I. G. Kuruma, G. Bartholini, R. Tissot, and A. Pletscher, "The Metabolism of L-3-0Methyldopa, a Precursor of DOPA in Man," Clin. Pharmacol. Ther. 1^(4), 678 (1971). 40. G. C. Cotzias, P. S. Papavasiliou, and R. Gellene, "Modifications of Parkinsonism - Chromic Treatment with L-DOPA," New Engl. J. Med. 280, 337 (1969). 41. G. Bartholini and A, Pletscher, "Effect of Various Decarboxylase Inhibitors on the Cerebral Metabolism of Dihydroxyphenyl&laxiine," J. Pharm. Pharmacol. 21(5), 323 (1969).

42. G. Bartholini, W. P. Burkard, and A. Pletscher, "Increase of Cerebral Catecholamines Caused by 3,4-Dihydroxyphenylalanine After Inhibition of Peripheral Decarboxylase," Nature 215, 852 (1967).

43. C. C. Porter, L. S. Watson, D. C. Titus, J. A. Totaro, and S. S. Byer, "Inhibition of DOPA Decarboxylase by the Hydrazine Analog of a-Methyldopa," Biochem. Pharmacol. 11, 1067 (1962).

44. D. L. Dunner, H. Keith, H. Brodie, and F. K. Goodwin, "Plasma DOPA Response to Levodopa Administration in Man: Effects of a Peri- pheral Decarboxylase Inhibitor," Clin. Pharmacol. Therap. 12(2), 212 (1971).

45. F. K. Goodwin, D. L. Murphy, H. K. H. Brodie, and W. E. Bunney, Jr., "Administration of a Peripheral Decarboxylase Inhibitor with L-DOPA to Depressed Patients," Lancer 1(7653), 908 (1970). 46. G. Bartholini, I. Kuruma, and A. Pletscher, " 3-0-Methyldopa, a New Precursor of Dopamine," Nature 23, 533 (1971). 47. G. Bryson, "Biogenic Amines in Normal and Abnormal Behavior States," Clin. Chem. 17, 5 (1971). 48. J. M. Hicks, D. S. Young, and I. D. P. Wootton, "Abnormal Blood Constituents in Acute Remal Failure," Clin. Chim. Acta 7, 623 (1962). 49. M. L. C. Hare, "Tyramine Oxidase. I. A New Enzyme System in Liver," Biochem. J. 22, 968 (1928).

50. S. Iskric and. S. Kveder, "A New Aspect of the in vitro Metabolism of Some Biogenic Amines," Croat. Chem. Acta 37, 233 (1965). 51. P. M. Tocci, J. Phillips, and R. Sager, "The Effect of Diet upon . the Excretion of Parahydroxyphenylacetic Acid and Creatinine in Man," Clin. Chim. Acta 4o, 449 (1972). -21+3-

52. A. A. Boulton, R. J. Pollitt, and J. R. Major, "Identify of a Urinary 'Pink Spot' in Schizophrenia and Parkinson's Disease," Nature 215, 132 (19^7).

53. J. R. Stabenau, C. R. Creveling, and J. Da3y, "'Pink-Spot1 3,4- Dimethoxyphenyl)ethylamine, Common Tean, and Schizophrenia," Ainer. J. Psychiat. 127, 611 (1970).

54. C. Van Der Heiden, et al., "Gas Chromatographic Analysis of Urinary- Tyro sine and Phenylalanine Metabolites in Patients with Gastro- intestinal Disorders," Clin. Chun. Acta 34, 289 >1971).

55. M. Kiese and H. Menzel, "Hemoglobin Formation in the Blood of Man After Intake of Phenacetin and N-Acetyl-p-aminophenol," Arch. Exp. Pathol. Pharmakol. 242, 551-54 (1962). 56. A. Klutch, M. Harfenist, and A, H. Coriney, "2-hydroxyacetophenetidine, A New Metabolite of Acetopheaetidine," J. Med. Chem. 9, 63-66 (1966).

57- H. Buch, K. Pfleger, and W. Rummer, "Investigation of the Oxidative Metabolism of Phenacetin in Rats," Biochem. Pharmacol. l6, 2247-56 (1967). ~~ 58. B. B. Brodie and J. Axelroc., "Fate of Acetophenetidin (Phenacetin) in Man and Methods for the Estimation of Acetophenetidin and Its Metabolites in Biological Material," J. Pharm. Exp. Ther. 97, 58-67 (1949).

59. J. N. Smith and R. I. Williams, "The Metabolism of Phenacetin (p-Ethoxy-acetanilide) in the Rabbit and a Further Observation of Acetanilide Metabolism," Biochem. J. 239-42 (1949).

60. C. A. Burtis, W. C. Butts, and W. T. Rainey, Jr., "Separation of the Metabolites of Phenacetin in Urine by High-Resolution Anion Exchange Chromatography," Amer. J. Clin. Path. _53, 769-77 (1970).

61. L. S. Goodman and A. Gilman (Eds.), The Pharmacological Basis of Therapeutics, Macmillan, New York, (1970). -244-

5. MACROMOLECULAR SEPARATIONS

The Macromolecular Separations Program was initiated ten years ago at the suggestion of Alvin M. Weinberg to provide an interdisciplinary group that would interface between the biologists in the Biology P.'vision and chemists and engineers in the Chemical Technology Division. Reversed- phase chromatography methods were developed for the purification of transfer ribonucleic acids, and samples were distributed to over 50 investigators in a dozen countries. The technology that was developed was later applied to other nucleic acids separation problems, such as tENA sequencing. More recently, this technology has been applied to the separation of other types of biological macromolecules, such as erythropoietin and colony stimulating factor.

5.1 Purification of Erythropoietin

R. L. Pearson

The hormone erythropoietin (Epo) is the humoral factor that stimu- lates the development of red blood cells in normal animals. It is a glycoprotein containing sialic acid and has a molecular weight of about

,000. Despite extensive research, only minute quantities of highly purified Epo have been prepared.1 Larger quantities are necessary to investigate the biochemical properties and chemical composition, as well as to obtain clinical evaluation of the effectiveness of Epo for the relief of anemia due to renal failure.

A project for the preparation of large quantities of Epo from anemic sheep plasma has been under way in the Unit Operations Section (UNOP) of the Chemical Technology Division. The recovery flowsheet employs -245- conventional ion exchange chromatography and precipitation techniques as 2 3 reported previously.

5.1.1 Preparation of Affinity Columns

It would be desirable to have an alternate purification step which would combine a higher degree of purification and better recovery of activity than has been attainable with conventional steps. Affinity chromatography, the use of selective adsorbents having biological specificity, could offer such a step. Conventional biochemical separation techniques depend on differences in the physiochemical properties of biological molecules. Since these differences are Gften quite subtle, purification is generally laborious and often incomplete. In contrast, affinity chromatography is an approach that exploits the most unique property of the macromolecules—their biological specificity.

Perhaps the most important class of reactions which have been used' in affinity chromatography is the interaction of antigens with their

specific antibodies. Specific antibodies with high affinity can be 4 developed against many molecules. Antibodies to Epo capable of neutral- izing the biological activity of the hormone have been reported.*^ We propose to form antibodies either to Epo or to the impurities contaminating the Epo, isolate the antibodies, attach them to Sepharose,* and thus to purify the Epo by either direct cr indirect antibody affinity chromatography.

Indirect affinity chromatography tests were designed in an effort to remove the contaminating proteins. In the first test an attempt was made

Registered trade name of the Pharmacia Company, Uppsala, Sweden. to induce a rabbit to develop antibodies to all the components in normal

sheep serum. The antibodies to normal sheep serum were isolated from

the rabbit serum by affinity chromatography, and then bound to Sepharose

activated with cyanogen bromide to prepare a column for the removal of

contaminating sheep proteins from the Epo solution.

One milliliter of a 1:1 mixture of normal sheep serum and Freund's

adjuvant was injected subcutaneously in each hip of a rabbit, followed

by 1/2 ml of normal sheep serum weekly for a month. Beginning with the

5th week, blood samples (10 to 50 ml) were dra,wn from the rabbit weekly

for processing. Booster shots of normal sheep serum were continued every

other week.

In order to separate the antibodies to normal sheep serum from the

rest of the proteins in the rabbit serum, 830 &2Qq u*1^8 (one ^280

is the amount of material that when dissolved in 1 ml will give an

absorbance of 1.0 at 280 nm with a 10-mm pathlength) of normal sheep <7 serum were coupled to 70 ml of Sepharose 4b using a published procedure.

A 2.3 x 10 cm column was prepared from this material. The rabbit serum

components passed through this column, leaving the sheep serum anti-

bodies bound to the column. A potassium thiocyanate (KSCN) solution was

then used to uncouple the bound antibodies. The antibodies were dialyzed

to remove the KSCN and then concentrated. An Ouchterlony test demonstra-

ted the presence of sheep antigen antibodies to the sheep serum in this

recovered sample.

The antibodies to normal sheep serum, 900 ^Qo were then

attached to 70 ml of the activated Sepharose 4B, and a 2.3 x 70 cm column

was prepared. -2V7-

5.1.2 Purification by Indirect Affinity Chromatography

A process solution containing Epo and some contaminants (Step 2 Epo), prepared by UNOP personnel, which assayed 2.5 I.U./A2g0, was passed through the column described in the previous section. About 10 to 15$ of the A^qq units were retained on the column in the first pass, and there was no additional retention in the second pass through the column. The remainder of the Agg^ units and 100$ of the Epo were accounted for in the column effluent. The experiment was encouraging in that Epo was not denatured by the column or reagents involved; however, there was little removal of contaminating proteins, presumably due to lack of antibodies for most of the proteins rather than low capacity.

In a second experiment, polycythemic sheep serum (free of Epo and thus unable to stimulate anti-Epo) was obtained and processed through the same two recovery steps used in the preliminary purification of the

Epo sample so that the antigens for injection in the rabbit would more closely parallel the contaminants in the Epo sample. To render a sheep polycythemic, one pint of blood was withdrawn from each of two sheep and injected into a third sheep on day 1, and again on day 3' 0*1 d&y 1> after the hematocrit had increased from 35 "to 65, the blood was withdrawn from the polycythemic sheep and processed through the first two Epo purification steps. The recovery steps removed 99-9$ of the A^g absorbing material. The remaining 0.1$ should be similar to the contaminants in the Epo sample. The sample was analyzed and shown to be free of

Epo, thus no anti-Epo would be formed.

Following the same procedure for inoculation and collection of rabbit serum, antigen from the serum of polycythemic sheep was then used -248- to inoculate a rabbit. The IgG fraction, containing all the antibodies, was isolated by DEAE-cellulose chromatography. The rabbit IgG fraction

was then bound to activated Sepharose and a 2.3 x 12 cm column was pre-

pared. A sample of 100 units of Step 2 Epo was repeatedly passed

through the column, with appropriate regeneration of the column between

each pass. The results (Fig. 5»l) show a repeated, but decreasing,

removal of contaminants. This is probably due to the fact that the

quantity of antibodies for specific antigens varies greatly, as a result

of variable responsiveness of the rabbit to different antigens. Thus,

different proteins are being removed in different amounts in each pass.

A linear extrapolation of the number of passes suggests thab after ten passes no additional contaminants would be removed. This should result

in a calculated 8-fold purification of the Epo. Analysis of the Epo

sample after the fourth pass showed a purification of 13-fold (from 2.5

to 33 I.U./A q ). This higher-than-expected value may be due to removal

of Epo inhibitors or to analytical precision.

These preliminary experiments with indirect affinity chromatography

were encouraging since partial removal of the contaminants could be

achieved with no loss of Epo. It may be, however, that the rabbit did

not develop antibodies to some of the contaminants, or it is possible

that the polycythemic sheep serum does not contain all the contaminants

present in the serum from sheep rendered anemic with phenylhydrazine.

Thus, only partial purification may be achieved by the indirect route.

It appears that direct antibody affinity chromatography may be the

more valid approach, despite the lack of highly purified Epo for antibody

stimulation. This problem will be overcome by recycling the upgraded ORNL DWG 73-9932RI 0 g 10

2 20 2 3 _l 30 O O 40 > OQ 50 O CN) tu 60 -r > \L> o s 70 LJ cr 80 o 80 CM 90 <

! 23456 789 10 NUMBER OF PASSES THROUGH ANTIBODY-SEPHAROSE COLUMN

Fig. 5*1* Removal of Contaminating Proteins by Indirect Antigen

Antibody Affinity Chromatography. -250-

Epo sample for subsequent cycles and "boot-strapping1' to higher-purity

Epo samples. The highest-purity Epo available will be injected in a rabbit

and the antibodies recovered on each cycle for the preparation of anti-Epo,

the contaminating sheep antigen antibodies will be removed on a sheep

antigen column, and the anti-Epo then bound to a cyanogen bromide activated

Sepharose column. This column will yield a more highly purified Epo

sample which will again be used for injection. Eventually, highly puri-

fied Epo antibodies should be obtained. Experiments are currently under way to evaluate this approach.

5.2 Purification of Colony Stimulating Factor

R. L. Pearson

Bone morrow progenitors of macrophages and granulocytes can proliferate in agar cultures to form large colonies of differentiating macrophages and/or granulocytes (immature forms of white cells) if the cultures are stimulated by a humoral factor called colony stimulating factor (CSF). Q This property establishes the assay for CSF activity, which involves incu-

bation at 37°C in 10% COg in the presence of mouse bone marrow cells. CSF

appears to be a glycoprotein which is found in the serum and urine of

normal humans and mice. Serum levels of CSF were shown to become elevated

in mice and humans following viral or bacterial infections. Thus, CSF

seems to be directly related to the response to disease, and some evidence

also suggests that elevated levels of CSF may be related to leukemia.

Enough purified CSF for clinical testing and for elucidation of the bio-

chemical and physical properties has not been obtained. The purpose of

this project (being carried out in collaboration with Dr. Richard Stanley -251- of the Ontario Cancer Research Institute) is to develop an improved flow-

sheet for the purification of CSF to allow scale-up procedures for the production of purified CSF for clinical testing. 9 Published procedures for the isolation of CSF from human urine involve

dialysis, batch DEAE-cellulose chromatography, and gel filtration as the

initial steps. To prepare sufficient CSF for clinical evaluation and k 5 chemical determination, it may be necessary to process from 10 to 10

liters of urine per year. Each processing step must therefore be evaluated with regard to scale-up capability.

The first step involves the dialysis of large volumes of urine. Since

conventional dialysis tubing is not feasible on a large scale, two alternate approaches were evaluated. If human urine is diluted with distilled

water so that its salt content is less than that of the wash solution used

in the first DEAE-cellulose chromatography step (0.0U M NaCl with a

conductivity of 11,500 micromhos/cm), then it should be possible to load the

diluted urine directly onto the DEAE-cellulose without dialysis. A urine

sample was diluted with water to a conductivity of 8,200 micromhos/cm. This

was then applied to a pretreated DEAE-cellulose and the CSF activity in

the sample was essentially all sorbed on the DEAE-cellulose. The dialysis

step could thus be eliminated simply by dilution.

Another alternate to conventional dialysis is the use of the recently

developed hollow fiber equipment. A Dow Chemical Company (Midland, Mich.)

Bio-Fiber 50 Miniplant was used to dialyze 3^00 ml of urine. The conducti-

vity decreased with a half time of 25 min; thus, this small unit could

dialyze 90 liters of urine overnight prior to the DEAE-cellulose column.

This method of reducing the conductivity appears more attractive than

dilution, since it does not entail a large volume expansion. -252-

9 Steps I and III in the flowsheet developed by Stanley and Met calf involve DEAE-cellulose chromatography, first batchwise and then with gradient elution. A logical approach is to combine these steps. In one experiment, 2 liters of urine were dialyzea in the hollow fiber unit until the conductivity was 10,000 micromhos/cm. One hundred milliliters of pre- equilibrated DEAE-cellulose was then added to the urine, and the mixture was stirred at room temperature for 30 min. This was then poured to form a 4.6 x 8 cm column. The column was washed overnight with 0.04 M NaCl (containing a 0.1 M Tris pH 7.0 and 0.02% NaN^), and then gradient elution was employed with eluent NaCl concentrations from 0.04 to 0.U M and an eluent flowrate of 120 ml/hr. The results are shown in Fig. 5.2. Most of the material contributing to the 280 nm absorbance, representing the majority of the contaminating proteins, passed directly through the column. Analysis of eluate fractions collected during the 3radient elution showed the CSF activity to be eluted at an elution volume of about 500 ml after the NaCl gradient was begun. A potential purification of 2700-fold was achieved, based on the removal of 99»96$ contaminating 2f3 nm absorl material. Thus, it appears that the two DEAE-cellulose steps can be combined while maintaining the same overall high purification at a considerable reduction in effort.

5.3 Isolation of Purified Transfer Ribonucleic Acids

A. D. Kelmers and D. E. Keatherly

Transfer ribonucleic acids (tRiiAs) are macromclecular components of the protein biosynthetic pathway in all living organisms. They bind specific L-ansino acids and contain the anti-codon, the "mirror*-'mage" of ORNL DWG 73 -9931 Rl

Fig. 5-2. Combined Batch-Gradient DEAE-Cellulose Chromatography

Using a 4.6- by 8-cm Column Preloaded with Dialyzed Urine. -254- the trinucleotide genetic code word, whicn controls the correct sequential insertion of the amino acids during protein synthesis. In addition to their primary function in protein synthesis, specific tENAs have "been

shown to exhibit, control and regulatory functions within the cell. In order to understand the biochemical and physiochemical functions of tRNAs on a molecular level, it has been necessary to purify individual tREAs.

Reversed-phase chromatography systems have been previously developed

for the isolation of tRNAs,^ ^ and these methods were used in the large-

scale production and distribution of purified tRNAs. However, the methods have been limited to only two groups of tRNAs: (l) those which are in a

favorable region of the chromatographic separation (free of overlapping adjacent peaks), or (2) those for which unique rechromatography conditions

could be established by manipulation of the experimental parameters so

* that tRNAs that cochromatograph in one column can be resolved in a subsequent chromatographic step. It would be desirable to have a separation method based on the unique biological properties of tRNAs, rather than chromatography cally exploiting minor chemical differences. Such a method would be equally applicable to any tRNA, and would allow purification of minor tRNAs observed during tumor formation or cell differentiation. Affinity chromatography columns, which employ enzymes chemically bonded to an inert support, could offer such a general tRNA purification system. Selection of the enzyme, the bonding reaction, and the solid support are critical to the successful development of such a chromatography system. Evaluation of these points is under way, as discussed in the following sections. -255-

5.3*1 Selection of Factor Tu

Several enzymes are involved in the interactions of aminoacyl-tKNAs with the growing peptide chain on the ribosome. These are referred to as

"elongation factors," identified as Tu, Ts, and G. Factor Tu reacts only with aminoacylated tRNAs (those with the amino acid bonded to the 3' terminus) and not with others. Si* ~.e tRNAs can be selectively aminoacylated by adding only one amino acid to the sminoacylation reaction mixture, factor Tu could thus afford a suitable enzyme for affinity chromatography columns for the separation of aminoacylated tRNAs from other tRNAs.

The reaction between Tu and aminoacyl-tRNA (aa-tRNA) occurs in two

steps, as shown:

Tu + GTP - Tu-GTP + aa-tRNA - Tu-GTF-aa-tRNA , 1 , (Complex I) (Complex II) * 1 ""'

The first reaction involves the interaction of factor Tu and guanosine 5'- triphosphate (GTP) to form the intermediate complex I. Actually, since

Tu is unstable, it is utilized as a stable duplex with either Ts or

guanosine 5'-diphosphate (GDP). Ts, if present, would catalyze the

exchange

Tu-Ts + GTP - Tu-GTP + Ts , (2)

and GDP, if present, would have to be removed simultaneously with Tu-GTP

formation since it has a 100-fold higher binding coefficient for Tu than

does GTP. Complex I, then, reacts with any aminoacyl-tRNA to form the

ternary Complex II. The formation of Complex I and II can be followed

by a simple Millipora filter assay.Complex I binds to Millipore

filters in the presence of 10 mM MgCl0, and can be assayed by counting radioactive GTP. Complex II does not bind and can be measured by

difference (i.e., the decrease in bound radioactivity). -256-

The affinity chromatography system thus visualized would Involve factor Tu chemically bonded to a suitable inert support, such as glass, diatomaceous earth, or Sepharose. The tRWA would first be aminoacylated with the selected arainoacid and recovered (from the synthetase and other reactants) by DEAE-cellulose chromatography. It would then be mixed with

GTP and suitable cofactors, and passed through the Tu affinity chromatography column. Only the aminoacylated tENA would bind to the column; all others would pass through. The column would be washed with additional binding solution to complete the removal of the other tKNAs, and subsequently a suitable discharge solution would be passed through the column to release the aminoacyl-tKNA. If mild conditions can be employed, it should be possible to reuse the column.

5.3.2 Isolation of Factor Tu

E. coli bacteria are an easily obtained source of elongation factor

Tu, and methods for the purification of Tu to homogeneity have been published.^ Partially purified factor Tu should be adequate for the preparation of the affinity chromatography columns, at least in the preliminary experiments. coli Q-13, a ribonucleose deficient strain, was grown overnight in the 350L fermentor, and 1260 g (wet weight) of cells was harvested in a Sharpies AS-26 centrifuge. Four hundred grams was further processed immediately, and the remainder was frozen in liquid nitrogen and stored at -170°C. The cells were disrupted by passage through a Gaulin press at 8,000 to 10,000 psi and then centrifuged at 10,000 rpm in a Sorvall RC-2-B to remove cell debris. This was followed by high-speed centrifugation at 32,000 rpm in a Beckman L2-65B -257- to yield a clear supernate free of ribo somes. The factors Tu-Ts and G were recovered by ammonium sulfate fractionation in the precipitate from

37 to 6b°/o saturation The recovered precipitate was dialyzed to remove the ammonium sulfate, and then incubated with guanosine 5'-diphosphate

(GDP) to dissociate Tu-Ts and form Tu-GDP. DEAE-cellulose chromatography was employed to resolve the Ts, Tu-GDP, and G. The Tu-GDP peak was identified by the Millipore disc assay after incubation of samples with radioactive GDP. Finally, precipitation with 55$ saturation ammonium sulfate was used to recover the Tu-GDP. It was dissolved in a Tris (pH

7.5) buffer containing magnesium acetate, p-mercaptoethanol, and guanosine

5'-diphosphate to stabilize the factor Tu, and stored in liquid nitrogen.

To obtain a measure of the quantity of active Tu in the samples, GDP exchange was measured by the Millipore disc method. A maximum value of approximately 200 pmoles per 20 p.1 of Tu solution was obtained, as shown in Fig. 5.3' The data suggest the existence of two exchange reactions: one rapid, complete in one minute or less; and the other slow, showing increased GDP binding for 15 min. Also, greater binding was obtained at

37°C than at 0°C, indicating that the factor Tu is not readily denatured at 37°C. In the complete system, Ts catalyzes the exchange of GDP with 17 Tu-GDP. Since a separate Ts peak could not be demonstrated following the DEAE-cellulose chromatography, the Tu-GDP sample may contain Ts; the rapid exchange may, therefore, be a. measure of the Tu-Ts present, while the slow exchange could indicate a measure of Tu-GDP exchange. The presence of Ts, or other proteins, should not interfere with the affinity chromatography columns. -258-

ORNL DWG 73-9930 R I

Fig. 5.5^ GDP Binding to Tu-GDP Determined by 3H Labeling. -259-

5.3-3 Complex I and II Formation

A series of experiments was carried out to investigate the operational parameters for the formation of Complex I (Tu-GTP) and Complex II

(Tu-GTP-aa-tRM.) using the Millipore disc assay method, before operating affinity chromatography columns.

Factor Tu exhibits a 100-fold stronger binding coefficient for GDP than GTP; thus, the solutions must be free of GDP in order for Complex

I to form. Since GTP is a relatively unstable compound and readily dissociates to form GDP and other products, and since the factor Tu is stored as the stable Tu-GDP complex, an energy system was added to constantly regenerate any GTP that dissociated to GDP. Phosphoenolpyru- vate (PEP) and pyruvate kinase (PK) were used as the energy system.

Figure 5.4 shows the Complex I formation at two temperatures in the presence or absence of the energy system. The apparently higher values at 37°C in the absence of energy are probably due to the higher binding constant for the GDP formed by hydrolysis of GTP at that temperature. Wo significant difference was noted at 0°C. It was concluded that the addition of an energy system is necessary, and PEP and PK were added to all subsequent experiments.

In Fig. 5«5> the results of a series of tests to define the temperature range for Complex I formation are shown. The rate increased from

0 to 37°C. Higher values were also obtained at higher temperatures, but this may simply be a reflection of the rate. Since the enzyme denatured rapidly at 50°C, it becomes apparent that temperatures ranging from 0 to

37°C could be employed. -260-

Fig. 5*4. Effect of the Addition of an Energy System (Phosphoenol- pyruvate and Pyruvate Kinase) on the Formation of Complex I. Fig. Effect of Temperature on Complex I Formation. -252-

The pH is another important operating parameter that must he

established. Since the aminoacyl-tRNA bond is rapidly hydrolyzed in

neutral or basic solutions, it is desirable to operate at as low a pH

as is consistent with Complex I and II formation. As shown in Fig. 5.6, both the value and pH optimum for Complex I formation are dependent upon

temperature. At 0°C the pH optimum is near 6.5^ while at 37°C it is near

8. These results would suggest the use of lower temperatures as a trade-

off of capacity for stability. If aminoacyl-tRNA hydrolysis is a problem

in affinity chromatography column operation, practical operation could

possibly be achieved at 0°C at a pH as low as 5.5=

The length of time that Complex I is stable, once it is formed, is

an important operational parameter. The series of experiments summarized

in Fig. 5.7 show a complex relationship of quantity of Complex I and time.

At short times (15 to 30 min) the results are as expected from Fig. 5.6;

chat is, initial formation at 0°C and pH 6.5, and at 37°C and pH 7.5* gave

higher values. At 37°C, at either pH, longer times first gave a decrease

and then an increase in the apparent value of Complex I. This can be

explained as the exhaustion or denaturation of the energy system after

1 to 2 hr, followed by formation of the more stable Tu-GDP. The complex

initially formed increased steadily with time, perhaps representing con-

tinued slow formation of Tu-GTP in the absence of stoichiometric amounts

of Ts. In any case, these experiments indicate that Complex I Is

sufficiently stable to allow a variety of experimental steps to be used

in the formation of Complex II and isolation of aminoacyl-tRNA.

Attention was then directed to the formation and stability of Complex

II, as measured by the Millipore disc assay. The series of experiments ORNL DWG 73-3929RI

Fig. 5*6. Complex I Formation as a Function of pH and Temperature. -603-

TIME AFTER INITIAL FOR M AT 10 N ( hr)

Fig. 5-7* Stability of Complex I at 0 and 57°C After Initial Formation. -265- suinmarized in Fig. 5.8 show that at a pH of 6.5 and 37°C Complex II is stable up to 2 hr. As expected, mixtures at pH 6.5 were more stable due to decreased hydrolysis of the aminoacyl-tRNA bond. At pH 7.5 a rapid decrease of Complex II was noted after 20 min, presumably due to hydrolysis.

E. coli factor Tu was shown to be equally effective in homologous or heterologous reaction mixtures for the formation of Complex II (Fig.

5-9). Phenylalanyl-tRNA, from either calf liver or E^ coli, was equally efficient in Complex II formation with E^ coli Tu. In both cases, a 1:1 molar relationship is shown between the phenylalanyl-tRNA added and the

Complex II formed. This is a most favorable finding with regard to the applicability of affinity chromatography columns since one column, prepared with Ej_ coli factor Tu, could be used to recover aminoacyl-tRNAs from many sources.

5.4 Distribution jf Products

A. D. Kelmers and D. E. Heatherly

The distribution of samples of purified products for evaluation or experimental use by others of this program has been an important function.

The purified tRNAs are being distributed in accordance with the guidelines established by the National Institute of General Medical Sciences, the contracting agency. Samples are made available to scientists throughout the world, including iron curtain countries, on request after evaluation by a review committee. Supplies of two highly purified tRNAs from E^ coli, formylmethionine and glutamic acid, have been distributed to 15 investigators in the preceding six-month period. These shipments totaled 2788 mg. -266-

•x (O CD I in r* o* a z QC O

o o o K> CM ~ (ossp J94IU/saioaid) n xaidlNOO -606-

ORNL 0W6 73-99I7RI

Fig. 9» Homologous and Heterologous Formation of Complex II

Using E^ coli Tu and Either Calf Liver or E^ coli Phenylalanyl-tRNA. -268-

The project for the purification of erythropoietin is scientifically coordinated with Dr. E. Goldwasser of the Argonne Cancer Research Hospital.

A number of small samples have been shipped to him for standardization of the analytical procedure. In addition, 180,000 I.U. of Step 2 Epo has been shipped to Dr. Goldwasser. This material will be further purified by him and then utilized in a series of limited clinical tests. It is anticipated that injection of this material in selected patients will demonstrate a temporary relief of anemia.

5.5 Separation and Comparison of Sequence of Three FormyMethionine tRNAs from E^ coli B. Z. Egan and J. P. Eubanks

Polypeptide synthesis on the messenger RNA is initiated by the trinucleotide codon AUG. This is recognized by a unique transfer RNA, identified as the "chain indicator" tRNA. This tRNA is aminoacylated with methionine and then enzymatically formylated to formylmethionine, thus effectively blocking any potential for peptide bond formation on the amine side of the amino acid. Considerable interest has been expressed in this formylmethionine tRNA because of its potential regulatory role in protein synthesis. The primary nucleotide sequence of one formylmethionine tRNA l8 has been published, and the effects of chemical modifications on this molecule have been interpreted in terms of this structure. The superior resolution of reversed-phase chromatography systems, however, resulted in the resolutiofMpf n of three chromatographicallf Mpf y fMpdistincf t formylmetr Lcuine tRNAs (tRNA ). Samples of tRNA^ and tRNA^ have been distributed to many other investigators. By further rechromatography on reversed-phase chromatography system 5 (RPC-5)^ the tRNAfMe t was -269-

fMet fMet resolved into two peaks, designated tRNA^ and tRNA^ . It was fMet deemed important to determine the structural relationships among tRNA^ , fMet fMet tRNAg , and tKNA in terms of the published sequence. The existence of multiple chain initiator tRNAs offers the possibility of selective

control of the initiation on specific messenger ENAs through availability of specific tRNAs . 5.5.1 Separation of tRNA^ and tRNA^

Samples of tRNA^6^ were rechromatographed on RPC-5 columns.1^ In

a typical run, 500 A^q units was applied to a 1 x ll6 cm column and

eluted with a NaCl gradient ranging from 0.U0 to 0.60 M. Complete sepa-

ration of two peaks was obtained. The pooled fractions were concentrated

about 20-fold using a 2.5 x 19 cm DEAE-cellulose column. The eluent

contained 0.3 to 0.8 M NaCl, 0.01 M MgClg, 0.02 M Tris-HCl, pH 7.0,

0.002 M Na^SgO^, and 0.1 ml/liter isoamylacetate.

5.5.2 Comparison of Ribonuclease Tj_ and Pancreatic Ribonuclease Digests

In order to resolve the primary nucleotide sequence of these tRNAs,

a series of digestions was carried out with the specific ribonuclease

T^, which hydrolyzes phosphodiester bonds only on the 3' side of guanosine

residues, or pancreatic ribonuclease, which hydrolyzes phosphodiester

bonds on the 3' side of any pyrimidine. Comparison of the composition

of the oligonucleotide fragments generated by these digestions yields

comparative information concerning the sequence. Reversed-phase chromatography was employed to separate oligonucleo- 20 fMet tides obtained from ribonuclease digests of the three tRNAs . In

a typical digestion, 50 to 100 \±1 of the T:. (10,000 units/ml) or pancreatic -270- ribonuclease solution (5.0 mg/ml) was added to 0.5 to 1.0 ml of solution containing about 100 and incubated for 4 to 5 hr at 37°C. Digests were then applied directly to the RPC-5 column.

Figure 5.10 compares the pancreatic ribonuclease digests, and Fig. 5.11 shows chromatograms obtained from the ribonuclease T^ digests of the three t,RNAs'^e't. The peaks were identified by base (nucleoside) composition, spectra, and comparison to the published structure. In addition, larger oligonucleotides were rechromatographed following further enzymatic fMet digestion. The chromatograms of the three tRNAs are very similar, with some significant differences in the presence or absence of specific fragments. Table 5-1 summarizes the observed differences in terms of the presence or absence of related oligonucleotides. fMet 5.5.3 Relationships Among Three tRNAs fMet The results can be summarized as follows: (l) tRNA^ appears to fMet fMet correspond to the published sequence for tRNA ; (2) tRNA^ differs fMet 4 from tRNA^ in that the s Up in nucleotide position 8 has Interacted with 21 fMet Gp in position 13 to form a cross-linked product; (3) tRNA^ differs from tRNA^Met in that m^Gp in position k-7 has been replaced by Ap.

5.6 Sequence Studies on Phenylalanine tRNA from Calf Liver

B. Z. Egan and J. P. Eubanks

Most tRNA primary base sequence studies have been concerned with tRNAs derived from bacterial sources. Only recently have these studies been extended to mammalian systems. The mammalian tRNAs are generally more complex in that they contain larger families of isoacceptors, making the purification of a specific tRNA more difficult, and these tRNAs also contain more modified bases. -271-

ORNL DWG. 73-2000R2

2.0 TIME (hr)

Fig. 5.10. Comparison of Pancreatic Ribonuclease Digests of

(a) tRNAf*et, (b) tMAfet, and (c) tRM^et. Column, 0.63 cm diam by

98 cm long; eluent, 250 ml of 0.5 to 2.0 M ammonium acetate—acetic acid gradient, pH 4.5; temperature, 37°C. -272-

ORNL-DWG 73-2487R2

TIME (hr) fMet Fig. 5.11. Comparison of Ribonuclease Tx Digests of (a) tRM^ , (b) tRHA?^6^, and (c) tRM?4^. Column, 0.63 cm diam by 98 cm long; B D eluent, 600 ml of 0.6 to 1-9 ^ ammonium acetate—acetic acid gradient, pH 4.5; temperature, 37°C» -P'VO _

Table 5-1- Comparison of Oligonucleotides Relating Three tRNAsfMe t

fMet et fMet tRNA'A tENAf tRNA:

Pancreatic Ribonuclease Fragments

GAAGm7GUp GAAGm7GUp GAAGAUp

GGAGCp GGAGCp

Ribonuclease T^ Fragments k k s UGp s UGp

CAGp CAGp m7GUCGp m7GUCGp AUCGp CAGp U *Gp -274-

By combining various chromatographic techniques, including BD- cellulose, reversed-phase, and Sephadex A-50 chromatography, we have separated and purified milligram quantities of several tRNAs from calf 22 Ser Phe liver. We have prepared sufficient quantities of tRNA and tRNA

to begin investigating the primary base sequence of these tRNAs. Extensive rechromatography on reversed-phase (RPC-5) columns indicated that the S er tRNA contained two isoacceptors which would be extremely difficult to Phe separate, while the tRNA preparation appeared more homogeneous. The tRNAPh e product was therefore chosen for sequence studies. Total base composition was obtained by nucleoside analysis,2 3 and oligonucleotides from a pancreatic ribonuclease digest were separated 24 by reversed-phase chromatography. Preliminary results were obtained on the base composition of the oligonucleotides, and some of the oligonu-

cleotides were further digested with ribonuclease T^ to aid in structure

assignment. pVip 5.6.1 Base Composition of Calf Liver tRNA Phe

Approximately 12 A2gQ units of calf liver tRNA was digested to the nucleoside level with snake venom phosphodiesterase and bacterial

alkaline phosphatase. The resulting digest was separated in.o its con-

stituent nucleosides by chromatography on an Aminex A-6 column. The

results are summarized in Table 5.2. The values shown are the average

for two analyses. In addition to the nucleosides listed, at least two

unidentified peaks were obtained. Dihydrouridine was not determined.

The values of nanomoles per milliliter were calculated from the

peak areas, and these values were normalized by assuming that 31.7 Phe Phe Table 5.2. Base Composition of Calf Liver tRNA and Comparison to Rabbit Liver tRNA

Rabbit Liver Amount ,™ Phe tRNTAA ^ nmoles/ml residue/moleculea (residues/molecule )

Uridine (u) 31.7 9.0 9

Pseudouridine (Y) 13.4 3.8

Ribothymidine (T) 2.31 0.7 1

Guanosine (G) 71.9 20.k 19

Cytidine (c) 63A 18.0 17

Adenosine (A) 55.6 15.8 16

St Calculated assuming the uridine residues are identical (9.0 per molecule) for calf liver and rabbit liver tRNAPhe.

1 Base d on a tentative structure proposed by Petrissant and Keith, 25 -276- nmoles/ml for uridine represents 9.0 uridine residues per tRNA molecule.

This allows comparison of the results with the tentative structure of Phe 25 rabbit liver tRNA determined by Petrissant and Keith. Phe The major base composition of tRNA from calf liver and rabbit

liver is very similar.

5.6.2 Pancreatic Ribonuclease Digestion of tRNA^*16

Approximately 60 &26o units of calf liver tENAPiie in 1.0 ml of 0.05

M Tris-HCl, pH 7.3, was digested with 50 i~il (5.0 mg/ml) of pancreatic

ribonuclease for k hr at 38°C. The digest was chromatographed in two 24 separate runs on an RPC-5 column to separate the oligonucleotides.

A typical chromatogram is shown in Fig. 5.12. The peak fractions were pooled, neutralized, and the eluent removed by lyophilization. Base

composition of the peak fractions was then determined after digestion with snake venom phosphodiesterase and bacterial alkaline phosphatase to nucleosides, as described above. In addition, aliquots of several of

the peak fractions were further digested with ribonuclease T^; these

digests were chromatographed on an RPC-5 column, and the base composition

of the resulting products was determined.

Results of these analyses are incomplete. However, preliminary data Phe offer supporting evidence for the similarity in structure of tRNA from

calf liver and rabbit liver.

Peak 2 contains Cp and Yp mononucleotides in a molPhe erati o of 11.7:2. A ratio of 11:2 would be expected for rabbit liver tRNA . yp also

occurs in peaks 6 and 17. Peak 3 is yp.

6, 13, 17 6 Peaks 10, 12, and contain2 modified bases. Peak is probably a mixture of dinucleotides, mp Gyp and ACp. Peak 12 may contain ORNL DWG 73-8634

Fig. 5*12. Reversed-Phase Chromatcgram of a Pancreatic Ribonuclease

Digest of Calf Liver tRNAPhe. Column, 0.63 cm diam by 97 cm long, RPC-5; eluent, 500 ml of 0.40 to 2*5 M ammonium acetate—acetic acid gradient, pH b.k; temperature, 37°C; flow rate, 1.15 ml/min. 9 -278- the trinucleotides Am GCp and GGCp. Peaks 13 and 14 were not completely- resolved, resulting in mixtures of oligonucleotides which were not readily identifiable. However, the minor base in peak 13 appears to be 7-methyl- 7 guanosine. This would be comparable to the suggested structure, AAAGm Gyp.

Ribothymidine is present in peak 14, probably as the trinucleotide GGTp.

The minor base In peak 17 is believed to be the base "Y", which occurs in tRNA^e from several sources.2^ ^ The base composition of this fragment is consistent with the formulation, AAYAYp. Peaks 15 and 16 have been tentatively identified as AGACmyp and GAAAYp, respectively. Dihydrouridine was not determined for any of the samples, and could possibly influence the apparent structure of the oligonucleotides.

Further analysis of ribonuclease T^ digests and more quantitative base composition data are required to conclusively characterize the oligonucleotides and complete the assembly into the total structure of the tRNA molecule. However, many areas of similarity with the sequence for rabbit liver phenylalanine tRNA are apparent.

5.7 References for Section 5

1. E. Goldwasser and C. K.-H. Kung, Proc. Natl. Acad. Sci. USA 68, 697-98 (1971). 2. E. Goldwasser, W. F. White, and K. B. Taylor, Biochim. Biophys. Acta 64, 487-96 (1962).

3. E. Goldwasser and C. K.-H. Kung, Ann. N.Y. Acad. Sci. 149, 49-53 (1968).

4. D. Pressman and A. L. Grossberg, The Structural Basis of Antibody Specificity, W. A. Benjamin Inc., New York, N. Y., 19btf.

5. J. C. Shooley and J. F. Garcia, Proc. Soc. Exper. Biol. & Med. 109, 325 (1962). -279-

6. R. D. Lange, E. Gardner, C. S. Wright, and N. I. Gallagher, Brit. J. Haemat. 10, 69-74 (1964).

7. P. Cuatrecases, J. Biol. Chem. 245, 3059 (1970).

8. D. Metcalf, J. Cell. Physiol. 76, 89-IOO (1970).

9. E. R. Stanley and D. Met calf, Aust. J. Exp. Biol. Med. 47, 453-66 (1969). 10. A. D. Kelmers, C. W. Hancher, E. F. Phares, and G. D. Novelli, Methods in Enzymology XX, 3-8 (1971).

11. A. D. Kelmers, H. 0. Weeren, J. F. Weiss, R. L. Pearson, M. P. Stulberg, and G. D. Novelli, Methods in Enzymology XX, 9-34 (1971). 12. C. W. Hancher, R. L. Pearson, and A. D. Kelmers, Methods in Enzymology (in press).

13. R. L. Pearson, J. F. Weiss, D. W. Holladay, and A. D. Kelmers, Methods in Enzymology (in press). 14. D. E. Millar and H. Weisshach, Arch. Biochem. Biophys. l4l, 26-37 (1970).

15. R. Ertel, B. Redfield, N. Brot, and H. Weisshach, Arch. Biochem. Biophys. 128, 331 (1968). 16. K.-I. Arai, M. Kawakita, and Y. Kaziro, J. Biol. Chem. 249, 7029- 37 (1972).

17. H. Weisshach, D. L. Millar, and J. Hachmann, Arch. Biochem. Biophys. 137, 262 (1970). 18. S. K. Dube, K. A. Marcker, B. F. C. Clark, and S. Cory, Nature 2l8, 232 (1968).

19. R. L. Pearson, J. F. Weiss, and A. D. Kelmers, Biochim. Biophys. Acta 228, 770 (1971). 20. B. Z. Egan, Biochim. Biophys. Acta 299, 245 (1973).

21. F. Berthelot, F. Gros, and A. Favre, Eur. J. Biochem. 29, 343 (1972). 22. R. L. Pearson, C. W. Hancher, J. F. Weiss, D. W. Holladay, and A. D. Kelmers, Biochim. Biophys. Acta 294, 236 (1973).

23. R. P. Singhal and W. E. Cohn, Biochim. Biophys. Acta 262, 565 (1972).

24. B. Z. Egan, Biochim. Biophys. Acta 299, 245 (1973). -280-

25. 0. Petrissant and G. Keith, personal communication. 26. U. L. RajBhandary, S. H. Chang, A. Stuart, R. D. Faulkner, R. M. Hoskinson, and H. G. Khorana, Proc. Nat. Acad. Sci. U.S. 57; 751 (1967). ~~ 27. K. Nakanishi, N. Furutachi, M. Funamizu, D. Grunherger, and I. B. Weinstein, J. Am. Chem. Soc. 92, 7617 (1970).

28. B. S. Dudock, G. Katz, E. K. Taylor, and R. W. Holley, Proc. Nat. Acad. Sci. U.S. 62, 9^1 (1969).

29. B. C-. Barrell and F. Sanger, FEBS Letters 3, 275 (1969).

30. L. M. Fink, K.. W. Lanks, T. Goto, and I. B. Weinstein, Biochem. 10, 1873 (1971). -281-

6. HYDROGEN PRODUCTION

Increasing interest is being shown in a "hydrogen economy" based on 1 2 the use of hydrogen as a medium of energy transport. 3 Other aspects of the use of hydrogen as a fuel, such as economy and uses other than in 3-9

power production, have been discussed elsewhere.

Present methods for the production of hydrogen are primarily

electrolysis of water and variations of steam reforming of natural gas.

It is desirable to have other methods available as alternatives and for

application in specialized situations.

6.1 Enzymatic Hydrogen Production

B. Z. Egan and J. P. Eubanks

It is well known that certain bacteria (e.g. Clostridium pasteurianum)

in the proper environment will produce hydrogen. Such systems have been

investigated,^ and two necessary enzymes, ferredoxin and hydrogenase,

have been isolated.rp^ hy^og^.pro^c^g process has been

demonstrated in an aqueous environment combining cellular extracts

containing the enzymes, and using sodium dithionite, Na^SgO^, as the

reducing agent. Hydrogen was formed as the reduction product from the

water, and presumably Na^SgO^ was oxidized to a higher oxide such as

Ha SO-, Na S 0^, or Na_S 0 ; the oxidation product has not been identi- CM dt P Ct Ci O Ui CA [ fied. One possible reaction may be expressed as

H O + Na„S„0] fe^red°Xin > H„t + NaJ3.n . (l) 2 2 2 4 hydrogenase 2 2 2 5

In order to make such a production scheme more attractive, it will

be necessary to regenerate the reactant, Na^SgO^, from the oxidation -282-

l8 product. It has "been reported that the higher oxides can be reduced to the dithionite by thermal decomposition:

Na.2S207 ^ - > l/2 O2T + Na2s206 550 - 630°C > (2)

630 71Q G 1/2 Ogt + Na2S205 " ° > 1/2 02t + Na2S20^ .

It has also been reported that the thermal decomposition of the oxidation products may produce oxides of sulfur. Nevertheless, a cyclic process can be considered such that Na S_0) + E 0 redoxin > t + NaJ3_0_ „ 2 2 4 o2 hydrogenase H 2 2 2 5-7 buffer (3)

Na2S2°5-7 630 " 71Q°C > V + Na2S2°4 •

The net reaction would therefore result in the production of hydrogen

(and oxygen) from water.

It is expected that for such a process, the necessary enzymes would eventually be immobilized on fixed, solid particles that can be used in a fixed-bed reactor. Such a process scheme is illustrated schematically in Fig. 6.1. Before such a process can be seriously considered, additional information must be obtained on the stability of the dithionite and enzymes, the composition of the reaction products, the reaction kinetics and efficiency of both the enzymatic reduction and the regeneration cycle, and elucidation of the chemistry of the regeneration cycle.

A program has been initiated to study the basic kinetics of the enzyme-catalyzed reaction using crude extracts of the enzymes, and the chemistry associated with the regeneration of the dithionite. -283-

ORNL DWG. 72-6893

Na2S204 , PHOSPHATE, H20

WATER ENZYMATIC REACTOR (FIXED BED OF IMMOBILIZED ENZYMES REACTANT 30-50° C) STORAGE AND MAKE-UP

BY-PASS LINE

Na2S205 H 0, PHOSPHATE 2 PHOSPHATE |TnnrYrrrvrj

COOLANT

Na2S204 prrrrnrvvrri

COOLANT

700° C °f THERMAL [ REGENERATOR 40°C

Fig* 6.1. Process Steps in Hydrogen Production by an Enzyme-

Catalyzed System Using Dithionite as the Reductant. Methods and conditions were determined for growing several grams of Clostridium pasteurianum bacteria as a source of crude enzymes or cell extracts. Freeze-dried cultures of the bacteria were obtained, and an initial inoculum was grown in the recommended beef liver medium for anaerobes. Small pieces of beef liver were soaked in tap water overnight in the refrigerator, autoclaved, and filtered through cheesecloth.

Peptone (10.0 g/liter) and KgHPO^ (1.0 g/liter) were added, the pH was adjusted, to 8.0, and the broth was filtered through paper. Solid CaCO^, meat, and broth were added to the culture tube, and the subculture was grown at 37°C in a nitrogen atmosphere.

After 2b to 36 hr, 5 ml of the subculture was aseptically transferred to 500 ml of a 37°C sterile medium through which nitrogen was 13 bubbled continuously. The culture medium contained the following quantities of materials per liter of distilled water: sucrose 20.0 g

calcium carbonate 10.0 g

potassium hydrogen phosphate, KgHPO^ 0.5 g

potassium dihydrogen phosphate, KH«-,P0^ 0.5 g

magnesium sulfate, MgS0^*7H20 0.1 g

sodium chloride 0.1 g

ferric chloride 0.1 g

sodium molybdate, NagMoO^^HgO 0.01 g

d-biotin 1.0 pg

p-aminobenzoic acid 1.0 pg

After about 2b hr of growth, 100 ml of this culture was transferred to a 5-liter fermentor containing the same medium. A nitrogen flow rate -285- of approximately 30 liters/hr was maintained. Growth was measured "by

periodically determining the ahsorbance at 650 nm (A^q). NO growth

was observed for 30 hr, and it was found that the pH had increased to

8.1. After the pH was lowered to 6.8 by the addition of KHgPO^, growth

resumed. In one 7-hr period, the A^^ increased from O.29 to 1.5^ and

the pH had dropped to 5*3• The cells were harvested with a centrifuge

and stored frozen (-20°C) under a nitrogen atmosphere.

In another fermentation run, the medium was supplemented with

(0.8 g/liter), the quantity of CaCO^ was reduced to 5-0 g/liter,

and the pH was periodically adjusted to maintain a value between 6.2 and

6.8. In one 6-hr period, the Ag,_0 increased from 0.29 to lA. The growth

rate therefore appeared to be similar for the different media. The

maximum attained was 2.2, corresponding to 2 to 3 g of wet cells per

liter.

6.2 Thermal Hydrogen Production

C. E. Bamberger

Analyses of the economy of hydrogen generation presently suggest

that a reduction in the price of hydrogen may be possible if thermal

energy could be used directly to decompose water, rather than indirectly

to generate electricity as an intermediate. The projected availability

of thermal energy at low cost from nuclear reactors is a very attractive

possibility for the decomposition of water. Since any chemical scheme

with a recycle mode of operation and only thermal energy for a driving

force is, in practice, limited by the Carnot cycle, the operating

temperature of the reactor sets the upper limit. -286-

We are currently using available thermochemical data to evaluate

several chemical reactions that can be combined in a cycle where the

net result is the decomposition of water. One of the selection criteria

for a reaction that appears to be thermodynamically favorable at the

temperatures of Interest is that the reactants do not involve elements

that are in relatively short supply. Some reactions are being examined with the components at other than unit activity. Although reducing the

activity by use of a solvent may provide additional driving force for

the reaction, it also adds complexity to the system. Presently, solvents

are being evaluated on the basis of their inertness toward the reactants

and their capability for dissolving some of the products of the reaction.

The solvents under examination range from organic compounds to molten

salts.

6.3 References for Section 6

1. J. O'M. Bockris, "A Hydrogen Economy," Science 176, 1323 (1972).

2. Anon., "Hydrogen: Likely Fuel of the Future," Chem. Eng. News, p. l4 (June 26, 1972); "Hydrogen Fuel Use Calls for New Source," Chem. Eng. News, p. l6 (July 3, 1972); "Hydrogen Fuel Economy: Wide Ranging Changes," Chem. Eng. News, p. 27 (July 10, 1972).

3. Hydrogen and Other Sypthetic Fuels, A Summary of the Work of the Synthetic Fuel Panel, USAEC TID 26136, Sept. 1972.

C. Marchetti, 'hydrogen and Energy," Chem. Econ. Eng. Rev. (Tokyo) _5, No. 1, 7 (1973).

5. G. DeBeni and C. Marchetti, "Hydrogen, Key to the Energy Market," Eurospectra IX, No. 2, 46, (1970).

6. N. Chopey, "Hydrogen: Tomorrow's Fuel?," Chem. Eng. p. 2h (Dec. 25, 1972).

7. W. E. Winsche, K. C. Hoffman, and F. J. Salzano, "Hydrogen: Its Future Role in the Nation's Energy Economy," Science l8o, No. 4093, 1325 (1973). -287-

8. J. Funk and R. M. Reinstrom, "Energy Requirements in the Production of Hydrogen From Water," IEC Process Design and Devel. _5, No. 3* 336 (1966).

9« F. F. Blankenship, S. S. Kirslis, and J. Braunstein, "Feasibility of Thermochemical or Radiolytic Production of Hydrogen as a Fuel From Non-Fossil Sources," p. 113 in Reactor Chem. Div. Ann. Progr. Rept. May 31, 1972, 0RNL-4801 (1972).

10. W. D. McElroy and B. Glass, eds., A Symposium on Inorganic Nitrogen Metabolism, Johns Hopkins Press, Baltimore, 1956.

11. H. C0 Peck, Jr., and H. Gest, "A New Procedure for Assay of Bacterial Hydrogenases," J. Bacteriol. 71, 70 (1956).

12. H. C. Peck, Jr., and H. Gest, "Hydrogenase of Clostridium butylicum," J. Bacteriol. 73, 569 (1957).

13. J. E. Carnahan and J. E. Castle, "Some Requirements of Biological Nitrogen Fixation," J. Bacteriol. 75, 121 (1958).

14. J. E. Carnahan, L. E. Mortenson, H. F. Mower, and J. E. Castle, "Nitrogen Fixation in Cell-free Extracts of Clostridium pasteurianum," Biochim. Biophys. Acta 44, 520 (1960).

15. L. E. Mortenson, R. C. Valentine, and J. E. Carnahan, "An Electron Transport Factor from Clostridium pasteurianum," Biochem. Biophys. Res. Commun. 7, 448 (1962).

16. R. C. Valentine, L. E. Mortenson, and J. E. Carnahan, "The Hydrogenase System of Clostridium pasteurianum," J. Biol. Chem. 238, ll4l (1963).

17. E. Knight, Jr. and R. W. F. Hardy, "Isolation and Characteristics of Flavodoxin from Nitrogen-fixing Clostridium pasteurianum," J. Biol. Chem. 24l, 2752 (1966).

18. M. Boros and B. Lorant, "Derivative Study and Determination of Alkyl Sulfate and Alkylarene Sulfonate,Seifen-Oele-Fette-Wachse 89, 555 (1963). ~~

19. L. Erdey, J. Simon, S. Gal, and G. Liptay, "Thermoanalytical Proper- ties of Analytical-Grade Reagents-IVA," Talanta 13, 67 (1966). 7. WATER POLLUTION STUDIES

Degradation of the nation's water resources by biochemicals and other organic materials is a crucial environmental problem, complicated by the variety, highly diverse nature, and concentration of such contaminants.

Instrumentation to determine the individual molecular contaminants is being developed and used in order to understand and combat this problem.

Automated, high-resolution liquid chromatographs, which were previously developed for the analysis of the molecular biochemical constituents in human body fluids (Sect. 2.1), have been applied to the analysis of various polluted waters. Samples of polluted natural waters were collected at different sites, concentrated up to 10,000-fold, and chromatographed on a high-pressure anion exchange column. Monitoring of the column eluate for ultraviolet absorbance and cerate oxidizability revealed the presence of numerous organic contaminants at concentrations less than 1 ^g/liter.

•The high-resolution chromatographs are also being used to investigate the effects of chlorination of sewage plant effluents and condenser cooling waters.

7.1 Automated Analysis of Dissolved Organic Compounds in Polluted Waters

R. L. Jolley, M. D. McBride, W. W. Pitt, S. Katz, and G. Jones

Water quality is currently a subject of considerable national concern.

The discharge of stable organic compounds in effluents from domestic and industrial sewage treatment plants to surface waters, and the possible buildup of the stable organic constituents by recycling, may adversely 1-3 affect water quality. The potable water used by people whose water supply is obtained from surface waters, particularly in heavily populated v

2 areas, is partially composed of reconditioned sewage. In order to determine the toxicological significance of such waters for consumers, it is necessary to identify and quantitatively measure the concentrations of stable organic compounds which exist in polluted waters. Assignment of the ultimate source of stable organic pollutants also requires that baseline data be accumulated for both sewage treatment plant effluents (domestic and industrial) and natural waters.

High-resolution chromatographic systems for analysis of organic constituents in the complex biological body fluids urine and blood have If been developed at the Oak Ridge National Laboratory. These systems have been successfully used for analyses of UV-absorbing constituents, car- 5-7 bohydrates, and oxidizable organics. Using these chromatographs and ancillary identification procedures, over 120 separated constituents have g been identified in body fluids. Katz et al. utilized the UV-Analyzer and the carbohydrate analyzer to separate and identify several organic compounds 9 in concentrates of effluents from sewage treatment plants.

Sewage treatment plant effluents are routinely chlorinated to milligram-per-liter concentrations of chlorine for disinfection purposes.

Only limited information has been available concerning chlorination effects

on the organic constituents in such effluents. There is an increasing

concerr as to whether stable chlorine-containing organic compounds are

formed during the chlorination process, concern which has evolved as a

result of the ecological problems attributed to chlorine-containing organic compounds present as almost ubiquitous pollutants in the 11-13

environment. For this reason, the effects of chlorination on

organic constituents in sewage effluents were examined. A methodology -290- for determining chlorination effects was developed which, combined 36 chlorination with CI radioactive tracer and high-resolution chromatographic separation of the chlorinated organic compounds using sensitive radioactive tracer monitoring.

This progress report describes results from the continued application of these analytical procedures to the problem of identification and measurement of individual organic compounds in polluted water, both process effluents and natural waters, and also, the results of the CI tracer experiments.

7.1.1 Effluents from Sewage Treatment Plants

Thirty-seven (13 during this report period) stable organic compounds have been identified in effluent samples (both chlorinated and unchlorinated) from the primary stage of domestic sanitary sewage treatment plants. The identities, concentrations, and procedures used for identification of these compounds are given in Table 7«1« A standard reference uv chromatogram of primary sewage treatment plant effluent is shown in Fig. 7-1, with the identified organic compounds indicated at their respective elution positions.

The high-resolution chromatographic separation techniques have been discussed in detail in several previous reports.The identification techniques have been integrated into a routine procedure involving lyophilization to remove buffer salts and water, followed by additional chromatography

(liquid and gas), uv spectroscopy, and mass spectrometry. Details of the routine identification procedure have been given previously. In addition to the 37 identified compounds, 78 compounds have been char-

acterized with respect to molecular weight, gas and liquid chromatographic retention times, uv spectra, and mass spectra. As more experience Table 7«1« Identification of Molecular Constituents in the Effluent of the Primary Stage of a Domestic Sewage Plant

Concentration Q Compounds Identification Method (^g/liter)

Maltose AC, GC

Galactose AC, GC

Glucose AC, GC

Glycerine AC, GC, MS

N"*" -Me thy 1 - 4 -pyr i done - 3 - c a rb oxamid e AC, UV, GC 10

Phenylalanine AC, GC, MS 90

Uracil AC, CC, UV, GC, MS ho

5 -A c e tylamino-6-amino-3-me thylura c i1 AC, CC, UV, GC iko

N^"-Methyl-2-pyridone-5-carboxamide AC, CC, UV, GC 20

Thymine AC, CC, GC, MS -7

Theobromine AC, CC

7-Me thy lxan thine AC, CC ^90

Inosine AC, CC, UV, GC, MS 50 Table 7.1(continued )

Q Concentration Compounds Identification Method (|jbg/liter)

Hypoxanthine^ AC, GC, MS 25

Xanthine AC, CC, UV, GC 70

Q MS Copper(ll) acetate (binuclear) AC, CC, UV, GC, MS Adenosine AC, CC 1,7-Dime thy Xanthine AC, CC 3-Methylxanthine AC, CC, UV, MS Caffeine AC, CC, UV, GC, MS 50 Guanosine AC, CC, UV 70 1 -Me thy lxa n th ine AC, GC, MS 20 Uric acid AC, UV, GC, MS 5 Orotic acid^ AC, GC, MS Succinic acidb AC, GC, MS 6 Phenol^ AC, UV, GC, MS 10 3-Hydroxyphenylhydracrylic acid Table 7.1 (continued)

Concentration Compounds Identification Method (|i,g/liter)

Phenylacetic acid AC, GC ~10

4-Hydroxyphenylacetic acid AC, UV, GC, MS 190

Benzoic acid AC, GC, MS

2-Hydroxybenzoic acid AC, GC, MS 7

4-Hydroxybenzoic acid AC, GC

3-Hydroxybenzoic acid AC, GC, MS

3-Hydroxyphenylpropionie acid AC, CC, MS ~20

Indican*3 AC, GC, F o-Phthalic acidd AC, UV, MS 200 h p-Cresol AC, GC, MS 20

AC - anion exchange chromatography; CC - cation exchange chromatography;

UV - ultraviolet spectrum; GC - gas chromatography on two columns;

MS - mass spectroscopy; F - fluorescent spectrum. Table 7.1 (continued)

From chlorinated effluent.

Identified by A. W. Garrison, Ph.D., at Southeast Water Laboratory.

^Mill Creek sewage effluent.

i ro vo •tr I -295-

ELUTION TIME.hr. 0 r r~ T ~T ELUTION VOLUME, ml 0 25 50 75 100 150

I- 2 "o^ S5 ONNO Z> SZ UJ11ZZJ U X ua m>- >->->mca C-L >- 4 X XX X a.§ O o oo o z a Qa. Q Oo: Q£ cc < >• Mo 5- >->- > o 5 X• Xi X• X• cvi kit ro z YT 20 T 1 1 1 1 1 1 1 r 1.0 ANION EXCHANGE CHR0MAT0GRAPH RUN CONDITIONS > .8 0.94-cm -ID t 150-cm STAINLESS STEEL WITH 8-12 0 TEMPERATURE PROGRAM, AMBIENT TO 55*0 AT 11.5 hi LINEARILY IN CONCENTRATION FROM 0.015 M TO 6.0] UJ ELUENT FLOW RATE 89 ml/hr; COLUMN PRESSURE I zu C4O Otr

ELUTION TIME.hr. 30 31 32 33 34 35 36 37 38 33 40 41 42 43 44 45 46 47 48 49 50

ELUTION VOLUME,ml 250 300

280 MO OF EFFLUENT C0NCEN1RATED 1000 X ABSORBANCE *T 254 nm ABSORBANCE AT 280 nm

Fig. 7.1. Reference Chromatogram from UV-Analyzer of Primary Sewage Trea

Concentrated by a Factor of 1000, Showing Elution Positions of Identified Comp ORNL DWG 72- II705R4

100 150 200

o «o e ©

n 1 1 1 1 1 1 r~ T T T T T ANI0N EXCHANGE CHROMATOGRAPH RUN CONDITIONS' 0.94-cm -ID x 150-cm STAINLESS STEEL WITH 8-l2pDIAM. AMINEX A-27 RESIN ; TEMPERATURE PROGRAM, AMBIENT TO 55°C AT 11.5 hr ,- ELUENT GRADIENT INCREASING LINEARILY IN CONCENTRATION FROM 0.015 M TO 6.0 M AMMONIUM ACETATE, pH 4.4; ELUENT FLOW RATE 89ml/hr; COLUMN PRESSURE 1800 psig .

L_ 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 5U

400

ENTRATED 1000 X ABSORBANCE AT 2G4nm {- ABSORBANCE AT 280 nm irom UV-Analyzer of Primary Sewage Treatment Plant Effluent, g Elution Positions of Identified Compounds. !v

i BLANK PAGE -2 o6- and information is obtained on the interpretation of the mass spectra of trimethylsilyl (TMS) derivatives, additional identifications may result from reexamination of the TMS-mass spectra determined previously.

Nine uv-absorbing organic compounds (two during this report period) have been identified from samples of unchlorinated effluent from the

secondary stage of domestic sanitary sewage treatment plants. These

constituents have been quantified, and their concentrations and anion

exchange elution positions are shown in Table 7«2. A standard reference

chromatogram of secondary sewage plant effluent for purposes of reporting

elution positions and identifications is given in Fig. 7.2.

Five samples of effluents from the secondary stage and one from the

primary stage of the Oak Ridge East Sewage Treatment Plant (ORESP) were

chlorinated with Cl-tagged chlorination agents. The primary effluent

sample was chlorinated to 2 ppm chlorine residual (orthotolidine method)

using chlorine gas, and the secondary effluent samples were chlorinated to

1 ppm using chlorine gas in two cases and hypochlorite solution in the

others. Essentially the same results were obtained -with each reagent -with

one major exception. One of the chlorine-containing constituents

subsequently separated by anion exchange chromatography was considerably

higher in concentration after chlorination with hypochlorite solution than

when chlorine gas was used. Details of experimental procedure ana results 17 are presented elsewiiere. The major results are summarized below.

After chlorination, the samples were concentrated and chromatographed

on an anion exchange column. Analysis of the resulting chromatograms

(an example is shown in Fig. 7.3) showed the separation of at least 62

chlorine-containing constituents, some of which possibly were stable -297-

Tahle 7-2. Identification of Molecular Constituents in the Effluent of the Secondary Stage of a Domestic Sewage Plant

Concentra Compounds Identification Method (pg/lite

Glycerine AC, GC, MS 4 j

Uracil AC, CC, UV 30 j

5-Acetylamino-6-amino-3-methyluracil AC, CC, UV 30 j

l-Methylinosine AC, CC, UV 8o;

Inosine AC, CC, UV 20 j

7-Methylxanthine AC, CC, UV 5 j i

1-Methylxanthine AC, GC 6 j

? 1,7-Dimethylxanthine AC, CC, uv

p-Cresol AC, GC, MS 90 J

AC - anion exchange chromatography; GC - gas chromatography on two columns;

CC - cation exchange chromatography; UV - ultraviolet spectrum;

MS - mass spectroscopy. Concentration (^g/liter)

20 BLANK PAGE 4

2.0 tf i r n r _1 1 1 1 [— 1.0 ANION EXCHANGE CHROMATOGRAPI; .8 0.94-cm-ID x 150-cnri STAINLE TEMPERATURE PROGRAM^ .6 LINEARILY IN CONCENT RATI] ELUENT FLOW RATE 89 (toft <£ .4 m (K O (Q z ; —C^-——-—^A^ —j —_ <

i.vv;

ELUTION TIME.hr 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 ELUTION VOL.,ml 250

280 fiQ. OF EFFLUENT CONCENTRATED 2000 X ABSORBANCE:& ABSORBANCE: •wi'/i Fig. 7.2. Reference Chromatogram from UV-Analyzer of Second;

Concentrated "by a Factor of 2000, Shoving Elution Positions of li

r BLANK PAGE r-T'

-298-

ORNL DWG 72 - II7I9R4

15 16 17 18 19 20 21 22 23 24 25 26 27 28 C9 30 31 I I ISO 200

o UinJ o(E 1

—1 | 1 1 1 1 r— T T T IANGE CHR0MAT0GRAPH RUN CONDITIONS s ^m-lD x 150-cm STAINLESS STEEL WITH 8-12 ft DIAM. AMINEX A-27 RESIN; PERATURE PROGRAM, AMBIENT TO 55°C AT 11.5 hr •, ELUENT GRADIENT INCREASING ARILY IN CONCENTRATION FROM 0.015 M TO 6.0 M AMMONIUM ACETATE, pH 4.4; ENT FLOW RATE 89ml/hr , COLUMN PRESSURE 1800 psig.

46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 4i)0

. ABSORBANCE AT 254 nm ABSORBANCE AT 280 nm

fzer of Secondary Sewage Treatment Plant Effluent,

>sitions of Identified Compounds. -299-

350 375 400 425

280/a OF CHLORINATED EFFLUENT (1.0 mg/liter COMBINED CHLORINE RESII

280^9 OF EFFLUENT CHLORINATED IN LABORATORY ( 1.0 mg / liter COMBIN

Fig. 7.3. Dual-C<

of a Sewage Treatment ] 1

z 14 ELUTION TIME.hr 20 22 24 26 ; I i i 1 1 1 125 ELUTION VOLUME, ml 225 250 27] t i i l l l l i I l I I l i i i i

® ® ©

s > /• \ A / 7 >S\ /

48 ELUTION TIME , hr 54 56 58 60

450 ELUTION VOLUME, ml 525 550

5IDUAL.0RTH0T0LIDINE ) CONCENTRATED I680X, 254 nm

36CI ACTIVITY INED CHLORINE RESIDUAL , ORTHOTOLIDINE ) CONCENTRATED 1040 X 254 nm

Column Chror/iatograms of the Chlorinated Effluent from the Secondary Stage

l Plant.

y c BLANK PAGE t-i 5 ORNL DWG. 73-2467R2

2 24 26 28 30 32 34

f j 1 1 1 1 5 250 275 300 325

id-. >6 58 60 62 64 66 68

550 575 600 625

36CI ACTIVITY ) 1040 - 254 nm

>ndary Stage

I BLANK PAGE -300- chloro-organic compounds. The concentrations of these constituents ranged from 0.02 to lb of CI per liter of original effluent.

Tentative identifications of 17 of the chlorine-containing constituents

"were established by comparison of their anion-exchange elution volumes with those determined for reference standards. Reasonable agreement of elution volume is presumptive evidence that the unknown may indeed be the same compound as the reference standard. The tentative identifications are given in Table 7*3* along with concentrations of the organic compound, as calculated from the radioactive tracer experiments. Since the extent of chlorination of this sample is approximately equivalent to that expected at ORESP, the concentrations of the tentatively identified compounds would approximate those expected at that sewage treatment plant. These concentrations were calculated from data which were based on the reasonable assumption that complete isotopic dilution of the CI occurred rapidly during the chlorination process. Therefore, these values may be high if the chlorination reaction forming the compound proceeds at a rate equivalent to, or faster than, the chlorination rate for the formation of chloramines.

The magnitude of the chlorination effect is dependent upon the chlorine

dosage and total reaction time. In one experiment with chlorination reaction

time (45 min) and conditions (l mg/liter chlorine residual) approximately

equivalent to those at ORESP, 0.6% of the chlorine dose is associated with stable chlorine-containing organic constituents separated chromato-

graphically from the chlorinated effluent; an additional 0.b% is associated

with chlorine-containing constituents which did not elute from the resin.

Thus, the total chlorination yield was about 1$ of the chlorine dose. The Table 7.3* Tentative Identifications and Concentrations of Chlorine-Containing Constituents in Chlorinated Sewage Plant Effluents

Concentration Peak Elution Reference Standard of Organic Volume Tentative Elution Volume Compound Constituent (ml) Identification (ml) (|4,g/liter)

16 72.4 ± 2.7 5-Chlorouracil 72a 4.3

18 80.2 ± 4.0 5-Chlorouridine 81 1.7

19 86.2 ± 4.2 8-Chlorocaffeine 86 1.7

22 102 ±3.3 6-Chlorogu.anine 109 0.9

32 218 8-Chloroxanthine 218 1.5 b 1+2 302 ±5.9 2-Chlorobenzoic acid 307 ± 9 0.26

43 312 ±5.9 5-Chlorcsalicylic acid 310 0.24

45 334 ±6.4 4-Chloromandelic acid 338 1.1

52 403 ±1.7 2-Chlorophenol 4oo 1.7

53 415 ±3.8 4-Chlorophenylacetic acid 4ll 0.38

c 55 436 ±5-4 4-Chlorobenzoic acid 434 ± n 1.1

56 444 ±4.4 4-Chlorophenol 446 0.69 Table 7*3 (continued)

Concentration Peak Elution Reference Standard of Organic Volume Tentative Elution Volume Compound Constituent (ml) Identification (ml) (^g/liter)

d 57 i+61* ± 8.7 3-Chlorobenzoic acid ij-55 ± 10° o.62

and/or 3-Chlorophenol J+56 o.5id

59 lj-96 ± 8.7 1+-Chlororesorcinol 1.2

6l 527 ± 15-6 3-Chloro-4-hydroxy- 5^0 1.3 benzoic acid

62 5^7 ± 20.2 l+-Chloro-3 -me thylphenol 550 1.5e

g Average of two determinations.

Average of seven determinations ± standard deviation. c Average of four determinations ± standard deviation.

'Values based on assumption that chlorine is present as the pure compound of either, and not as

a mixture of both.

Concentration in H-2. -303- chlorination yield calculations are "based on the assumption of complete isotopic dilution during the chlorination of the CI in the chlorinating agent with the "pool" of nonradioactive chlorine (as chloride) in the effluents. The remainder of the available chlorine (99%) was apparently utilized in oxidation reactions.

Since the experimental conditions are similar to the field conditions at ORESP, chlorination at the sewage treatment plant should produce similar chlorination yields (i.e., about 1°j0 of the chlorine dose should ae associated with stable chlorine-containing organic constituents). Table summarizes the chlorination yields obtained in the experimental series.

7.1.2 Natural Waters

The identification and quantification of trace organic compounds present in various natural waters are being studied using high-resolution, anion exchange chromatography. The sites chosen for obtaining natural waters for the study, and the basis of selection, are listed in Table 7*5*

During this report period the technique for concentrating the water samples prior to analysis was further evaluated; samples were taken at least once from each of the sample sites; the samples taken were concentrated and run on the analytical system; known samples of humic acid were obtained and run as reference material on the analytical system; and

development of a group of suitable internal standard compounds was begun.

The concentration technique that had been adequate for sewage plant

effluents was found to be unsatisfactory for natural waters because of

the greater concentration factor needed, and the consequent larger quantity

of insoluble salts that were produced. It was found that the concentration

technique, in addition to the filtration, vacuum distillation, and Table Chlorination Yield of Chlorine-Containing Constituents "with Respect to Reaction Time for the Chlorination of Effluents

Chlorination Yield Reaction Time of Eluted Total Chlorination (min) Constituents^ (%) Yield0 (%)

15 0.55 0.93

0.58 O.98

90 0.75-0.87 1.27

1.6-1.8

The yield is expressed as percent of chlorine doss which, after

chlorination, is associated with stable chlorine-containing

constituents.

Chlorination yield of chlorine-containing constituents sepa-

rated as chromatographic peaks in the radioactive tracer

experiments.

Chlorination yield of the total chlorine-containing constituents

separated from the chlorinated effluent. The total is defined as

the sum of both eluted constituents and those which did not elute

from the resin during the chromatographic separation. Table 7« 5» Initial Sampling-Sites for the Measurement of the Molecular Organic Contaminants in Polluted. Water

Geographic Location Physical Location Reasons for Selection

Oak Ridge, Tennessee Walker Springs Watershed Well characterized ecologically and relatively undisturbed

Watts Bar Reservoir, Cooling water inlet to TVA Reasonably unpolluted water Kingston, Tennessee steam plant routinely chlorinated for antifouling purposes

Fort Loudon Lake, Knoxville, USN & MC Reserve Training Downstream from major city sewage Tennessee Center treatment plant

Lake Marion, South Carolina Santee Dam Example of waters with high natural organic content

g Mississippi River, Memphis, Cooling water intake of Nation's major watershed and site Tennessee Allen steam plant of samplss taken for other studies

South Fork of Holston River, Just above junction with Downstream from large industrial Kingsport, Tennessee North Fork organic chemical plant

g Inlets chosen to provide baseline for later studies of effects of chlorination. -306- freeze-drying steps, required the filtration of the natural water through an ion exchange column to remove those cations which form the bulk of the inorganic insoluble compounds. The concentration technique adopted involved:

(l) filtration through a filter to remove large particles, some colloids, and bacteria; (2) passage through a 2-liter bed of weak cation exchanger (Rexyn 102) at about 3 liters/hr; (3) concentration in the vacuum still by a factor of about 100; (4) freeze-drying; (5) redissolution in dilute acetic acid to remove carbonates; (6) freeze-drying; and (7) dissolution to desired volume of dilute acetate buffer, and separation from remaining solids by centrifugation.

The concentrated samples were analyzed on the anion exchange chromatograph having the cerate oxidative detector in series with the uv detector described in Sect. 4.1, except that a 0.31-cc sample was normally used. Tabulated in Table "J.6 are the numbers of peaks obtained with the various samples. This is only a qualitative comparison between samples, as nothing is said about peak sizes on the chromatograms. Figure 7*4 shows representative chromatograms for the water samples from Watts Bar

Lake and Lake Marion, South Carolina. These indicate the diversity of sample types and concentrations to be found in natural waters.

One material that is known to be present in some natural waters is humic acid. Chromatograms of humic acid have been prepared on the anion exchange chromatograph used in this study, and also on the cation exchange system used for identification of fractions normally obtained after separation on the preparative anion exchange study. These chromatograms are presented in Figs. 7-5 and 7*6. -307- j

..1

j s \

Table 7-6. Qualitative Summary of Chromatographic Analysis Results i

Concentration Ultraviolet- Cerate Common Tota'i

Sample Factor (10^) Absorbing Peaks Oxidative Peaks Peaks Diffj•p 3 Walker Branch 10 21 2b 11 a 1'1 Walker Branch, plus 10 bQ b5 18 extracts of solids N Watts Bar Lake 10 53 b6 21

Fort Loudon Lake 1.3 93 b7 19 i 'i Lake Marion 0.5 35 35 19

Holston River (light) 0.5 10 13 7 ? (dark) 3.0 31 26 21

Q Mississippi River 0.^ 6 111 3 j

3, Not desalted before concentration; consequently, it had very large quantity of solids ever

low concentration factor. ]

-A 5

mmary of Chromatographic Analysis Results "MIi V-'.J

* jLtraviolet- Cerate Common Total Number of ' :' ' '3-i sorting Peaks Oxidative Peaks Peaks Different Peaks II 21 24 11 34

k8 45 18 75

®1 53 46 21 78 "V'i? . 'i 93 47 19 119

35 35 19 51

10 13 7 16 31 26 21 36 i-.vss

ill 17 H

.ently, it had very large quantity of solids even at Hi 'iW mna

xn^Ai iH -^sS »

W0 w I M fr3 lis Hi BLANK PAGE r i

NI C E ABSORBANCE ABSORBANCE ABSORBANCE o o O S Ai> JiA P. J P C <0 O <0CO O O

f- i \

8QH niMHPZ> s iglsl —! 5 zc aim m IBSSS ll.^i ?5 •niiE * o J oj

8o H m H SCO « o> £ u H , - —< « -t o <2 o

,'VWNOn-MNIl tccvc \ ACID

o jO

o> a p

-L. -LJ 111 _1_ o — Nwioioisaini — ro ESCENCE CERIUM (III) FLUORESCENCE v n N a (A 5 o CERIUM (IS) FLUORESCENCE (ARBITRARY UNITS) "Nl >IT3) (ARBITRARY UNITS) CERIUM (II) FLUORESCENCE (ARBITRARY UNITS) O)

-go£-

\ n o •f ASSOR0AMCE «+ o o 2 : • 6 s" • ss o » »•#.*« o •r * I Is "C r/f o > a t1 * \ O OO GS £3 KJ o f5 o ff. rat * • ° «•; o *•* * § O I i»ot

f>P tn ft> i O ci \ h> P * D f> o tfj —4 w in ct- m tt c C P

& c*P - ft> *re< ti o *t o c-t- o 3 3 wc* O ti ^ P 3 t •X3 > M s jo1

03 w- ra O — fO tM oi w a> o o — »»» ET CERIUM (LIT) FLUORESCENCE P S» (ARBITRARY UNITS) o n OFNL DWG. 73-6940

6 TIME (hr)

Fig. T'5* Analysis of Humic Acid in Soil Leach Solution by the

UV-Analyzer with Cerate Oxidlmetric Detector. -310-

ORNL DWG. 73-6941

W H Z > CC< oz H 00 a: 4

UJ o z UJ o UJ o: o3

i 2 or UJ o

4 5 TIME (hr)

Fig. 7*6. Cation Exchange Chromatographic Separation of Humic

Acid in Soil Leach Solution. -311-

As a further aid in identification, a set of internal standards is being developed for use in more exact positioning of peaks. These internal

standards are being selected from compounds not likely to occur in natural waters and from those which provide sharp peaks at relatively low concentrations. Such compounds should be stable when stored in solution, and

their uv and cerate oxidative responses should be sufficiently unique so

as to facilitate easy recognition. Compounds being considered and tested

are tyrosine, pyruvic acid, fumaric acid, hippuric acid, p-bydroxy-

phenylacetic acid, kynurenic acid, and m-hydroxycinnamic acid; these

compounds elute at approximately 10$ intervals in the chromatogram.

Tyrosine, pyruvic acid, and p-hydroxyphenylacetic acid are easily recognized

by their strong cerate oxidative responsesT whereas fumaric and hippuric

acids provide relatively strong uv signals at 25*+ nm. The remaining

compounds have characteristic ratios of cerate oxidative to the uv

responses at 25U and 280 nm.

7.2 Environmental Effects of Antifoulants

R. L. Jolley and J. E. Thompson

".Tie expected increased demand for electrical power in the future

will result in greater numbers of nuclear and fossil fuel power plants,

and therefore greater use of natural surface waters in the cooling

systems of such power plants. Because biological organisms grow in or

on the surfaces of the cooling systems and reduce the efficiency of heat

transfer, it is necessary to clean the surfaces periodically. An accepted

practice is to add chemical antifoulants such as chlorine to the cooling

W8ters. These biocides destroy the film of biological organisms covering

the heat transfer surfaces of the cooling systems. The primary reaction -312- of chlorine with pollutants in natural waters is thought to he with ammonia and organic amines and amides to form chloramines. Many other reactions such as halogenation of phenols may possibly occur. Chloramines are known biocides; however, other chlorinated organic compounds may also have biological effects.

Prior to studying the chlorination of organic compounds in natural waters per se, chlorination effects on process waters or effluents entering surface water are being investigated. The methodology for determining chlorination effects was developed as discussed in Sect. 7.1. Preliminary

investigations of the biotoxicities of the chlorine-containing organic compounds tentatively identified as present in chlorinated sewage effluents

(Table 7.4) is currently being conducted in a cooperative effort with the

Environmental Sciences Division. The zooplankton Daphnia magna and

Diaptomus clavipes are being used as test organisms in this study, the

source of which includes not only acute toxicity, but also chronic effects

(e.g., survivorship and fecundity).

The methodology developed with process effluents for identification of

chlorinated organic compounds and determination of biotoxity will be applied

to determination of chlorination effects on cooling waters.

7.3 References for Section 7

1. F. M. Middleton and A. A. Rosen, "Organic Contaminants Affecting the Quality of Water," U.S. Public Health Reports 71, 1125 (1956).

2. R. L. Bunch, E. F. Barth, and M. B. Ettinger, "Organic Materials in Secondary Effluents," J. Water Pollut. Contr. Fed. 33, 122 (1961).

3. G. E. McCallum, "-Advanced Waste Treatment and Water Reuse," J. Water Pollut. Contr. Fed. 35, 1 (1963). -313-

k. C. D. Scott, R. L. Jolley, W. W. Pitt, and W. F. Johnson, "Proto- type Systems for the Automated, High-Resolution Analyses of UV- Absorbing Constituents and Carbohydrates in Body Fluids," Amer. J. Clin. Pathol. 53, 701 (1970).

5. R. L. Jolley and C. D. Scott, "Preliminary Results from High Resolu- tion Analyses of Ultraviolet-Absorbing and Carbohydrate Constituents in Several Pathologic Body Fluids," Clin. Chem. l£, 687 (1970).

6. R. L. Jolley, K. S. Warren, C. D. Scott, J. L. Jainchill, and M. L. Freeman, "Carbohydrates in Normal Urine and Blood Serum as Deter- mined by High-Resolution Column Chromatography," Amer. J. Clin. Pathol. 53, 793 (1970).

7. S. Katz, W. W. Pitt, Jr.,, and G. Jones, Jr., "Sensitive Fluorescence Monitoring of Aromatic Acids After Anion Exchange Chromatography of Body Fluids," Clin. Chem. 19(8), (1973).

3. The Molecular Anatomy (MAN) Program First Semiannual Progress Report, March 1 to August 31, 1972, ORNL-^815 Special, p. 59.

9. S. Katz, W. W. Pitt, Jr., 7. D. Scott, and A. A. Rosen, "The Determination of Stable Organic Compounds in Waste Effluents at Microgram Per Liter Levels by Automatic High-Resolution Ion Exchange Chromatography," Water Research 6, 1029 (1972).

10. W. J. Weber, Jr., Physiochemical Processes for Water Quality Control, Wiley Interscience, New York, 1972, p. "w:

11. R. W. Risebrough, P. Reiche, D. B. Peakall, S. G. Herman, and M. ri. Kirven, "Polychlorinated Biphenyls in the Global Ecosystem," Nature 220, 1098.(1968).

12. E. Hirst and H. Bank, "Striking the Balance." Environment 13(9), 3^ (1971).

13. F. Coulston and F. Korte (eds.), Environmental Quality and Safety, Vol. I. Global Aspects of Chemistry, Toxicology, and Technology as Applied to the Environment, Academic Press, Inc., New York, 1972, p. 2(dT.

1*4-. J. E. Mrochek, W. C. Butts, W. T. Rainey, Jr., and C. A. Burtis, "The Separation and Identification of Urinary Constituents Using Multiple- Analytical Techniques," Clin. Chem. 17, 72 (1971).

15. W. W. Pitt, Jr., C. D. Scott, and G. Jones, Jr., "Simultaneous Multi- column Operation of the UV-Analyzer for Body Fluids," Clin. Chem. 18, 767 (1972). -314-

16. Oak Ridge National Laboratory, "Automated Analysis of Individual Refractory Organies in Polluted Water," report prepared for the Water Quality Office, Environmental Protection Agency, Project #16020 EGY, Contract #ll}-l2-833, October 1972.

17. R. L. Jolley, Chlo^-ination Effects on Organic Constituents in Effluents from Domestic Sanitary Sewage Treatment Plants, ORNL-TM- 4290, (August 1973). 8. RADIOACTIVE WASTE DISPOSAL

Experimental investigations in support of three different methods of radioactive waste disposal were conducted. The amounts of stored radiation energy in the bedded salt in a radioactive-waste repository are being determined; the shale fracturing process for ultimate disposal of intermediate-level waste solutions is being studied; and the Pressurized

Aqueous Combustion (PAC) process for converting combustible contaminated wastes to ash was investigated.

8.1 Radiation Effects in Salt Mine Waste Repositories

G. H. Jenks and C. D. Bopp

Experimental studies which were aimed at establishing the amounts of stored radiation energy in the salt in a radioactive waste repository were continued. The bedded salt in a repository will be exposed to gamma radiation under varying sets of conditions of dose rate, dose, and salt temperature. The maximum dose will be about 2 x 10^" rads, and this will occur adjacent to a 5-kW waste can where the dose rate at 5 1 burial will be about 7 x 10 rads/hr. The temperature at this location 2 will rise to a maximum of about 350°C. At other locations and with other cans, the maximum dose rates, doses, and temperatures will be below these.

It was known that gamma radiation energy can be stored in salt, and there were reasons to believe that the amounts stored would depend not only upon dose but upon the rate at which the dose is accumulated and upon the salt 3 temperature. Some of the sets of exposure conditions which will prevail in a repository could not be duplicated in laboratory tests of reasonable duration. Accordingly, our experiments were designed to provide information -316- on the effects of the different exposure variables that would enable extrapolation to and prediction of stored energy at waste repository conditions.

The experiments were primarily ones in which samples of salt were exposed to gamma rays at various combinations of dose rates, doses, and salt temperatures, and subsequently analyzed for accumulated stored energy.

The sets of dose rates, doses, and temperatures that were employed in experiments conducted during this report period are shown in Tables 8.1 and 8.2. Other information on specimens is found in Tables 8.3 and 8.4.

(Experiments prior to this report were made with Harshaw salt that had been 7 8 irradiated at 95°C and at 10 rads/hr with doses in the range of 8.1 x 10 to 4.7 x 10"^ rads.) The irradiations in the experimental work were accomplished using either a cobalt source or a spent HFIR fuel element.

The work with the spent HFIR fuel elements made use of a facility previously if designed and constructed for this purpose. The samples of salt used in the irradiations were comprised either of Harshaw synthetic salt crystals or of bedded salt from the mine at Lyons, Kansas. The analyses for stored energy were accomplished using the drop calorimeter that was designed and5 constructed for such work, and which was described in a previous report.

This calorimeter operates at a high temperature at which the stored energy is quickly annealed. The sample is dropped into the calorimeter from a temperature in the neighborhood of 100°C.

In other experiments, portions of irradiated specimens were heated prior to measurement of stored energy in order to obtain information on the thermal annealing characteristics (re-formation of crystalline NaCl

Work done by E. Sonder, Solid State Division, ORNL. -3"l7-

Tabls 8.1. Dose Rates, Doses, and Temperatures for HFIR Gamma Exposures of Salt

Exposure Temperature 7 Time No. (°C) (rad/hr x 10~') (hr) (rad x 10~y)

1 160 3.6 211.5 7-6

2 160 6.1 212.0 13.0

3 173 6.6 264.5 17.^

k 173 3.1 148.6 4.6

5 144 7-7 128.4 9.9

6 144 4.0 244.1 9-15 -671-

Table 8.2. Stored Energy in Harshaw Salt Irradiated "by Cobalt Gamma Rays at l65°C

Observed Average Stored Dose Rate Total Dose Sample Reference Energy^ No. (rad/hr x 10"^) (rad x 10-9) (cal/g)

c-10 1.0 1.2 0 to 0.3

c-11 1.0 8.6 1.5 ± 0.3

c-12b 1,0 16.8 0.9 ± 0.3

c-13 0.16 2.0 < 0.1

c-ib 0.16 3.1 < C.l

c-15 1.0 3.6 l.k ± 0.15

Average of two measurements. The spread of the two values is shown.

Irradiated at temperatures ranging from 162 to l80°C. -319-

Table 8.3- Exposure Temperatures and Stored Energy Results for HFIR Gamma Exposures

Observed Stored Energy from Average Stored Sample Reference No. Prior Irradiation Temp. Energy 8 and Description (cal/g) (°c) (cal/g)

7-1-1 L-B - 160 7.6 ± 0.2

D,E 7-1-2 L-PT - 160 7.2 ± 0.5

7-2-1 L-B - 160 13.1 ± 0.3°

D,E 7-2-2 L-T - 160 9.1 ± o.6

D 7-2-3 L2-T - 160 10.8 ± O.O

7-2-4 H-T

7-3-1 L-B - 173 8.3 ± 0.1

7-3-2 L-T 173 8.2 ± 0.26

7-3-3 H-B face - 173 5.5 ± 0.0

7-3-4 H-B center - 173 5.1 ± 0.2

7-3-5 H-T - 173 3.9 ± 0.1

7-3-6 (7-2-1; L-B)-T 13*1 ± 0.4 173 12.2 ± 0.9

7-4-1 L-B 173 3*3 ± 0.1 e 7-4-2 L-T - 173 3.6 ± 0.0

7-4-3 !i-T «» 173 0.4 ± 0.1

7-4-4 (7-3-1; L-B)-B 8.3 ± 0.1 173 7.9 ± 0.1

7-4-4 (7-3-3,4; H-B)-T 5^3 ± 0.2 173 3.2 ± 0.0

7-5-1 L-B - 144 10.3 ± 0.2

7-5-2 L-C - 144 10.2 ± o.ie

7-5-3 (7-4-1; L-B)-C 3.3 ± 0.1 144 13.7 ± 0.2®

7-5-4 (7-2-1; L «B) -C 12.7 ± 0.4 144 18.3 ± 0.3E -320-

Table 8.3 (continued)

Observed Stored Energy from Average Stored Sample Reference No. Prior Irradiation Temp. Energy^ and Description3 (cal/g) (°c) (cal/g)

7-5-5; H-T - 144 8.1 ± 0.2

7-5-6; (7-3-3,4; H-B/-T 5-3 ± 0.2 144 12.5 ± 0.4e

S 7-5-1; L-C - 144 10.8 ± 0.2

7-6-1; L-B - Ikk 11.1 ± 0.4

7-6-2; H-T - 144 l.k ± 0.3

7-6-3; (7-5-1; L-B)-B 10.3 ± 0.2 144 15.0 ± 0.4

7-6-4; (7-i>-5; H-T)-T 8.1 ± 0.2 Ikk 10.6 ± 0.0

e 7-6-5; (7-3-3,U; K-B)-T 5.3 ± 0.2 Ikk 10.8 ± 0.2 aSee Table Q.k.

Average of two measurements. The spread of the two values is shown. e Lyons block material sampled in a manner to be representative of the

samples used in the thermal annealing studies- d

Two highest of three measurements. Values considered less reliable

than other values since the particle size in the calorimetric samples

exceeded that used with other Lyons specimens. Also, the preheat

Thestemperature pointes wa nost 105° plottedC rathe. r than 125°C. -321-

Table 8.4. List of Symbols Used for Sample Identification and Description in HFIR Gamma Exposures

Symbol Meaning

L Lyons salt.

Lp Lyons salt which was from a different location and contained fewer traces of clay than the material labeled L.

H Harshaw synthetic crystals of NaCl; billet number 100.

B Refers to specimens which were in the form, of rectangular blocks of l/2 in.-square cross- section.

T Refers to pieces of salt located in an 8.3-mm-ID aluminum tube during irradiation. These pieces were cubiform with dimensions in the neighborhood of 1/8 to 3/16 in.

C Refers to pieces of salt in a 9»3-nim-ID cylindrical hole in the aluminum block in which all specimens were positioned. These pieces were somewhat larger than those in the tube.

PT Refers to pieces of salt held within a 7-nan Pyrex tube which was placed within the aluminum tube mentioned above. -322- from radiation defects) of the stored energy. Experiments of the same type had been conducted previously using samples of Harshaw salt 4"hat had •7 been irradiated at 95°C and 10 rads/hr. The results of these experiments showed rapid annealing of 2 to 3 cal/g in the temperature range of 115 to

120°C. The remainder annealed only at temperatures above about 200°C.

The rate of this second-stage, high-temperature annealing was found to be independent of the amount of stored energy. Measurements over a range of temperatures showed an activation energy of about 33 kcal/mole, and extrapolation indicated that thermal annealing will be important at the low dose rates and concomitant long exposure times that will prevail in a waste repository.

In other experiments, additional investigations were made of the amounts of H^ and 0^ evolved upon aqueous dissolution of irradiated salt.

We previously performed analogous investigations on specimens of irradiated

Harshaw NaCl and used these results, along with other experimental and theoretical information, to qualitatively explain the apparent difference between drop-calorimeter measurements of the amount of stored energy in

NaCl and solution-calorimeter results reported by others for the same specimens.^ The present work had two objectives: verifying that the amount of H2 evolved is proportional to the amount of stored energy in the NaCl, and improving the accuracy of the experimental relationship between stored energy and H^ evolution.

In other work which will be mentioned, we obtained analyses of the amounts of H^ and 0^ that were contained within brine inclusions in two _

The gas measurements were made by J. R. Walton, ORNL Analytical Division, using a mass spectrometer for the identification and qualitative analyses of evolved gases. -323- different samples of irradiated Lyons salt and within one sample of unirradiated Lyons salt. Also, a facility for irradiation of samples of synthetic radioactive wastes in the core of the ORR was designed and part of the construction of this facility was completed.

8-1.1 Radiation and Stored Energy of Salt Specimens

The results of stored energy measurements with the drop calorimeter are listed in Tables 8.2 and 8.3. The symbols used for sample identification and description in Table 8.3 are explained in Table 8.4. The symbols 7-1 through 7-0 in Table 8.3 refer to HFIR gamma exposures 1 through 6. (In several cases, as shown, samples which were irradiated in one KFIR gamma exposure were irradiated 8n additional amount in a subsequent exposure.) The total dose and dose rate values listed in

Table 8.1 were estimated from the calibration data of Compere, Savage, and Baker. The specimens (Table 8.3) were located within 1-3/4 in. of the center line of the spent- HFIR fuel elements. The temperature of

specimens during HFIR gamma exposure was controlled by adjusting the voltage to a Calrod heater located within the aluminum block inside of which specimens were positioned. Adjustment was required to compensate for the decrease of gamma heating from radioactive decay of the gamma

source. The temperatures decreased approximately 3 to 5°C between successive

adjustments, and were measured by sheathed thermocouples at four points within the block (three points for 7-5 and 7-6). The temperatures listed have been adjusted to account for the estimated temperature gradient due Q to gamma heating. Sonder supplied information on dose rates, doses, and

temperatures for the cobalt gamma irradiations (Table 8.2). -324-

The technique followed in the calorimetric measurement minimized

release of geologic brine when a sample was dropped into the high-temperature

region of the calorimeter. We observed that bedded-salt specimens that had dimensions greater than about l/l6 in. often shattered into fragments,

including much fine dust, when dropped into the calorimeter. The stored

energy value that was recorded when this happened was always low. The

following procedural steps largely eliminated the shattering and improved

the precision of the measurements:

(1) The Lyons salt was cleaved to a very small particle size

before measurement (about 500 pieces per gram).

(2) The calorimeter temperature was held as low as possible (at

380°C rather than the 460 to 500°C used in earlier work).

(3) The cleaved specimens were heated for 1 hr or more at 125°C

and then dropped directly into the calorimeter from this

temperature. In earlier work, the preheat temperature was

95 to 105°C.

We believe that the success of these modifications in technique was due

to release of much of the occluded brine in the cleaving process and its

subsequent vaporization at 125°C.

Wo problems were encountered in calorimeter measurement of the Harshaw

specimens. The samples had a 2 to 3 mm square cross section and were about

1 cm long. The preheating treatment and the calorimeter temperature were

the same as those used with Lyons specimens.

The estimated uncertainty in the observed value for stored energy in

a Harshaw specimen of ± 5$ resulted from the measurement of the area under

a curve plotted by a Brown recorder. Our estimate for the uncertainty in -325- the amount of stored energy in Lyons salt that is exposed to a given set of test conditions, but without representative sampling, is ±10%.

8.1.2 Results of Thermal Annealing Studies

The specimens and the annealing treatments employed in these studies are listed in Table 8.5* The table also gives values for the observed decrease in stored energy that resulted from an annealing treatment, and for the corresponding zero-order rats constant. The data are shown graphically in Fig. 8.1.

The data obtained from the Harshaw salt specimens irradiated at l60°C support the previous finding that annealing of the higher-temperature stage g in Harshaw salt irradiated at 95°C follows a zero-order rate process.

They also support the values and the activation energy found in the previous work. The present data and most of the previous data can be correlated by the equation

k = 1.30 x 1013 [exp (-1.67 x IO^/t)] , (l) where k is the zero-order annealing constant, cal/(g.hr), and T is the temperature in °K.

The data obtained with the irradiated Lyons salt specimens also

support zero-order kinetics. However, they indicate that the activation energy and rate constants may differ somewhat from those for the Harshaw

salt and probably depend on the radiation dose. These data that were

obtained at the higher temperatures (and higher annealing rates) showed

appreciable scatter. We judged that the lower rate-constant values among

a group of scattered values are the most nearly correct, the straight lines — JL.O-

Table 8.5. Thermal Annealing of Stored Energy in Harshaw and Lyons Salts

Annealing „ ^ , m j. 1. . Zero Order Treatment Decrease in , ^ 7. \ \ Sample Reference No. Temp. Time Stored Energyb Rate Constant and Description8 (°C) (hr) (cal/g) (cal g hr )

7-2-4; H-T f 185.0 278.3 5.7 ± 0.3 (2) 0.020 c 225-2 15.8 6.3 ± 0.4 (2) 0.40 d 227.4 6.8 2.9 ± 0.6 (2) 0.42

7-2-1; L-B a 192.7 70.5 4.9 ± 0.3 (2) 0.070 b 195.2 134.3 8.3 ± 0.1 (2) 0.062 c 212.5 4.8 6.6 ± 0.6 (2) 1.4 d 217.6 9.4 5.3 ± 0.5 (2) 0.56 e 218.3 12.0 6.6 ± 0.4 (2) 0.55 f 226.2 5.65 6.6 ± 0.9 (2) 1.2

7-1-1; L-B i 220.0 47.8 5-9 (1) 0.102 a 238.8 2.4 3.8 ± 0.2 (2) 1.6 e 244.8 2.8 2.2 (1) 0.79 d 258.4 3.6 4.2 ± 0.5 (2) 1.2 c 262.6 1.0 4.2 ± 0.6 (2) 4.2

7-6-2; H-T 228.9 4.5 4.7 ± 0.8 (2)C J- • 1

aSee Table 8.4.

10The number of measurements on the annealed sample is shown in

parentheses. The spread of the results of two measurements is shown

as a plus-minus value.

Q This point not plotted. ORNL DWG. 73-9924

Fig. 8.1- Zero-Order Thermal Annealing Constants as a Function of

Irradiation Temperature for Harshaw and Lyons Salt. -328- were drawn accordingly. It is notable that the line drawn through the data for Lyons specimen 7-2-1 extrapolates, at 160 to 173°C, to rate constant values which are near those of the extrapolated Harshaw-line.

In the correlation subsequently described, we assumed that the k for Lyons salt in this temperature range is two-thirds that for Harshaw salt.

8.1.3 Correlation of Energy Storage Data

With a few exceptions to be mentioned later, our data are expressed and correlated reasonably well on the basis of the empirical Eq. (2) and its integral, Eq. (3):

dE "dt = V " k2Es " 8k > (2)

E« = ^I ^ t1 " + Ese"k2D > ^ where

Eg is stored energy, cal/g,

E° is stored energy at time t = 0,

I is the dose rate, rads/hr,

t is the irradiation time, hr,

D is the dose, rads,

a is a constant having a value > 1 depending on temperature, and

k is a parameter which for the purposes of this correlation is

identified with the parameter defined by Eq. (l).

The data obtained previously with Harshaw salt irradiated at 95°C also

are correlated reasonably well with Eqs. (2) and (3)« (However, as

stated above, two-stage annealing was found with the 95°C material, while only a single stage -was found with the specimens irradiated at the higher temperatures.) The values of the constants in Eq. (3), which were determined from the data at the different irradiation temperatures, are listed in Table 8.6. Curves obtained from these constants, together with experimental data points, are plotted in Figs. 8.2-8.5* The cardinal number adjacent to a point identifies the HFIR gamma exposure in which the specimen was irradiated. A number pair signifies that the given specimen was irradiated in two different HFIR gamma exposures with the irradiation in exposure 2, 3> or 5 followed, respectively, by the irradiation in exposure 3> or 6. The data obtained with the cobalt-gamma-irradiated specimens are identified by sample reference numbers c-10, c-11, c-l'3> c-l4, and c-15 (Table 8.2). The correlation curve (and exposure conditions) with which a data point is associated is identified by an arrow when the point falls away from the curve.

In evaluating the constants in Eqs. (2) and (3)? we worked initially with the Harshaw salt data; thus these evaluations will be discussed first.

Since the correlation of the data taken at l44°C showed that the value of ak [in Eqs. (2) and (3)] was negligible compared with the values of k^I in tests 7-5 and 7-6, it was therefore neglected in correlating these data, aa well as those obtained at 95°C. The empirical value of k^ was the same

at 95 and l44°C, but the value of k2 at l44°C was about three times that

at 95°C. The datum at l60°C was adequately represented with a value of k^ equal to that at l44°C, and with a value of k^ slightly greater than that

at l44°C. This representation was improved slightly by setting ak = 0.0024, which is equal to the calculated value of k in Eq. (l) at l60°C. The

value of k„ at 173°C was determined from the result obtained with the -330-

g Table 8.6. Values of Constants in Equations (2) and (3) for Harshaw and Lyons Salt

k (10~9 cal -1 -1, k Temperature x g rad ) 2 a (°c) Harshaw Lyons (10"11 rad"1) (dimensionless)

95 1.20 3.0 b

144 1.20 1.66 9.3 b

160 1.20 1.70 10.5 1

165 1.20 1.70 11.0 2.2

173 1.20 1.70 12.0 4.4

Q The value of k for Harshaw salt is given by Eq. (l); k for the Lyons

salt is two-thirds that for the Harshaw salt. The value of ak was negligible relative to the value of k.I. -JJJX-

ORNL DWG. 73-9925RI

Fig. 8-2. Stored Energy in Harshaw and Lyons Salt That Was Gamma-

Irradiated at and at and 77 Mrads/hr. -332-

ORNL DWG. 73 -9926RI

DOSE (rads) Fig. 8«3« Stored Energy in Harshaw Salt That Was Gamma-Irradiated at 160, 165, and 173°C and at O.65 to 66 Mrads/hr. -333-

ORNL DWG. 73-9927RI

DOSE (rads)

Fig. Q.b. Stored Energy in Lyons Salt That Was Gamma-Irradiated at 160 and 173°C and at 31 to 66 Mrads/hr. ORNL DWG. 73-9919

DOSE (rads) Fig. 8«5« Stored Energy in Harshaw Salt That Was Gamma-Irradiated at 95 ± 5°C and at 107 rads/hr. -335-

specimen which was exposed in both 7-3 and 7-4. Using this value, it was apparent that the values of the terms in Eq. (3), k^I - ak, were

substantially less than those at lower temperatures for the same values

of I. Using the value of k^ obtained at lower temperatures, the value of

the constant a was Finally, for the l65°C data, it was assumed that

the same value of k^ prevailed and that the value of k^ fell between the

values found for 160 and 173°C. Thus, the fitted value of the constant

a was equal to 2.2. £ For the low-dose-rate experiments (10 rads/hr), c-13 and e-l4, it -9

can be noted that zero stored energy would be expected with k^ = 1.2 x 10

cal g rad""1", and with the value of k equal to that given by Eq. (i);

that is, thermally activated annealing would preclude any accumulation of

stored energy. This was found to be true.

From data at l44°C on irradiation of Lyons salt, it was established

that the appropriate value of k^ was about the same as that for the Harshaw

salt at this temperature, while the value of k^ was about 40% greater than

that for the Harshaw salt. Adequate fits to the higher-temperature data

were obtained using: (l) the higher value of k^; (2) values of k^ equal

to those found from the Harshaw data; and (3) an ak value of 173°C equal

to two-thirds that used for the Harshaw salt. The lower value of ak for

Lyons vs Harshaw salt may relate to the differences found in the thermal

annealing experiments.

The data that comprise the principal exceptions to these correlations

are those for samples 7-5-3, 7-5-4, and 7-5-6 (Table 8.3). These samples

were irradiated at l44°C after prior irradiations at 160 or 175°C. The

stored energy that was produced in the prior irradiations apparently had -336- little or no effect on the accumulation of additional amounts of stored energy during the subsequent exposure. On the other hand, the Harshaw sample, 7-3-3,4, which was irradiated at 173°C, did exhibit a reduction in the amount of stored energy accumulated during subsequent exposure.

(Some of the same material comprised one of the specimens showing no effect in the irradiation in 7-50 Also, the samples in 7-5 tests that were re- irradiated in 7-6 tests exhibited reductions in the amounts of stored energy accumulated during this latter exposure. It seems likely that the phenomenon observed with tests 7-5-3, 7-5-4, and 7-5-6 is connected with the types and

sizes of defect-agglomerates which are formed at different temperatures, but the detailed explanations are unknown. Other possible exceptions to the applicability of the correlations can be seen for some of the data plotted in Figs. 8.2 and 8.3*

Work involving interpretations of the experimental data and the

analysis of their significance in a waste repository is still in progress.

Temperature-dose relationships in a waste repository "will be examined more

closely as part of these evaluations. The interpretations and evaluations

indicate that only small amounts (i.e., < 20 cal/g)1 of stored energy will

accumulate at the low dose rates and high temperatures prevailing in a

repository.

8.1.4 Evolution of Hg upon Salt Dissolution

The results of these measurements and the values of some quantities

calculated from the results are presented in Table 8.7- As before, other 10 gases were sought but only traces of 0^ were found, and the pH of the

solutions formed by the dissolutions was about S.O.1^'11 Table 8.7. Results of Measurements of Hydrogen Evolved upon Aqueous Dissolution of Irradiated NaCl

Apparent Energy per Equivalent Trapped Electron Amount of. (eV/trapped electron) Amount H Difference Between 2 Trapped Drop Soln. Evolved Electrons Calorimeter, Calorimeter, eD aUd % Sample No. (y-mole/g NaCl) (^uaole/g NaCl) (eV/trapped electron) eD %

c-4 48.8 97.6 5.31 2.7^ 2.57

c-8 119.2 238.4 5.26 2.89 2.37 iiQ

The sample material was Harshaw salt that had been irradiated at 7 95°C and 10 rads/hr. The stored energies in these samples were measured 6 12 previously. fJ~ Values for test c-4 were 12.0 and 6.19 cal/g from drop- and solution-calorimeteo r measurements/ * , respectively. Analogous values for test c-o were 29.0 and 15*9 cal/g.

Possible ionic species which may be formed upon aqueous dissolution of radiation defects, and the effects of their formation upon the stored energy value which is calculated from a heat-of-solution measurement, were discussed in a previous report.11 We have since recognized another likely reaction of the radiation defects: + Na° + 1/2 Cl2 + H20 1/2 H2 + Cl~ + 1/2 HgO + Na . (b)

When other products are formed, as in reaction (4), H^O is a reactant,

and, because of this, the heat of solution is below that of the reference

reaction. The values for stored energy which were reported from the

results of solution-calorimeter measurements were calculated by assuming

that the species in solution are the same as those formed by unirradiated

salt.

As shown in Table 8.7, the values, as well as the e values, of JJ o the two specimens, c-b and c-8, are in near agreement. The average value

of for c-l+ and c-8 is near the energy that might be associated with a

Frenkel pair.1^1^ However, it is about 25% greater than the heat of

formation of crystalline NaCl from elemental Na and Cl^, and it seems

likely that the energy state of colloidal defects approximates the energy

* The solution calorimetric work was done by S. Lindenbaum of the University of Kansas in collaboration with E. Sonder."^ -339-

state of the elements more nearly than it does that of Frenkel defects.

Also, the average difference between g-p ai*d e^ f°r c-4 and c-8 is greater than the 1.6 to 1.95 eV which would be expected as the maximum difference between the actual heat of solution of defects and the assumed heat of

solution in which only Na+ and CI are formed. The simplest explanation for the small discrepancies between inferred and expected values is that

the reported amount of H^ is about 20$ less than the amount which was

actually evolved.

8.I.5 Loss of Water from Lyons Salt During Radiation Exposure

Several block-type specimens which were irradiated in HFIR gamma

exposures 2 and 3 were examined for changes in weight during exposure.

It was known that bedded salt from the Lyons mine contained < 0.25 $ water, which was held in small brine cavities within crystals and on

crystal boundaries."^ It was expected that some of the water would be

lost during irradiation due to thermal-gradient-induced migration of brine

cavities and/or due to expulsion of brine or steam along crystal boundaries.

Table 8.8 provides information on the specimens, together with the results

of weight measurements. The surfaces of the irradiated specimens were

examined for deposits which would result from evaporation of brine at a

surface, and none were detected.

Most of the pertinent information concerning the sample identification

symbols and exposure conditions for the tests on Lyons salt is given in

Tables 8.1, 8.3, and 8.4. The B^ samples were 1.2 in. thick and located

2-3/4 in. above the center line of the fuel element. The B^ samples were

1 in. thick and were located about 6 in. above the center line. The dose

rate at this location was 50 to 60$ of that at the center line. All Table 8.8. Change in Specimen Weight Daring Radiation Exposure with Exposure Conditions, Calculated Brine Migration Distances, and Observed Weight Changes

Calculated Average Dose Average Migration Rate Observed Sample rv Average Temp. Temp. Gradient Distance Weight Change Identification (rad/hr x 10"' (°c) (°C/cm) (cm) (mg) (%)

7-2-5; L2-Bm 6.0 160 3.4 0.21 -6.1 o.l6

7-2-6; L-B^ 3-4 160 3-8 0.24 -22.4 0.22

7-3-7; L-Bu 3-7 173 4.1 0.32 -17.2 0.15

7-3-3,4; H-B 6.6 173 3-7 0.29 -0.6 0.01 -341-

B and B , as well as B, specimens were in contact with aluminum conductors m u' ' on two opposite faces. The other sides were in contact with thick layers

of gas (helium) and were thus effectively insulated. Accordingly, the highest temperature was at the midplane of the specimen.

The values for migration distance are those calculated using Fig. 6.4

of ref. 3> and apply to brine-filled cavities within a crystal. These move up a temperature gradient. Cavities having both a gas and liquid

phase would move down a temperature gradient, and the rate of movement 17 would exceed that of brine-filled cavities.

The results indicated that a significant fraction of the original

water was lost from the lyons specimens during irradiation. The

insignificant loss of weight observed for a Harshaw specimen supported

the assumption that the weight loss of Lyons specimens resulted from water

loss; the Harshaw salt had no brine inclusions. The absence of deposits

on the surfaces of the irradiated Lyons specimens and the possibility that

brine migration distances were less than sample thicknesses indicate that

the water was lost as vapor which diffused out along crystal boundaries.

8.1.6 Gases in Unirradiated and Irradiated Lyons Salt

A mixture of gases and water were extracted from several samples of

Lyons salt, and quantitative analyses were made of the molecular constit-

uents in the off-gas (Table 8.9). These data were obtained using a

crystal-crusher device which was attached to a mass spectrometer through

a gas handling system. The sample was placed within the metal crystal-

crusher, outgassed at about 100°C, and then crushed by hammering on the

crystal through a metal rod which was bellows-sealed into the crusher.

The released water was trapped at liquid nitrogen temperature and Table 8.9. Gases in Irradiated and Unirradiated Samples of Lyons Salt with Calculated Values of G-H~

Quantity of Gas G-H. 2 H 0 H N + CO 0 Weight Gamma Dose No, molecules/ 2 ^ ~ o Salt Sample (g) (rads) 100 eV) (mg) (cnr) (cnr) (cm15)

PSV-2 2.403 5 x 108 0.26 0.8 2.19 x 10~3 4.53 x lO-^ 1.71 x 10"^

10 ^ J -if. J 7-2-5; L0-B_ 2.524 1.3 x 10 0.016 2.0 9.9 x 10 2.02 x 10 2.17 x 10

2.224 0 0 0.8 0 8.9 x 10~5 1.3. x 10~5 subsequently weighed. The gases were analyzed in the mass spectrometer making use of a known amount of Ne tracer -which was mixed with the gases within the gas handling system. The salt particle size after crushing was mostly less than 0.5 nam- A few larger particles had dimensions ranging up to maximum of about 2.8 mm.

The PSV-2 sample (Table 8.9) consisted of salt which had been l8 irradiated in Project Salt Vault. It was taken from a core sample of the salt around Hole VI, Room 1, and was located about 2.3 in. from the edge of the hole at the midplane. The maximum temperature attained in this location was estimated to be 150 to 155°C. Brine migration took place both during and following irradiation exposure. The migration following irradiation occurred because the salt remained at high temperature for an appreciable period after electrical and radiation heating was terminated. Cooling of the brine cavities must have produced gas-vapor phases, and the resulting two-phase droplets very likely migrated rapidly down a temperature gradient toward the edge of the hole for several days during which air was passed through the hole. 17 The y-2-5; sPec^men was described previously. Brine droplet migration in this sample took place only during irradiation. The unirradiated specimen, L, was a sample of the Lyons salt used in the irradiation experiments (Table 8.3).

The amounts of H2 found with the PSV-2 and 7-2-5 specimens correspond to 0.13 and 0.22 m respectively, within the brine that remained in the salt.

The low value of G-H2 for 7-2-5 clearly indicated that the amount of H2 19 in this sample was near a limiting value. A previously reported analysis based on known radiation chemistry led to the prediction that the steady- state amount of H^ "would exceed 0.05 m, but the actual value could not be predicted because of lack of information on rate constants for pertinent reactions at elevated temperatures. With the available experimental values, it may be possible to refine our estimates of reactions and rate constants in brine under irradiation. However, this has not yet been attempted.

The ratio of 0^ to H^ in specimen 7-2-5 was somewhat less than 0.5, and this indicated that other oxidized species such as C10o were radiolysis products. In specimen PSV-2, the ratio of 0^ to H^ was very much less than 0.5* This behavior may have been associated with the migration of brine after termination of irradiation and the concomitant dissolution of irradiated salt within brine droplets. As mentioned previously, dissolution of irradiated salt subsequent to irradiation produces H^ but very little 0^. Presimiably, H^O^ and/or ClO^" are oxidized species which balance the H^.

The amounts of air indicated by the amounts of N^ + CO exceed those 20 present in the gas bubbles within the brine cavities. Accordingly, most of the air must have been trapped on crystal boundaries.

8.1.7 Future Work

A report of the work on stored energy and other radiation effects in

salt is currently in preparation. No other work on this problem will be

done until core samples of New Mexico salt are received, which may be

as late as 1974.

8.2 Waste Disposal by Hydraulic Fracturing

H. 0. Weeren and J. G-. Moore

The shale fracturing process is the method presently in use at ORNL

for the safe, ultimate disposal of locally generated intermediate-level -345- waste solution (ILW). In this process, concentrated waste solution is mixed with a blend of cement and other solids and injected under pressure into a bedded shale formation at a depth of about 800 ft. The pressure of the injected grout is high enough to initiate the formation of a crack between adjacent layers of shalek As the injection continues, the grout fills this crack and then gradually spreads out to form a thin, approximately horizontal, sheet several hundred feet in extent. Shortly after the injection is complete the grout sets, thereby permanently fixing the radioactive wastes in the shale formation. Subsequent injections form new grout sheets parallel to the first and a few feet above it.

Major recent and current activities of the shale fracturing program have included: (l) the disposal of about 300,000 gal of ILW solution in a series of four injections; (2) the design and engineering of modifications to the shale fracturing facility that will improve the operability of this facility for the next several years; and (3) the preliminary design of a new shale fracturing facility that will have the capability of disposing

of waste solutions with a higher specific activity than can now be handled.

The shale fracturing research program has been directed toward providing

assistance to the overall program where needed, establishing the effect

of certain variables on mix characteristics, and determining some of the

properties of the sludge in the G-unite waste tanks at the 0KNL tank farm

so that a tentative estimate of its pumpability could be formed.

8.2.1 Waste Injections

Several modifications to the shale fracturing facility are scheduled

to be installed in 1973* and detailed information on the performance of

the existing facility was needed for the design of the proposed -346- modifications. A particular effort -was made to obtain these data from a series of four injections of intermediate-level waste solution made during the fall of 1972.

Prior to the first injection in the series, analyses of the waste solutions to be injected were obtained, and simulated waste solutions were prepared for testing with the solids blended for the injection. It was found that the properties of grouts made with plant-mixed solids were significantly different from those of grouts made with laboratory-mixed solids; more of the plant-mixed solids had to be used to yield a product with acceptable properties. A recommended solids-to-liquid ratio that would result in an acceptable grout was prepared for the solids mix stored in each storage bin.

Injection ILW-8, the first injection in the series, was made on

September 29, 1972; few difficulties were encountered. A total of

72,700 gal of ILW waste, 5100 gal of contaminated water, and 3600 gal of cleanup water was injected at an overall solids-to-liquid ratio of 7.3 lb/gal.

Injection ILW-9 proceeded quite smoothly. In this run, a total of

68,300 gal of ILW waste, 3900 gal of contaminated water, and 3400 gal of cleanup water was injected at an overall solids-to-liquid ratio of

7.8 lb/gal.

Injection ILW-10 ran very well with only minor equipment difficulties being encountered. It was observed that some solids were retained in the bulk storage bins and would not feed freely to the mixer. About 6% of the total solids were held up. A total of 84,760 gal of ILW waste, 7230 gal of contaminated water, and 1580 gal of cleanup water was injected at an overall solids-to-liquid ratio of 7»1 lb/gal. -347-

Injection ILW-11 was marked by intervals of difficulty in initiating and maintaining an adequate flow of solids to the mixer. The solids tended to stick in the storage bins and, during the last part of the injection, bridged in the mixer hopper. The bridging restricted the attainable solids flow rate which, in turn, resulted in somewhat poorer control of the mix ratio during this injection than was the case in previous injections. About 100,000 lb of solids was left in the bins at the end of the injection. A total of 75,760 gal of ILW waste, 5820 gal of contaminated water, and 525 gal of cleanup water was injected at an overall solids-to-liquid ratio of 7-2 lb/gal.

Radiation exposures during the injections were •uniformly low (an average of 27 mR per person per injection). Radiation exposures during preinjection maintenance operations were several times greater than those during the injection (an average of 185 mR per person per injection).

About ten days after each injection, any unbound water that existed in the shale beds was bled back through the injection wall. The volume and rate of water bleed-back varied considerably between injections. An analysis of the bleed-back water indicated that virtually all of the radionuclides were retained by the grout. Strain and displacement measurements that were made on the wellhead during two injections indicated that the stresses were small and that failure due to fatigue is probably not a matter for serious concern. An analysis of pressure changes in monitoring wells during the injections indicated that these changes were closely correlated with grout sheet movement. This monitoring technique has promise as an instant indication of grout sheet orientation--something that is badly needed. A detailed account of the four injections has "been written and will

"be issued shortly.

8.2.2 Environmental Impact Statement

Various modifications to the ORNL liquid waste handling system have been proposed. These modifications include a new shale fracturing facility, two new 250,000-gal waste storage tanks for intermediate-level waste, and a new pipeline to connect the facilities in Bethel Valley with the waste disposal site in Melton Valley. An analysis of the probable environmental impact of these facilities, required as a funding prerequisite, has been made; and a first draft of the environmental statement has been written. No significant damage to the environment can be foreseen from

the new wasl^ facilities, and the overall environmental impact is expected

to be beneficial. The shale fracturing facility will remove large volumes

of potentially hazardous radioactive wastes from the existing surface

storage facilities and fix these wastes in impermeable shale formations, well removed from the biosphere.

8.2.3 Sludge Characteristics

The characteristics of the sludge that has accuralated in the Gunite waste tanks must be determined before serious plans for disposal of this

sludge can be made. Although each tank contains soirr- sludge, about Q0%

is found in Tanks W-5, W-6, and W-8. These three tanks were sampled, and

some preliminary determination of sludge properties was made.

The sludge from Tank W-6 consisted of a mixture composed of about

45 wt fo small particulates (about 1 to 5 pt in diameter) and 55 wt %

agglomerated particles (about 200 to 500 gu in diameter). The separated -349-

and dried agglomerated particles could "be crushed easily; the sludge density was 1.5 g/cc. The sludge from Tank W-5 was much like that from

W-6, except that the agglomerated partciles constituted about 75$ of

the sample; the sludge density was 1.6l g/cc. The sludge from Tank W-8

consisted entirely of small particles; the sludge density was 1.21 g/cc.

8.2.4 Leaching Studies

Several accident situations that were postulated for the environmental

impact statement involved the contact of waste-bearing grout with ground-

water. An analysis of these situations required a knowledge of the leach

rates of various radionuclides from grouts under a variety of conditions.

A study was set up to obtain the necessary information. The radionuclides 90 137 239 244 of particular interest were Sr, Cs, Pu, and Cm. The variable

of particular interest was the cure time of the grout prior to the start

of leaching. Values were obtained from grouts that had been cured 0, 7,

and 28 days. In addition, the leach rate was determined for the special

case of fluid grout dumped into well agitated water (corresponding to the

situation in which a vertical fracture that occurred during an injection

intersected a stream bed).

All leach tests were made with a grout prepared by using 6 lb of

standard.solids mix per gal of W-7 simulated waste. The standard solids

m^x is made up as follows:

Cement (Type I) - 2.5 parts

Fly ash - 2.5 parts

Attapulgite 150 drilling clay - 1.0 part

Illite - 0.5 part

Delta gluconolactone - 0.003 part -350-

The W-7 simulated, waste solution contains:

NaOH - 0.18 M

ai(NO3)3 - 0.0074 M

NH^MO - 0.003 M

NaNO^ - 0.8l M

NaCl - 0.093 M

Na^SO - 0.094 M

Na^CO, - 0.19 M

TBP - 400 ppm

The series of waste injections required solids-to-liquid ratios of at least 7 lb/gal; however, results of preliminary "cold" tests made with ratios varying from 5'5 to 7»0 lb/gal showed that 6 lb/gal is about the maximum that can be handled easily in laboratory leach studies. At 7

and 6.5 lb/gal the grouts were very thick, and the preparation of leach

samples without bubble or other surface flaws was difficult. At 6 lb/gal

the grout was pourable and produced acceptable solids for leach testing.

Results from a grout made at 5'5 lb/gal showed it would also be acceptable but the higher ratio (6 lb/gal) was chosen so as to more closely approach

actual operating conditions. Phase separation was low at both ratios,

0.631o and 0.47$ at 5-5 and 6.0 lb/gal, respectively.

Leach samples were prepared from the aforementioned grouts to be

compatible with IAEA requirements (i.e., they were cast into cylinders

50 mm in diameter and 50 mm high by use of 6-oz plastic bottles that had o been cut off 50 mm from the bottom). A total of about 100 cm of grout 2 was used in each sample. A surface area of 20 cm was exposed to the

leachant. The samples were cured for 7 days by placing the plastic containsrs in a tray of water inside a plastic bag. After curing, each sample was placed in another container and 150 ml of tap water was added.

The water was changed daily until the activity in the leachant was too low for accurate measurement; thereafter, weekly to monthly changes were made.

The. tests are continuing to determine the leach characteristics over longer periods of time.

Strontium Leach Results. — In a series of duplicate experiments made 85 with Sr-traced grouts, the average amount of strontium leached from the specimens in about 18 weeks was 6.3, 1«7, and 1.0% for 0-, 7-, and 28-day- cured grouts, respectively. The loss in the final 8 weeks accounted for less than 0.3% of the initial strontium. Preliminary estimates of the -9 -10 effective diffusivities were calculated and found to be 5 x 10 , 3 x 10 , and 3 x 10 -11 cm 2/sec , respectively, for 0-, 7-> and 28-day-cured specimens. 85 These values were estimated from plots of the cumulative fractions of Sr leached from the specimens as a function of the square root of the total 21 leach time. This is the data treatment method recommended by the IAEA.

More recently, it has been suggested that, in the case of limited solubility, the amount of material leached from a surface can be correlated with a theoretical model in which the rate of leaching is assumed to b2e2 controlled by diffusion of the liquid phase out of the leach samples.

The following differential equation describes this phenomenon:

II _ = Dg 9 x + ( c) j (5) whore

C = concentration of the chonical species of interest in the

liquid phase within the interstices of the solid matrix, C^ = concentration of the chemical species at saturation,

t = time,

= effective diffusivity of the leach solution within the

solid matrix,

x = distance within the solid matrix, and

k = leaching rate constant at the site of leaching within the

solid matrix.

Using this method, the data from the 7-day-cured specimens produced a Dg of 1 x 10 10 cm2/sec and a k of 3 x 10"^ sec"1. Both of these values 23 seem reasonable for a material such as strontium in cement.

Additional specimens have been prepared and are being cured for future leach tests involving long-cured material. To determine the effect of strontium precipitate on the leach characteristics, a simulated W-7 waste solution was prepared with a strontium concentration of 1 g/liter 85

(traced with Sr) and a calcium concentration of 0.25 g/liter to improve the handling qualities of the strontium precipitate. Approximately 90$ of the strontium precipitated in the alkaline ¥-7 solution. This mixture will be allowed to age 1 month and then will be incorporated into a grout for leach testing. p'7 afterCesiu curinm gLeac 0, hJ, Results and 28. day— sDuplicat have beee nspecimen leachesd witof grouh tatp watecontaininr forg 1. IC s and 15 weeks, respectively. The total quantity of cesium removed varied inversely with the curing time and amounted to 6.1, 2.4, and 0.68$, respectively. A plot of the total fraction lost as a function of the square root of the total leach time yielded estimates of the effective diffusion coefficients (D ) of 2 x 10 1 x 10 and 1 x 10 for (he 0—, 7-} and 28-day-cured specimens. Additional grout specimens have been prepared and are now curing in preparation for future leach tests involving long-cured material.

Curium Leach Results. — Curium leach tests have been in progress for the last nine weeks. There are insufficient data to calculate an accurate -14

De, but initial data indicate the coefficient will be in the range of 10 l6 2 to 10 cm /sec. The maximum amount of curium that was leached in the 9- week period occurred in the 0-day-cured material and amounted to less than

0.1% of the initial curium concentration.

Plutonium Leach Results. — The plutonium leach tests have also been in progress for 9 weeks, but even less material has been leached from these specimens. Only about 0.01% of the initial plutonium has been removed from the specimens, yielding a D€ estimate of approximately 10 -lo -17 2 to 10 cm /sec. As with the curium tests, it will probably be several months before sufficient data are collected to determine accurate coefficients.

Loss of Radioactive Materials from Freshly Prepared Grouts into

Water. — Information was required to assess the loss of radioactive

compounds into the aqueous phase i^. the event that a radioactive grout was dumped into a rapidly m ving body of water. Grouts at 6 lb/gal were prepared with W-7 simulated waste containing the desired radionuclides

and then immediately mixed for 1, 3, and 5 br with 10 volumes of tap water. Mixing time had no apparent effect on the loss of strontium or

uasium into the aqueous phase. About 0.7% of the strontium and 7% of the

cesium were present in the aqueous supernate in all three cases. There

wii.i no cliHiii'U in the aqueous cesium concentration when the mixtures were -354- allowed to stand 24 hr, were then remixed and the activity of the supernate subsequently measured. Under the same conditions, 2% of the strontium was found in the aqueous phase from the 5-hr mix test, and 1$ in the aqueous phase from the 1- and 3-hr tests. 244 239 Tests made with Cm and Pu showed that the amount of activity in the aqueous phase varied inversely with mixing time. The amount of 244 . Cm lost to the aqueous phase was 2, 1, and 0-3% after 1 volume of grout had been mixed with 10 volumes of water for 1, 3, and 5 hr, respectively. 239

For yPu, the losses were 2.4, 1.9, and 1.0% for the same mixing periods.

In both sets of experiments these values were obtained after the aqueous- grout mixtures had stood 1 hr. The activity of the aqueous phase decreased with time; after 24 hr only 0.02 to 0.06% of either the curium or plutonium remained in the supernate.

Analytical Method. — In the first few weeks of leaching, the cesium and strontium in the aqueous phase were determined by integral gamma counting

of a 5-ml sample. Although the background was quite high (400 to 500

counts/min), the activity of the aqueous phase was sufficiently high to

compensate for the background. Later, as the activity decreased, a narrow

differentiaDc l counting technique was use-1 dOr 7wit h a 0.525-MeV setting for the

Sr and a 0.670-MeV setting for the Cs. This gave a higher ratio of

aqueous activity to background and allowed the lower aqueous concentrations

to be more accurately measured.

The concentrations of plutonium and curium in leachants were determined by a special alpha scintillation counting technique suggested by W. J. 24

McDowell. Aliquots of the leachants were contacted with a small volume

of a scintillation liquid containing an extractant to remove the plutonium -3c;c;- or curium present. The scintillation liquid was then separated from the aqueous phase and counted using a scintillation spectrometer.

The scintillation liquid was made up in toluene as follows:

2-(4'-biphenylyl)-6-phenylbenzoxazole (PBBO) - 4 g/liter

naphthalene - 200 g/liter

di(2-ethylhexyl)phosphoric acid (D2EHPA) - 0.5 M 239 Simultated leach solutions spiked with known concentrations of Pu showed that essentially all of the plutonium was extracted on contact "by

100 ml of leachant (adjusted to a pH 11 to 12) with 15 ml of the scintillation liquid. Similar results were obtained with curium-containing leachants after the pH had been adjusted to about 3* Initial experiments using a scintillation liquid containing only 0.05 M D2EHPA were unsuccessful.

8.3 Wet Oxidation of Plutonium

W. E. Clark

The Pressurized Aqueous Combustion process (PAC) has been under intermittent development for several years at ORNL as a method of converting combustible contaminated waste to ash. It is an adaptation of the high- pressure wet oxidation process which has been used for such diverse purposes as the manufacture of vanillin from sulfite papermill waste and the partial oxidation of sewage sludge (e.g. "Zimpro"). In previous work on PAC at ORNL, the effects of temperature, stirring, slurry concentration, catalysts, and gas flow rate on the reaction rate have been investigated. The principal constituents in the aqueous effluent have been

identified and corrosion tests have led to the choice of a resistant material of construction for the high-pressure equipment. -356-

Our work during the past six months has "been primarily aimed at determining the effect of pH on the completeness of oxidation of an aqueous slurry containing 5$ ("by weight) of simulated glove-box waste

(paper, cloth, rubber, plastic). A high pH (>7) is desirable from a corrosion standpoint, but this has a slight negative effect upon the completeness of the reaction. For example, a 1-hr exposure at 28o°C in a solution with a starting pH of 5 resulted in about 88$ combustion, vs

86$ at pH 9.25, and 83$ at pH 11. The solids were completely solubilized

(except for ash) in all cases. Solutions from slurries at high pH were colored, while those from low pH slurries were clear. Sodium tetraborate was used to control the pH since it mc.y be necessary to add boron or some other soluble neutron poison. The off-gas from the process always had a very marked odor even when mass spectrometry could barely detect any organic material. A bed of activated charcoal removed the odor completely.

This program was can cancelled as of June 1, 1973' The final report is now in rough draft form.

8.4 References for Section 8

1. J. 0. Blomeke et al., An Analysis of Energy Storage and Its Effects in the Proposed National Radioactive Waste Repository, 0RNL-TM-3403 (June 1971).

2. R. D. Chevsrton and W. D. Turner, Thermal Analysis of the National Radioactive Waste Repository: Progress Through March 1972, ORNL- 4789 (Sept. 1972), p. 46.

3. G. H, Jenks, Radiolysis and Hydrolysis in Salt-Mine Brines, ORNL- TM-3717 (March 1972), pp. 13-20.

4. A. L. Boch et al., Radioactive Waste Repository Project: Annual Progress Report for~Period Ending September 30, 1972, 0RNL-4824 (December 1972), pp. 208-9.

5. A. L. Boch et al., ibid., pp. 202-8. 6. G. H. Jenks and C. D. Bopp, unpublished report, October 1972.

7. E. L. Compere, H. C. Savage, and J. M. Baker, "High Intensity Gamma Irradiation of Molten Sodium Fluoroborate-Sodium Fluoride Eutectic Salt," J. Nucl. Mater- 3^, 97-100 (1970).

8. E. Sonder, private communications, March-May 1973.

9. G. H. Jenks and C. D. Bopp, unpublished report, December 1972.

10. G. H. Jenks and C. P. Bopp, unpublished report, September 1972.

11. A. L. Boch et al., Radioactive Waste Repository Project: Annual Progress Report for Period Ending September 30, 1972, 0RNL-4824 (December 19721, pp. 211-16.

12. G. H. Jenks and C. D. Bopp, unpublished report, January 1973-

13. E. Sonder, private communications, October 1972 and May 1973.

Ik. P. D. Schulze and J. R. Hardy, "Frenkel Defects in Alkali Halides," Phys. Rev. B6, 1580 (1972).

15. H. A. Schweinler, 0FJML, private communication, June 1971.

16. R. L. Bradshaw and W. C. McClain, eds., Project Salt Vault: A Demonstration of the Disposal of High Activity Wastes in Under- ground Salt Mines, ORNL-4555 (April 1971), p. 170.

17. A. L. Boch et al., Radioactive Waste Repository Project: Technical Status Report for Period Ending September 30, 1971, OHNL-4-751 (December 1971), p. 228. "

18. R. L. Bradshaw and W. C. McClain, eds., Project Salt Vault: A Demonstration of the Disposal of High Activity Wastes in Underground Salt Mines, ORNL-4-555 (April 1971), pp. 161-73.

19. G. H. Jenks, Radiolysis and Hydrolysis in Salt-Mine Brines, 0RNL-TM- 3717 (March 1972), pp. 6-8.

20. A. L. Boch et al., Radioactive Waste Repository Project: Annual Progress Report for Period Ending September 30, 1972, 0RNL-4824 (December 1972), pp. 216-18.

21. E. D. Hespe, ed., "Leach Testing of Immobilized Radioactive Waste Solids, A Proposal for a Standard Method," At. Energy Rev., V 9(l), 195-207 (1971).

22. H. W. Godbee, Estimation of the Loss of Radioisotopes from Radioactive Waste Solids to the Environment During Long-Term Storage, 0RNL-TM Report (in preparation). -358-

Godbee, ORNL, private communication.

McDowell, OKNIi, private communication. -359-

9. ENVIRONMENTAL SURVEYS

Comprehensive surveys to assist in the assessment of environmental effects associated with milling of uranium, production of power at nuclear power plants, and usage of toxic metals and compounds are in progress, or approaching completion. The uranium milling study will be of direct use to the AEC Directorate of Regulatory Standards in formulating appropriate

,ras low as practicable" guidelines for application to the nuclear fuel

cycle. The studies on liquid and solid radioactive waste (radwaste) treatment systems at nuclear power plants will provide the AEC Directorate

of Licensing with operating data to assist in evaluating the efficiency

and adequacy of these systems. The studies on mercury usage and chemicals used in cooling water treatment are of value in the overall national programs on pollution abatement.

9.1 The Nuclear Fuel Cycle - Milling of Uranium Ores

M. B. Sears

A new program was initiated in January 1973 to make comprehensive

technical studies of the effectiveness and costs of alternative radwaste

treatment systems, and to correlate the costs with the impact of these

wastes on the environment. These studies are for the use of the Directorate

Regulatory Standards in establishing "as low as practicable" guidelines for

licensing nuclear installations under the Code of Federal Regulations.

Chemical Development Section B staff members are actively participating

in Part 3 of this program on the milling of uranium ores. Members of the

Unit Operations Section and the Environmental Sciences Division are also

contributing to Part 3- -360-

The engineering phase of the study will result in the development of incremental capital and operating costs for changes and additions to systems, and in the development of corresponding source terms for radioactive emissions and noxious effluents. The.se systems will cover the range from present practice to the foreseeable limits of available technology, on the basis of expected typical and normal operation over the life of the facilities.

A program plan and preliminary survey of the uranium milling industry were compiled as a guide for the detailed radiological impact study. This report includes a brief description of the radiological problems and a discussion of the rationale behind the preliminary selection of "model" mill flowsheets, "model" mill sites, and generic case studies for radwaste treatment. These were intended as general guidelines, subject to modification, as the in-depth studies proceeded.

Present practices of effluent control and waste management in the industry have been surveyed. Most of the literature (over 200 publications) pertinent to uranium mill processes has been abstracted and catalogued.

A questionnaire with particular emphasis on effluents and the environment was distributed to all active uranium mills in the United States. Replies have been received from half the mill operators. Visits were made to six operating uranium mills representing the different flowsheets in use today and the newest, most modern mill designs. In addition, three stabilized tailings piles were inspected, and discussions were held with members of the Region IV Office of AEC Regulatory Operations and the Grand Junction

Office of the AEC. The calculation of radioactive source terms for the airborne dusts from the mill stacks, and for liquid effluents from a solvent extraction mill, is now in progress. -361-

9*2 Nuclear Industry Wastes

H. W. Godbee and A. H. Kibbey

9*2.1 Use of Evaporation for the Treatment of Liquids in the Nuclear Industry

The purposes of this study were to collect, collate, and report information on the performance of evaporators at nuclear installations and, in particular, nuclear power stations. Information collected includes quantities of each kind of liquid evaporated, and the efficiency with which radionuclides are removed from the stream treated, as well as a comparison of design and operating capacities. In the first analysis, the results of this study are needed to provide operating data to assist the AEC

Directorate of Licensing and the nuclear power Industry in their evaluation of the efficiency of evaporators used in liquid radwaste treatment systems at nuclear power stations. In a broader sense, they are needed to evaluate the role of this unit operation in lowering the discharge of radioactivity in the liquids from any nuclear installation.

Operating and design data on evaporators were collected by direct contacts with 30 organizations, including suppliers of nuclear reactor systems, architect-engineers for these systems, equipment manufacturers, operators of present and proposed nuclear power stations, and selected

AEC-operated facilities. The principal emphasis was placed upon data concerning the system decontamination factors (ratios of feed concentration to condensate concentration) which were achieved. The report1 summarizing the results of this study includes values of decontamination factors for different classes of radionuclides based upon their known behavior in boiling solutions, literature data, and plant experience. The basic theoretical and design factors influencing the decontamination factor (DP) -3^2- of evaporators are also presented to facilitate application of the results to the design and evaluation of evaporators for treating radioactive wastes.

The results of our survey show that:

(1) An average system DP of 103 to 10^ can be expected under

routine operating conditions for nonvolatile radioactive

contaminants treated in evaporators. / n 3 b (2) An average system DF of 10 to 10 can be expected under

routine operating conditions for ruthenium in alkaline,

but not acidic and oxidizing2 , solutions3 . (3) An average system DF of 10 to 10 can be expected under

routine operating conditions for iodine in alkaline, but

not oxidizing or acidic, solutions.

(4) Oil, soaps, and detergents mixed with aqueous wastes reduce

the DF1s by a factor of about 10.

(5) All of the DF values above assume that the evaporator is

well designed, adequately sized, and operated with

reasonable skill. 5 (6) Higher decontamination dactors, average system DF's of 10

to xo6 can be achieved but retire simultaneous optimization

of all conditions, for example: (l) feed conditions to

ensure that all solutes are nonvolatile, (2) evaporator

design and operating conditions to minimize entrainment,

and (3) operator skill and attention to ensure that these

conditions are maintained.

Among the recommendations presented are: -363-

(1) Liquid radwaste evaporators should he tested before use on

actual waste streams. This is probably the only reliable

method of demonstrating that the desired DF values can be

achieved over the extremes of conditions expected. Stable

isotopes and tracer levels of radioactivity can be used in

these tests.

(2) The behavior of iodine in liquid radwaste during evaporation

is complicated and poorly understood. Laboratory studies

are needed to better define its vapor pressure as a function

of pH value, redox potential, and other parameters -which

determine the physicochemical behavior of iodine. Predicted

improved operating conditions should then be confirmed in

large-scale evaporator tests.

9.2.2 Solid Waste Practices at Nuclear Power Plants

The evaluation of solid radwaste treatment systems and in-plant protection programs requires a knowledge of the handling and treatment methods, together with the type, quantity, and radioactive content of the

solid was be generated at nuclear power plants. An estimate of the quantity

and radioactive content of solid waste generated is needed in the preparation

of environmental statements and safety evaluation reports. A comprehensive

review which can be used as a basis for evaluating systems and for making

a reliable estimate of quantities and radioactive content of solid radwaste

is now available.

Limited surveys have been performed by private industry in the United

States, the AEC Directorate of Regulatory Standards, and ORNL. In addition,

the NUKEM Company (Hanau, Germany) has carried out a comprehensive survey -364- of the West German nuclear industry. The available reports show large variations both in the quantity of solid radwaste produced at operating reactors and in the estimates presented by applicants for new reactors.

Present AEC guides and directives do not require operating reactors to routinely report this type of information. The methods for collecting and treating the solid wastes also show great variations.

All of the available information in AEC dockets has been collected and is being summarized in a generic report. Emphasis will be placed on correlating the amounts of waste collected with the type of operations in the plant and with the waste treatment methods.

9-3 Mercury Studies

W. E. Clark

For the past two years, the National Science Foundation has funded a small program aimed at studying the use and disposition of mercury in

U.S. society as a part of a larger program of pollution abatement studies.

Prior work consisted of a survey of the mercury reprocessing industry in 2 the United States and the initiation of a survey of mercury usage and disposal practices by agencies of the U.S. government. During the past six months, the emphasis has been centered on completing the latter survey.

A minor effort was expended on updating the reprocessing study, and some effort was also put forth on paper studies of battery recycle as a means of recycling mercury and other elements. The entire program was terminated as of June 30, 1973. The battery work is not discussed herein because of

its preliminary nature. Work on the other two projects during the past

six months is summarized below. -365-

9«3»1 Survey of Mercury Reprocessors

Response to our new (1971 information) questionnaire to reprocessors was too poor to allow for quantitative revision of the numbers given in

0RNL-NSF-EP-22. However, telephone conversations with a number of the

larger reprocessors indicate that the volume of mercury handled by commercial

reprocessors is probably no smaller than that in prior years, in spite of

the drop in the price of mercury from more than $600 per flask to slightly

more than $300. Restrictions imposed on mercury usage by the Environmental

Protection Agency have apparently increased the pressure for recovering

mercury from residues which would previously have been discarded. There

are many new sources of scrap mercury, although most of them are of small

volume. These effects have apparently offset the decrease in volume

resulting from the considerably diminished use of virgin mercury. Fewer

mercury batteries are now being handled by commercial reprocessors because

of a more aggressive campaign by a major battery manufacturer to collect

used batteries and recycle the mercury. We have never been able to obtain

even approximate recycle figures from battery manufacturers.

9-3«2 Usage of Mercury by Agencies of the U.S. Government

Twenty-three major agencies of the U.S. government were questioned

concerning the amounts of mercury and mercurials used, and about their

methods of disposal of waste materials containing mercury. Of these, nine

reported using negligible amounts, eleven submitted detailed questionnaires

directly from subagencies or individual sites, two submitted limited returns,

and one which promised a limited response has not yet reported. Our final

summary includes data from 679 individual sites and subagencies. The total

usage was slightly more than 101,000 lb/year, or about 2.5% of the national -366- usage. Of this, about 71$ was consumed as metal, 2-7$ was in other chemical forms, and the remaining 26.3$ was in various manufactured items.

A minimum of 30,700 lb of mercury was recovered from scrap, mostly from used metal with smaller amounts from batteries and chemicals. About 200 sites recycled mercury either internally or through external reprocessing services. The final report, now in rough draft form, will take note of disposal and safety practices as well as presenting quantitative data on usage.

9.4 Biological Response to Proprietary Chemicals Used in Cooling Water

R. H. Rainey

0RNL and the Cooling Tower Institute (CTl) are cooperating in a program in which the environmental impact of proprietary chemicals used in cooling water may be evaluated. Many formulations are added to the water used to cool power plant condensers to decrease corrosion, the deposit of solids, and the growth of bacterial organisms. Most of these formations contain chemicals which are potentially harmful to environmental organisms.

However, the composition must be known before the environmental impact can be evaluated. CTI has agreed to assist in a program in which the companies that market the cooling water chemicals will make available to 0RNL the proprietary formulations of their products. 0RNL will then conduct literature searches and use other means of determining the response of various organisms to these chemicals, while maintaining the confidentiality of the products.

Contacts have been established with the cooling water chemical companies that are members of CTI, and several companies, including three of the largest, have given oral agreement to cooperate with the program. A -367- contract agreement has "been drawn up by the Union Carbide Law Office, and negotiations are under way in an attempt to make it acceptable to both the chemical companies and AEC. Proprietary data will not be transmitted to

ORNL until the agreements have been signed by representatives of the chemical companies, Union Carbide, and AEC.

In order to expedite the program, a number of the CTI Cooling Water

Treatment Committee has provided a list of corrosion inhibitors and biocides.

These lists, while not current, are useful for preliminary studies. The list of corrosion inhibitors includes 177 materials from 20 companies, but identified by generic type only. The biocides are a sampling from 19 companies and include 96 products which contain at least 56 different chemical compounds. The list of biocides has been used to explore computer search techniques for identifying the chemical compounds and for obtaining references to biological responses to these compounds. It is anticipated that the lists of proprietary cooling water chemicals would be much larger than these preliminary lists. Therefore, efficient computerized

search programs would be required.

9.5 References for Section 9

1. H. W. Goabee, Use of Evaporation for the Treatment of Liquids in the Nuclear Industry, 0RNL-4-790 (October 1973).

2. Walter E. Clark and W. Fulkerson, Survey of the Mercury Reprocessing Industry, I968-I97O, ORNL-NSF-EP-22 (October 1972). -368-

10. SEPARATIONS PROCESSES

Goals in this program are to develop new and improved separations systems for use in radiochemical processing and in the extractive metallurgy field.

In prior years, this has resulted in development of solvent extraction processes for recovering uranium, thorium, vanadium, beryllium, and cesium from ores, and fission products from reactor fuel reprocessing wastes.

The most recent effort has been concerned with: (l) by-product recovery of uranium from wet-process phosphoric acid; (2) separation of radium from uranium ore tailings; (3) separation of alpha emitters from reprocessing wastes; and (4) study of the applicability of high-pressure ion exchange to certain hydrometallurgical separations.

10.1 Recovery of Uranium from Wet-Process Phosphoric Acid

D. J. Crouse and F. J. Hurst

During this report period, laboratory tests were completed on an

alternative first-cycle extraction process1 for recovering uranium from

wet-process phosphoric acid. The purpose of this study has been to

develop an economical method for recovering the 0.1 to 0.2 g of uranium

per liter present in wet-process phosphoric acid that is produced from

Florida phosphate rock as an intermediate step in the production of

fertilizer. The acid represents a potential source of about 2000 tons

of uranium per year, which at the present time is lost when the acid is

converted to fertilizer and dispersed to the soil.

The new first-cycle process uses a commercial mixture of mono- and

dioctylphenylphosphoric acid (0PPA) as the extractant, rather than the

synergistic combination of di(2-ethylhexyl)phosphoric acid (D2EHPA) and -369-

2 trioctylphosphine oxide (TOPO) used previously. This reagent is much

less expensive and has a higher extraction power for uranium than the earlier extractant. In addition, it extracts uranium(iv), the prevailing

oxidation state of uranium in wet-process acid; this eliminates the liquor

oxidation step required in the earlier process. The D2EHPA-T0P0 solvent

is used in the second cycle, as "before, to produce a high-grade U^Og.

10.1.1 Process Flowsheet

In the first cycle extraction process, the acid is cooled to 40-45°C,

and the uranium is extracted with a 0.3 to 0*4 M mixture of mono- and

dioctylphenylphosphoric acids in an aliphatic diluent. Uranium is

recovered from the solvent by contacting it with a 10 M s°l'u-tion

containing sodium chlorate; the chlorate oxidizes the uranium to the less-

extractable hexavalent state and effects its transfer to the aqueous phase.

A convenient source of strip solution is the 45 to 55$ P2°5 P^duct acid from

the evaporators. The strip solution can be loaded with uranium to 15 to

20 g/liter and, after dilution to 6 M E^PO^, can be fed directly to the

second-cycle extraction system.

In the second cycle, the uranium is extracted with 0.3 M D2EHPA—

0.075 M TOPO in an aliphatic diluent. The organic extract is scrubbed

with water to remove extracted phosphoric acid and then stripped with an

ammonium carbonate solution under conditions that result in direct

precipitation of the uranium as rapid-filtering ammonium uranyl tri-

carbonate (AUT). The AUT is calcined to U„0A. -370-

10.1.2 Solvent Composition

The choice of diluent for the OPPA extractant can significantly affect extraction performance and phase separation characteristics. Amsco

Odorless 450, a refined high-boiling, high-flash-point aliphatic solvent, was selected as a suitable diluent for process use and was used in most . -X-X- test work. On the basis, of cursory tests, Napoleum 4-70 appears equally suitable.

The commercially available OPPA extractant is an approximately equimolar mixture of mono- and dioctylphenylphosphoric acids. This is fortunate, since tests have shown that uranium extraction coefficients obtained with the mixture are higher by a factor of about 15 than with either component alone. Of course, the extraction coefficient is also dependent on the total extractant concentration; for th® equimolar mixture, it is proportional to the 1.5 power of the extractant concentration.

Although higher extraction efficiency is attained by increasing the extractant concentration, this is done at the expense of higher solvent costs. In most cases, an OPPA concentration in the range

of 0.3 to 0.4 M gives a uranium extraction coefficient of about 10, which is adequate for good extraction efficiency.

10.1.3 Effects of Process Variables on Extraction

Variables that have a major influence on uranium extraction are the

H^POacid.^ concentrationThe uranium ,extractio the Fe(llln coefficien) concentrationt decrease, and bthye a temperatur factor ofe abouof tht e

* American Mineral Spirits Co., Atlanta, Ga. ** Phillips Petroleum. -371-

1.5 as the phosphoric acid concentration was increased from 5 to 6 M, the

typical concentration range of wet-process acid produced in the United

States.

Without cooling, the temperature of wet-process acid feed to a solvent

extraction plant would be approximately 6o°C. By cooling the acid to about

40°C (which phosphate producers say can be done economically), the uranium

extraction coefficient is increased by a factor of about 2. Cooling below

40°C does not appear advantageous due to higher cooling costs, and has the

added disadvantage of poorer phase separation encountered at lower tempera-

tures. The optimum temperature for the process, therefore, appears to be

near 4-0°C.

Freshly produced wet-process acid usually contains from 6 to 10 g of

iron per liter, of which 0.5 to 3 g/liter is in the ferrous state. Under

these conditions, essentially all of the uranium is present as U(l"V), and

no reduction is necessary for efficient extraction of the uranium. On the

other hand, ferric iron is appreciably extracted by the OPPA solvent and

interferes with uranium extraction; the uranium extraction coefficient can

be increased by a factor of 2 to 3 by reducing most of the ferric iron.

This procedure does not appear attractive, however, because of the high

cost of the reduction step. Usually, a uranium/iron decontamination factor

of about 50 is achieved in the first cycle. This Is adequate since the

D2EHPA--T0P0 solvent separates the uranium efficiently from iron in the

second cycle.

10.1.4 Stripping of Uranium

OPPA is such a powerful uranium extractant that stripping of the

uranium from the extractant is difficult. Oxidizing the uranium to the -372- hexavalent state enhances the stripping efficiency since U(VI) is extracted much less strongly than U(lV). However, even with a change in valence, use of a stripping solution vith strong complexing power is necessary to obtain favorable distribution of U(Vl) to the aqueous phase. Efficient stripping is obtained by using NaClO^ as the oxidant in wet-process acid that has been evaporated to ~10 M H^PO^ concentration. Table 10.1 shows data for stripping 0.32 M OPPA—Amsco 450 solvent. In batch contacts at an organic/ aqueous phase ratio of 5/1, 95$ of the uranium was stripped when the NaClO^ concentration was 4 g/liter, whereas only 34$ was stripped in the absence of the oxidant.

10.1.5 Process Demonstration

A continuous demonstration of the process was made in bench-scale mixer-settler units with four extraction and three stripping stages. The solvent, ^0.32 M OPPA in Aniseo 450, was subjected to about 80 complete extraction-stripping cycles. In order to simplify the initial operation,

"pure" 6 M H^PO^ (containing only Fe2+ and was processed over the first 35 cycles of operation. We then successfully processed approximately

200 gal of "green acid," obtained from two commercial phosphate plants.

("Green acid" is produced from phosphate rock that has been calcined to eliminate organic matter). Finally, about 100 gal of regular wet-process acid ("brown acid") was processed successfully over the last 20 cycles.

Uranium recoveries were > 90$, and the strip product solution contained

15 of 20 g of uranium per liter.

Figure 10.1 shows typical stage data obtained by treating a sample of "green acid" that contained 0.07 g of uranium per liter. Uranium recovery in four extraction stages was about 91$ with an aqueous/organic flow ratio -373-

Table 10.1. Effect of Sodium Chlorate on Stripping Uranium from OPPA with H^PO^

Operating Conditions:

Organic — 0.32 M OPPA in Amsco 450 containing 1 g of U(IV) per liter.

Aqueous — wet-process acid evaporated to 10 M H PO^ and containing NaClO^*

Procedure — solutions mixed vigorously at an organic/ aqueous phase ratio of 5/1 for 5 min at . 4o°c.

Concentration Concentration of Uranium Uranium of NaClOq (g/liter) Stripped (g/liter) Organic Aqueous (%)

0 0.75 1-95 34

1 0.48 2.58 52

2 0.14 4.09 85

4 0.06 5-99 95 ORNL DWG. 73-9921

0.32 1 OPPA—AMSCO 450 {15 ml/min)

EXTRACTION (45°C) STRIPPING ( 30°C)

0.071 U 0.I6U 0.29 U 0.45 U 0.081 U 0.019 U 0.004 U 2.4 Fe ^l.2Fe 12.8 U 2.7U 0.65 U 0.006 U 0.015 U 0.026U 0.047U 25.6 Fe

1 RAFFINATE COMPANY D STRIP EVAPORATED ACID PRODUCT WET-PROCESS 0.079gU/liter SOLUTION ACID 7.5 gFe/liter 10 M. H3PO4 (105 ml/min) NqC103 600 g/liter 7g NaC!03/liter (0.007 ml/min) 13 g Fe/liter (0.5 ml/min)

Fig. 10.1. Demonstration of First-Cycle Uranium Recovery from

Phosphoric Acid with OPPA Extractant. The numbers in the blocks show

the metal concentrations in grams per liter at steady state. -375- of 7/1* More than 99% "the uranium was stripped in three stages to give a product solution containing 12.8 g of uranium per liter; this corresponds to a concentration factor of 180 across the extraction-stripping system. In later operation with this same acid feed, hut at an aqueous/ organic extraction ratio of b/X} uranium recovery was 98%, and the strip product solution contained 19 g of uranium per liter (concentration factor of 270). The decontamination factor of uranium from iron when treating, this acid was ~50.

In the demonstration test, the minimum sodium chlorate consumption

achieved was about 0.8 lb per pound of uranium recovered, a factor of 5

to 6 higher than the stoichiometric requirement. We found that; the major

consumption of chlorate was not in oxidizing uranium, but rather to a side reaction in which chlorate reacts with chloride in the acid to produce

chlorine and/or chlorine dioxide. If these gaseous products are contained

(by sealing the stripping units from the atmosphere), much more efficient utilization of the chlorate is achieved. Also, by operating the stripping units at 30°C and by adding most of the chlorate in the form of an almost

saturated water solution to the first stripping stage (the balance was

added to the third stage), more efficient utilization of the oxidant was

obtained.

The extraction units were operated in two modes: with the aqueous

phase continuous, and with the organic phase continuous. Although much

higher throughputs were attained during operation with the aqueous phase

continuous, the amount of solvent entrained in the raffinate was excessive.

Entrainment was moderate, however, when three stages were operated with

the aqueous phase continuous and the fourth (raffinate) stage was operated -376- with the organic phase continuous. With this method of operation, the raffinate contained about 0.2 gal of solvent per 1000 gal when treating

"green acid;" passing the raffinate through a 1.5-hr holdup tank decreased the entrainment to <0.1 gal. When treating "brown acid," the solvent

entrainment of about 0.4 gal per 1000 gal was decreased to 0.2 gal by passing the raffinate through the holdup tank, and finally to <0.1 gal by

passing it through a column that contained Amsco 450 diluent and l/4-in.

Berl ..addle packing.

During the processing of "green acid," operation was smooth and there was essentially no buildup of solids at the aqueous-organic interface.

With "brown acid," results were much more variable. When the feed contained

dispersed solids, most of the solids accumulated at the liquid-liquid

interface and the settlers flooded rapidly. However, with filtered feed,

operation was satisfactory. Pilot plant tests with fresh acid at the

plant site are needed to determine whether the process can successfully

treat "brown acid."

Samples of the solvent, taken periodically, showed no appreciable

loss of extraction power with cycling (Table 10.2). Solvent titrations

showed that the sum of the concentrations of the mono- and di-OPPA

components remained relatively constant, although the mono- di- ratio

changed. Preferential loss of mono-OPPA would be expected since other

measurements have shown that its distribution loss to the aqueous phase is

about 25 parts per million parts of aqueous, whereas distribution loss of

the di-OPPA is negligible. Interpretation of the titration data is diffi-

cult, since the extent of diluent evaporation is not known, and since an

appreciable volume of fresh solvent was added during the run to compensate -377-

Table 10.2. Effect of Cycling on Solvent Composition and Uranium Extraction Power

Uranium Number of Type of Extraction OPPA Concentration (M) Organic Acid Coefficient3 Cycles Processed (Eg) Mono- Di- ' -> ;alb

Start 11. 4 0.175 0.151 C.326

35 Pure 11.3 0.148 0.158 0.306

50 "Green" 10.7 0.154 0.174 0.328

80 "Brown" 10.5 0.175 0.170 0.345 a Solvent sample was subjected to a standard extraction test with

wet-process phosphoric acid at 25°C.

Sum of the mono- and di- components. The original solvent and the

solvent that was added to compensate for volume losses over the

first 60 cycles anolyr d 0.175 K mono- and 0.151 K di-0P?A; the

replacement solvent over the last 20 cycles analysed 0.187 M

mono- and 0.157 M di-OPPA. -378- for solvent losses from sampling, evaporation, entrainment, and spillage.

Nevertheless, these data do serve to indicate that the reagent is stable enough for process use and that distribution losses to the raffinate are low.

10.2 Separation of Radium from Uranium Ore Tailings

D. J. Crouse and F. G. Seeley

We are studying certain aspects of the radium pollution problem

associated with uranium milling. The purpose of this work is to characterize

the leaching of radium from uranium ores and sulfate-leached ore tailings, with the eventual objective of decreasing the leachability of radium from

the tailings of present processes or removing the radium (and other radio-

nuclides) from the ore wastes by a special process.

Most of the experimental tests have been made with a sandstone-type uranium ore from the Ambrosia Lake district (New Mexico). Leaching of

this ore three times with 3 M HNC or 3 M HC1 at 85PC and 20 to 2% pulp

density dissolved of the radium (Table 10.3)' However, leaching with sulfuric acid, which is used in all acid-leach uranium mills,

dissolved <1% of the radium. From 80 to 90% of the radium in the sulfate-

leached tailings is found in the slime (-150 mesh) fraction, which repre-

sents only about 20-;C of the tailings weight. Leaching with hot 3 M HNO^

removed 97 to 9&% of the radium from both the sand and slime fractions. » A number of salt solutions were tested including some that were previously

found to be effective for removing radium from ore tailings. From 65 to

of the radium was removed from the sulfate slime fraction by leaching

for 16 hr et room temperature and about 3-3^ pulp density with 1 U

solutions of CaCl„r::, BaCl c0> Ua£,» SDTA, and :;ap cDTPA (Table 10.4). Another tsst series showed that leaching of radium from the tailings with MaCl solution

is ropid, with equilibrium being reached within 0.5 hr. Table 10.3« Leaching Radium from Ambrosia Lake Ore and Sulfate-Leached Tailings

Leach Type Pulp Time Leach Radium Dissolved ($') of Leaching Density per Contact Temp. First Second Third Sample Solution (%) (hr) RO Leach Leach Leach Total

Ore (-35 mesh) 3 M HN03 20 2 85 93.8 1.5 0.2 95-5

3 M HC1 25 1 85 90 3.3 0.5 93.8

0.5 M H2S0U 50 17 60 0.12 0.21 - 0.3

0.75 M HGSO^ 50 17 50 0.25 0.51 - 0.8

8 Sand tailings 3 M HK03 25 1 85 9h 3.0 - 97 g Slime tailings 3 M HNO^ 25 1 85 72 25. 0.8 97.8

The residue (tailings) from a sulfuric leach of -35 mesh ore was screened to give a sand

fraction (+150 mesh) and a slime fraction (-150 mesh). The sand fraction, which constituted

8l$ of the residue by weight, contained 16$ of the radium; the slime fraction contained 84$. -380-

Table 10.4. Leaching of Radium from Sulfate Slime Solids with Salt Solutions

Leaching conditions: Ambrosia Lake slime solids (derived from H2S01+ leach of ore) were leached for 16 hr at room temperature with indicated solution; pulp density was about 3-3$ (30 ml of solution per gram of solids).

Salt Radium Leached Solution Final pH (%)

H20 4.0 <10

1 N FeCl3 0.9 59

1 N Na^EDTA 11.7 93

1 N Na-DTPA 11.8 94 — 5 1 N NaCl 3-8 36

1 K KC1 3-8 12

1 N CaCl2 5.0 73

1 N BaCl2 3.4 65

1 N NH^Cl 3.6 15

1 N A1C13 2.3 20

1 N LiCl 4.0 11

1 N HC1 0.6 34 -381-

In further examination of sodium chloride as a leaching agent, more

than 90% "the radium was dissolved by repeated leaching with large volumes

of 1 M NaCl solution (Table 10.5)' Radium leaching was apparently inefficient until most of the soluble calcium and barium had been leached from the ore

solids. When leach solution from the first contact was used to leach fresh

tailings, radium removal was slightly poorer than in the original contact.

A series of precipitation tests was made to test the observation of k

S. C. Lind concerning the essential nonseparability of radium, and barium

in the sulfate system (i.e., the radium/barium ratio is approximately the

same in the solution and precipitate phases). On addition of 22substoichio6 - metric amounts of sulfuric ecid to BaCl^ solutions containing Ra/ the

amount of radium precipitated was 30 to 50% higher than the calculated

amount of barium precipitated, indicating only slight preferential pre-

cipitation of the radium (Table 10.6). Addition of calcium to the system

had only a minor effect on the amount of radium precipitated. Determination of 226R a is accomplished by using a 400-channel gamma ^ 2lk spectrometer and integrating channels under the 0.609-MeV peak of Bi. 2P2 The samples are counted approximately 38 days (10 half-lives of "* Rn) after

the solids are sealed in counting tubes. The method is valid down to 226 approximately 25 pCi of Ra.

10.3 Separation of Alpha Emitters from Reprocessing Wastes

F. G. Seeley

The high-level waste generated by the processing of spent fuels by

a .standard Purex solvent extraction procedure contains most of the fission

products, americium, curium, and neptunium, as well as small amounts of

plutonium and uranium that were not recovered in the solvent extraction

processing. The actinides constitute only a small fraction of the hazard -332-

Table 10.5* Leaching of Sulfate Slime Solids with 1 M NaCl Solution

Procedure: Ore slimes from sulfuric acid leach were leached four times at room temperature with successive volumes of 1 M NaCl solution at ~ 3«3% pulp density.

Total Radium Leach Leached Analysis of Leach Solution (mg/liter) Contact W) Calcium Barium

1 38 1465 95

2 88 36 2.8

3 91 4.8 1.6

4 93 1.7 <0.5

a lB 29 1595 110

£ Solution from the first contact was contacted with fresh slime

solids at 3-3% pulp density. -383-

Table 10.6. Precipitation of Radium from BaClg and BaCl^-CaCl^ Solutions

Procedure: H^SOLj. was added to a 0.2 M BaCl2 solution that contained 600 pCi of -b^g mixture was diluted to 20 ml, allowed to stand 16 hr, and centrifuged. In some tests, CaCl^ was also added to the system.

Solution Volume (ml) „ ,. - Radium — 2 11 \r tt o/^s rv o n m Sulfate Precipitated _ 00(22C6 0.4 M H^SO, 0.2 M CaCl. _ (pluf s Ra N) —2b - 2 AddeA da (%)

10 1 0 20 29

10 2 0 40 52

10 3 0 6o 92

10 5 0 100 >100

10 l 2 17 31

10 2 4 29 53

10 3 6 37 80

"Percent of the stoichiometric amount needed to precipitate all of

the barium plus calcium. -384- associated with the waste at the time of discharge from the reprocessing plant; however, because of their long lives, they become the controlling hazard after several hundred years. Removal of these actinides from the waste could simplify the long-term storage problem associated with such wa ste.

We have initiated a program to search for separation procedures that will provide efficient removal of these alpha emitters. Emphasis in the initial tests has been centered on the solvent extraction of americium and curium, since these elements appear to be most difficult to separate from the waste. Several organic extractants were screened for their ability to extract cerium and europium (stand-ins for americium and curium) from nitric acid and nitrate salt solutions. None of the extractants examined, with the exception of octylphenylphosphoric acid (approximately equimolar mixture of mono- and di- acids) and a "pyro" phosphoric acid (reaction mixture of octanol and Po0 ), extracted cerium(lll) significantly from ^ z> nitric acid solutions that were 1 M or higher in concentration (Fig. 10.2a).

The extraction coefficient for cerium from 1 M HNO^ "with 0.3 M octylphenylphosphoric acid was 60, but decreased to <1 as the acid concentration was increased to 5 M. The "pyro" reagent, although an efficient extractant, is unstable and is not suitable for process use.

Reasonably efficient extraction of cerium(lll) from lithium nitrate solutions that were 0.2 M in nitric acid (Fig. 10.2b) was obtained with several solvents. The extraction power of octylphenylphosphoric acid decreased as the salt concentration was increased, but the opposite effect was observed for most of the other extractants. The extraction coefficients with trioctylphosphine oxide (T0P0) ranged from 70 to 1900, and over most ORNL DWG. 73-9920

Fig. 10.2. Extraction of Cerium(III) from (a) HN03 Solutions and

3+ from (b) 0.2 M m03-LiN03 Solutions. Aqueous phase: 0.005 M Ce traced with 144Ce. Organic phase: (l) 0.3 M Adogen k6k (quaternary) in 95^ diethylbenzene—5$ tridecanol; (2) 0.5 M diamylamylphosphonate in dodecane; (3) 0.3 M methyltrioctylphosphonium nitrate in 95$ diethylbenzene—5% tridecanol; (4) 0.3 M octylphenylphosphoric acid in dodecane; (5) 0.5 M di(tridecyl)phosphoric acid in dodecane; (6) 0.3 M

TOPO in benzene; (7) 0.9 M "pyro" phosphoric acid in dodecane (reaction mixture of 1 mole of octanol and l/2 mole of P20B). Phase ratio: l/l. of the nitrate concentration range were considerably higher than those obtained with the other extractants.

Di(2-ethyIhexyl)phosphoric acid (D2EHPA) is an extremely efficient extractant for Pu(lV), and could conceivably be used for scavenging plutonium from the waste. In tests of the extraction of fission products with 0.02 M D2EHPA from 1 to 8 M HNO^, 95Zr-Kb extraction coefficients ranged from 30 to 77, while those of Ce(lll), Eu, and Ru were less than 95

0.01. The Zr-Nb can be stripped from this extractant with oxalic acid, but the stripping is slow. The stripping coefficient with 0.5 M oxalic

acid increased from 0.3 at 1 min to 4-200 at 16 hr.

10.4 High-Pressure Ion Exchange Studies

F. J. Hurst

The use of ion exchange resins of very small particle size and high column pressures (termed "high-pressure ion exchange") has been shown to have a number of important advantages over conventional ion exchange in

some applications. These advantages include higher throughput capacity,

suppression of gas bubble formation (gases derived from radiolysis) in the

resin bed, and much more efficient resolution in chromatographic separations.

A disadvantage is a negligible tolerance for solids which would rapidly

plug the fine resin bed. High-pressure ion exchange has been applied 5 6 successfully in analyzing body fluids, separating actinides, and 233 7 purifying U. We have initiated studies to determine if exp3.oitation of

the unique advantages of high-pressure ion exchange might greatly broaden

acceptance of ion exchange in metal separations and other separations

processes. Initial consideration is being given to difficult hydro-

metallurgical separations such as the separation of nickel from cobalt, zirconium from hafnium, and niobium from tantalum. Experimental work has been started on separating nickel from cobalt.

In the industrial Caron process , the nickel-cobalt oxide ore is subjected to a reductive roast to convert the nickel and cobalt to the metallic state. The calcine is leached under oxidizing conditions with an ammonium hydroxide—ammonium carbonate solution to dissolve the metals as their ammine complexes. The nickel and cobalt are recovered from solution and separated from each other by precipitation techniques. 9 Several cation exchange resins were tested previously by the Bureau of Mines for separation and recovery of nickel and cobalt from ammoniacal liquors. It was found that the resins had a moderately good loading capacity for nickel and cobalt, provided the ammonia and carbonate concentrations were not high. Also, the resins sorbed nickel more strongly than cobalt, but the selectivity was not adequate to allow a clean-cut separation.

Our initial tests were made with a sample of leach liquor obtained from a commercial

company. The solution had a pH of 9«7 3nd contained the following constituents at the indicated concentrations (g/liter):

11.2 Ni, 0.4-2 Co, 0.044 Cu, 72 Mly and 55-6 C02- 2+Th e nickel is present as the blue divalent hexammine complex, [^(WH^)^! , whereas the cobalt exists in almost equal amounts of the yellow trivalent hexammine complex,

4 + [Co(NH^)^]^ " and the red trivalent pentammine complex, [CotM^)* H20]^ „

A preliminary column test with Dowex 50 (8% cross-linkage) resin

showed that the red cobalt pentammine complex is sorbed less strongly

than the blue nickel hexammine species. The yellow cobalt hexammine complex, however, is sorbed much more strongly than the other species and -388- is very difficult to elute. A number of attempts, all without significant success, were made to convert all of the cobalt to the yellow form since this conversion would greatly simplify separation of the metal ions.

Our present approach has been to remove the strongly sorbed yellow cobalt species in a preliminary column. The column effluent is then pumped to the main column where the nickel and red cobalt complexes are sorbed and separated by chromatographic elution with a concentrated ammonium carbonate solution. Figure 10.3 shows the results of a column test in which ^2 ml of the commercial leach liquor, sufficient to load the resin to about 5% of its capacity, was pumped onto the top of a l6-cm-deep bed of 15- to 25-ix resin and then eluted with 4 M (KH^JgCO^ solution at a

-2 -1 flow rate of 3.2 ml cm min . The pressure developed in the column at this flow rate was about 180 psig. Both the cobalt and nickel products were recovered at a purity exceeding 99a similar run in which

2 M (MH^gCO was used to elute the column, some overlapping of the bands occurred.

Nickel and cobalt distribution coefficients calculated from the column test data agree with coefficients determined in batch tests for various concentrations of ammonium carbonate (Fig. 10.4). The batch data were obtained by eluting Ni and Co from samples of 100-mesh Dowex 50 resin containing about 50 mg of Ni and 1 mg of Co per gram of resin with (NH^^CO^

solutions (the yellow cobalt species had been removed prior to the

tests). Both the batch and column data indicate that the Ni/Co separation factor is about 1.6, which is adequately high to permit efficient separation

of the metal ions. -389-

ORNL DWG. 73-9922

Fig. 10.3. Chromatographic Separation of [CO(NHq )sH20p"'' from

s+ rNi(NH3)6] by Elution with k M (NH*)2C03 Solution. Column: 16- by

0.63-cm with Dowex 50W-X8 resin (15-25 NH** form). Flow rate:

3.2 ml cm"2 min""1. Sample: 23 mg of Ni and 0.^5 mg of Co were sorbed on the top of column before elution. ."3 0Q-.

ORNL DWG 73- 9923 200 1

JiOO

H- Z UJ 50 o y. u. Uoi o 20

10 E A,o BATCH TESTS H C/) A,® COLUMN TEST

UJ 2 X X 0.1 0.5 I 5

(NH4)2C03 CONCENTRATION (M)

Fig. 10.4. Effect of Ammonium Carbonate Concentration on the

2+ 3+ Sorption of [NiO'JHa )s] and [C0(m{3 )SH20] by Dowex 50W-X8 Resin. -391-

Although efficient separation of nickel from cobalt was obtained in these tests, thv flowsheet as demonstrated would not be attractive for use in a large tonnage industry. Data are needed at higher flow rates and for conditions that make more effective use of the sorption capacity of the column. Also, a more efficient eluting solution is needed.

10.5 References for Section 10

1. Chem. Technol. Div. Annu. Progr. Rep. May 31, 1973, ORNL-4883, p. 56.

2. F. J. Hurst, D. J. Crouse, and K. B. Brown, "Recovery of Uranium from Wet-Process Phosphoric Acid," Ind. Eng. Chem., Process Design Develop. 11(1), 122-28 (1972).

3. S. D. Shearer, "The Leachability of Radium-226 from Uranium Mill Solids and River Sediments," Ph.D. thesis, University of Wisconsin, Madison, 1962.

S. C. Lind et al., "The Solubility of Pure Radium Sulfate," J. Am. Chem. Soc. 5o(JJ, (1918).

5. C. D. Scott, "Automated High Resolution Analysis for the Clinical Laboratory by Liquid Column Chromatography," Advan. Clin. Chem. 15, 1 (1972).

6. D. 0. Campbell, "Chromatographic Separation of Transplutonium Elements Using High Pressure Ion Exchange," Ind. Eng. Chem., Process Design Develop. 8, 95 (1970).

7. R. H. Rainey, Laboratory Development of g. Pressurized Cation Exchange Process for Removing the Daughters of 2from QRNL-4731 (December 1972).

8. J. R. Boldt, Jr., The Winning of Nickel, p. 425, Van Nostrand, Princeton, N. J., 1967.

9. I. W. Nicholson, P. T. Brooks, and J. B. Clemmer, "Innovations in Processing Nickel-Cobalt Ores," presented at the Rocky Mountain Minerals Conference, sponsored by the A.I.M.E., Salt Laks City, Utah, Sept. 17-19, 1958. 11. CHEMICAL APPLICATIONS OF I.UCLEAR EXPLOSIVES

The Chemical Technology Division has participated in the Plowshare Program for several years oy studying the potential behavior of radionuclides in underground applications of nuclear explosives. Studies of radionuclide behavior in the recovery of oil from oil shale are being made at the present time. The oil shale deposits of the western iJnited Stages represent the world's largest undeveloped petroleum resource, and the currently developing energy crisis has increased interest in recovering this oil. The feasibility of fracturing the shale with nuclear explosives, and subsequently retorting the oil in place, has been studied for a number of years by several governmental agencies and commercial companies. We are studying the oil recovery system with regard to the process behavior of the radionuclides that would contaminate the fractured shale.

11.1 Contamination by Tritium W. D. Arnold Our present studies are concerned primarily with tritium. In earlier studies it was shown that some tritium contamination of the oil can be expected.

Recent tests were made in a large retort that allows the generation of a moving retorting front. This permited the behavior of the radionuclides to be examined under conditions that more closely resemble those expected in a nuclear-broken shale chimney. In these tests, a 1500-g charge of 0.2-in- to 1-in. shale pieces formed a 30-in bed in the 2-in.- ID retort. Oil was produced by heating the outside of the glass column -393- with a clamshell furnace while passing a gas stream (150 cc/min each of nitrogen and carbon dioxide) through the shale bed. The furnace was lowered down the column to generate a moving retorting front. Oil and water were collected in receivers at the bottom of the column. The gas stream passed from the retort through a bed of CuO (750°C) to convert any hydrocarbons to CO^ and H^O, two cold traps (-70°C) to recover the water, snd a flowmeter before leaving the system to a filtered hood.

The shale samples were contaminated with tritium by heating for kO to 50 days at 85°C in a sealed vessel with 0.25 ml of water containing about 100 nCi otf tritiated water. The contaminated shale was transferred to the retort at the end of the exposure period. Five oil fractions collected during a retorting run at a shale temperature of 36o°C contained essentially the same tritium concentration (Table 11.1). A total of 70

H,Ci of tritium was found in the retort products. The distribution was

about the same as that observed in the earlier test series.1 About 30% was found in the oil, about 50% in the water produced, about 10% in the off-gas, and about 10% in the depleted shale. The furnace was lowered manually (every 10 min) at the rate of 6 in./hr in this test. The system has now been modified to provide mechanical lowering of the furnace at a uniform rate. Retorting over a longer period of time (by decreasing the

furnace travel rate from 6 to 2.3 in./hr) had essentially no effect on

the tritium contamination of the oil.

Adding water vapor to the entering gas stream by sparging it through

water at 50°C decreased the tritium contamination of the oil to about

0.06 vvCi/ml (Table 11.2), or about half that obtained previously. In

this test, the furnace was lowered past the shale column four times: -394-

Table 11.1. Tritium Concentration of Shale Oil Fractions

Procedure: 1500 g of 0.2-in. to 1-in. oil shale was contaminated with tritium by heating for 50 days at 85°C with 0.25 ml of water containing about 100 |iCi of tritiated water. Oil was retorted at a shale temperature of 360°C by lowering the furnace manually at the rate of 6 in./hr.

Oil Tritium Fraction Volume in Oil No. (ml) (MCi/ml)

1 36.8 0.141

2 32.5 0.137

3 34.0 0.137

4 22.0 0.136

5 2.13 0.125 -395-

Tat>le 11.2. Tritium Distribution in Retorting Test with Water Vapor Added to Entering Gas Stream

Retorting procedure: Furnace was lowered four times down the shale at 2.3 in./hr and at the indicated temperatures. The entering gas was saturated with water at 50°C.

Tritium Tritium Shale Oil in in Temp. Volume Tritium in Oil Water Off-gas (°c) (ml) (UiCi/ml) CuCi) (li-Ci) (M-Ci)

130 0 - - 29.9 0.9

2 45 0 - - 24.0 0.5

350 59 0.064 3.8 8.3 1.0

485 14-9 0.059 8.8 3.3 2.0 at 130, 2^-5, 350, and 485°C. About 30 g of -water vapor was introduced in each of the four passes. About 55% of the tritium was removed from the shale in the first two passes prior to any oil production. About 28$ of the oil was produced at 350°C, and the remainder at 485°C.

Additional testa are being made to better define tritium behavior in the oil -recovery system and to study possible methods of decontaminating the shale chimney before oil recovery.

11.2 Contamination by Fission Products D. J". Crouse and W. D. Arnold Tritium concentration can be minimized by using fission explosives designed to yield low amounts of tritium. In this case, knowledge of the process behavior of the fission products becomes more important. We have studied fission product behavior only briefly by heating shale oil with underground test shot debris.1 The results indicated that the most likely pathway for contamination of the oil is by physical entrainment of extremely fine radioactive solids. When contaminated shale oil was distilled, the fission products concentrated in the residuum and did not contaminate the overhead fractions. We plan to study fission product behavior during retorting, as well as methods for separating the radionuclides to allow use of the products obtainable from the residuum. Other potential problems needing study include decontamination of retort off- gases, disposal of contaminated water that is produced with the oil, and a general evaluation of radioactive hazards to operating personnel at the facility. - 397-

11o 3 References for Section 11 1. W. D. Arnold and D. J. Crouse, "Radioactive Contamination of Oil Produced from Nuclear-Broken Shale," Proceedings of the ANS-AEC Symposium on Engineering with Nuclear Explosives, Las Vegas, Nevada, January 14-16, 1970, C0NF-700101, Vol. 2, pp. 1597-1611. -398-

12. CHEMISTRY OF RADIOIODINE

The purpose of this program is to obtain a better understanding of the complex chemistry of iodine in chemical processing systems as a basis for providing highly effective containment of radioiodine in nuclear fuel processing plants. Information from the program is particularly applicable to the processing of short-cooled high-burnup reactor fuels, especially in the dissolution, solvent extraction, and off-gas treatment steps. The studies include identification of the different iodine species present in these systems under various conditions, and examira tion of the types and kinetics of reactions that these species undergo to form other iodine species. The most recent work has been concerned primarily with development of a mercuric nitrate scrubbing process for removing radioiodine from off- gases and investigation of chemical reactions that affect the efficiency of this process.

12.1 Mercuric Nitrate Scrubbing Process D. J. Crouse, J. M. Schmitt, and W. B. Howerton In the process flowsheet for the removal of iodine compounds in gas streams by mercuric nitrate scrubbing, the high-acid mercuric nitrate scrub solution is circulated in closed cycle from the bottom to the top of the scrub column. To prevent excessive accumulation of iodine, part (or all) of the solution is passed through an iodine take-off system in which the solution is evaporated to concentrate and oxidize the iodine, causing most of it to precipitate as mercuric iodate. The supernatant solution and the distillate are combined and returned to the scrub system. The radioactive mercuric iodate conceivably could be stored in this form, or -399- converted to sodium iodate "by reaction with caustic, to allow complete recycle of the mercury.

Most of the iodine in the head-end off-gas stream will be in the form of Ig, which is reedily trapped, but a high-efficiency process must also be capable of handling the alkyl iodides that are inevitably present in the off-gases. Early studies of the process1 showed that the efficiency in scrubbing methyl iodide is limited by its slow reaction with mercury in the aqueous phase to form the stable mercuric iodide complex. The reaction rate is directly proportional to the mercury concentration and increases by a factor of about 100 as the nitric acid concentration is increased from

0.2 M to 12 M. Other alkyl iodides (butyl, octyl, dodecyl) behave similarly and can be scrubbed with an efficiency that is about equivalent to or higher than observed for methyl iodide. Aromatic iodides are much more stable and penetrate the system. These, however, should be present at very low concentrations, if at all.

12.1.1 Methyl Iodide Removal in Packed and Bubble-Cap Columns

Methyl iodide scrubbing data were obtained over a range of gas flows in a 1.85-in.-diam column packed with l/4-in. Berl saddles, and in a 1-in.- diam column packed with protruded stainless steel packing (0.l6-in. size).

The decontamination efficiency in the packed columns, measured as the decontamination factor (DF), was approximately constant over the total length of column. As expected, the scrubbing efficiency was much higher with 0.4 M Hg(N03)2—12 M HNO^ than with 0.4 M HgtNO^--10 M HN03 solution (Table 12.1). In tests with the former solution using Berl saddle packing, the DF per foot of packing decreased from 12.1 to 4.78 as the superficial

gas velocity was increased from 19 to 38 fpm. Much higher efficiencies Table 12.1. Scrubbing Methyl Iodide from Air in Packed Columns -with Mercuric Nitrate—Nitric Acid Solutions at 25°Ca

Scrub Solution DF per Hg2+ HNO Gas Flow Scrub Flow Foot of (M) (M) Column (liters/min)(fpm) (liter/min) Iodine DF Packing

0.4 10 b 7-5 14 o.l 4370 5-35

10 19 0.1 664 3.67 517 3.^9

10 19 0.3 2047 4.59

10 19 0.5 1954 4.55

15 29 0.1 92.9 2.47

18 34 0.1 50 2.19

0.4 12 b 10 19 0.3 2.6 x 105 12.1

15 29 0.3 1.7 X 10 7.02

20 38 0.3 3110 5.00

20 38 0.1 2490 4.78 Table 12.1. (continued)

Scrub Solution Dp per Hg + HNO Gas Flow Scrub Flow Foot of (M) (M) Column (liters/min) (fpm) (liter/min) Iodine DF Packing

4 0.4 12 c 6 39 0.1 2.7 x 10 30.0

8 52 0.1 875^ 20.6

10 65 0.1 2475 13-5 2753 14.0

12 78 0.1 959 9.86

£L

Scrub solution circulated in closed cycle; countercurrent flow.

1.85-in.-diam column packed with 5 ft of l/4-in. Berl saddles. Q 1-in.-diam column packed with 3 ft of Pro-Pak stainless steel packing. -402- were obtained using the protruded stainless steel packing. For a given decontamination performance, the gas throughput capacity of the column was higher by a factor of about 3 "with this packing than when using the l/4-in. ceramic saddles. Data for scrubbing methyl iodide from air in an eight-stage, 1-in.- diam simulated bubble-cap column are shown in Fig. 12.1. The depth of liquid per stage was 4 in., and the gas was dispersed through holes in a glass tube at a depth of about 3 in. With 0.4 M Hg(N03)2»-12 M HNO^ solution, the DF per stage was about 4.4 at a gas flow rate of 3'3 liters/ min (superficial gas velocity of 21 fpm), and about 3'25 at 5*3 liters/ min; the former value corresponds to a total DF of 1.4 x 10 across the eight stages. The liquid inventory per stage during gas flow was about 35 nil* The methyl iodide concentration of the air feed in these runs was approximately 0.14 mg/liter.

12.1.2 Decomposition Rates of Various Organic Iodides Prior studies with methyl iodide showed that the decontamination efficiency for one stage in a "perfectly mixed" system can be predicted if the values are known for the liquid/gas distribution coefficient and the rate constant for reaction of the methyl iodide with mercury in the aqueous phase. The expression used for zero aqueous flow, an assumed first- order reaction rate, is:

DF - 1 = ^ K^k , (1) where

VT = liquid holdup, liters, -403-

ORNL DWG 73-808R1

GAS VELOCITY (fpm)

Fig. 12.1. Scrubbing Methyl Iodide from Air with Hg(N03 )2— HN03

Solutions at 25°C in a Simulated Bubble-Cap Column. -404-

F = gas flow rate, liters/min, K^ = liquid/gas distribution coefficient, (jr k = first-order reaction rate constant, min \

We have started a series of measurements with a number of different organic

iodides to determine the value of the product KG^ x k, which is a measure of the relative sorption of each compound when using a specific scrub solution. The measurements are made by dispersing air containing the organic iodide through a fine glass frit into the scrub solution (which is vigorously agitated in a baffled mixer) and then determining the iodine DF. Table 12.2 shows some values determined for methyl, n-butyl, n-octyl, and phenyl iodides. The difficulty in trapping phenyl iodide as compared with trapping the alkyl iodides is evident.

12.1.3 Process Demonstrations

A series of eleven 4-hr runs was made in the eight-stage simulated bubble-cap column to study removal of methyl Iodide and I from air streams and the effects of gas-phase contaminants on the efficiency. The iodine concentration in the feed stream was relatively high (0.5 to 0.6 mg/liter); this corresponds to the concentration that might be expected in processing the head-end off-gases in a plant designed to minimize in-leakage of the cell atmosphere into the vessel off-gas system. All of the runs were made with an air flow of 2.5 liters/min and a scrub solution [0.4 M Hg(N0 — 11-12 M HNO flow rate of 4 ml/min. With such a high iodine input rate ~ 3 and low scrub solution flow rate, considerable iodine accumulates in the solution in a single pass through the column. Under such conditions, it is preferable to pass the total recycle stream through the iodine take- off system, rather than only & small fraction of the stream as was done in Table 12.2. Relative Measure of Sorption of Different Organic Iodides in Mercuric Nitrate Solutions

Constituents of Scrub Solution 2+ Liquid Gas Flow Rate Hg HNO- Volume Species 3 (liter) (liter/min) DF (M) (M) &

Methyl iodide 0.05 10 0.4 0.2 6020 r 3010 1.3 X 10 65OO

Methyl iodide 0.05 10 0.4 0.4 3700 3700 1700 1700

4 Methyl iodide 0.4 10 0.4 0.3 > 105 > 7 x 10

Methyl iodide 0.4 10 0.4 0.4 9600 9600

Butyl iodide 0.1 1 0.4 0.3 1050 790

Butyl iodide 0.2 1 0.4 0.3 2550 1900 4 Butyl iodide 0.2 8 0.4 0.15 > 105 > 4 x 10

5 Butyl iodide 0.4 10 0.4 0.15 > 10 > 4 x lo14" Table 12.759. (continued)

Constituents of Scrub Solution Liquid Gas Flow Hg2+ HN0 Volume Pate Species (M) (M)^ (liter) (liter/min) DF KGk

4 4 Octyl iodide 0.4 10 0.4 0.3 4.4 x 10 3-3 x 10

Phenyl iodide 0.2 8 0.4 0.3 2-8 1.3 Phenyl iodide 0.4 10 0.4 0.3 10.6 7.2 -1+07-

2 other demonstrations of the process using a packed column contactor. The column effluent was pumped to an evaporator and concentrated four- to fivefold to precipitate Hg(l0^)2, which was then removed on a sintered glass filter. The filtrate and the distillate were recombined in a surge tank and recycled to the top of the column. The effluent gas from the first column was passed through a four-stage bubbler column that initially contained 0.15 M Hg(N0^)2~0.2 M HNO^. The latter column was used to scrub nitric acid vapors from the gas and to scavenge any (this dilute scrub solution would not be effective for CH^l or other organic iodides) that might have bled off from the first column scrub solution. Actually, no iodine was detected in the second-cycle scrub soltuion in any of the runs.

The air feed contained methyl iodide in the first two runs and in the second two runs. However, this change had no effect on the efficiency,

as shown by the close agreement of the IF's in these four runs (Table 12.3).

Addition of N^O^, octane, and dodecane contaminants to the gas stream had no detrimental effect. On the other hand, addition of aromatic organic vapors (xylene, diethylbenzene) decreased the DF by two to three orders

of magnitude.

Following the test with xylene contamination, the column was operated

for 11321 h r with normal air flow, but without iodine or xylene in the air. The I concentration in the effluent air decreased slowly from about

560 counts min"1 liter-1 during the first hour to 3.7 counts min-1 liter"1

during the 12th hour. At this point, another test "(Run 10) was made with

iodine input. The DF was 1.7 x 10^ (the effluent gas had an concen-

tration of 2.7 counts min"1 liter"1), which was about a factor of 5 lower

than had been obtained (Run 5) before the system was contaminated with

xylene. -4o8-

Ta"ble 12.3' Results Obtained by Using Mercuric Nitrate Solutions to Scrub Ig and CH^I froni Air in an Eight-Stage Bubbler Column

Procedure: Air containing indicated concentrations of iodine and contaminants was passed, at the rate of 2.5 liters/min, through eight-stage simulated bubble-cap column countercurrent to 4 ml/min of 0.4 M Hg(^3)2—11-12 M HNO3 at room temperature; effluent air was passed through four- stage bubbler column initially containing 0.15 M Hg (1^)3)2 —0.2 M HNO^ and then through charcoal traps. All of solution effluent3 from first column was passed through iodine take-off system (evaporator, condenser, filter) to remove Hg(l03)2 before being recycled to top of scrub column. Solution in second column was circulated in closed cycle at the rate of 3 ml/min; 8 ml of solution was withdrawn every 2 hr and replaced with fresh solution.

Feed Air CH I Organic Run T2 3 N2O3 Contaminant Overall No. (mg/liter) (mg/liter) (vol i) (mg/liter) npb h. 1 - 0.48 - - 4.5 x 10

4 2 - 0.46 - - 5.7 x 10 4 3 0.6 - 1 - 4.9 x 10 4 4 0.6 - - - 3.6 x 10 4 5 0.6 0.036 0.4 - 8.0 X 10

^ 4 6 0.6 0.038 0.4 0.3° 7.6 x 10

14 7 0.6 o.o4o 0.4 0.8° 8.1 X 10

d 4 8 0.5 o.o4o 0.4 0.035 5.2 x 10 f 9e 0.6 0.04 0.4 0.15 21 4 10 0.6 o.o4 0.4 - 1.7 X 10

11 0.6 o.o4 0.4 -^0.01g 126

Q The column effluent contained ~0.4 g of iodine per liter; the

solution recycled to the top of the column contained "MD.03 g of

iodine per liter. -409-

Table 12.3• (continued)

bNo iodine accumulated in the second column; therefore, it did not

contribute measurably to the decontamination efficiency. c0ctane.

^Dodecane. eAfter run 9, the system was operated for 12 hr without iodine input to eliminate the effects of xylene contamination. fXylene. cr &Diethylbenzene. The results suggest that the mercuric nitrate scrubbing system has an extremely low tolerance for aromatic organic contaminants since they react readily with iodine in a nitric acid system to form organic iodides that 3 are relatively stable. In this scrubbing system, the organic iodides bleed off slowly from the recycle solution and control the overall IF that is achieved. Although they are decomposed by the mercuric nitrate solution, the decomposition is slow.

12.2 Formation of Organic Iodides in Process Systems D. J. Crouse and W. B. Howerton In the course of our iodine studies, we have become more and more impressed with the effect of small concentrations of organic impurities on the experimental results obtained. As described in the preceding section, this is particularly true in nitric acid systems where the combination of organic compounds, I^, and nitric acid readily produces organic iodides which are more difficult to trap than I^. Some of these organic iodides are much more stable to hydrolysis than, for example, methyl iodide, and their generation can significantly reduce the efficiency of the iodine removal process. In certain of these systems, care must be exercised in experimental work to eliminate organic impurities that are normally present in ordinary distilled water and in some laboratory chemicals; otherwise, anomalous results may be obtained. Besides being an experimental problem, these effects are of obvious potential importance in a processing plant where the amounts and types of organic impurities in the processing system will be subject to less-rigid control than in the laboratory. -411-

We have initiated a study of the formation of organic iodides from organic compounds of different types under a range of conditions, and of the behavior of these iodides in various process sytems. The rate of reaction of I with toluene (presumably to form a mixture of o- and p-iodotoluene) in nitric acid was examined by mixing an I^-H^O solution with toluene-saturated nitric acid (toluene/iodine mole ratio of about 4).

The amount of iodinated toluene that was formed was detemined by an extraction method. In 7 M HNO^, about 25% of the iodine reacted in 2 hr at room temperature, and >99$ reacted within 19 hr. In 11 M HNO^j 33% reacted in 10 min and 95% in 45 min. Preliminary examination of the rate of reaction of mercury with the iodotoluene showed that the reaction is 2+ extremely slow; its estimated half-life was 135 min in 0.1 M Hg —8 M

HNO^ solution. 12.3 References for Section 12

1. W. E. Unger et al., LMFBR Fuel Cycle Studies Progress Report for November 1971, No. 33, QRNL-TM-3624 (December 1971).

2. W. E. Unger, et al., LMFBR Fuel Cycle Studies Progress Report for January 1972, No. 35, 0RNL-TM-3724 (February 1972).

3- R. L. Datta, "Halogenation. Direct Iodination of Hydrocarbons by Means of Iodine and Nitric Acid," J. Am. Chem. Soc. 39, 435 (1917). -412-

13. CTR SUPPORTIVE RESEARCH

In current conceptual designs for deuterium-tritium fusion reactors, the plasma is confined within a refractory metal vessel that is sur- rounded by a liquid lithium-bearing blanket maintained at 600 to 1000°C.

Heat is ultimately removed from the blanket through a steam generation

system. Lithium metal is currently the leading candidate for the blanket material, although molten LigBeF^ is also a distinct possibility. Some of the lithium in the blanket is transmuted by neutrons released in

the fusion reaction, yielding tritium that must be recovered and recycled to replace tritium consumed in the fusion reaction. Knowledge of the thermodynamics of the Li-LiT-Tg system at low LiT concentrations (and correspondingly low pressures) is necessary1 if reasonable evaluations of efficient separation and recycle processes are to be made. Such

studies are in progress. Permeation of hydrogen isotopes through hot,

clean metals and through metals coated with an oxide film is also being

studied. Research designed to provide information pertinent to the

evaluation of LigBeF^ as a blanket-coolant is also in progress. Studies

in this area include determining the effects of strong magnetic fields on

aqueous analogs of molten salts, and measuring the solubilities of hydrogen isotopes in molten LigBeF^.

13.1 Equilibria in the Hydrogen Isotope—Lithium Systems

F. J. Smith and 0. K. Tallent 2 Equilibrium data for hydrogen isotope-lithium systems are sparse.

Recent studies of the Li-LiH-Hg system have been made by Veleckis and

Van Deventer, whereas the Li-LiD-D^ system has been investigated by -4J.3-

Goodall and McCracken. However, most of the experimental measurements have "been made in systems in which the hydrogen or deuterium concentration

-2 in lithium was greater than about 10 mole fraction (1500 to 3000 wt ppm).

No useful data are available for the Li-LiT-T^, system. Consequently, we have initiated an experimental program to measure the equilibrium pressures of hydrogen isotopes when they are present in lithium at low concentrations. Our primary interest is in the Li-LiT-T^ system, but most of the initial work has been done with hydrogen to allow us to become facile with the experimental procedures. Equilibrium measurements are being made by both a total pressure (Sieverts') and a partial pressure method.

13.3.. 1 Total Pressure Method

The modified Sieverts1 apparatus being used to study the Li-LiH-Hg system is shown schematically in Fig. 13.I. The pumping system consists of a mechanical vacuum pump coupled to an oil diffusion pump. The pressure measurement instruments include a mercury manometer, a McLeod -3 -6 gage (0-5 torr), a Hastings gage (10-1 torr), and an ion gage (10

10 torr). Hydrogen is purified by passage through a palladium diffuser.

The volume of each portion of the apparatus, except the pumping system, must be calibrated separately before making an experiment. This includes the internal volumes of the pressure measuring instruments. The mercury manometer is not used for experiments in which the equilibrium pressures are less than 5 torr. The volume of the envelope that contains the sample is purposely kept small compared with the total volume of the system.

The temperature of the sample is measured with a Chromel-Alumel thermo*- couple located in a well in the quartz envelope, and the temperature is

-415- recorded by a recording potentiometer. The temperature of the remainder of the system is maintained at 28 ± 1°C.

The procedure used in a typical experiment is as follows: A weighed

amount (~1 g) of pure lithium is sealed under vacuum in a small type

304 stainless steel capsule which is placed in the quartz envelope. The

envelope is then connected to the rest of the system, and its volume is

calibrated. The system is evacuated at ambient temperature, heated to a

preselected temperature, and evacuated further by pumping until the

pressure is reduced, to about 10 torr. The quartz envelope containing

the capsule is subsequently valved off, and hydrogen is admitted to the

rest of the system until a predetermined pressure is obtained. Finally,

the valve to the quartz envelope is opened and the change in pressure with

time is recorded until apparent equilibrium is reached. Additional

equilibrium points at a given temperature are obtained by adding more

hydrogen and repeating the above procedure. The concentration of hydrogen

in the lithium at each equilibrium point is calculated using the initial

weight of lithium, the total amount of hydrogen added, and the amount of

hydrogen remaining in the system after equilibration.

Data for two experiments — one in which the initial hydrogen pressure

was high (~1 atm), and another in which this pressure was low (~0.0034

atm) — are given in Figs. 13• 2 and 13*3. The temperature in each case

was 888 ± 2°C. The data given in Fig. 13.2 are typical of those obtained

when the equilibrium hydrogen pressure was greater than 0.5 torr. As

can be seen, the hydrogen pressure decreases systematically until a

constant equilibrium pressure is attained. The data given in Fig. 13.3

are plotted semilogarithmically for convenience in showing the total -4l6-

ORNL DWG. 73-9996RI

Fig. 13.2. Rate of Diffusion of Hydrogen Gas into a 304 Stainless Steel Capsule Containing Molten Li-LiH at 888°C. Surface area of capsule, 13»19 cm3; thickness of capsule wall, 0.051 cm; initial hydrogen pressure inside capsule, 0.191 atm; initial hydrogen external pressure, 1.00 atm; volume of external hydrogen, 257 206 cc. -4-17-

ORNL DWG 73-9997RI 10 -I 1 1 1 i

O DATA OBTAINED WITH H2 IN SYSTEM • DATA OBTAINED WITH NO H2 IN SYSTEM 1.00

a. i«-

o 1.0 . oo, UJ UJ tr - z UJ X- 0.01

0.01 100 300 400 500 600 TIME (min)

Fig. 13-3- Rate of Diffusion of Efy-drogen Gas into a 304 Stainless

Steel Capsule Containing Molten Li-LiH. Surface area of capsule,

13.19 cm2; thickness of capsule wall, 0.051 cm; volume of external hydrogen gas, 161.15 cc; initial hydrogen pressure, 2.6 torr; temperature, 888 ± 2°C. -4l8- pressure change. The upper curve is typical of those obtained in experiments in -which the equilibrium hydrogen pressure was less than about

0.2 torr. The pressure of the system initially declines, but ultimately passes through a minimum. This behavior will be discussed in more detail later in this section.

The experimental data are considered in terms of the equilibrium

1/2 H2(g) + Li(a) = LiH(d) , (1) in which (g) denotes gas and (d) denotes dissolved in the liquid Li-LiH phase. The equilibrium constant for reaction (l) can be expressed as

IJ 7 TC ^V) LiH(d) LiH(d) 1/2 ~ ~ 1/2 ' ^(a) Pn2(g) aLi(d) Pn2(g) in which a, N, p, and y denote activity, mole fraction, pressure (torr), and activity coefficient, respectively. As N_ approaches zero, LlH(d) 'LiH^ Li(d) ^ecome Poetically constant. The so-called Sieverts1 constant,

pl/2

L1H(d) is therefore a special limiting case of the inverse of Eq. (2). In our first series of experiments, equilibrium data were obtained for the Li-LiH-Hg system at 888 ± 2°C. These data are given in Fig. 13.4 1/2 as plots of p ' vs L . The equilibrium pressures shown in the main 2 LlH(d) part of Fig. 13.4 were greater than about 0.5 torr. These values were -419-

ORNL DWG 73-9998

0.0 0.10 0.20 0.30 LiH CONC.lmols %) SEE INSERT

10 20 30 LiH CONCENTRATION IN LiH- LI SOLUTION i mola %)

Fig. 13.4. Isothermal Pressure-Composition Data for the Li-LiH-H2

System at 888 ± 2°C. -420- obtained graphically from plots similar to that shown in Fig. 13.2.

The-Sieverts1 constant calculated from the limiting slope of the plot in Fig. 13.4 was 90• 7, which is in excellent agreement with the value of

9O.3 extracted from the data of Veleckis and Van Deventer.

As shown in the insert in Fig. 13.4, apparent equilibrium points obtained at pressures less than about 0.2 torr were somewhat higher than

expected from the Sieverts' constant derived from the data obtained at higher pressures. These apparent equilibrium pressures were the minimum values obtained from plots similar to that shown in Fig. 13.3. The deviation from Sieverts' behavior illustrated in the insert in Fig. 13.4

cannot be explained reasonably by thermodynamic considerations, but has 4 5 been observed experimentally by other workers. This behavior at very low pressures may be due to experimental errors such as desorption of gas from the metal capsule because the capsule was not thoroughly

evacuated initially, desorption of gas from the walls of the system, inert impurities in the hydrogen, and diffusion of gases into the quartz

envelope. Errors in pressure that could result from these difficulties

are not significant at pressures above about 1 torr.

When the equilibrium hydrogen pressure was very low, the change in the pressure of the system with time was similar to that shown in the

upper curve in Fig. 13.3* The cause of this behavior is not known, but

it was noted that a steady increase in pressure of the system with time

occurred after evacuating the system to ~10 J torr. This pressure

change at 888°C is shown as the lower curve in Fig. 13.3. The change is

almost linear with time, and it is approximately parallel to that obtained

in attempting to determine the equilibrium hydrogen pressure. If it is -4-17- assumed that the increase in pressure of the system represented by the lower curve in Fig. 13.3 is due to outgassing of the capsule (or a similar effect), and that the equilibrium between gaseous hydrogen and liquid lithium is established when the upper pressure-time curve passes through its minimum, the equilibrium hydrogen pressure can be estimated as simply the difference in pressures represented by the two curves. The Sieverts1 constants estimated on this basis at 888°C are in reasonable accord with the value obtained by extrapolating the higher pressure data to

Px/ = 0. 2 In each experiment, hydrogen had to permeate the walls of the 304 stainless steel capsule before it could dissolve in the lithium and form

LiH. The permeation constant can thus be estimated from the rate of change of hydrogen pressure with time as the system strives to attain equilibrium, if it is assumed that the reaction to form dissolved LiH is very rapid compared with the rate of diffusion of hydrogen through the stainless steel. The permeation constant, k, is estimated using the following equation:7

FL = 2 2 ^ A(P^ - PV ) ' ^ where

F = the rate of flow of hydrogen into the stainless steel capsule, cc(STP)/sec,

L = the thickness of the stainless steel wall, cm, 2 A = the surface area of the capsule, cm ,

P^ = the pressure of the hydrogen outside the capsule, atm,

P = the pressure of the hydrogen inside the capsule, atm. u The permeation constant, k, is the quantity of hydrogen in cc(STP) 2 passing per second through a -wall of l~cm area and 1-cm thickness, •when a pressure difference of 1 atm exists across the wall. Data showing the change in the amount of hydrogen present in the gas as a function of time during equilibration in two experiments at 888°G are given in Figs. 13.2 and 13-3• It should be recalled that the system had been brought to equilibrium before these data were obtained. More hydrogen was then admitted to establish a new external pressure, P^, in the known volume of the system. Values of F were estimated from the initial slopes, assuming Pg to be the equilibrium dissociation pressure determined just prior to the admittance of hydrogen. Data for the experiment given in Fig. I3.2 were obtained with P = 1 atm and P2 = 0.19 atm, whereas in the other experiment (Fig. 13.3) P1 = 0.0034 atm and Pg was assumed to be zero. The values of F estimated from the initial slopes (and other required information) of Figs. 13.2 and 13.3 are 2.88 xlO J and 1.9 x 10 cc(STP)/sec, respectively. The corresponding values of k are 2.0 x 10 -5 2 -5 2 and 1.1 x 10 cm /sec, compared with the value of 4.5 x 10 cm /sej 7 estimated from data reported by Redhead et al.

13.1.2 Partial Pressure Method

In the technique using total pressure measurements described above, the calculated concentration of hydrogen isotope in the lithium is dependent on an accurate material balance for the isotope. Hydrogen

"sinks" in the system become more important and are more difficult-, lo correct as the pressure of the hydrogen isotope is decreased below ih jut -3 10 torr. Therefore, an alternative method is being tested for the study of the Li-LiT-Tg system at low equilibrium pressures. In this technique, small amounts of tritium gas are mixed with an inert gas

(helium or argon), and the resulting gas mixture is equilibrated with molten lithium. Our initial experiments were conducted with an apparatus

similar to that depicted in Fig. 13.1. The lithium was encapsulated in

type 304 stainless steel and analyses of the gas mixture before and after

extended exposure to the capsule were expected to provide the information

necessary to calculate the amount of tritium dissolved in the lithium

for a given partial pressure of tritium in the gas. However, at tempera-

tures below 100°C and tritium partial pressures of about 10"^ torr the time

required to reach equilibrium is extremely long because of the very low

external tritium pressure. After some initial experimentation, it was

concluded that meaningful equilibrium data could not be obtained using

this apparatus.

We are currently conducting experiments with the apparatus shown

in Fig. 13.5. The lithium is contained in a molybdenum crucible which

'is open at the top. The circulating tritium-containing gas is contin-

uously bubbled through the liquid lithium. Both the gas and lithium

are sampled at the temperature of uhe experiment and are analyzed for

tritium. Preliminary results from these experiments are encouraging in that

most of the tritium admitted to the system transfers from the gas phase to

the lithium and the apparent equilibrium partial pressure of tritium in

the gas varies with temperature in the expected manner.

13.2 Phase Equilibria in the Li-K System

F. J. Smith and J. F. Land

Some conceptual designs for thermonuclear reactors use a potassium ORNL DWG 73-9999

VACUUM

Fig. 13.5. Tritium Sorption Apparatus. coolant in a Rankine cycle that is interposed between the lithium blanket 8 9 and a steam generation system. It has also been suggested^ that lithium be added to the potassium in concentrations as high as k at. to permit tritium to be recovered from the potassium as lithium tritide in a cold trap at a temperature of about 82°C. Adequate solubility of lithium in potassium is crucial to the use of this method for removing tritium from liquid potassium. Preliminary considerations indicated that this scheme was impractical since data from the literature1^'and regular solution 12 theory, using the solubility parameters of Hildebrand and Scott, strongly suggested that lithium and potassium were immiscible at low temperatures. This expectation was borne out in a study of the Li-K system at temperatures between 80 and 550°C.

The apparatus and experimenta-jo l procedure were basically the same as those described previously . Analyses of filtered samples taken at different depths in the crucible indicated the presence of two liquid phases above l80°C, only one liquid phase between 63 and l80°C, and no detectable liquid below 63°C. The measured mutual solubility data for the Li-K system are shown in Fig. 13.6, in which the logarithm of the solubility of lithium or potassium in the other component is plotted against the reciprocal of the absolute temperature. The solubility of lithium in potassium, and of potassium in lithium, can be represented by:

log S(L. .n K)(wt ppm) = 6.09 - 1837/T(°K) (5)

log S(K in Li)(wt ppm) = 5.50 - 1362/T(°K) . (6)

As indicated in Fig. 13.6, a potassium-rich liquid phase that is in equilibrium with practically pure solid lithium is formed at temperatures -426-

ORNL DWG 73-3953 TEMP (°C) 100,000 600 500 400 300

~ 10,000

2o SOLID Li , SOLID K

CO w n is 1,000 H Zo a. z OZ y- Z S2 z < 100 SgE °z Z hi Oo

<1 St 10 OL _i I I < o

I I I I I I I • I ' ' > > > ' i» « ' I 1.0 1.5 2.0 3.0 lO'/T (*K)

Fig. 13.6. Composition of the Two Immiscible Phases in the Li-K System. -427- between 63 and l8o°C. At temperatures above l8o°C, two liquid phases which have limited miscibility are present. No depression of the melting points within ±2°C for the Li-K mixtures was observed by thermal analysis, which is in agreement with the investigation of Boehm and Klemm.^1 It is impossible to extrapolate our data and accurately predict the consulate temperature for the lithium-potassium system. However, the large miscibility gap below 500°C indicates that it is probably greater than 1000°C.

The observed solubility of lithium in potassium is only about 9 wfc ppm at 82°C, the temperature proposed for operation of the cold trap. This solubility is markedly lower than the 73^0 wt ppm required for a solution that is 4 at. % Li. Therefore, the results of this investigation show that the proposal^ involving use of K-Li %) as the coolant in an intermediate Rankine cycle is probably not feasible.

All solubility measurements were made with the condensed phases in equilibrium with their own vapors. The observed apparent partial pressures are shown in Fig. 13.7. Within analytical uncertainty, they are 17 the same as the vapor pressures of lithium and potassium. This is to be expected, however, since one liquid phase was almost pure lithium and the other almost pure potassium. 13.3 Permeation of Deuterium Through Structural Metals at Low Pressures

R. A. Strehlow and H. C. Savage

The permeation of hydrogen through metals at elevated temperatures has been an active research area for many decades. Permeation of hydrogen through hot platinum was studied to pressures as low as a few torr by 18 Richardson et al. in 1904. They observed a square-root pressure -428-

ORNL DWG 73-3952

TEMP. (*C)

Fig. 13«7« Apparent Partial Pressures for the Two-Phase Liquid-

Liquid Immiscible Li-K System. The lines are the reported vapor pressures of pure lithium and potassium. •In- dependence and considered it evidence for the dissociation of hydrogen into atoms when dissolved in platinum. This hypothesis led to the following equation for hydrogen permeation:

F = KT1/2 e-W*) (P^2 - p£'/2)/a , (7) where

F = the permeation rate per unit area of surface,

P^ and Pg = the upstream and downstream gas pressures,

T = the temperature,

d = the thickness of the metal,

K and b = the constants for the gas-metal system.

Square-root dependence [i.e., adherence to Eq. (7)] was also found by 19 20-24 24 25 Sieverts and others; however, Smithells and Ransley ' pointed out that no data, including their own, obeyed Eq. (7) when the pressures were below about 25 torr. This assertion led to the prevailing general 26 view that the pressure dependence departs from the Richardson equation at low pressures. The present work was undertaken to obtain data over a very wide range of pressures to test the validity of Eq. (7). The data were also needed so that tritium releases from fusion and fission nuclear reactors could be estimated with more accuracy than was possible previously.

We used a steady-state method to study the permeation- 4o f deuterium through several metals from pressures ranging from 9 x 10 to 750 torr.

The apparatus contained a specimen tube (typically 25 cm long x 1.27 cm

I.D., with a wall thickness of O.89 to 1.6 mm) of the metal to be tested.

The tube was assembled inside a 5-cm-I.D. steel jacket in an electrically -MO- heated furnace in such a way that deuterium or an inert gas—deuterium mixture could flow along the outside of the specimen. The total pressure of this permeating gas, measured "by 6-in-diam absolute pressure gauges of 0-20 and 0-800 torr ranges, was maintained at the desired level in the range 0-750 torr. Inside the specimen tube, argon admitted at a rate of 15 cc(STP)/min swept the permeated Dg to a mechanical vacuum pump, and a portion of this flow was then admitted to a quadrupole mass spectro- -6 meter to produce a pressure of 3 x 10~ torr. Deuterium at mass k was -12 readily measurable down to a pressure of 10~ torr; this pressure corresponded to a permeation flow rate of about 10 cc(STP)/hr. Deuterium was used to avoid the protium background in th-9e mass spectrometer that became appreciable at a pressure of about 10 torr. Wo significant amounts of mass 3 (HD) were observed.

Pure (99.982$>) Dg, argon containing k.2and helium containing wa0.08s prepare$ Dg werde i usen and evacuateas the permeatind heliumg tan gasesk (volume. The, He—0.08 2800 cm$o )D bg ymixtur admittine g

•3

75 cm (STP) of D u and pressurizing to 500 psig with mass-spectrometer- grade helium. Dissolution of D in the metal of the tank can change the -r ^ concentration of such dilute solutions; for the above conditions, a loss of about 1°}o per day is estimated. Accordingly, experiments using this dilute mixture were completed within a few days after preparing the -k mixture, and it was found that Dg partial pressures as low as 5 x 10 torr in the permeating gas were readily maintained in the steady-scate determinations. Several hours were required to reach steady-state conditions in many of the experiments. -431-

The results from experiments with metals that were free of thick oxide films are summarized in Table 13.1. At a given temperature, log- 1/2 log plots of permeation rate vs P^' were linear with slopes of about

0.5 in experiments with stainless steel, Hastelloy N, and the chromium-

plated nickel. With these metals, the values of Pg were negligible

compared with P ; thus, the data are seen to be consistent with the behavior expected from Eq.. (7). Data were also obtained with nickel

over a very wide range of pressures at 623°C (Table 13.1). At very low

pressures, the values of P^ were significant compared with the values

of P^. Consequently, thesl/2 e dat1/2a are shown as a plot of the measured permeation rates vs (P^ - P^ ) in Fig. 13.8. The data from experi-

ments with nickel yield a straight line over the entire pressure range,

and it is significant that the line extrapolates to the origin. This

behavior is precisely that expected from Eq. (7) at a given temperature.

Our data (obtained with three metals at pressures less than about 30

torr) show no systematic departure from the square-root dependence on

pressure by the permeation rate down to pressure24 25 s as low as 10 torr. Prior to the work of Smithells and Ransley ' numerous explanations

for the variance from the square-root dependence of permeation on pressure had been suggested in the literature. Borelius and Lindbloom, 22

for example, had proposed a threshold pressure below which no diffusion 24,25 took place. Smithells and Ransley attributed the departure to a

sorption and surface coverage effect. They assumed that the permeation

flow was not proportional to the area of specimen but to an "effective

area" defined as that part of the area covered by hydrogen atoms. We believe this assumption to be in error because it assume» s that both Table 13.I. Pressure Dependence of Deuterium Permeation Through Unoxidized Metals

Value of n Permeation Rate in the Pressure No. of Temp. at 1 torra Pressure Dependence8, Range Data n Metal (°c) [cc(STP)-mm/hr-cm ] Term, p (torr) Points

2 Nickel 623 (1.8 ± 0.2) x 10" 0.53 ± 0,02 0.0009-750 33

304L SS 697 (2.6 ± 0.5) x 10"k 0.54 ± 0.05 0.03-31 14

Hastelloy N 605 (3-4 ± 0.2) x 10~3 0.52 ± 0.u2 0.004-31 16

3 Hastelloy N 688 (8.9 ± 0.4) x 10" 0.4-9 db 0.02 0.05-31 8

2 Nickel—0.025 616 (1.0 ± 0.1) x 10" 0.51 ± 0.03 1-750 5

mm Cr plate i. ji . Determination of exponent in the term p with a 95% confidence level as calculated from a least-squares program. n is the exponent on the pressure terms that give the best correction of data as per Eq. (7). ORNL-DWG 73-6324 V^i-V^'torr) 2 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.50 o 0.020

/ • 0.45 D? -Ni (62!3°C ) 0.018 o /

0.40 0.016

CM CM E — Ixl S 1 oc of cr UJ 0.20 • 0.008 Ua.J CL / ° 3 UrJ> UoJ o

0.10 0.004

0°• /

* 4 0.05 0.002 / 0 5 10 15 20 25 30 35

Fig. 13'8« Deuterium Permeation Through Nickel at 623°C. Dual scales are used to encompass the extended pressure range from 0.0009 to

750 torr. -k3b- equilibrium and nonequili"brium conditions apply simultaneously. If the equilibrium solubility "was not reached and the diffusion process was limited by surface coverage, permeation might then be proportional to the coverage instead of to the solubility with its dependence on the square root of pressure.1 9 Since the surface coverage also varies with the 27 square root of pressure, no change in functional relation at low pressures need be expected. There will, of course, be a (very low) pressure range where the dissolved concentration of hydrogen, or the surface coverage, becomes limited by the arrival rate and sticking fraction of hydrogen molecules at the surface. Under these conditions, one might indeed expect a first-power relation between pressure and permeation rate (modified appropriately to account for the dissociation of hydrogen). The only theoretical treatment of the pressure dependence of permeation found in the literature since the work of Smithells and Ransley 28 in 1935 is that of Ash and Barrer, w who developed a set of relation

strips by considering the rates of all forward and reverse reactions for the various transport processes involved in hydrogen permeation. In their treatment, however, the equilibrium solubility was introduced in such a way that the conclusions based upon the model are probably invalid.

Specifically, the third of six rate equations given,

R3 = k3 6 [1 - (u/us)] , (8) requires that the rate of transport into the metal be zero at saturation, which certainly cannot occur for anything less than full occupancy of

solution sites. In this equation, R„ is the rate of transport of -435- hydrogen from a surface, where 9 is the fractional coverage "by the hydrogen atoms; u is the dissolved quantity of hydrogen "just inside the metal;" ug is the saturation solubility; and k^ is a constant.

Since no theoretical reason presently exists for the permeation rate to vary with pressure in any fashion other than the half-power dependence we observed down to below 10 torr, we conclude that Eq. (7) for hydrogen permeation through metals may be used without major qualification.

13.4 Permeation of Deuterium Through Metals Coated with Oxide Films

R. A. Strehlow and H. C. Savage

Tritium permeation rates from a fusion reactor through the large metal areas of an associated steam generator probably need to be limited -8 2 to less than 10~ cc(STP)/hr-cm . This assumes that 5 nCi/liter is the maximum permissible concentration of tritium in liquid effluent from a 7 2 reactor, the steam generator has an area of about 8 x 10 cm , a dilution 5 (cooling) flow of 6.4 x 10 gpm is used, and containment of the steam is not practicable. Since at least one surface of the steam generator will have an oxide film, we attempted to determine whether an oxide barrier could be applied or formed on a metal to provide tritium containment.

We expected permeation rates that varied with the first power of pressure instead of the half power, as was found with clean metals (Sect. 13.3)*

The experimental apparatus and method described in Sect. 13.3 were used to study the effect of oxide films on permeation.

Oxidized materials present an experimental difficulty in that reduction of the oxides by permeating gas can occur during an experimental determination. Accordingly, we used a permeating gas mixture that -436- was of a moderate and controlled oxidation potential. In order to obtain steady-state permeation values as a function of pressure (to test for first-power pressure dependence), we used a mixture of deuterium and deuterium oxide, in the ratio of 38 to 1, that had been made by bubbling D2 through D^O at 1 atm and 23°C. We found the permeation rate for the oxidized materials to be dependent on pressure to a greater degree than the square root, and to be proportional to the first power of pressure in four cases (Table 13.2). The oxides formed on stainless steel and on Incoloy 800 were found to be about 0.5 m|i thick and they were analyzed by Auger and electron spectroscopy to be MnCrgO^. The materials, Croloy T9, Croloy T22, and a duplex tube of Ni—Incoloy 800 were studied most recently. It appears that Croloy T9 with approximately 9% Cr formed an oxide with significantly lower permeability than did the T-22 with 2.25% Cr. The duplex tube was studied in order to determine whether an oxide such as the highly- effective MnCrgO^ on Incoloy 800 could be formed on the inside of a tube specimen. The inner part of the tube was Incoloy 800, and the outer part was nickel. We used deuterium alone as a permeating gas after the inner surface was oxidized.

As seen in Table 13.2, we found substantially the same behavior as with the Incoloy 800 tube studied earlier using Dg-DgO mixtures for permeation. The permeation rate at 1 torr was higher by a factor of 6, but this was a rate through only a single film. The direction of flow (whether the deuterium permeates the metal or the oxide first) does not

present any significant differences for this case. 29 Based on a permeation model derived earlier, significant differences -437-

Table 13.2. Pressure Dependence of Deuterium

Permeation Rate Temp. at 1 torr, 1-mm-thick c< 2 Material (°C) [cc(STP)/hr-cm ]

Type 304L-SS (oxidized) 785 (4.5 ± 0.5) X

Sanicro 31 672 (1.6 ± 0.1) X 10"3

Incoloy 800 (initially) 649 (1.6 ± 0.2) X 10 1

b Incoloy 800 (after 649 (4.7 ± 0.4) X 10" ] extended oxidation with D2--2.6f0 D20) -4: Incoloy (exposed to 53i (6.3 ± 0.4) x 10 H20 at 538°C, 3500 psig, 10^ hr) c -5 Croloy 573 (6.6 ± 0.5) x 10 (oxidized D2—2.6% DgO, 72 hr) -6 Croloy T9 (same specimen 570 (5.7 ± 0.4) x 10 after 650° oxidation) c, d Croloy T22 285 (1.8 ± o.l) x 10

Croloy T22 580 (8.0 ± 0.1) x 10- 3 I •-0> Croloy T22 .'oxidized 580 (6.0 ± 0.2) x 10' D2—2.6% D20)

Duplex tube, 0.750 in OD x 0.50 in. ID, l/l6 in. Ni outside, l/l6 in. Incoloy 800 inside* (1) Initially 654 (1.6 ± 0.2) x 10"3

5 (2) After 196 hr D2-DgO 745 (2.8 ± 0.4) x 10" oxidants

(3) Tube temp, reduced 541 (1.0 ± 0.3) x 10"5

9* n Determination of exponent n in the term p with a 95% confidence level from a least-squar bASTM-A-213-T9. c Material studied during period March 1 - August 31, 1973. dASTM-A-213-T22. en is the exponent on the pressure terms that give the best correlation of data as per Eq. am Permeation Through Oxide-Coated Metals

e Pressure coating Dependence,a Pressure Range No. of ] (n)e (torr) Observations

0.70 ± 0.05 5 x 10~2-31.U • 9 -3 i 0.63 ± 0.02 3-750 10 -3 i 0.61 ± OoOk 1-750 5 .6 i O.96 ± 0.05 10-750 8

0.51 ± 0.02 1.3-750 9

f5 0.68 ± 0.05 7-750 5

r6 0.85 ± 0.03 10-750

0.67 ± 0.01 1-750 7

r3 0.56 ± 0.01 1-750 12

r3 0.59 ± 0.02 3-750 6

r3 0.28 i O.kk ± 0.55-2.0 3

1.00 ± 0.16 1-18 11 if -5 i 0.85 ± 0.2 1-90 13 t i

uare fit. BLANK PAGE -438- would not have heen anticipated. We do not "believe that a preapplied or preformed film -would necessarily serve to impede tritium flow sufficiently in a fusion reactor. Chemical compatibility and many other factors will be involved in making that determination. This work has shown that oxides can serve to reduce the permeation rate by signii:cant amounts.

For example, our data for Incoloy 800 at 649°C, when extrapolated to torr, yield a permeation rate of 5.8 x 10- 6 cc(STP)-mm/hr-cm 2 . For the same material oxidized to produce a 0.5-pm-thick film, a value of 6.8 x -10 2 10 cc(STP)-mm/hr-cm results, which is a decrease of four orders of magnitude in the permeation rate.

13.5 Chemisorption of Tritium on Graphite R. A. Strehlow The presence of significant amounts of tritium on the graphite from 30 the MSRE^ prompted a study of the chemisorption of tritium on graphites.

The MSRE graphite data suggested that the bonding of the tritium was tenacious even at the high temperatures of the reactor, and that the kinetics of sorption were fast in comparison with the fuel salt circu- lation time.

An apparatus was constructed and operated to permit exposure of graphite at elevated temperatures to a gaseous mixture of tritium in helium. The apparatus is shown schematically in Fig. 13«9- A titanium purification furnace was not used because of the long time required to

equilibrate with the tritium. The copper oxide furnaces for analysis

(F2 and F3 in Fig. 13-9) were operated in conjunction with the apparatus to allow water analyses as well as continuous tritium analyses. The ORNL—DWG 72-7515

Fig- 13.9" Apparatus for Study of Tritium Sorption on Graphite. -440- moisture impurity in the tritium-helium mixture was always less than 0.9 ppm. The exposure of 1-cm samples to the tritium-helium mixture was carried out in the sample furnace (F1 in Fig. 13.9)* which was made of quartz. The tritium concentration was in the range of 1 to 10 ppm by -3 volume (partial pressures ~ 10 torr). Data were obtained on three types of graphite: COB (MSRE-type) , P0C0, and H-327. After the usual exposure conditions of 4 hr at 750°C, the graphite pieces were sampled by removing successive 0.006-in.-thick samples which were analyzed for tritium. The tritium contents of the samples were substantially the same as those found in the examination of MSEE graphite, with values of several times 109 dis min -1 g -1 for the outermost cut, decreasing to values below 10^ dis min"1 g"1 for interior samples. Tritium contents of 6 to 8 x 1010 -1 -1 dis min g were found for some samples of high sruface area.

The gradient of tritium concentration within the graphite reflects qualitatively the microstructure of the specimens (Table 13.3)' Graphite sample H-327, with a ratio of 2 between external and internal tritium, appeared under the microscope to be the most porous, whereas type GGB was highly impregnated. The data in Table 13.3 indicate a square-root dependence of sorption amount with Tg pressure, and a dissociative process is therefore suggested: V2*8(g,*C(fl)-T-C(a). (?)

Rinsing with water and alcohol did not remove significant amounts of tritium; hence the chemical form of the tritium is uncertain. However, the large amount adsorbed at these very high temperatures is surprising.

A sample of OREL lamp-black-based graphite was treated by oxidation in steam to increase its specific surface area. After oxidation, the -Mi-

Table 13.3. Tritium Sorption on Commercial Graphites

Type of Amount of T2 Sorbed Graphite Exposure Conditions (mGi/g)

POCO 8 |j.Ci T2/CC He > 6.7 (surface), 750°C 0-7 (internal)

CGB 8 iiCi T2/CC He > 10 (surface), 750°C 0.3 (internal)

POCO 0.77 (.iCi/cc He 3 (surface), 780 °C 0.4 (internal)

CGB 0.77 nCi/cc He 4 (surface), 780 °C 0,l6 (internal)

H-327 6.4 nCi/cc He 1.8 (surface), 790 °C 0.68 (internal) surface area had increased to 2 m2 /g . A total of 5 x 10 0 tritium atoms per square centimeter (about 50 mCi/g) had sorbed after the usual exposure. The inventory of tritium in the moderator of a fusion reactor cannot yet be estimated with precision because of the uncertain effect of neutron irradiation damage of the graphite, the probable temperature, and the designed tritium pressure. From the data obtained so far in this work, a few milligrams of tritium per kilogram of graphite would be expected. A. topical report3 1 covering the details of this work is in preparation.

13.6 Effects of Strong Magnetic Fields on Aqueous Analogs of Molten Salts

S. Cantor and D. M. Moulton*

Lithium-bearing molten salts are being considered as alternatives •32 to liquid lithium for the blanket of a CTR. The leading candidate is molten LigBeF^. The purpose of this study is to determine the magnitude of the emf induced in such a relatively poor electrically conducting fluid when it crosses a strong magnetic field. An electrical conductor

(e.g., metal) traversing a magnetic field induces an electric field in accordance with Faraday's law:

E = U x B , (10) where the magnitude of the electric field, E, is equal and perpendicular to the vector product of velocity, U, and magnetic flux density, B. In our initial studies, the emf expected to be induced by molten-salt flow in the blanket of a fusion reactor is being simulated by using aqueous

^Present address: 2283 Hearst, Apt. 23, Berkeley, Calif. 94709. solutions flowing at room temperature. The emf is given by

V = Ed , (11)

where V is the emf and d is the diameter of the pipe carrying the liquid.

The apparatus consists basically of an electromagnet surrounding a

pipe, a pipt loop with a pump, and a flowmeter. The electromagnet

contains tapered pole pieces across which a field of up to 21.2 kG can be obtained. The field, when measured with a gaussmeter (Rawson Model

820), was found to be homogeneous within 1$ across the pole pieces. The

pipe loop was constructed mostly of 1.5-in.-0D copper pipe. The portion

of the loop passing across the magnetic pole pieces ^a 4-0-cm-length of

Pyrex tubing, 1.7 cm ID) contained a pair of flat, thin copper electrodes

(9 cm long) cemented to opposite internal surfaces of the glass tube.

Thin insulated wires from the electrodes led from one md of the glass

tube through a rubber tube cemented onto the glass, and also clamped to

a copper pipe section. These electrode wires were connected to a recording

potentiometer. The glass tube was positioned between the pole pieces so

that the electrodes were perpendicular to both the magnetic flux and the

direction of flow. The aqueous solutions used in the study were trans-

ported by means of a centrifugal pump with a capacity of 4-0 gpm. A

flowmeter in the pipe circuit was calibrated to read up to 78 liters of

water per minute. Three liquids, which were not expected to corrode the

cast-iron pump impeller housing, were used in this study: distilled

water, tap water, and aqueous 0.6 M

The results obtained thus far are presented in Fig. 13.IO. The

solid lines indicate the expected emf if Faraday's law is obeyed. Most of ORNL DWG 73-9994R1 VELOCITY IN GLASS PIPE Im/sec) 2.86 4.29 5.72

I -f 4r I

11.5 KILOGAUSS

30 40 50 60 70 80 90 100 PERCENT OF FULL FLOW

Fig. 15.10. EMF vs Flow Rate for Three Liquids Crossing Strong

Magnetic Fields. The solid lines, at constant field intensity, indicate results expected if Faraday's lav; held. -Ul+5- the data -were obtained at 20 kG (the upper line in Fig. 13.10). Although there is some scatter, the data from all three liquids are consistent with

Faraday's law. Moreover, the electric conductivity of the liquid does not appear to affect the results. The conductivity was varied by a factor -6 -1 -1 of about 10,000, from about 6 x 10~ Q~ m for distilled water to 6 x 10- 2 Q -1 m -1 for 0.6 M KNO^ solution. It was noted that, even under c highly turbulent conditions (Reynolds number of about 10 ), Faraday's law held (Fig. 13.10).

The results suggest that the emf induced in molten LigBeF^, the most promising molten-salt for a CTR blanket coolant, will be in accord with

Faraday's law. The electrical conductivity of this molten salt at 600°C -1 -1 34 , is 220 ft m , or about a factor of 3700 higher than that of 0.6 M KNO^ solution at room temperature. The induced emf reduces the chemical stability and, equivalently, increases the corrosivity of the salt. Since the emf also depends on the diameter of the pipe that crosses the magnetic field [see Eq. (11)], it would be advantageous to reduce this dimension by either "honeycombing" a large pipe or subdividing the flow through a series of small-bore pipes. Other methods for countering the effects of induced emf?s in molten-sa_t~blanketed fusion reactors are discussed in 32 another report.

13.7 Thermodynamics and Mass Spectrometry of Vanadium Fluorides J. D. Redman and S, Cantor

The chemical thermodynamics of vanadium fluorides are being investigated in order to predict equilibria (e.g., corrosion, Tg-TF partial pressures) that will be important if vanadium or its alloys are used in -446- conjunetion with molten fluorides in fusion reactors. Vanadium is one of three refractory metals under consideration as the structural metal

in a fusion reactor blanket; the other two are niobium and juolybdenum.

Mass spectrometry of effusing vapors is a very useful tool for

obtaining data for the lower-oxidation-state fluorides of vanadium that

are likely to be stable in a molten solvent of LiF and BeF,-,. A Bendix

time-of-flight mass spectrometer is being used to study the thermal

decomposition of crystalline VF and VF . It is hoped that mass spectro- 3 p graphy can be used to study the equilibrium,

3VF2(g) - 2VF (g) + V(s) . (12)

By combining this equilibrium data with sublimation data for VF2(s) and

VF^(s), and with the free energy (or enthalpy) of formation of VFp (see M/ Sect. 14.2), the free energy of formation of VF^(s) can be determined.

The vapor effusing from VF^(s) between 898 and 1173°K in a

graphite effusion cell was found to be predominantly (Tempera-

tures were measured in the cell by means of Chromel-Alumel thermocouples •f* sheathed in stainless steel.) From an analysis of VF^ intensities, the

following equation was obtained for the equilibrium sublimation pressure

of VF^3 : log p(atm) = 8.728 - 14,750/T (°K) . (13)

This equation agrees reasonably well with the re suit of a Russian mass-

spectrometric study-3'5 for the sublimation reaction:

VF (s) = VF.(g) . (14) -447-

• The data were correlated by the equation

log p(atm) = 9.476 - 15.603/T (°K) . (15)

From Eq. (l4) we obtain an enthalpy of sublimation of 67.5 kcal/mole,

some h kcal/mole less than that reported by the Russian investigators.

A mass spectrometric pesk due to about as intense as the

VFg+ peak;, was also observed from the same sample. The Russian workers + reported VF^ , with an intensity at 1000°K about four times as great as

the values that we detected. The parent molecule for the VF^+ is un-

certain since either VF^ or VF^ may be present as an impurity in the

samples used. However, one cannot dismiss the possibility that VF^ also

undergoes some form of disproportionation reaction.

The thermal decomposition of VF„(s) is more complex than that of < • VF^(s). The initial study in a graphite effusion cell showed that the

VF„ disproportionated to VP and vanadium metal, the latter probably

reacting with the graphite container to form a stable carbide. Study

of VFg decomposition in a vanadium effusion cell again showed that VFg

disproportionated to VF^ and V. However+ , in this test, VF also seemed to be a parent gaseous species since VF exhibited a lower appearance + +

potential than VFg and VF^ . Tentative values for the appearance potentials for VF^ , VFg , and VF , obtained from ionization efficiency

curves at 1348CK, were 12.5, 11.0, and 10.5 eV, respectively. Peak

intensities were measured at 16 eV to avoid excessive fragmentation; at

+ + 1273°K , the intensities (in scale divisions) were: VF , 0.24; VF2 , 0.13; VF 0.02. Based on intensities observed between 1223 and 1348°K, the -kkQ- enthalpy for the reaction

VF2(g) + V(s) a 2VF(g) (16)

•wars found to be approximately 9 kcal/mole. This enthalpy seems to be much too small compared with the enthalpy of about 108 kcal/mole for the analogous reaction3^'^

CrF2(g) + Cr(s) ^ 2CrF(g) . (17)

It seems likely, but by no means certain, that some of the VF+ intensity is obtained from a VF2 gaseous precursor.

In a second test using VF2(s) in the vanadium effusion cell that was not heated above 1223°K, VF+ was not observed near its expected appearance potential of 10.5 eV. Moreover, at 16 eV (5 eV above the appearance

+ v + + + potential of VFg ), there was no VF peak, but only VFg and VF^ . In this test, vaporization ceased in about 30 min, which was an insufficient period of time to obtain an ionization curve to verify the absence of a VF species. Inspection of the cooled cell showed a dark blue condensate (the same color as crystalline VF2) on the lid, but the orifice was not plugged. We observed a metallic residue which appeared to cover unreacted VFg in the cell bottom. Volatilization probably ceased when VFg in the bottom of the cell became covered with vanadium metal produced by a dispropcrtionation reaction. Analysis of the cell contents has not been completed. Experiments are continuing in an effort to sort out the puzzling behavior of VFg decomposition. -hk9-

13.8 The Solubilities of Hydrogen, Deuterium, and Helium in Molten Li2BeF)+ A. P. Malinauskas* and D. M. Richardson As noted in Sect. 13«6, the molten compound Li^BeFi is a candidate for use in thermonuclear reactors, both as a blanket material to breed 32 tritium and as the primary heat transfer fluid. Knowledge of the solubility of hydrogen isotopes in this liquid is required for the evaluation of the salt as a heat-transfer fluid, and for devising methods for recovering tritium from the salt. Our measurements of the solubilities of Hg and were made because practically no data existed in the literature; the results, which are being prepared for publication, can be used to estimate the solubility of tritium in molten LigBeF^.

The apparatus and experimental procedure employed in this study, along with preliminary datoQa on the solubility of Hg in molten Li^BeF^, are described elsewhere. These data showed that the solubility of

H obeyed Henry's law at 873°K in the range of 1 to 2 atm, as do the 39 solubilities of the noble gases. ^ Additional data have been obtained

with hyorogen and deuterium. Within experimental uncertainty, the

apparent solubilities of these isotopes were identical. The data obtained

in the temperature range 773 to 973°K can be represented by 3 log10 (10 Kc) = log10 T - (1535/T) - 0.7684 , (18)

where

K = the distribution coefficient, concentration of the gas in the salt divided by its concentration in the gas, moles/liter,

* Chemical Technology Division, Chemical Development Section B. -450-

T = temperature, °K.

The solubilities that were determined for helium over the temperature range of 773 to 1073°K are described by the expression

3 log1Q (10 Kc) = log10 T - (1177/T) - 0.7954 . (19)

, 39 These results are 20% lower than those obtained by earlier investigators, but cause of the differences has not been established.

The initial 70 experiments were determining either hydrogen or helium solubility determinations, and all measurements were conducted with the same batch of salt. In experiments designed to determine the solubility of deuterium, salt samples that had been saturated with deuterium- were sparged with an inert gas in the usual manner to strip out the dissolved deuterium. Analyses of the stripped gases showed that 15 to

4o$> of the dissolved species was hydrogen, present either as HD or H^.

Interpretation of the data suggested that the solubility of deuterium was practically identical to that of hydrogen. After several experiments in which the salt was saturated with deuterium, studies with hydrogen were resumed. Wo deuterium was detected as a dissolved species.

The above results indicat-3 e that an exchangeable reservoir of hydrogen of at least 3 x 10 mole per kilogram of salt was present in the system. 40 The probable impurity is hydroxide that originated from oxygen or moisture contamination early in the experimental sequence. 13.9 References for Section 13

1. J. S. Watson, An Evaluation of Methods for Recovering Tritium from the Blankets or Coolant Systems of Fusion Reactors, ORWL-TM-3794 (July 1972). -451-

2. E. Veleckis and E. Van Deventer, Physical Chemistry of Liquid Metals and Molten Salts Semiannual Report, January-June 1971, ML-7823 (1971), P. 14. "

3. E. Veleckis and E. Van Deventer, Physical Inorganic Chemistry Semi- annual Report, July-December 1972, ANL-7978 (1973), p. 6.

4. D. Ht J. Goodall and G. M. McCracken, Proceedings of the 7th Symposium on Fusion Technology, Grenoble, France, October 22-27, 1972, EUR-4938, p. 151-56.

5. S. A. Meacham, E. F. Hill, and A. A. Gordus, The Solubility of Hydrogen in Sodium, APDA-241 (June 1970), p. 21.

6. S. Dushman, Vacuum Technique, p. 539, Wiley, New York, 1949.

7. P. A. Redhead, J P. Hob son, and E. V. Kornelsen, The Physical Basis of Ultra-High Vacuum, p„ 102, Chapman and Hall, London, 1968.

8. W. R. Chambers, A. P. Fraas, and M. N. Ozisik, A Potassium-Steam Binary Vapor Cycle for Nuclear Power Plants, ORNL-3584 (May 1964).

9. A. P. Fraas, Comparison of Two Tritium Removal Systems Designed to Minimize Contamination of Steam Systems in Full-Scale Thermonuclear Power Plants, ORNL-TM-2932 (April1970). 10. P. I. Shamri, Lithium and Its Alloys, AEC-tr-3436 (1952).

11. B. Boehm and W. Klemm, Z. Anorg. Allgem. Chem. 243.- 70 (1939).

12. J. H. Hildebrand and R. L. Scott, Solubility of Non-electrolytes, 3rd ed., Reinhold, New York, 1950.

13. L. M. Ferris, J. C. Mailen, J. J. Lawrance, F. J. Smith, and E. D. Nogueira, J. Inorg. Nucl. Chem. 32, 2019 (1970).

14. C. E. Schilling and L. M. Ferris, J. Less-Common Metals 20, 155 (1970). 15. F. J. Smith, J. Less-Common Metals 27, 195 (1972).

16. F. J. Smith, J. Less-Common Metals 29, 73 (19"

17. Daniel R. Stull, Ind. Eng. Chem. 39- " 1

18. 0. W. Richardson, J. Nicol, and T. nell, 1. Mag. 6(8), 1 (1904)

19. A. Sieverts, Z. Physik Chem. 60, 129

20. M. Lombard, Compt. Rend. 177, ll6 (19^3). -452-

21. H. G. Deming and B. C. Hendricks, J. Am. Chem. Soc. 45, 2857 (1923).

22. G. Borelius and S. Lindbloom, Ann. Physik 82, 201 (1927).

23. W. R. Ham, J. Cnem, Phys. 1, 476 (1933).

24. C. J. Smithells and C. E. Ransley, Nature 134, 8l4 (1934); Proc. Roy. Soc. A150, 172 (1935).

25. C. J. Smithells, Cantor Lectures, 1938; J. Roy. Soc. Arts 86, 939-83 (1938).

26. F. J. Norton, pp. 570-81, in S. Dushman, Scientific Foundations of Vacuum Technique, 2nd ed., ed. "by J. M. Lafferty, Wiley, New York, 1962.

27. R. H. Fowler, Proc. Cambridge Phil. Soc. 31, 260 (1935).

28. R. Ash and R. M. Barrer, Phil. Mag. 8(4), 1197 (1959).

29. R. A. Strehlow and H. C. Savage, The Permeation of Hydrogen Through Structural Metals at Low Pressures, 0RNL-4881 (in press).

30. S. S. Kirslis and F. F. Blankenship, MSR Program Semiannu. Progr. Rep. Aug. 31, 1971, ORNL-4728 (1972), p. 46.

31. R. A. Strehlow and H. E. Robertson, The Sorption of Tritium on Graphite at Elevated Temperatures, 0RNL-4882 (in preparation).

32. W. R. Grimes and S. Cantor, pp. 165-68, The Chemistry of Fusion Technology, ed. by D. M. Gruen, Plenum Publishing Corp., New York, 1972.

33. Lange's Handbook of Chemistry, 9th ed., pp. 1208, 1210 (1956).

34. S„ Cantor et al., Physical Properties of Molten-Sa.lt Reactor Fuel, Coolant and Flush Salts, ORNL-TM-2316 (August 1968).

35. L. N. Sidorov, M. Y. Denisov, P. A. Akishin, and V. B. Shol'ts, Russ. J. Phys. Chem. 40, 620 (1966).

36. R. A. Kent and J. L. Margrave, J. Am. Chem. Soc. 87, 3582 (1965).

37. D. D. Wagman, W. H. Evans, V. B. Parker, I. Halow, S. M. Bailey, and R. H. Schumm, Selected Values of Chemical Thermo dynamic Properties, U.S. Nat'l Bur. of Stds. Publication Tech. Note 270-4 (1969), P. 120.

38. A. P. Malinauskas, D. M. Richardson, J. E. Savolainen, and J. H. Shaffer, Ind. Eng. Chem. Fundamentals 11, 584 (1972). 39 • G. M. Watson, R. B. Evans III, W. R. Grimes, and N. V. Smith, J. Chem. Eng. Data 7, 285 (1962).

40. A. L. Mathews and C. F. Baes, Jr., Inorg. Chem. 7, 373 (1968). -454-

14. THERMODYNAMICS OF MOLTEN-SALT SYSTEMS

It was pointed out in Sect. 13 that molten Li^BeF^ is a potential

"blanket material for a CTR. Molten salts can also he used as heat- transfer media, catalysts for coal conversion processes, and in the purification of metals by electroslagging. Our present research effort in this area is aimed primarily at providing basic thermodynamic data for selected solutes in molten fluoride mixtures, particularly LigBeF^. Some theoretical work related to molten oxides is also being done.

l4.1 Chemistry of Tellurium in Molten LigBeF^ •if -if-if C. E. Bamberger, J. P. Young, and R. G. Ross

Because tellurium has recently been associated with the corrosion 12 3 of metals in nuclear reactors, 9 including the MSRE, the present study was undertaken to determine whether soluble tellurium species can exist in significant concentrations in molten LigBeF^. Absorption spectrophoto metry was the principal technique used. Silica was chosen as the con- 4 tainer material because of its inertness toward tellurium-lithium alloys 5 6 and because its equilibria with molten Li^BeF^ have been determined.

The tellurium and LiTe were used as received, whereas the lithium 3 telluride, Li^Te, was further purified by heating at 600°C under vacuum.

The solvent and solutes were always handled in an inert-gas dry box; the silica tubes containing the solids were sealed under vacuum with a torch.

* Analytical Chemistry Division.

Present address: Babcock and Wilcox, Lynchburg Research Center, Lynchburg, Va. 24505. 1. r-r- —fJJ-

The absorbance spectrum of tellurium vapor in equilibrium with liquid tellurium was determined between 195 and 750 nm at 525 to 600°C; the results are shown in Fig. 14.1. The spectrum of molten LigBeF^ in equilibrium with tellurium in both the liquid and gaseous states at temperatures up to 620°C did not reveal any absorbance attributable to tellurium or its compounds.

The spectrum of as-received LigTe in molten LigBeF^ showed a significant absorbance peak at 4-78 nm (Fig. l4.l), which disappeared with time when the solution was maintained in a thermal gradient. When purified

LipTe was equilibrated at 620°C with Li^BeF, that had previously been contacted with silicon, the salt did not show any absorbance. In order to confirm that the absorbance at 478 nm was due to a complex, probably formed by a reaction such as

(1) in which m and n denote stoichiometric constants associated with the Li-

Te compound, and (s), (g), and (d) denote solid, gas, and dissolved in salt, respectively, Te, LigTe, and LigBeF^ were isothermally equilibrated at 600°C. A moderate absorption peak was observed at 4-78 nm. The same spectrum was obtained by partially reducing tellurium with silicon in the presence of Li„BeF. ,

| Te2(g) + £ Si(s) + m LiF * Lije^d) + SiF^(g) } (2)

or by partially oxidizing Li?Te with air or FeF

n Li2Te(s) + ) FeF2(d) * Lije^d) + (2n-m) LiF(d) + (3) -45b-

ORNL-DWG 73-1449

1.25

1.00

LU 0 1 0.75 CQ otr cn

0.25

WAVELENGTH (nm)

Fig. 14.1. Absorption Spectra of Tellurium Species Contained

in 1-cm-ID Silica Cells, (i) Gaseous tellurium (Tea) at 525°C.

(II) LimTen(LiTe3) dissolved in molten Li2BeF4 at 600°C. -45b- have been measured, whereas, there are essentially no data for VF„ and VF_ 2 3> the fluorides that are likely to be stable in fluoride media significant to fusion reactors (see Sect. 13.7).

The free energy of formation of a compound is, of course, the most valuable thermochemicai datum for predicting the degree of reaction of the compound. Lacking this, the next best datum is the enthalpy of formation since it is relatively easy to predict an entropy of formation.

A theoretically sound method for predicting enthalpies of formation of ionic compounds is derived from lattice energies. The heat of formation of a divalent metallic fluoride is related to lattice energy by the following equation:

AH° --- U(Rq) •» AH° + A+ (Ix + I2) + 2[-EF3 - 4.4 , (4) where

AH° = the standard heat of formation at 298°K (Kcal/mole),

U(R ) = the lattice energy [i.e., the enthalpy associated with 0 + the process: Me* (g,0°K) + 2F~(g,0°K) - MeFg(cr,298°K),

AH° = the standard heat of sublimation of the metal, Me, s at 2o8°K,

AH° - the standard heat of dissociation of fluorine at a 298%

(I + I„) = the enthalpy absorbed when the metal is ionized 1o Me2+ at 0°K,

[-E„] - the electron affinity of fluorine, [i.e., the enthalpy i? of the process: F(g) + e'(g) - F~(g) at 0°K].

All terms in the right-hand side of Eq. (4) can be accurately measured except U(Rq), which must be calculated by combining theoretical considerations with crystallographic and other data (see Sect. 14.3).

U(R ) may be expressed as a sum of energies: U(R ) = U + U + U + U + tf , (5) o e r w z c where

U s the electrostatic energy, the energy calculated by applying Coulomb's law to the charges and positional coordinates of atoms in the crystal lattice,

U = the energy associated with the repulsive forces operating r between electron-charge clouds of neighboring atoms,

U^ = the energy associated with van der Waals forces,

U - the zero-point vibrational energy,

U = the energy of cohesion that may be attributed to covalent 0 bonding.

In Eq. (5)j U is by far the largest term and can be accurately calcu- e lated from crystallographic data. The other terms are calculated with much less accuracy. In fact, U^ is almost never calculated but often represents the energy unaccounted for by the other terms.

Of the MeFg compounds, six have the rutile structure: VFg,

CoF MnPg, FeF2, 2> ^iF^, and ZnF^. For these it can be assumed that differences in U(Rq) reflect differences in Ue» This is equivalent to saying that the sum IJ„ + U^ + + U^ is identical for this isomorphous series of six compounds. Using this assumption in conjunction with Eqs.

(4) and (5) leads to the expression

AH° - U - AH° - (I. + I ) = constant, (6) x e s id0

in which the constant includes the thermodynamic terms associated only with the anion (AH°, E^). Values for AH° and (1^ + Ig) were taken from 7 8 critically compiled tables ' , and U values for all six MeF„ compounds €t u Q were computed from a program which uses Ewald's method. The enthalpy BLANK PAGE -46o- of formation, is accurately known for CoF^, NiFg, and ZnFg (see

Table lU.l for the references); therefore, the validity of Eq. (6) may

he tested with these compounds. If the constant is identical when Eq.

(6) is applied to CoF^, NiF^, and ^nFg' "t^ien assumption of constancy

of the nonelectrostatic energies in Eq. (5) would appear to he correct.

The data for evaluating the constant are listed in Table lU.l. As

shown, the agreement among the three values of the constant is excellent.

Usyig the average value of the constant; we estimate the heats of forma-

tion of the other three difluorides as follows:

298°K (kcal/mole)

FeFg(cr) -1?6

MnF2(cr) -213

VF2(cr) -210

The estimated AH° for FeF„ can be compared with the value of -168.7 f c. 1 0 ± 10 kcal/mole given in the JANAF tables;" * the value of AH° for MnF2 is

much more negative than the -190 i 5 kcal/mole reported by Pitzer and

Brewer.11 It will be interesting to see if other experimentally deter-

mined AH° values for FeF2 and MnFg will support the correlation. No

experimental data are available to permit even a rough determination of

AH° for VF2; the estimate given above is probably as valid as can

presently be obtained.

lb.3 Lattice Energies of Cubic Alkaline-Earth Oxides. Affinity of Oxygen for Two Electrons

S. Cantor

New experimental determinations of the compressibilities of alkaline- 12-In- earth oxides have recently been reported. These values can be used -46i-

Table 14.1. Calculation of the Constant Defined by Equation (6)

(units of kcal/mole)

Metal Fluoride AH, U AH, (Ix + I2) Constant

e CoF. -165.4 -782.8 101.5 574.7 -58.8

NiF, -157.2" -795.3 102.7 595.0 -59.6

ZnF, -182.7° -785.1 31.2 630.9 -59.7

Average -59-^

NBS Tech. Note 270-4 (see ref. 7).

5E. Paidzitis, E. K. Van Deventer, and W. N. Hubbard, J. Chem. Eng. Data 12, 133 (1967).

'NBS Tech. Note 270-3 (see ref. 7). -462- to improve the accuracy of the value for the enthalpy [-E] of the reaction

0(g,0°K) + 2e-(g,0°K) - 02~(g,0°K), (7) which is referred to as "the affinity of oxygen for two electrons." This reaction is highly endcthermic; recent evaluations of [-E] are 162 ± 15

(ref. 15) and 179 ±8 (ref. 16) kcal/mole. Although it is unlikely that

experimental means will he found for directly measuring its value, accurate knowledge of [-E] is desirable since it can he used in predicting other enthalpy changes such as heats of formation of ionic oxides and

ionization potentials of elements in the Born-Haber thermochemical cycle. 12-14

The compressibility data alluded to above are necessary for the accurate evaluation of the lattice energy [denoted as U(Ro)], which in this report refers to the enthalpy change for the process

Me2+(g,0°K) + 02"(g,0°K) - MeO(cr,298.15°K) . (8)

In Eqs. (7) and (8), g and cr refer to gas and crystal, respectively.

Lattice energies for MgO, CaO, SrO, and BaO were calculated using the 17 Ladd and Lee equation after it had been modified to include a correction l8 term computed from the Hildebranc • tion of state. The correction term is necessary to account for L i'act that the interionic distance

is evaluated from crystal data at room temperature; the theory underlying

lattice energy calculations refers to a process at 0°K. The three basic equations are:

Qf r h- Q/ K £ o (9) D 1 R h C - R /p ") + u R o' ^ y z o -463-

6 8 R (9V/p)F(T,P) - 2Ue + 42C/R° + 72D/R° — = / —o (10) p -(u + 3VTa/3) + 6C/Ro + 8D/Rq and

In these equations

U(R ) is the lattice energy,

R is the interionic distance, o U is the electrostatic energy, which for cubic alkaline-earth oxides is—-- -•—kcal/mole, provided -R^ is given in A, •Ko

V is the molar volume at the temperature in interest,

T is 298.15°K in all of our considerations, a is the volume expansivity of the crystal at the temperature of interest = 1 ( > , V V 5T J XT

p is the isothermal compressibility at temperature T,

p the so-called "hardness parameter" is obtained for Eq. (10),

and C and D are the van der Waals constants.

For these salts

C = 6.5952C+_ + + C__) , (12) in which

c. . = 3 E. E. a. a./2(E. + E.) , (13) ij i j i D i J and

D = 6.l457d, + 0.40006(d_ + d ) . (lb) +— ++ —— <

in •which

a.E. a.E . _ d. . = C. . Ue- L ni nJ J

In Eqs. (12)-(15), E^ and E^ are characteristic energies for ions i and j;

a.nd a^. are electronic polarizabilities, e is the charge on the electron, and n. and n. are the numbers of electrons on each ion that contribute to i J the refractive index. In these calculations n. and n. were each assumed i J to be 6, the number of outer p electrons. U is the zero-point energy, that is, the potential

energy of the crystal at 0°K. All of the quantities necessary to solve Eqs. (9)-(15) are either based on crystal data or may be estimated by well-established procedures.

The calculated values for the lattice energies [U(Rq)] are given in Table lb.2; the stated errors are due primarily to uncertainties in the values of the compressibilities.

The negative of the enthalpy change for reaction (8), when added

to the sum of the enthalpy changes for the reactions

Me(cr,298°K) + l/2 02(g,298°K) - MeO(cr,298°K) AH°

Me(g,298°K) - Me(cr,298°K) -AH°s

0(g,298°K) - 1/2 02(g,298°K) -l/2

Me(g,0°K) + 0(g,0°K) - Me(g,298°K) + 0(g,298°K) AH = 2.96 kcal

Me2+(g,0°K) + 2e"(g,0°K) - Me(g,0°K), -(^ + yields [-E] = -U(RQ) + AH° - AH° - 1/2AH£ + 2.96 - + IG). (l6)

The values of [-E] obtained from Eq. (l6) using the lattice energies for

the alkaline-earth oxides estimated above are also presented in Table

lb.2. Values for the thermochemical quantities (£H°, AH°, AH^, +

were taken from tables critically compiled by the National Bureau of -U65-

Tahle lU.2. Calculated Values for the Lattice Energies of the Alkaline-Earth Oxides, U(Eo), and the Affinity of Oxygen for Two Electrons, [-E]

U(Ro) [-E1 Oxide (kcal/mole) (kcal/mole)

MgO -905 ± 3 146

CaO -815 ± 8 14-9

SrO -767 ± 10 ikb

BaO -736 ± 15 153 -466-

Standards. 19^ 20 Uncertainties in each value of [-E] are almost entirely the uncertainties in U(Rq). The "best" value is

149 ± 8 kcal/mole , which is much lower than previously derived. 3 A much more detailed report of this work is given in a publication that has been accepted for publication in the Journal of Chemical Physics.

14.4 Relationship Between Sonic Velocity and Entropy

in Molten UOg

S. Cantor

Compressibility is an important property in determining the stresses of fuel elements following partial or total melting of the solid fuel. 21 The compressibility of molten UOg has been reported, and is based on measurement of sonic velocity at four temperatures: 2865, 2890, 2910, and 2923°C. At such high temperatures, reliable measurements are difficult and expensive. In fact, the program at HEDL for measuring sonic velocities of molten ceramic fuels was discontinued 22afte r obtaining two sets of results, one on AlgO^ and the other on UOg. Thus, a means of correlating the data for molten U02 for the purpose of predicting the compressibility of similar melted fuels (PuOg, U02~Pu02 mixtures, etc.) is of value. 23 It has recently been shown J that sonic velocities can be predicted from entropies and from volume by applying the Debye theory of specific heats to molten salts (predominantly single-component halides). The correlating equation is: 1 C = 10~5 (const) (V/n)3 T exp (-S/3nR) , (17) where

C = sonic velocity, km/sec,

const = a combination of basic physical constants, equal to 1508.87 in cgs units,

V = molar volume, nil,

T = temperature, °K,

n = the number of atoms in a molecule (n = 3 for TJOg),

S = the entropy associated with atomic displacements (the absolute entropy minus any entropy associated with the occupancy of more than one electronic energy state,

R = the gas constant.

The terms S and R must be expressed in identical units.

Application of Eq. (17) to molten U02 requires accurate values of the entropy because of the exponential dependence. The molar volume does not need to be known accurately because of its cube-root dependence.

Absolute entropies of molten UOg are available from the sum of entropies 24-27 derived from various calorimetric investigations. From these absolute entropies, we subtract the entropy associated with a magnetic ordering transition which occurs near 30°K-1 ;- 1th e latter entropy is approximately R In 3, or 2.18 cal mole °K (ref. 24). Over the limited temperature range of interest (see Table 14.3), the molar volume of UOg was assumed to be 31 ml. The constan28 t volume is justified by the precision of the density measurements.

Sonic velocities at four temperatures, calculated from Eq. (17), are listed in the last column of Table 14.3, The agreement between the measured and calculated values is striking, especially if one considers that the experimental sonic velocities have admitted uncertainties of -468-

Table 14.3. Entropies and Sonic Velocities of Molten UO

a Tempm . Entropy*3 Sonic Velocity (km/sec) (°K) (cal mole""1 °K""1) Measured2"1" Calculated0

3138 73.77 I.793 1.68

3163 74.02 1.785 1.67

3183 74.23 1.776 1.66

3196 74.36 1.75q 1.65 a ci The temperatures at which sonic velocities were measured.

The absolute entropy minus the electronic entropy (see text).

Calculated from Eq. (17). -469-

M 21 3%. Thus, the correlating equation, originally applied to more conventional molten salts, seems to apply quite well to molten UC„. There is little reason to doubt that, with reliable calorimetric and/or estimated entropies, Eq. (17) is an excellent means for estimating sonic velocities in molten fuels, especially ceramic oxides.

14.5 References for Section 14

1. F. Dekeroulas, R. Lebeuze, D. Calai, A. Van Craeynest, and M. Conte,

J. Nucl. Mater. 43(3), 313 (1972).

2. B. F. Rubin, E. A. Aitken, and S. K. Evans, Trans. Am. Nucl. Soc. 15

(1), 218 (1972).

3. Ho E. McCoy, The Development Status of Molten Salt Breeder Reactors,

ORNL-4812 (August 1972).

4. M. S. Foster, C. E. Johnson, K. A. Davis, and J. Peck, Argonne

National Laboratory, Chemical Engineering Div. Semiann. Rept. for

July-Dec. 1967, A2TL-7425 (1968), p. 174.

5. C. E. Bamberger, C. F. Baes, Jr., and J. P. Young, J. Inorg. Nucl.

Chem. 30, 1979 (1968).

6. C. E. Bamberger and C. F. Baes, Jr., J. Am. Ceram. Soc. 55.(11), 5^4

(1972).

7. D. D. Wagman, W. H. Evans, V. B. Parker, I. Halow, S. M. Bailey, and

R. H. Schumm, Selected Values of Chemical Thermodynamic Properties,

U.S. Nat'1 Bur. Std. Publications Tech. Note 270-3,4,5 (1968-1971).

8. C. E. Moore, Ionization Potentials and Ionization Limits Derived

from the Analysis of Optical Spectra, U.S. Nat'l Bur. Std. Publi-

cation NSRDS - NBS 34 (1970). -470-

9. The computer program was supplied by W. Busing, Chemistry Division,

Oak Ridge National Laboratory.

10. D. R. Stull and H. Prophet (eds.), JANAF Thermochemical Tables, 2nd

ed. U.S. Nat'1. Bur. Std. Publication NSRDS - NBS 37 (1970).

11. G. N. Lewis and M. Randall, Thermodynamics (revised by K. S. Pitzer

and L. Brewer), 2nd ed., p. 674, McGraw-Hill, New York, 1961.

12. P. R. Son and R. A. Bartels, J. Phys. Chem. Solids 33, 819 (1972).

13. R. A. Bartels and V. H. Vetter, J. Phys. Chem. Solids 33, 1991 (1972).

14. 0. L. Anderson and P. Andreatch, J. Am. Ceram. Soc. 49, 4o4 (1967).

15. M. L. Huggins and Y. Sakamoto, J. Phys. Soc. Japan 12, 241 (1957).

16. M. F. C. Ladd and W. H. Lte, Acta Cryst. 13, 959 (1960).

17. M. F. C. Ladd and W. H. Lee, J. Inorg. Nucl. Chem. 11, 26b (1959).

18. M. P. Tosi, pp. 1-66 in Solid State Physics, Vol. 16, Academic,

New York, 1964.

19. V. B. Parker, D. D. Wagner, and W. H. Evans, Selected Values of

Chemical Thermodynamic Properties, U.S. Nat'l. Bur. Std. Publication

Tech. Note 270-6 (November 1971).

20. C. E. Moore, Ionization Potentials and Ionization Limits Derived

from the Analysis of Optical Spectra, U.S. Nat'l. Bur. Std.

Publication NSRDS-NBS 34 (1970).

21. 0. D. Slagle and R. P. Nelson, J. Nucl. Mater. 40, 349 (1971).

22. 0. D. Slagle, Hanford Engineering Development Laboratory, Richland,

Washington, personal communication.

23. S. Cantor, J. Appl. Phys. 43, 706 (1972).

24. J. J. Huntzicker and E. F. Westrum, J. Chem. Thermodynamics 3, 6l

(1971). -471-

25. D. R. Frederickson and M. G. Chasanov, J. Chem. Thermodynamics _2,

623 (1970).

26. L. Leibowitz, M. G. Chasanov, L. W. Mishler, and D. F. Fischer,

J. Nucl. Mater. 39, 115 (1971).

27. R. A. Hein, P. N. Flagella, and J. B. Conway, J. Am. Ceram. Soc. 51, 291 (1968). 28. J. A. Christensen, J. Am. Ceram. Soc. 46, 607 (1963). -U-72-

15. COAL CONVERSION STUDIES

L. M. Ferris, M. R. Bennett, and C. T. Thompson

The current energy crisis in the United States has stimulated an increase in research relating to the conversion of coal to liquid and gaseous fuels. However, one aspect that presently is not receiving much attention is the in situ gasification of coal. In situ gasification methods hased on partial oxidation of the coal have been considered previously.1 Recently, Dr. J. W. Larsen of the University of Tennessee 2 suggested to us that gaseous HF might be an effective catalyst for the underground hydrogasification of coal. The scheme, if feasible, would involve injection of H^-HF mixtures into a coal seam and withdrawal of the gaseous (and/or liquid) hydrocarbon products. As described below, this possibility was explored in a series of batch autoclave experiments at about 300°C. The results were very discouraging in that no detec- able hydrogenation of the coal occurred using hydrogen pressures up to about 500 psig and gas mixtures containing up to 20 mole HF.

The studies were made using a 1.5-liter stainless steel autoclave that was equipped with continuous temperature and pressure recording devices. The coal used was a highly agglomerating, low-sulfur, bitu- minous type obtained from TVA's Bull Run Steam Plant. This coal had an H/C atom ratio of about 0.9; it also had an ash content of about 17%, comprised mostly of aluminum, silicon, and iron. The coal was used without pretreatment other than air-drying and grinding to about -100 mesh. In a typical experiment, about 6 g of the dried coal was loaded into a stainless steel crucible, which was also placed in the autoclave. -473-

Sufficient NaHFg to ensure a constant HF pressure at 300°C was loaded into a separate crucible, which was also placed in the autoclave. After the air had been swept out of the autoclave with pure hydrogen, the autoclave was pressurized with hydrogen to about 225 psig. (This gave an overall Hg/C mole ratio of about 3.) The system was subsequently heated to 300°C where the HF pressure from the dissociation of NaHFg was about

4-5 psig (ref. 3) and the hydrogen pressure was about U35 psig. The system was maintained at temperature for several hours (with the pressure being continuously recorded) before it was cooled to room temperature.

The gases from the autoclave were then vented through a train consisting of a NaF trap held at 125°C to collect HF, a Drierite trap to collect water, three gas sample bulbs arranged in parallel, and, finally, a Wet

Test Meter. The gas bulbs were valved off in sequence in order to obtain samples of the gas at different stages of the venting procedure.

The residual coal was removed from the autoclave and weighed before analysis.

No evidence for hydrogenation of the coal was found in each of several experiments in which the HF concentration of the gas was 10 to

20 mole Gas-chromatographic and mass-spectrographs analyses of the gas at the end of each experiment typically showed less than 0.2% hydrocarbon gases. In each case, the weight of the coal had decreased about

0.5 g> but the H/C atom ratio remained essentially unchanged from its original value. The residual coal analyzed 5 to 15 wt % fluorine. X- ray analysis of the residues revealed the presence of AlF^. Qualitatively, therefore, both the presence of fluorine in the residue and the weight loss of the sample can be explained on the basis of HF reacting with inorganic constituents of the coal. The most likely reactions (over-

simplified) are:

Si02(s) + 4HF(g) > SiFJ+(g) + 2H20(g) , (l)

AlgO (s) + 6HF(g) > 2A1F (s) + 3H20(g) , (2)

in which (s) and (g) denote solid and gas, respectively. This mechanism was supported "by results of experiments in which water was added to the

system in addition to HF and hydrogen. The fluorine concentrations of

the residue were much lower in these experiments than in those where water was absent.

15.1 References for Section 15

1. E. S. Donath, Chemistry of Coal Utilization, ed. by H. Lowry, Wiley,

New York, 1965.

2. J. W. Larsen, University of Tennessee, personal communication, Dec.

13, 1972.

3. J. Fischer, J. Am. Chem. Soc. 79; 6363 (1957). -475-

APPENDIX I

PROCEDURES FOR CHEMICAL TESTS ON THE MINIATURE FAST ANALYZER -476-

APPENDIX I: PROCEDURES FOR CHEMICAL TESTS ON THE MINIATURE FAST ANALYZER

A. Uric Acid

Reagent: a. 1.0 M borate buffer, pH 9.1. b. Nova industry bacterial uricase: reconstitute by adding 10.0

ml of a 1 M borate buffer, pH 9-1. Refrigerate at 4°C. (Does

not need to be made up fresh every day, should be used unitl

gone.)

Sample: a. Blood serum b. Uric acid standards: Thaw a 2-ml vial of stock uric acid standard,

(l mg/ml), and dilute to 2, 4, 6, 8, and 10 mg % by pipetting

0.1, 0.2, 0.3> 0.4, and 0.5 ml into 5-ml volumetric flasks and

filling to mark with distilled H^O.

Wavelength: 290 nm

Temperature: 30°C

Loading volumes: a. Reagent pipette

Reagent pump, 50 at 30% = 15 Mil

Diluent pump, 50 at 100% = 50 M»1 b. Sample pipette

Sample pump, 20 at 25% = 5 n,l

Diluent pump, 50 p.1, at 100% = 50 M<1 -477-

Procedure:

a. The automatic rotor loader is set in the rrrulti-sample/single

reagent mode, and a quartz rotor is loaded as follows:

Cuvet No. Sample

1 H2O 2 2 mg % standard 3 4 mg % standard 4 6 mg % standard 5 8 mg % standard 6 10 mg % standard 7-17 Controls and samples

h. The nonlinear fixed-time program is used with the following

parameters:

Reaction time; 240 sec

Readings in interval: 8

Readings/cuvet: 20

Sample interval: T-. = 1

Delay time interval: 3 sec

c. The rotor is placed in the analyzer, allowed to reech 30°C, and

the analysis is started.

B. Alkaline Phosphatase

Reagent: Calbiochem Alkaline Phosphatase Kit: reconstitute by adding

2 ml of Vial A (buffer) to Vial B (substrate). Cap, and dissolve

contents by gentle inversion.

Sample: blood serum

Filter: 400 nm Temperature: 30 °C

Loading volumes (sample-reagent loader): a. Reagent pipette

Reagent pump, 20 Hi, at 100% = 20

Diluent pump, 200 Hi, at 25% = 50 Hi b. Sample pipette

Sample pump, 20 HI, at 50% = 10 pi

Diluent pump, 50 Hi, at 100% = 50 y.1

Procedure: a. Set sample-reagent loader in single-chemistry:multi-sample mode.

(1) Sequentially place samples into the cups in the sample carousel.

(2) Place reagent into reagent cup. b. Place rotor on turntable c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer, and proceed as indicated in the

Analyzer Manual.

Calculations:

a. Activity (I.U. liter 1 min-1) = x 1380. J v 1 mm

C. Acid Phosphatase

Reagent: Use Smith-Kline Instruments Acid Phosphatase Kit.

a. Substrate tablet: dissolve four tablets in 2 ml of distilled water.

b. Sodium tartrate, supplied in vials, is reconstituted with 10 ml

of distilled water. -479-

2. Sample: Use fresh serum or plasma free from hemolysis; avoid oxalate

and fluoride; sample should he less than 1 hr old. The prostatic acid

phosphatase -will remain stable for a longer period of time if the pH

is immediately brought below 6.0. A suggested serum preservative is

acetate buffer, 5 moles/liter, pH 5.0. Add 20 ill of this buffer per

milliliter of serum.

3- Filter: 400 nm

4. Temperature: 30°C

5* Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 |a,l, at 100% = 20 p.1

Diluent pump, 200 pi, at 25% = 50 p1

b. Sample pipette

Sample pump, 20 M<1, at 50% = 10 \ul

Diluent pump, 50 p.1, at 100% = 50 ubl

6. Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample mode.

(1) Sequentially place duplicate sample aliquots into the cups

in the sample carousel.

(2) Place reagent into reagent cup.

b. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. After loading, 5-P«l aliquots of distilled water are placed into

the reagent cavities of the even-numbered cuvets (i.e., 2, 4,

6, 8, 10, 12, 14, 16). Five-microliter aliquots of the tartrate -480-

reagent are added to the odd-numbered reagent cavities. (Note:

an Oxford pipette can be used to perform this function.)

e. Place loaded rotor into its holder in the Analyzer, and proceed

as indicated in the Analyzer Manual.

7- Calculations:

a. Total activity = x 2090 (even cuvet). * mm A b. Tartrate inhibitable activity = x 2090 (odd cuvet). , . , . , ., . , . Total Activity - Tartrate Activity c. Percent tartrate inhibition = , -,——rr — x 100. mTotal Activity

D. Creatinine Phosphokinase (CPK)

1. Reagent: Calbiochem CPK Reagent Kit: reconstitute by adding 2 ml

of H20 to Vial B (NADP), swirl to dissolve, and add to Vial A (CPK

reagent). Cap, and dissolve contents by gentle inversion.

2. Sample: blood serum

3. Filter: 34-0 nm

k. Temper a ture: 30 ° C

5. Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 Hi, at 100% = 20 Hi

Diluent pump, 200 Hi, at 25% = 50 Hi

b. Sample pipette

Sample pump, 20 HI, at 50% = 10 Hi

Diluent pump, 50 Hi, at 100% = 50 y.1

6. Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample mode,

(l) Sequentially place samples into the cups in the sample

carousel. 1, O-I

(2) Place reagent into reagent cups.

b. Place rotor into reagent cups.

c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer, and proceed as indicated in

the Analyzer Manual.

7. Calculations:

a. Activity (i.U. liter"1 min"1) = x 4180. v mm

E. Lactic Dehydrogenase—Lactate Substrate (LDH-L)

1. Reagent: Calbiochem LDH-L Reagent Kit: reconstitute by adding 2 ml

of HpO to Vial B (MB), swirl to dissolve, and add to Vial A (LDH-L

reagent). Cap, and dissolve contents by gentle inversion.

2. Sample: blood serum

3. Wavelength: 340 nm

4. Temperature: 30°C

5. Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 JJUI, at 100% = 20 p»l

Diluent pump, 200 M-l, at 25% = 50 (Jkl

b. Sample pipette

Sample pump, 20 Hi, at 50% = 10

Diluent pump, 50 ubl, at 100% = 50

6. Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample mode,

(l) Sequentially place samples into the cups in the sample

carousel. -482-

(2) Place reagent into reagent cups. t>. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer, and proceed as indicated in

Analyzer Manual.

Calculations:

a. Activity (I.U. liter-1 min-1) = x 4l80. v 1 mm

F. Serum Glutamate Oxaloacetate Transaminase (SGOT)

Reagent: Calbiochem GOT Reagent Kib: reconstitute by adding 2 ml

of HgO to Vial B (NADH), swirl to dissolve, and add to Vial A (GOT

reagent). Cap, and dissolve contents by gentle inversion.

Sample: blood serum

Filter: 340 nm

Temperature: 30°C

Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 Hi, at 100% = 20

Diluent pump, 200 HI, at 25 % = 50 HI

b. Sample pipette

Sample pump, 20 HI, at 50% = 10 H-1-

Diluent pump, 50 Hi, at 100% = 50 pbl

Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample mode,

(l) Sequentially place samples into the cups in the sample

carousel. -ifOJ-

(2) Place reagent into reagent cup.

b. Place rotor on turntable.

c. Move samples and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer, and proceed as indicated in the

Analyzer Manual.

7. Calculations:

a. Activity (l.U. liter"1 min"1) = ^r— x 4-180. 0 v ' mm

. G. Serum Glutamate Pyruvate Transaminase (SGPT)

1. Reagent: Calbiochem GPT Reagent Kit: reconstitute by adding 2 ml

of H20 to Vial B (NADH), swirl to dissolve, and add to Vial A (GPT)

reagent). Cap, and dissolve contents by gentle inversion.

2. Sample: blood serum

3. Wavelength: 340 nm

4. Temperature: 30°C

5. Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 p,l, at 100% = 20 j-ul

Diluent pump, 200 p.1, at 25% = 50

b. Sample pipette

Sample pump, 20 M-l, at 50% = 10 nl

Diluent pump, 50 at 100% = 50 \i±

6. Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample

mode. -484-

(1) Sequentially place samples into the cups in the sample

carousel.

(2) Place reagent into reagent cup.

b. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer, and proceed as indicated in the

Analyzer Manual.

7* Calculations:

a. Activity (I.U. liter"1 min"1) = x 4l8o. J v ' mxn

H. Serum Glucose

1. Reagent: Calbiochem Glucose Reagent Kit: reconstitute by adding 2

ml of buffer to Vial B (NADP), and swirl to dissolve. After dissolving

the combined contents of two vials of B, add to a single vial of A

(glucose reagent). Cap, and dissolve contents by gentle inversion.

2. Sample: blood serum

3. Wavelength: 340 nm

4.' . Temperature: 30°C \ 5. Loading volumes:

a. Reagent pipette

Reagent pump, 20 HI, at 100% = 20 pi

Diluent pump, 200 pi, at 25% = 50 H1

b. Sample pipette

Sample pump, 20 pi, at 10% = 2 pi

Diluent pump, 50 Hi, at 100% = 50 pi Procedure: a. Set sample-reagent loader in single-chemistry:multi-sample mode.

(1) Sequentially place diluted samples into the cups in the

sample carousel.

(2) Place reagent into reagent cup. b. Place rotor on turntable. c. Move sample and reagent probes into position, and depress start

button. d. Place loaded rotor into Analyzer and proceed as indicated in the

Analyzer Manual.

Calculation: a. Glucose concentration (mg/dl) = AA x 373*

I. Blood Urea Nitrogen

Reagent: Calbiochem BUN Reagent Kit: reconstitute by adding 2 ml of

H20 to Vial B (NADH), swirl to dissolve, and add to Vial A (BUN reagent).

Cap, and dissolve contents by gentle inversion.

Sample: serum

Wavelength: 3^0 nm

Tempe r a ture: 30 ° C

Loading volumes (sample-reagent loader): a. Reagent pipette

Reagent pump, 20 m>1, at 100% = 20 jil

Diluent pump, 200 p.1, at 25% = 50 b. Sample pipette

Sample pump, 20 nl, at 10% = 2 (j,l

Diluent pump, 50 y.1, at 100% = 50 -486-

6. Procedure:

a. Set sample-reagent loader in the single-chemistry.multi-sample mode.

(1) Sequentially place the diluted samples into the cups in the

sample carousel.

(2) Place reagent into reagent cup.

b. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer and proceed as indicated in the

Analyzer Manual.

7. Calculations:

a. Urea concentration (mg/dl) = AA x 500.

J. Serum Triglyceride

1. Reagents:

a. Reagent 1: Calbiochem 3-Vial Triglyceride Reagent Kit:

reconstitute by adding 7 ml of H^O to the contents of "Vial A

(ATP, PEP, MDH, PK, LDH, buffer), swirl to dissolve, and add

to contents of Vial C (lipase). This reagent becomes the diluent

used in the sample pipette; consequently, the diluent pump of

this pipette needs to b^ flushed and primed with this reagent.

b. Reagent 2: Vial B, containing the glycerol kinase, is made up

to a volume of 1.0 ml and is subsequently refrigerated at 4°C

until use.

2. Sample: blood serum

3. Wavelength: 3^0 nm

4. Temperature: 30°C -487-

Loading volumes: a. Reagent pipette (H^O diluent)

Reagent pump, 20 Hi, at 50% = 10 Hi

Diluent pump, 200 Hi, at 35% = 70 b. Sample pipette (Reagent 1 as diluent)

Sample pump, 20 Hi, at 10% = 2 Hi

Diluent pump, 50 Hi, at 100% = 50 Hi

Procedure: a. Set sample-reagent loader in single-chemistry:multi-sample mode:

(1) Prime diluent pump of sample pipette •with reagent 1.

(2) Sequentially place samples into the cups of the sample

carousel.

(3) Place reagent 2 into reagent cup. b. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. After loading, the rotor is allowed to stand covered with a Lucite

cover for 10 min to ensure complete enzymic hydrolysis of the

serum triglyceride.

e. After hydrolysis, place the rotor in the Analyzer and proceed as

indicated in the Analyzer Manual.

Calculations:

a. Triglyceride concentration (mg of triolein/dl) = AA x 1880.

K. Calcium

Reagent: either obtain reagent kit from Pierce Chemical Company

(Rockford, Illinois), or prepare your own as follows: -488-

a. Methylthymol blue, 8.3 x 10mole/liter

b. 8-hydroxyquinoline, 250 mg/dl reagent

c. Polyvinylpyrrolidone, 5% w/v

d. Ethanolamine, final reagent concentration of 1 mole/liter

e. HC1, 6.25 ml of 0.6 N HC1 per 100 ml of reagent to adjust pH

of reagent to 11.5

f. Sodium sulfite, 0.5 mole/liter

2. Sample: serum diluted 1:10 with calcium-free water

3- Standard solutions: using a reference-grade CaCl^, prepare calcium

standards in the concentration range 6 to 12 mg/dl. These standards

are diluted 1:10 before analysis.

4. Wavelength: 620 nm

5. Temperature: 30°C

6. Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 2.0 pi, at 100% = 20 pi

Diluent pump, 200 pi, at 25% = 50 pi

b. Sample pipette

Sample pump, 20 pi, at 50% = 10 pi

Diluent pump, 50 pi, at 100% = 50 pi

7• Procedure:

a. Set sample-reagent loader in the single-chemistry:multi-sample

mode.

(l) Sequentially place diluted calcium standards representing

solutions of 6, 8, 10, and 12 mg/dl of calcium, respectively,

into cups 2, 3, 4, and 5 of the sample carousel. -489-

(2) Sequentially place diluted samples into the remaining cups

in the sample carousel.

(3) Place reagent into reagent cup.

b. Place rotor on turntable.

c. Move sample and reagent probes into position, and depress start

button.

d. Place loaded rotor into Analyzer and proceed as indicated in the

Analyzer Manual.

8. Calculations:

a. Determine the calcium concentration of each sample by correlating

final absorbance with concentration.

L. Total Bilirubin

1. Reagents:

a. 1.0% of 2,4-dichloroaniline in 1.25 M HC1

b. % NaN02

c. Methanol—ethylene glycol diluent (k:l v/v)

d. Preparation - prepare diazotized reagent by mixing 5 nil °f

the 2,k-dichloroaniline solution with 100 pi of the 5% NaNO^;

stir solution for 15 min at 6°C. Add 25 ml of methanol-

ethylene glycol while maintaining the temperature at or near

6°C. When not in use, this reagent should be stored at -10°C.

When stored at this temperature, the reagent should be stable

for up to 10 days.

2. Sample: blood serum diluted 1:10 with 0.1 M Tris, pH T*b, containing

0.9% NaCl. -490-

3. Bilirubin standards: using Hyland Q,-Pak control serum II as a stock

reference standard, prepare four standard solutions that are equivalent

to 1:10 diluted solution containing 1.2, 2.4, 3-6, and 4.8 mg/dl of

bilirubin, respectively.

4. Wavelength: 550 nm

5. Temperature: 30°C

6. Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 20 |jbl, at 100% = 20 g,l

Diluent pump, 200 ^1, at 25% = 50 Hi

b. Sample pipette

Sample pump, 20 Hi, at 50% = 10 pi

Diluent pump, 50 Hi at 100% = 50 Hi

7. Procedure:

a. Set sample-reagent loader in single-chemistry:multi-sample mode.

(1) Sequentially place diluted bilirubin standards representing

solutions of 1.2, 2.4, 3-6, and 4.8 mg/dl, respectively, into

cups 2, 3, 4, and 5 of the sample carousel.

(2) Sequentially place diluted samples into the remaining cups

of the sample carousel.

(3) Place bilirubin reagent into reagt.i tips.

b. Place rotor on turntable.

c. Move sample and reagent probes into position, nrirl HRPTPM! «JI.P* I,

button.

d. Place loaded rotor into Analyzer and [.. ucbtnj uu J.mJicy i.«.<) iu

Analyzer Manual. -491-

Caleulations: a- Determine bilirubin concentration for each sample by correlating

final absorbance with bilirubin concentration.

M. Multi-Enzyme:Multi-Sample Analysis

Reagents: Calbiochem CPK, LDH-L, SGOT, and SGPT Reagent Kits: reconstitute each by adding 2 ml of H20 to Vial B, swirl to

dissolve, and add to contents of Vial A. Cap, and dissolve contents by gentle inversion.

Sample: blood serum

Wavelength: 340 nm

Temperature: 30°C

Loading volumes (sample-reagent loader):

a. Reagent pipette

Reagent pump, 2Q |J.l, at 100% = 20 Hi

Diluent pump, 200 Hi, at 25% = 50 Hi

b. Sample pipette

Sample pump, 20 Hi, at 50% = 10 Hi

Diluent pump, 50 Hi, at 100% = 50 h1

Procedr.e:

a. Set sample-reagent loader in multi-chemistry:multi-sample mode.

(1) Place aliquots of sample 1 in cups 2, 6, 10, and 14; aliquots

of sample 2 in cups 3, 7, 11, and 15; aliquots of sample 3 in

cups 4, 8, 12, and 16; and aliquots of sample 4 in cups 5, 9,

13, and 17 of the sample carousel, respectively.

(2) Place aliquots of LDH-L reagent in cups 2, 3, 4, and 5;

aliquots of SGPT reagent in cups 6, 7, 8, and 9; aliquots of CPK reagent in cups 10, 11, 12, and 13; and aliquots

of SGPT reagent in cups ±k, 15, 16, and 17 of the

reagent carousel, respectively. b. Place rotor on turntable. c. Move sample and reagent probes into position, and depress

start button. d. Place loaded rotor into Analyzer and proceed as indicated

in Analyzer Manual.

Calculation: a. Activity (I.U. liter-1 min"1) = ^r— x 4l8o. -493-

APPENDIX II

ON LINE COMPUTER FOR HIGH-RESOLUTION CHROMATOGRAPHY ' > *

APPENDIX II: ON LINE COMPUTER FOR HIGH-RESOLUTION CHROMATOGRAPHY"

A. Computer Operation

A general operating routine for the new computer systems is given "below.

1. Loading the baseline routine into the computer:

E A

L G 5

Adjust aperture screws on photometer so absorbance reading is > - 0.005-

2. Starting the data taking process on unit UN (UN = 1 or 2):

a. If the previous data-taking rate was not once every 20 sec:

T FRTE(20), FINI(UN),! (UN = 1 or 2).

b. If the previous rate was once every 20 sec:

T FINI(UN),.' (UN = 1 or 2).

The unit begins acquiring data once the return key is struck.

3. Stopping the data-taking process on unit UN:

T FTRM(UN),.' (UN = 1 or 2).

k. Starting the computer analysis of the data being acquired:

E A

L G 1

5. Stopping the computer analysis of the data before It is complete:

a. Unconditional halt:

CTRL/C [striking control (CTRL) key and C key simultaneously].

b. Conditional halt:

Turn unit 1 key switch on until motion light on cassette No. 0 is

activated; then turn it off. After the Teletype outputs a return -495-

instruction, the computer is available for general use. The

analysis program can be started at the exact point it was

terminated by doing a library return command (e.g., L R l).

6. Acquiring two chromatograms on the same tape for unit UN" (UN = 1 or 2).

Do not terminate after the first chromatographic analysis is complete.

When the injection valve is switched, initiating the second run, turn

the appropriate unit key switch (l or 2) on for 5 min; then turn it off.

This run is chromatogram No. 2 with respect to the data analysis program.

7. Removing the data tape and analyzing the data at a later date:

L U 1 (Library unload command)

Remove tape after motion light begins flashing.

Replace tape:

T FACC (UN),! (UN = 1 or 2)

E A

L G 1

B. FOCAL Programs

The computer programs written in the FOCAL Language are listed in Tables

II-l through II-3- -496-

Table II-l

EATA ANALYSIS PROGRAM w C-F0CAL.J3.B FLUID 31513W

01 • 01 A "NUMBER OF DATA POINTS TO BE ANALYZED"N,! 01 .02 A "UNIT NUMBER (1 OR 2>"UN,!,"CHANNEL NUMBER <1=254,2=280>"CH, I 01 .03 A "WHAT TIME IS THE COLUMN HEATER ACTIVATED?"TH,! 01 .04 A "DO YOU VISH PEAKSTRIP TO OPERATE? CYES OR NO>"ANS,!JG 1.56 01 .05 A "CHROMATOGRAM NUMBER CI OR 2)"CN,!,"PHOTOMETER PATHLENGTH"PL , ! 01 .06 A "VOLUME CF FLOW RATE WONLTOR"VS,!JI <1-CN>15.05%S BP=0;S BZ" 1 01 .11 S DL=.0002JS BL=.0055; S DX=.005556;S DR=.00004 01 .12 S DP=.000045S DA=FFET;I ,!!! 01 .16 S DY=1.0£S DQ=.008;S I=4;S N1=0JS M1=0JS VF=-1 01 .19 S A.1=0JS CJ=6*UN-5;S CJ=3-CHJS MI=-3JS BM=05S HT=0JS AE=05S K= 0 01 .20 F L= 1, 7.»D 125S Z(L)=FCHNCL-1 +BZ,CH) S S ZCL+105=FCHNCL-1+BZ,CJ> 01 .21 S YP=-FLOGCZ<4)/4000)/2.303;S YQ=-FLOG/4000>/2.303JS YL= YP 01 .22 s BZ=BZ+BP*153 01 .27 s D2=1.GJS DW=0IS A=0JS DK=DRJS DM=05S DF=0JS DG=0JS A2=0 01 .30 D 13SI C Y2 +DL-Y3)1.45 SI 1.33 01 .31 I (Y3-Y2)1.555 I CAE>1.5,1.36 01 .33 I <5-DG)1.6;S DG=DG+1JS DK=DK+.00004;S AE=0JD 1.65,'D 11JG 1.3 01 .36 S YL=Y3;S DK=DR;S A=0;S DG=0 01 .37 I CBL-FABS CY3-YP)51.43 01 .38 I *DX 01 .61 s KB=1;S YN=Y7iS DB=IJF J=L,3*KE+L5S XCCJ>=0JS YCCJ)=0 01 .62 s KE=1;S KF=2*KE5S YM=Y2;S A2=ALJS AE=0JS A1=05S RM=0;S UM=0 01 .65 s A=A+< CY3+Y2)/2-YP)*DX 01 .90 s D=FSTRC129+1-DB,Y3> 01 .93 s DD=C2*V2-VL-2*YL-Y2+2*Y3>/7 01 .94 s D3=FABSCDD);I 2.7 9 01 .95 s D4«FABS*S D5=FABSCD2+DD>;I CD3+D4-D5-.000005)2.75,2.75 1. * —ry I -

Table Ii-l (continued)

02.20 I CDW-l>2.22,2.22,2.25 02.22 S DW=DW+1JI CBM)2.79;S BM=-lJS BN=I-2JS Y0=Y1JG 2.79 02,25 I CXCCKF-l>>2.3,2.3,2.4 02.30 S XCCKF-1>=BN*DXJS YC=Y0 JG 2»71 02v40 I CXCCKF)>2.5,2.5,2.7 02.50 S XC=BN*DXJS YC=Y0JS BM=0JG 2.75 02.70 S KE=KE+1JS K=K+lJS KF=2*KEjS XCCKF-l>=BN*DXjS YCCKF-1>=Y0 02.71 S BM=0 02.75 S D$=DDJS BM=0JS DW=0 02.79 I CY7-YN)2.80;S YN=Y7 02.80 I CDM>3.12JI 2-83jS UM=0JS YM»Y3JG 3.1 02.83 I C4-UM32.84JS UM=UM+1JG 3.1 02.84 I *DXJS PV=M1*VS+LR*T1*DX 02.88 S KR=KE 03.10 D li;D 13JG 1.65 03.12 I CYO-YN>3.13jS R=0JG 3»15 03.13 S R= CYM-YP)/(YN-YQ) 03.15 I CYP+BL/2-Y3) 3 . 18 J I (Y3+DR/2-Y2>3.3JG 3.22 03.18 I CY3-Y2>3.3 03.21 S AE=1JS A1=A1+CCY3+Y2>/2-YP>*DX 03.22 I C2+2*DF-RM>3.35jS RM=RM+lJG 3.1 03.30 S RM=0jS AE=0JS A1=0JG 3.1 03.35 I CYP+PY-Y3>3.80 03.40 S ET=-1JS A3=CA+A2-A1>/PLJS DE=I-4JI CSR>3.45JI <1-KE>3.7 03.45 I C.0005-A3>3.65S K=K-KEjS YL=Y3JD 11JG 1.27 03.60 S K=K-KE+ljD 10 03.61 D 11JS YL=Y3JI 3.65J I CI+DX-TH-1.0>1.27 03.64 I (HT>1.27JS YP=Y3JS YQ=Y7jS HT=-1JG 1.27 03.65 D 1.27JD 13JG 1.6 03.70 I C-XCCKF>>3.80JS KE=KE-US K=K-1JI 3.45 03.80 I <3-KE>3.45>F L*KB,KEJS XBCL>=CXC(2*L>+XCC2*L-1>>/2 03.82 F L=KB,KEJS YB=CYCC2*L>+YCC2*L-1>>/2-YP 03.84 F L=KB,KEJS XRCL>eXC<2*KR>-XCC2*KR-l>JS XS=2/XRCL> t2 03.86 S YBCKH)=YM-YP;I CET>3.87JS KB=KE+1JS YM=Y3;S DM=0JS UM=0JG 3.3 03 . 87 S D=FSTR(1,K)J S D=FSTRC2,KE>JS D=FSTR(3,DB>JS D=FSTRC4,DE> 03.88 S D=FSTRC5,Ml*VS>JS D=FSTR(6,YP>JS T>=FSTRC7,LR> JS D=FSTRC8,Nl*DX> 03 . 89 F J=1,KE;.S D-FSTRC 10+J,XB( J> > JS D=FSTR(10+KE+J,YB(J>> 03.90 F J=1,KEJS D=FSTR(10+2*KE+d,XRCJ>>JS D=FSTR(10+3*KE+J,XS> 03.91 S D=FSTR(9,PL>JS D=FSTRC10,MI> 03.92 S ET=0JL S 1,0JG 5.01 03.93 S DA=FFET(UN,BP,0> J I C-DA>11.63 03.94 I C-FSTRC2>>3.97JS MI=FSTR<1> 03.95 T XI0.07j" ENVELOPE DATA",YK,PV,A3,!JG 3.61 03.97 S K=K-KE+ljD 10JG 3.61 10.02 S MI=MI+4JS D=FSTRC768+MI,A3>JS D=FSTRC769+MI,PV> J0.<*4 S D=FSTR<770+MI,R>JS D=FSTRC771+MI,LR> 10.10 T X10.07xK,YK,PV,R,LR,A3,! 10.15 T •» ",PS,DE*DX,YP,! -498-

Ta"ble II-l (continued)

11.02 I >11.055L S l,l5G 11.6 11.03 S DA=FFETCUN,BP,0>5I <-DAM 1.63 11.05 S 1 = 1 + 1 JfF L=1,65S ZCD = ZCL+1)5S ZCL+10>=ZCL+11) 51.10 I (N-I)11.5;i CBS+BP*153-I-2-BZ>11.4 11.15 $ BC=I+2+BZ-BP*153;S Z<7)=FCHNCBC,CH>JS ZC17>=FCHNCBC,CJ> 11.20 Z {FCI)11.25,11.355Q 11.25 I CVF5U.38JS VF=-l;S M1=M1+1JS L2=LR5S LR=VS/*DX5S N1=I 11.26 J CM1-1>11.38,11.3851 CFABSU .295S LR=L25S Ml=Ml-l5»3 11.3 11.29 S LR=L2JS M1=M1+1 11.30 T !,"PROBLEM ENCOUNTERED IN DETERMINING FLOW RATE",!!jR 11.35 3 VF=05R 11.38 H 11.40 I C156-BP>'11.5JS BP=BP+l 11.47 S DA=FFETCUN,BP,0 5 5S BS=FPRMC2>5I CDA> 1 ! = 73.-!! « 15, i I .61 11 .50 L 5 1, l;G 11 .55 11.55 Q II =655 T "TO CONTINUE, LIBRARY RETURN FILE #1",!5Q 11.61 I Cl-DA)11.65,11.63JT "PROGRAM MALFUNCTION",!50 11.63 T "TAPE TRANSFER ERROR",!5G 11.6 11.65 T "NO ADDITIONAL DATA AVAILABLE FROM THIS CHANNEL",!50 11.70 I (FCNI(0))11.475L S I,15G 11.6 11.75 S DA=FFET(UN,BP,0)51 C-DA>11.635G 11.47

12.05 I CBS-BZ-L+1512.105R 12.10 S BP=BP+15S DA=FFETCUN,BP,0)5S BS=FPRMC2)

13.10 S V2 =V15S VI=Y15S Y1=Y25S Y2=Y3 13.20 S B3=C-2*Z<1>+3*Z<2>+6*ZC3>+7*Z<4>+6*Z<5>+3*Z<6>-2*ZC7>>/21 13.30 S B7 = C-2*Z<11)+3+ZC12)+6*ZC13 > +7*Z{14)+6*Z(15)+3*Z(113.4055 B7=l 13.40 S Y3=-FLOGCB3/4000)/2.3035S Y7=-FL0GCB7/40005/2.303 15.05 S BP=-1JS CI=6*UN-6 15.10 S BP=BP+1 15.12 S DA=FFET15.95S BC=0 15.15 S BC=BC + 1JI (BS-BCJ15.151 CFCI(BC,CI>>15.15 15.20 I (BS-BC-12)15.215 1 15. 15,15.2550 15.22 D 15.151 CFCI(BC+12-BS,GI>>15.1255 BP=BP-1 15.25 S BZ=BC5G 1.11 15.90 T "START OF SECOND CHROMATOGRAM WAS NOT LOCATED",!50 -1+99-

Table II-2

PEAK STRIP PROGRAM y C-F0CAL.J3.B FLUID 31513V

05.01 S KE=FSTR(2);S KP=FSTRC1>-KE;S DB=FSTRC3>;S DE=FSTR(4>;S P=FSTR<9> 05.02 S Ml=FSTRC5);S YP=FSTR<6>;S LR=FSTR(7>jS Nl=FSTRC8>JS MI=FSTRO0) 05.03 F J=lxKEJ?S XB=FSTRC10+J)SS YBCJ)=FSTRC 10+KE+J) 25.04 F J=1,KEJS XRSS XSCJ>=FSTRC10+3*KE+J> 05.05 T %10.06,"PEAKSTRIP ROUTINE",!SS DX=.002778SS LM=0JS IC=0JS LC=0 05®06 S Z=FITR(;S KF=2*KE+15S KG=2*KE 05.07 I C30-DE+DB>5.08fS Z~l 05.08 S NP=3*KE;S MP=NP+lJF J=1,S^DE-DB+1SS CCCJ+Z-1>/Z3=FSTRCl28+J) 05.09 S UM=FITR<(DE-DB)/Z>+i;F J«1,UM;S D«FSTRC384+J,C> 05.11 D 10JF J«I,NP;F K=I,MP;S C(MP*+K>=0 05.12 F J=I,KE;.F K=I,KEJF L=DB,Z,DEJS XC=L*DX;D 6 05.14 F J«1,KE;S S1=0JF L=DB,Z,DE;S XC=L*DX;D 6.3;D 6.37;t> 8 05.17 F J=I,NP;F K=I,J;S CCMP*CJ-l)+K)=CCMP*CK-l>+J) £5-I9 F J=i,NPjS XC=C(MP*CJ-l)+J> 05.20 F J=I,NP;F K=1,NPJS C(MP*+K>=C+K>/FSQT(XCCJ>*XC=CCMP*J>/FSQTCXCCJ>) 05.27 I 5.29,5.29JF J=1,NPJS C(MP*CJ-1>+J>=C+LM 03.29 S LA=NPJF M=l,NP-lJF L=M,NP-l;F J=M,NPJD 7 05.30 D 10JS SU=0JS LL=LA+l;S YCCMP>=1 CS.32 F J=LL,MPJS SU=SU-CCMP*(LA-1>+J)*YCCJ) 05.33 I CFABSCC(MP*-0.00001)5.79,5.79 05.34 S YCCLA)=SU/CCMP*CLA-1>+LA) 05.36 S LA=LA-1 J I CLA)5.39,5.39,5.3 05.39 F L=1,NPJS YCCL)=YC(L)/FSOT 05.51 F L=1,KEJS XR=XR-2*YCCL>,*S XS=2/XRCL>t2 05 . 53 S S2=0;F L=DB,Z,DEJS XC=L*DXJD 8*US S2-S2 + CCF-CD> t2 05 .55 D 10JI (-LM55.75JI CS2-S135.56JG 5.79SS LM=0.000lJG 5.71 05.56 I (8-IC)5.73.yS IC=IC+1JI (LS-.00001 )5.6,5.6,5. 11 05.60 F L=1,KEJS A=1.2533*YBCL)*XRCL5/P;S PV=LR*,S D=FSTR<2,35;T "LEAST SQUARES SUM",S2,! 05.70 L R 1 05.71 F L=I»NPSS YC(L>=-YCCL> 05.72 D 5.5JD 5.51;G 5.11 05.73 T "EXCESSIVE ITERATIONS",!;G 5.6 05.75 I

Ta"ble II-l (continued)

06.20 S XK=FEXPC-CXC~XBCK>>t2*XSCK>>;S YK=2*YBCK>*CXC-XBCK>)*XSCK> 06.30 S XJ=FEXP(-(XC-XB(J>)»2*XS(J));S YJ=2*YBCJ>*CXC-XB(J>)*XS(J) 06.35 D 10;S SK=XK*YK*CXC-X3CK>>*2/XRCK> 06.37 S SJ=XJ*YJ*CXC-XB*2/XRCJ> 06.40 S CCMP*CJ-1)+K)=CCMP*CJ-15+K>+SK*Sj 06 . 50 S CCMP*C J-1>+K+KE>=CCMP*CJ-1 >+K+KE>+SJ*XK*YK 06.60 S CCMP*CJ+KE-1>+K+KE)=CCMP*CJ+KE-1>+K+KE)+YJ*XJ*YK*XK 06.70 S CCMP*CJ-1>+K+KG)=CCMP*CJ-l>+K+KG)+SJ*XK 06.80 S CCMP*CJ+KE-1)+K+KG)=CCMP*CJ+KE-1)+K+KG)+XJ*YJ*XK 06.90 S CCMP*CJ+KG-1 )+K+KG>=CCMP*CJ+KG-l>+K+KG)+XJ*XK

07.10 D 10J5 YA=C.CMP*L+M>*CCMP*CM-1 3+J+l )/CCMP*CM-l)+M> 07.20 S CCMP*L+J+1)=C-YA

08.10 S CD=0;S CF=FSTR(384+CL+Z-DB>/Z5-YP;F K-1,KE;D 6.25S CD=CD+YBCK>*K 08.15 D 105S Sl=Sl+CCF-GD)t2 08.20 S CCMP*J)=CCMP*J)+CCF-CD)*SJ 08.25 S CCMP*CJ+KE>)=CCMP*CJ+KE)>+CCF-CD)*YJ*XJ 08.30 S CCMP*CJ+KG >)=C CMP**XJ

09.10 T KP+L,PV,A,! 09.15 S MI=MI+45S D=FSTRC768+MI,A);S D=FSTRC769+MI,PV> 09.20 S D=FSTR(771+MI,LR)SS D=FSTRC770+MI,0>

J0.10 I CFCNIC0M10.155L S 0,0;G 5.80 10.15 R * -1+99-

Table II-2

PEAK IDENTIFICATION PROGRAM

V C-F0CAL.J3.B FLUID 31513V

0.1.02 A "NUMBER OF PEAKS"MT, ! , "URINE VOLUME LOADED ONTO COLUMN"VC,!!! 0JL.05 S MVC1 )-0;S MV(2)=64.l;S MV<3)=259.2;S MV<4)=497.5;S MV<5)s750.8 01.06 S ARC2_)=2.93JS AR<3)=0.46;S AR<4)=3.73;S AR<5)=0.24 0.1.07 S C<2)=0.3;S C<3)=0.03;S C<4)=0.15;S C<5)=0.02 01.10 S MI=-3SS L=2SS LM=.01 01.15 S MI=MI+4;i <4*MT-MI )1.50*S EV=FSTR<769+MI.)5 I 1.20; I <3-,! 01.63 T "HIPPURIC ACID ",MV<4>,! 01.64 T "P-CRESOL ",MV<5),« 01 .65 T ! ! 01.70 S JKC1 >=7,'S JK< 2 ) =4SS JK<3)=4;S JK<4)=8

02 . 05 S MI=MI+4; I C4*MT-MI)2.95JS EV=FSTR<769+MI5SI CEV)2.05,2.05 02.08 I (EV-CMVCL+l>5)2.13,2.13JS SF=0JS L=L+l;i 2.15JS SF=-1S;F K=1,JK;S N2.19/S C1=C1+1;S IN3 . 09, 3. 10.' I <1-12)3.11,3.12 02 . 32 I <1-14)3.13,3.14;.! <1-16)3.15,3.16;.! <1-18)3.17,3.18 02.45 I <1-52)3.51,3.52;I <1-54)3.53,3.54/1 <1-56)3.55,3.56 02.46 I <1-58)3.57,3.58;I <1-60)3.59,3.60;I <1-62)3.61,3.62 02.47 I <1-64)3.63,3.64;J <1-66)3.65,3.66;I <1-68^3.67,3.68 02.60 I <1-102)4.01,4.02;.! <1-104)4.03,4.04;! <1-106)4.05,4.06 02.61 I <1-108)4.07,4.08;I <1-110)4.09,4.10SI <1-112)4.11,4.12 02.62 I <1-114)4.13,4.14;I <1-116)4.15,4.16;! <1-118)4.17,4.18 02.75 I <1-152)4.51,4.52;!

Table II-3 (continued)

03.01 T ' 'PSEUD OUR IDINE " J S C=244/6.32JG 5.2 03.02 T "N-METHYL-4-PYRID0NE-3-CARB0XAMIDE"JS C=152/10.7JG 5.2 03.03 T "URACIL"jS C=112/7.432JG 5.2 03.04 T "5-ACETYLAMINO-6-AMINO-3-METHYLURACIL" J S C=198/9.57;G 5.2 03.05 T "N<15-METHYL-2-PYRID0NE-5-CARB0XAMIDE,,;S C=152/14.3JG 5.2 03.06 T "7-METHYLXANTHXNE";S C=166/6.57JG 5.2 03.07 T "HYP0XANTHINE"5S C=136/10.01JG 5.2 03.51 T "XANTHINE"JS C=152/6.W34JG 5.2 03.52 T "3-METHYLXANTHINE'»;S C = 166/6.27;G 5.2 03 . 53 T "l-METHYLXAtfTHINE"JS C=J.66/5 . 28 J G 5.2 03.54 T "URIC AClD"JS C=163/3.4;G 5.2

04.01 T "2-AMIN0-3-HYDR0XYBEN?.0YLGLYCINE";G 5.1 04.02 T "2-FUROYLGLYCINE"JG 5.! 04.03 T "5-HYDR0XYMETHYL-2-FUR0IC ACID"JS C=142/12.859JG 5.2 04.04 T "HIPPURIC AC ID"JS C=179/2.85JG 5.2 04.51 T "VANILLOYLGLYCINE"JG 5.1 04.52 T "3-HYDROXYPHENYLHYDRACRYLIC ACID"JG 5.1 04.53 T "HOMOVANILLIC ACID"JG 5-1 04.54 T "4-HYDROXYHIPPURIC ACID"JG 5.1 04 . 55 T "3-HYDROXYHIPPURIC ACID"JG 5.1 04.56 T "215-FURANDICARBQXYLIC ACIDMJS C=156/11.51JG 5.2 04.57 T "KYNURENIC ACID"JG 5.1 04.58 T "P-CRESOL"JG 5.1

05.10 T !*" ",ER,EV,FSTR<768+MI),!JG 2.25 05 . 20 S A=FSTRC768+MI5 JS R=FSTRC771+MI) 05.25 T !," ",ER,EV,A*R*C/VC#" UG/ML",!JG 2.25 #