National Measurement Laboratory 1979 Technical Highlights

Home , Glinka (crater)

National Measurement laboratory

1 Q7Q Technical Ly,y Highlights

U.S. DEPARTMENT OF COMMERCE National Bureau of Standards NBS-SP 572

MATIONAL BU&SAV Or STANDASUS LIBBART

JUL 8 1980

National Measurement laboratory

1979 Technical Highlights

U.S. DEPARTMENT OF COMMERCE Philip M. Klutznick, Secretary

Luther H. Hodges, Jr., Deputy Secretary

Jordan J. Baruch, Assistant Secretary for Productivity. Technology and Innovation NATIONAL BUREAU OF STANDARDS Ernest Ambler, Director

April 1980 .

National Bureau of Standards Special Publication 572 Nat. Bur. Stand. (U.S.), Spec. Publ. 572, 126 pages (April 1980)

CODEN: XNBSAV

Issued April 1980

U.S. GOVERNMENT PRINTING OFFICE WASHINGTON: 1980

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402

Price $4.25 (Add 25 percent for other than U.S. mailing ) Foreword

This report of technical highlights of the National Meas- urement Laboratory is the first in a series of annual reviews. It covers selected scientific accomplishments of the Laboratory for the calendar year 1979. A general Laboratory Overview as well as more specific Center and

Program Overviews set the framework for the individual technical reports. These reports represent the wide range of Laboratory activities, the breadth of which should be apparent from the overviews themselves. Future editions of this report will present different, but nevertheless representative, selections of technical highlights reporting research activities of the preceding year. : , Contents

National Measurement Laboratory Overview 1 Center for Thermodynamics and

Molecular Science Overview . Center for Absolute Physical Quantities Overview ... 5 39 Bond Energies and Chemical Reactivity Rydberg Values for X- and y-Rays 7 43 Richard D. Deslattes and Wing Tsang

\ Ernest G. Kessler, Jr. Dynamic Thermophysical Measurements 44

i Laser Cooling of Stored Ions 10 A red Cezairliyan

David J. Wineland Pressure as a Probe of Molecular Interactions 46

Observation of Pure Rotational Transitions in the HBr + Vern E. Bean ! Molecular Ion with Laser Magnetic Resonance .... 11 The Catalytic Methanation Reaction: A Source of Richard J. Saykally and Synthetic Natural Gas 47 Kenneth M. Evenson John T. Yates, Jr. Extension of Absolute Frequency Measurements David W. Goodman, to the Visible 13 Richard D. Kelley, and Donald A. Jennings, Theodore E. Madey F. Russell Petersen and Multiphoton Chemistry 49 Kenneth M. Evenson 8 John C. Stephenson Fine Structure Constant Determined to an Quantum Calculations of Excimer Molecules 52! 7 Accuracy of 1 in 10 14 Morris Krauss P. Thomas Olsen and Edwin R. Williams New Developments in the Evaluation of Thermochemical Data 53 i Superconducting Thermometric Fixed Point Devices . . 16 Donald D. Wagman, Robert J. Soulen, Jr. and David Garvin, James F. Schooley Vivian B. Parker, Improved Laser Test of the Isotrophy of Space 17 Richard H. Schunun, and John L. Hall and Alain Brillet J. Brian Pedley Overview 21 Center for Radiation Research Center for Analytical Chemistry Overview 57i 24 Radioactivity Standardization Laser Enhanced Ionization in Flames 59) Wilfrid B. Mann John C. Travis, Spectra of Highly-Ionized Atoms 25 Gregory C. Turk, Joseph Reader James R. DeVoe, and Peter K. Schenck Polarized Low Energy Electron Diffraction (PLEED) A Technique to Study Surface Magnetism 28 Measurement of Trace Elements in Solar-Cell Daniel T. Pierce and Silicon 61

Robert J. Celotta Richard M. Lindstrom, Ronald F. Fleming, and X-Ray Image Information Theory 29 Paul J. Paulsen Joseph W. Motz and Michael Danos Quantitative Analysis of Individual Trace-Level Organic Constituents in Alternate Fuels 63 New Studies of Electric Quadrupole (E2) Willie E. May and Excitation in Probes 31 Stephen Chesler William R. Dodge N. Electrochemical Glueballs and Quark States 34 Computer-Controlled System for Sydney Meshkov Electrode Kinetics Research 6-! Lawrence M. Doane and Quantum Basis for Visible Region Radiometric Richard A. Durst Standards 35 Jon Geist, High Resistance Junction Formation in House 6( Edward Zalenski, and Wiring Circuits with Aluminum Conductors . . Dale E. Newbury A . Russell Schaejer Discrimination of Natural From Anthropogenic Development of Test Procedures to Determine Calorific Carbonaceous Pollutants Through Mini- Value of Refuse and Refuse-Derived Fuels 107 Radiocarbon Measurements 68 Eugene Domalski and Joseph G. Berke Lloyd A . Currie and George A. Klouda Nuclear Fuel Assay Using Resonance Neutrons 108 Center for Materials Science Overview 75 Roald A. Schrack,

James W . Behrens, and

Erosion of Brittle Materials by Solid Particle Impact . . 79 Charles D. Bowman Sheldon M. Wiederhorn, Improved Uses of X rays for Nondestructive Bernard J. Hockey, Evaluation 110 A . William Ruff, and Masao Kuriyama, Lewis K. Ives William J. Boettinger and Atomistic Theories of Crack Growth and Fracture ... 81 Harold E. Burdette Edwin R. Fuller, Jr., High Accuracy Particle Size and Particle-Fluid Robb M. Thomson, and Interaction Measurements Using the Particle Brian R. Lawn i Doppler Shift Spectrometer 113 Neutron Diffraction Profile Analysis David S. Bright and and the Structures of Electronic Ceramics 84 Robert A. Fletcher

- Antonio Santoro and Robert S. Roth

Recent NBS Advances in Physical Properties Measurements at Ultra High Pressures 86 Stanley Block, Ronald G. Munro, and

Gasper J. Piermarini

Piezoelectricity and Pyroelectricity in Poly (vinyl id ene fluoride) 89 Martin G. Broadhurst and George T. Davis Electrochemical Noise Measurements: A New Measurement Technique for Diagnosis and Study of Localized Corrosion 92 Ugo Bertocci

Migration Models for Polymer Additives 94 Isaac C. Sanchez, Shu Sing Chang, and Leslie E. Smith

Office of the Deputy Director for Programs Overview 99

Load Cell Mass Comparator 102 Randall M. Schoonover

Interagency Cooperation in Mass Spectral Data Base Dissemination 103 Lewis H. Gevantman

Standard Reference Material for Antiepilepsy Drugs in Serum 104 Robert Schaffer and Dennis J. Reeder

National Measurement Laboratory Overview

* John D. Hoffman Ian R. Bartky BS was created by Congress in 1901 to meet Director Edmund A. N Dimarzio the needs of a growing Nation for a unified measurement Lois J. Frolen system. For more than 70 years the Bureau has con- Donald R. Johnson David T. Goldman tributed to advances in science and technology, the Deputy Director Charles M. Guttman Ruth A. Haines for Resources & growth of industry, and efficiency in the marketplace. Huie, III Operations Robert E. Michael J. Kurylo, In so doing, the Bureau has built a scientific facility and III staff that stand with the world's best. Rosemary S. Cary C. Gravatt, Jr. Maddock Director Deputy George P. Lamaze for Programs John Mandel The basic enabling legislation of 1901, as amended, Arthur O. establishes these main purposes and functions of the McCoubrey Arthur O. McCoubrey Hans J. Oser Bureau: Associate Director for Robert C. Paule Measurement Services Gaylon S. Ross, Sr. Joe D. Simmons • The custody, maintenance, and development of the Gilbert M. Ugiansky Robert W. Zwanzig national standards of measurement, and the pro- David T. Goldman vision of means and methods for making measure- Associate Director for * Staff as of December Long Range Planning 31, 1979 ments consistent with those standards, including the comparison of standards.

• The determination of physical constants and properties of materials, of great importance to scientific or manufacturing interests "and not to be obtained of sufficient accuracy elsewhere."

• The development of methods for testing materials, mechanisms, and structures, and the testing of materials, supplies, and equipment.

• Cooperation with other Government agencies and with John D. Hoffman is studying, by means private organizations in the establishment of of a polarizing optical microscope, a standard practices incorporated in codes and speci- polymer sample which has been crystal- fications. lized under controlled temperature conditions while co-workers, Gaylon S. Ross and Lois J. Frolen look on. This • Advisory service to Government agencies on scien- work is part of an ongoing project tific and technical problems. directed toward an understanding of the nature of crystals and the crystallization • process in industrially important polymers. Invention and development of devices to serve special needs of the Government.

1 NATIONAL MEASUREMENT LABORATORY

Thus, NBS has both unique and special functions in along disciplinary lines and a number of offices tha relation to the Nation's science and technology, and very direct the scientific programs by drawing on the com broad responsibilities as well. As part of the U.S. Depart- petences in the centers and divisions. The centers cove ment of Commerce, the Bureau carries out its mission in absolute physical quantities, radiation research, thermc a variety of ways. dynamic and molecular science, analytical chemistry and materials science.

The cornerstone of the NBS mission is to serve, for i the United States, as the authoritative source of accurate, The cross-cutting programs are nondestructive evalua compatible, and useful physical measurements and, fur- tion, environmental measurements, standard referenc ther, to ensure their international compatibility. As the materials, standard reference data, nuclear technology world's leading technological Nation, the United States recycled materials, and measurement services. Thes has a vital interest in ensuring that the world system of centers and programs will be described in more deta measurement is not only internationally compatible but later. is also sufficiently sophisticated to meet U.S. needs. NML has a staff of 1400. Approximately 80 percer The National Measurement Laboratory bears a large are scientists and engineers, of which approximately 6 part of the responsibility for fulfilling this mission. The percent have earned doctorate degrees. The NML has Laboratory was created in 1978 as a result of a major total fiscal year 1980 budget obligation of $75.4 millior

NBS reorganization. The following major objectives of which $45.8 million is directly appropriated func have been established to guide its programs. (STRS money), $29.6 million is in return for technkc services rendered.

1. Provide the central basis, within the United States of a complete and consistent set of physical and chemical One of the most important events for NBS in recer measurements. years was the recognition by the Office of Managemei

and Budget and the Congress of the need to widen th_- 2. Provide reference measurement methods and refer- Bureau's scientific base (competence building). Th ence materials so that measurements performed in diverse- National Measurement Laboratory has invested almo< locations for industrial, commercial, or regulatory appli- $3 million in new programs in a variety of discipline;

cations are consistent. In the area of surface science, a program has begun t characterize the structures and reactions of atoms an,

3. Assure that technological development is not hin- molecules adsorbed on surfaces. This area will als dered by a lack of accurate physical or chemical infor- investigate other surface phenomena, such as catalysis mation. In electrochemistry, a study has begun of the mechanism and kinetics of complex electrochemical oxidatior 4. Apply NML capabilities in the physical sciences reduction processes involving organic compounds. Th! and measurement technology to solving specific problems electrochemistry program will also develop methodologie of national concern. and detector systems for analyzing selected organic an' 1 organo-metallic compounds.

5. Provide outstanding comprehension and competence in selected fields of advanced scientific research A small angle neutron scattering system has bee with long-term relevance to areas of critical importance funded for the NBS reactor. This facility will be use to NBS and the Nation. to study the structure and properties of polymers an magnetic and biological materials. Methods will also t

To meet these objectives, the laboratory is managed developed to probe bulk microstructures for inhomc

in a matrix form, with five centers organized generally geneities affecting materials performance. Another ne 1 NATIONAL MEASUREMENT LABORATORY

ompetence area will investigate fundamental mechan- :sms of chemical and physical processes induced by inter- ction of ionizing radiation with matter. The results of

his research will be used to develop a sound basis for i-losimetry measurements and assist those concerned with he problem of the effects of ionizing radiation on bio- ogical systems.

I new program involving offsite work by NML sci- i A

-ntists is the use of the National Synchrotron Light Source

t Brookhaven. In a joint effort with the U.S. Naval ftesearch Laboratory, Bureau scientists will begin real- ime in situ studies on materials transformation, analysis f crystal structure, and studies of material properties and phase transitions via electronic bonding properties of toms and molecules in condensed matter. A new com- etence area has also been funded to develop a state-of- /he-art picosecond pulse laser for studying effects of ionizing radiation, structure of liquids, surface science, ,:hemical kinetics, and molecular and atomic spectro- copy.

: In addition to these new competence areas, NML has received Congressional funding for a materials durability .program. The need for this research was established by

| 1978 NBS-sponsored study on the economic effect of yietallic corrosion in the United States. Areas to be emphasized include localized corrosion, elastic plastic •racture, and wear.

J The center and program overviews will give more Retails on these new research areas as well as their many .listing programs. Following each of these overviews, ..^embers of the scientific staff will make brief technical

. presentations describing ongoing research efforts. While

..fjhese presentations cannot give a complete description

. -»f NML activities, they do exhibit the far ranging breadth

tf its research, ranging from the theoretical studies of uarks to quantitative analysis of trace-level organic onstituents in alternate fuels.

Center for Absolute Physical Quantities

Karl G. Kessler George W. Burns David J. Glaze he NBS Center for Absolute Physical Quantities is J. Pirro* James E. T Ralph P. Hudson Michael Di Gray the consistency and integrity Robert E. Edsinger Noboru Hironaka responsible for maintaining John P. Evans David A. Howe John W. Cooper of the United States' legal physical measurement stand- George T. Furukawa Wayne M. Itano* Mary C. Cutkosky Sandra C. Greer** Stephen Jarvis, Jr. ards and for assuring that they are compatible with Ugo Fano Leslie A. Guildner Donald A. Jennings Walter G. Schweitzer international standards. The Centers central mission is James T. Hall* James L. Jespersen Quantum Metrology James Hines, Jr. George Kamas to carry out this responsibility in accord with the needs Richard D. Deslattes Harold J. Hoge** Gregory M. Kielian James C. Houck Judah Levine of this Nation. The demands of our highly technical Anthony J. Burek Wilbur S. Hurst Lindon L. Lewis Kenneth C. Harvey society for measurement accuracy can no longer be met Raymond W. Richard W. Hyland Howard E. Machlan Hayward Robert S. Kaeser Robert J. Mahler entirely with artifact primary standards such as meter Charles H. Manney Albert Henins Billy W. Mangum Emma Mendez- bars and kilogram masses. Consequently, as a result of Helen K. Holt Walter Markus Harvey Marshak Quinones Ernest G. Kessler, Jr. research to which NBS has made important contributions, Kenneth E. John B. Milton Robert E. La Villa McCulloh Karl B. Persson most basic standards of measurement are or soon may be Gabriel G. Luther Paul H. E. Meijer F. Russell Petersen William C. Sauder based not on artifacts but on properties of nature such as Allan S. Risley Thomas M. Shay* Earl R. Pfeiffer Harmon H. Plumb Hugh G. Robinson light. James J. Snyder atomic transitions or the speed of Martin L. Reilly Earl W. Smith Electrical John L. Riddle Samuel R. Stein As standards have become more sophisticated, the Measurements and Charles L. Gregory J. Rosasco requirements for improved measurement methods have Standards Margaret G. Scroger Trembath, Jr. Barry N. Taylor Robert J. Sou4en, Jr. Fred L. Walls intensified and our understanding of the physical world Donald D. Thornton Marc A. Weiss Norman B. Belecki is Joseph S. Wells has deepened. Our physical measurement system Marvin E. Cage Charles R. Tilford Giovanni Craig T. Van David J. Wineland bound together by the laws of physics, a detailed under- Costabile** Degrift Quantum Physics standing of which is necessary to maintain an accurate, Robert D. Cutkosky Deborah Van Gordon H. Dunn Raymond L. Vechten* consistent system, conversely subjecting these laws Steven L. Baughcum* and Driscoll** Bernard E. Welch Earl C. Beaty Ronald F. Dziuba Sharrill D. Wood to test. Peter L. Bender Woodward G. Eicke, Length and Mass James E. Faller \n order to meet its primary mission, the Center con- Jr. Measurements and Alan C. Gallagher Bruce F. Field ducts a research program at the forefront of physics. Standards Gebbie George M. Free Katharine B. Ralph P. Hudson Sydney Geltman Forest K. Harris** That program involves measurements of increasingly John E. M. Louis B. Holdeman Richard S. Davis Goldsmith* high accuracy, made possible by capabilities developed Paul T. Olsen Marvin Kroll John L. Hall William D. Phillips Howard P. Layer through the determination of important physical con- David G. Hummer John Q. Shields James E. Taylor Stephen R. Leone stants and the study of fundamental physical phenomena. June E. Sims and Frequency Time L. Linsky Jaan Toots Jeffrey James A. Barnes Frank Magnotta* Frequently the measurement expertise of the Center is Thomas E. Wells Earl R. Mosburg, Jr. Earl S. Williams David W. Allan applied to important problems relating to other national David Norcross Edwin R. Williams Richard L. Barger W. E. Beehler Arthur V. Phelps issues, such as those involved in the development of Thomas J. Witt** Roger James C. Bergquist* Patricia L. advanced energy systems. Temperature Joseph V. Cateora Ruttenberg Measurements and George J. Collins Stephen J. Smith The most strongly mission-oriented programs of the Standards John Cooper David A. Van Baak* Center are concerned with measurement for James F. Schooley Dicky D. Davis Richard T. Weppner standards Vern E. Bean Robert E. Drullinger mass, length, time and frequency, temperature, and elec- Theodor H. Lyman B. Elwell *Post Doctoral Staff trical quantities. They are multifaceted efforts requiring: Benzinger** Kenneth M. Evenson "Research Associates William R. Bigge Mark Feldman and Guest Workers basic research; development of sensors, instrumentation, and methodology; critical evaluation and documentation of data; international comparison of standards; and numerous measurement services to the public, including training sessions in measurement procedures, measure-

Gabriel Luther readies the apparatus for measuring the ment assurance programs (MAPs), and instrument Newtonian gravitational constant. The change in the period calibrations. Other programs, which concentrate on ex- of a torsion pendulum is caused by the gravitational attraction ploratory research into the basic properties of matter, of two spherical tungsten masses located near the bottom of the picture. focus on areas such as atomic and nuclear spectroscopy,

5 NATIONAL MEASUREMENT LABORATORY

areas that presently offer the best prospects for advancing The laser will continue to be an important tool fo our metrological capabilities. standards research. Work is underway to improve stable

Recently, in response to some difficult measurement high-accuracy laser systems and to utilize laser systems ii challenges, the Center has: provided frequency standards studying atomic and molecular properties and in investi 13 gating the details of molecular interaction processes. / stable to 1 part in 10 for navigation, lunar ranging, and military targeting; developed relatively simple but new technique developed by NBS, laser magnetic reso accurate thermometer calibration devices for improving nance, has yielded high resolution information on th the reliability of clinical and medical diagnostics through- structure of excited molecules. Further development o out the country; harmonized optical x-ray and y-ray this technique will, for the first time, allow studies o absorption wavelength standards to help resolve a discord in muonic properties of molecules in excited states. physics experiments; used state-of-the-art laser technology Pioneering NBS work on the measurement of very hig] to test the validity of Einstein's theory of special rela- frequencies is leading to the redefinition of the meter ii tivity; and more accurately measured the gyromagnetic terms of the speed of light. The technique of frequenc ratio of the proton and the fine-structure constant. measurement has now been extended by NBS scientists Research to improve our measurement capability and working in collaboration with scientists at the Nationa extend our knowledge of fundamental constants and Research Council of Canada, into the visible regioi 14 processes is underway or planned in a variety of areas. where the measured frequency of light is 5 X 10 hertz Frequency, the physical quantity that can be measured Since the length of a light wave is equal to the speed o with the highest accuracy, can now define our time scale light divided by the frequency of the waves, a very precis 1? to within 1 part in 10 '. However, further improvement length scale can be established by measuring the fre is required for modern global navigation systems and for quency of a stabilized laser source. Because a value fo measurement of movements of the earth's crust. Re- the speed of light is agreed upon, accurate frequenc searchers in the Time and Frequency Division are explor- measurements can be used to define the fundamenta ing new ideas for improved frequency standards and for unit of length, the meter, in terms of the speed of light. methods to coordinate time worldwide at an accuracy of A fundamental phenomenon of physics, the Josephso; a few billionths of a second. effect, has been used by the Electrical Measurements an< Recent work in the Center has established an accurate Standards Division to establish a new basis for accurat wavelength scale for x-ray and gamma-ray measurements. voltage measurement. Based on the Josephson effect That work is being extended to both lower energies and voltage can be measured in terms of frequency. Smal higher energies, with potential applications to atomic and devices incorporating this capability are now used in NB! nuclear spectroscopy. Until now, the main applications of laboratories to maintain the U.S. legal volt. They replac this new scale have been redeterminations of the scales banks of standard cells used in the past. Development for muonic and pionic x-ray spectra. Other work in the now under way will make these devices, Center has led to a proposed NBS/Naval Research Lab- called Josephsoi junctions, widely applicable, oratory research program using the synchrotron light more especially in industry source construction at now under Brookhaven. This effort Modern instrumentation capable of high accurac; will emphasize application of the synchrotron source to measurements must not only be automated, but fre descriptive materials science while also providing oppor- quently must also be capable of internal calibration tunities for further studies in scattering and spectroscopy. Studies planned in the Center will investigate basic physi In the temperature measurement area, current work is cal phenomena that can be utilized to develop self-cali establishing a firm foundation for the International Prac- brating devices. Such devices, when perfected, will mab tical Temperature Scale. The improvements in that scale it possible to maintain maximum accuracy in field meas and its extension to lower temperatures have been to a uring instruments without frequent calibrations by NBS large extent the result of NBS efforts. The future NBS program in thermometry will concentrate on the develop- The thrust of all Center programs is to build a stronger ment of methods for dynamic measurements on transient, more accurate foundation for physical measurements anc non-equilibrium systems such as those encountered in to improve understanding of the basic phenomena o; new energy system. which our physical measurement standards are based.

6 4 CENTER FOR ABSOLUTE PHYSICAL QUANTITIES

Rydberg Values for X- and y-Rays

Richard D. Deslattes and Ernest G. Kessler, Jr. Quantum Metrology Group

w,e have significantly im- using it as an invariant basis for the proved the accuracy with which x-ray establishment of secondary standards. and y-ray transitions are known in Today's best knowledge of the Bal- Source terms of the Rydberg constant, R mer « region of atomic hydrogen is These improved connections have derived from linking (by means of

sharpened critical comparisons be- Doppler-free spectroscopy) its non- tween theory and experiment for both overlapping components to certain normal and exotic atoms. In the case Doppler-free absorption lines in molec- of x-ray spectra from normal atoms, ular iodine. Of particular interest are our measurements have helped reveal those measurements carried out with

provocatively simple regularities in the respect to I2 absorption features within departures of experimental results from the 633 nm HeNe laser gain curve. the predictions of relativistic self- (Subsequent improvement in precision

consistent field calculations. Our y-ray of measurement with respect to I2 ab-

measurements contribute indirect in- sorption in another region is not of formation on the atomic x-ray problem significance for the present effort in just mentioned but, perhaps more im- view of limitations which are, for the portantly, serve to renormalize many most part, intrinsic.) spectra from muonic, pionic, and Our work therefore took as a token kaonic atoms. These measurements for R a visible 633 nm HeNe laser

have yielded enhanced tests of quan- locked to an I_. absorption feature tum electrodynamics in muonic spectra closely related to that used in the Bal- and improved particle masses in the mer a spectroscopy mentioned above. case of certain pionic and kaonic x-ray We used this source to establish a scale spectra. with respect to which the spatial period of an almost perfect silicon crystal was determined using x-ray interferometry Our strategy in this work followed along a [110] lattice direction. This the pattern of establishing secondary procedure produced a more accurate spectroscopic standards, well known (~ 0.1 ppm) local standard in the 0.2 in other spectroscopic regions. Al- nm (2 A) region than is otherwise though the measurement chain we available via classical x-ray methods. developed links the y-ray and x-ray transitions to the base units of wave- Figure 1. Schematic features of the measure length and frequency quite well, the ment chain established for linking visible main scientific interest derives from and 7-ray (or x-ray) wavelengths. Part a their linkage to the Rydberg constant, shows the x-ray/optical interferometer used Ry.. Specifically, one extracts from to determine a particular lattice spacing the theory of the hydrogen atom a relative to a visible standard (see text). Part b suggests the principle involved in description of the visible Balmer a comparing lattice spacings of various speci- spectroscopic complex. Since the theo- mens so as to reduce the effect of x-ray line

retical term differences are propor- uncertainty on the comparison. Part c is a tional to R R might be said to diagram of a classical transmission two- x , rx crystal instrument for x-ray and 7-ray wave- set the scale of the H-atom were it length measurement. Our realization of this otherwise known or R^ may be de- system provides for absolute angle measure- rived from the scale of the H-atom as ment with high precision thereby establishing ( + ) Detector (-) is presently the case. All other atomic (through parts b and a) direct connection to calculations are also homogeneous in certain visible reference wavelengths and through these to R the manifest importance of Figure 1 R y , hence By utilizing non-dispersive techniques less crucial to the quality of inter- ± 2.5° and ± 15° in first and second we were able to transfer the above crystal comparisons than they would generation instruments respectively. calibration to other crystals of Si and be in the case that we had to use fully Perfect crystal resolutions, typically 20

Ge having appropriate shapes and ori- dispersive techniques. It is also clear to 50 ppm, provide a basis for ade- entation for x-ray and y-ray diffraction that the accuracy required of angle quate pointing precision provided sta- measurements. The non-dispersive (or measurement schemes is significantly tistically required source strengths slightly-dispersive) character of these less in this approach than it would be (~ 10 to 100 Ci) are available. Ac- transfer measurements insured that the in a fully dispersive area. This step is curacy, on the other hand, requires that results were not degraded due to the not a limiting one at the 0.1 ppm level. the interferometric angle be calibrated width and shape of x-ray line sources Most recent work has focused on in terms of fractions of the cyclic angle, used. Finally, pairs of crystals cali- the third step, figure lc, wherein pairs 2tt. To this end, we have employed brated in this way were used in double- of crystals calibrated as just described optical polygons and sensitive nulling crystal x-ray and y-ray spectrometers are placed in the symmetrical trans- autocollimators which furnish indexing with accurate means for diffraction mission geometry shown. The interval precision well below 1 milli arc-sec. If angle measurement. These three steps between the two diffraction conditions Angles between all adjacent face pairs then relate in a rather direct way the indicated is twice the Bragg angle, 6, of a polygon sum to 2n regardless of i; x-ray and/or y-ray lines measured in where the wavelength to crystal-spac- the polygon's in-plane errors and with the last part to the I_. stabilized HeNe ing ratio, A/d is simply 2sin#. Evi- calculable (small) effects from out-of- lasers used in the first part and through dently, accuracy available for A follows plane errors. The process of marking them to the fundamental scale unit. that associated with d, provided mea- all external angles constitutes a calibra-

Rod. We proceed at this point to surement precision is adequate and tion; such calibrations have been estab- comment briefly on the three-step provided the scale for 0 is established lished with a precision of about 0.05 process. on an absolute basis. Measurement ppm but require periodic repetitions to

The overall measurement chain is precision has been obtained by means track secular changes in interferometer suggested by the three diagrams of of polarization encoded angle inter- geometry. Thus the calibration con-

1 y-ray figure . Figure la shows a diagram of ferometry which yields a typical sensi- stant operant during a particular the combined x-ray and optical inter- tivity of 0.05 milli arcsec for ranges of or x-ray diffraction measurement is in- ferometry of the Si 220 repeat distance. Our procedure yielded (by means of observations at a set of optical lock Ca Ba Sn Tl Pb Ba points) an effective number of x-ray 3d-2p 5g-4f 4f-3d 5g-4f 5a-4f 4f-3d intervals transited in a scan of approx- 20 - imately 100 optical half-wavelengths. 15 [- For the particular Si specimen used, T Q. 5 eV the summary of several hundred meas- CD 10 77, urements was d 22o(Si, 25 °C) = 192.01707 (0.1 ppm). adja- pm An JL. 0 ±M 7^<^M%.V Wfc 7. M 77777Z cent specimen was then used in a -5 z -7M2> A geometry such as that illustrated in -10 figure to calibrate UJ lb other specimens -15 and, indeed, other crystal types. Evi- -20 I 420.8 434.5 437.6 dently, the procedure suggested in 157.5 ^ 200.6 ^ 347.6347 fi 'V 'V \r figure lb gives a null result when E (keV) standard and unknown crystals agree, muonic independent of x-ray wavelength. In Figure 2. Comparison between theory and experiment for a class of outer orbit x-ray transitions. Hatched bands represent estimates of theoretical uncertainties and data the case of a small difference, one can have been renormalized as required by new 7-ray data arising from the measurement chain adopt a perturbation type approach in of figure 1. In the absence of this revised energy scale, hatched bands would have centered which it becomes rapidly evident that on horizontal dashed line segments shown. The resulting discrepancy pattern would have properties of the x-ray line used are been small but significant.

8 ENTER FOR ABSOLUTE PHYSICAL QUANTITIES

terpolated from a "curve of growth" including prior and subsequent calibration results. We feel that this process ' 3 -cf Sb Pu may be carried out with errors not Tn- | Mo -^-O"' greater than 0.1 ppm.

The above-described procedures Cu have been by now applied to frequently 10 v used y-ray secondary standards avail- r (E [keV]) able from the isotopes 198Au and 192Ir 170 [1] along with those from Tm and 169 -6 - Yb [2]. In separate exercises the Th Pu same technology has been applied to -4 - Sb the x-ray while related W Kai line [3], -2 Mo procedures were previously applied to C'J *Ag c- Ru Cu Kai and Mo K«i [4]. By also uti- 10 lizing otherwise available high preci- (E [keV])% sion results on ratios of K-series x-ray lines to a y-ray reference [5] together Figure 3. Comparison of experimental data with our measurement of the y-ray as obtained or modified by our work with same time, we have undertaken to relativistic self-consistent field calculations in broaden the available data base by a reference [1], it has been possible to the relaxed orbital limit. Part a of this appreciably expand this atomic x-ray systematic study of these discrepancies figure shows results for Kn, lines while part b as functions of principal and orbital data base [6]. In general, our x-ray shows those for Kq-. Further results presently results are significantly poorer owing to being accumulated show similar patterns for quantum numbers as well as Z. K/3 lines. Several the great spectral widths of these lines. the case of more prominent avenues suggest themselves for future work. will On the other hand, both direct and in- These be pursued direct measurements of these lines have as resources permit. Obviously, one the quality of this comparison is no wishes to extend the optically con- permitted us to approach the 1 ppm level. longer limited by lack of knowledge of trolled scale to higher energies. This is

the reference y-ray lines. particularly interesting as it becomes Two examples illustrating applica- As a second application area, we possible to reckon a y-ray cascade sum- tions of this work are shown in figures turn to x-ray spectra of normal atoms. ming to the order of AE = 10 MeV. 2 and 3. In figure 2 we show a com- past several years, relativistic In this case, the corresponding mass parison between theory and experiment Over the self-consistent field calculations have decrement, AM is measurable to ~ 1 for a class of muonic x-ray transitions become available [9]. These have been ppm. This combination of AE and the in moderate-to-high Z elements. This applied to the case of single inner shell associated AM is sufficient [12] to comparison is a significant test of vacancies in the relaxed-orbital limit establish a value for Faraday's constant quantum electrodynamics, especially [10]. When such "term-value" calcula- which is manifestly independent of the vacuum polarization terms, which tions are available with sufficient pre- chemical procedures. Significant nu- had previously showm significant dis- cision, predictions of emission transi- clear spectroscopy near 2 MeV awaits crepancies [7]. Such a comparison is tions follow from subtraction of terms. application of the high resolution capa- naturally sensitive to the numerical Comparisons between theory and ex- bilities of our instrumentation. A pre- values of the y-ray scale [8]. Were we periment for the Kd lines of several cision measurement of the electron- to neglect the scale the change then elements are shown in figure 3 [11]. positron annihilation radiation has been comparison points would lie on the The systematic discrepancies shown largely developed but remains to be horizontal segments appearing in figure (which are evidently linear in Z) were executed. Finally, the progress to date 2. These results are satisfying in two not evident in previous data because in atomic x-ray spectra suggests the ways: evidently we have now agree- of their lesser precision. Efforts toward utility and interest of considerably ex- ment between theory and experiment understanding this situation have begun panding their study which we propose in an area of some significance; and, on the theoretical side while, at the to undertake.

9 NATIONAL MEASUREMENT LABORATORY

References Laser Cooling of

A more complete description of this work Stored Ions is in press, Annals of Physics (1980). David J. Wineland,

[I] Kessler, E. G., Jr., Deslattes, R. D., Time and Frequency Division Henins, A., and Sauder, W. C, Phys. Rev. Lett. 40, 171 (1978).

[2] Kessler, E. G., Jr., Jacobs, L., Schwitz, W., and Deslattes, R. D., T T e have been able to reduce in a fundamental way—by substantially Nucl. I n.st r. and Meth. 160, 435 w + lowering the temperature of atoms. the temperature of a sample of Mg (1979). ions to a value <0.5 K by using The "atoms" in our experiments Kessler, Jr., Deslattes, R. D., ions confined [3 1 E. G., radiation pressure. This technique will were actually atomic Mg and Henins, A., Phys. Rev. A 19, allow us to reduce in a fundamental in a room temperature electromagnetic 215 (1979). way the present limits to accuracy in "trap" which in turn was contained in frequency standards and precision a high vacuum. (The use of ion traps [4] Deslattes, R. D., and Henins, A., has another significant advantage in that P/iyy. Rev. Lett. 31, 972 (1973). spectroscopy, namely those imposed by the first and second order Doppler the confining forces are benign and do [5] Borchert, G. L., Z. Naturforsch. 31a, shifts. When combined with first- not cause significant frequency shifts 102 (1976). order Doppler cancellation schemes, ( < 1 0 ) . The cooling principle can first noting that the [61 Deslattes, R. D., Kessler, E. G., Jr., the method should yield orders of be understood by Jacobs, L., and Schwitz, W., Phys. magnitude improvement over present ions are harmonically bound to the oscillate Lett. 71 A, 41 1 (1979). techniques. center of the trap, hence they back and forth with an amplitude and [7] Rafelski, J., Muller, B., SofTard, G., In spite of the great accuracy and velocity governed by the temperature. and Greiner, W., Ann. Phys. 88, 412 stability achieved by today's frequency We can irradiate the Mg ions with (1974); Watson, P. J. S., and Sun- standards, the need for improved monochromatic light from a frequency- daresan, M. K., Canad. J. Phys. 52, in both 2037 (1974). standards continues to grow doubled dye laser whose frequency is the pure and applied sciences. The tuned lower than the ion rest frequency [8] Hargrove, C. K., Hincks. E. P., Mc- present uncertainty in the cesium beam but within the Doppler profile. The Kee, R. J., Mes, H., Carter, A. L., frequency standard which defines the ions will preferentially scatter photons Dixit, M. S., Kessler, D., Wad- length of the second is about 1 part in when they are moving toward the laser, den, J. S., Anderson, H. L., and 13 10 . Interesting tests of various gravi- and hence Doppler shift it into reso- Zehnder, A., Phys. Rev. Lett. 39, tational theories require clocks which nance. For each scattering event, the 307 ( 1977); Dubler, T., Kaeser, K., are several orders of magnitude more momentum of the ion is changed by Robert-Tissot, B., Schaller, L. A., accurate. In addition, substantially the of the absorbed photon. Schellenberg, L.. and Schneuwly, H., momentum 4 Nucl. Phys. A 294, 397 (1978). better clocks are needed, for example, After about lO scattering events, sub- in long baseline interferometry and in stantial cooling below room temper- [9] Desclaux, J. P., Atomic Data and various navigation systems. ature can be achieved. Figure 1 shows Nuclear Data Tables 72,311 (1973). Doppler shifts are the dominant a plot of temperature versus time. The

[101 Huang, K-N., Aoyagi, M., Chen, cause of uncertainty in the cesium temperature was monitored by observ- M. H., Crasemann, B., and Mark, H., standard and other existing or pro- ing the induced currents in the trap Atomic Data and Nuclear Data posed frequency standards. When in- electrodes. This experiment* showed Tables 18, 243 (1976). terrogating the frequency of an atomic that, although cooling could be ob- transition (which provides the basis tained, resolution was limited by noise [II] Deslattes, R. D., Jacobs, L., Kessler, for the standard) we must use some in the detection electronics. A more E. G., Jr., and Schwitz, W., Com- method to cancel first-order Doppler sensitive measu/e of temperature could parison of Relativistic Atomic SCF so is Calculations with Improved Experi- shifts. However, our ability to do be obtained if we monitored the Dop- mental Data (volume of papers done limited, and residual shifts always pler width of the optical transition. in honor of Professor Y. Cauchois occur. More importantly, we are al- This was complicated because under upon her retirement). ways required to account for the the optimum cooling conditions that second-order Doppler effect, or rela- were used (small'compact ion cloud, [12] Deslattes, R. D., and Kessler, E. G., tivistic time dilation shift of the moving focused laser beam) cooling rates and Jr., Precision Gamma- and X-ray atoms. In the Time and Frequency heating rates (as the laser was tuned Energies (to be published in Pro- Division, a group including J. C. Berg- through resonance) could as high ceedings of Sixth International Con- be ference on Atomic Masses (AMCO- quist, R. E. Drullinger, Wayne M. as 1000 K/s. Hence a second, fixed- 6) held September 18-20, 1979, in Itano, F. L. Walls, and D. J. Wineland frequency laser was used to keep the East Lansing, Michigan). has attacked the Doppler shift problem ions cold while the scattered light from

10 CENTER FOR ABSOLUTE PHYSICAL QUANTITIES

Observation of Pure Rotational + Transitions in the HBr Molecular Ion with Laser Magnetic Resonance

1 Richard J. Saykally and Kenneth M. Evenson, Time and Frequency Division

Laser on

JL T M olecular ions are extremely 5.6 min M I- important in the upper atmosphere Ions removed and are thought to dominate the chemical dynamics of the more dense Laser off gas clouds found in the interstellar medium. To study ions in these environments by spectroscopic meth- Temperature Time -350 K ods, such as radio or far-infrared astronomy, accurate laboratory spectra

must first be obtained. Although Figure 1. Ion temperature versus time when laser cooling is applied for fixed detuning of about 2 GHz. The ions were initially heated above equilibrium temperature with the optical emission spectra have been laser. Laser cooling was then applied for a fixed time until a temperature approaching 0 K observed for a fairly large number of was achieved. After the laser is turned off, the ions rethermalize to the ambient temperature molecular ions, the high-resolution owing to collisions with background gas. techniques required to yield the desired

level of accuracy (e.g., microwave the frequency-swept laser was moni- ties in atomic clocks should be even- spectroscopy, EPR, rf-optical double tored. Figure 2 shows a resulting trace tually lowered by several orders of resonance and, recently, Doppler-tuned where the ion temperature was about magnitude. laser spectroscopy) have been success- 2.5 K. Similar traces now indicate that fully applied to these species in onlv a temperatures < 0.5 K can be obtained. few cases. The difficulty is simply that The limit on temperature achievable of producing a detectable number of by this method is governed by the ions by a means compatible with the effects of recoil. For example, if the respective detection systems. We have ion comes to rest when the "last" cr recently observed the pure rotational photon is absorbed, it will attain the spectrum of a molecular ion HBr + — o — recoil energy of the photon upon re- u by the technique of laser magnetic c emission. We have studied this prob- o resonance (LMR). This first detection o lem theoretically and find that, for Mgr of is significant CL an ion by LMR ,ions, where the electric dipole transition because it makes the entire class of (is fully allowed, the limit on tempera- paramagnetic ions amenable to study. ture is about 1 mK. The cooling limit

on less strongly allowed transitions,

however, can be < < 1 ^K. Even with Within the last few years, LMR has the presently-attainable temperature, emerged as a powerful technique for we should be able to reduce the second- the study of pure rotational spectra of order Doppler shift correction in very Figure 2. Photon scatter rate versus laser transient species in the gas phase. It frequency showing the three isotopes of Mg- 'high resolution studies (e.g., frequency exhibits a considerably higher sensitiv- in the trap. The sample is maintained cold s:andard) by about three orders of ity than that of competing techniques, during the scan by irradiating the low fre- magnitude. extremely weak re- such as microwave spectroscopy and The quency side of the strong line with a second laxation of cold ions in a trap will pro- laser. From left to right the three lines are EPR, while the resolution attainable is duce line-widths that are limited only due to the roughly the same. The LMR method (3s2S = 1/2) by the natural lifetime, implying , M) ^ itself is intimately related to gas-phase 1/2 15 Q 10 . Used in conjunction with = -3/2) EPR. The appropriate paramagnetic > (3p2p , Mj 3/2 established techniques (confinement to energy levels of an absorbing sample 24 20 transitions in the Mg, ^Mg, and Mg iso- dimensions less than a wavelength, are tuned by a dc magnetic field until topes. At 300 K the Doppler width is 2 photon Doppler-free spectroscopy, approximately equal to the spacing between their difference frequency equals that 24 26 saturated absorption, etc.) uncertain- the Mg and Mg isotopes ( = 3 GHz). of a fixed-frequency source. The prin-

11 i ,

NATIONAL MEASUREMENT LABORATORY

A= -79.494 THz cipal distinction between these two near 50 W transversely pumps the gain B = 0.23845 THz methods is: in EPR the relevant transi- cell of a FIR laser. The FIR radiation J = 9/2 5 - tion is between different magnetic sub- oscillates in the optical cavity defined r + = levels (Mj) of the same angular mo- by a fixed mirror at one end and a 280 MHz J 5/; 4 - mentum state (J), typically occurring movable mirror attached to a microm- 140.3 urn in the microwave region for normally eter drive at the other. The detection j = 3/; accessible laboratory magnetic fields volume of the cavity is separated from :I J = 7/2 (2 tesla) ; in LMR the relevant transi- the gain cell by a thin polypropylene — = p=TJ+ J 1/: tions are between rotational states (in window set at the Brewster angle. A - 180.3 [im Q = V2\Q\1/2 -g.-cM 24 GH. molecules) or fine-structure levels (in fraction of the FIR laser power is 79.26 THz atoms), and occur in the far-infrared coupled out of the cavity and into a + J = 5/2

( FIR) . In both techniques, the sample liquid-helium cooled germanium bol- 252.3 urn is contained inside a resonant cavity. ometer by an adjustable 45° coupling ±J = 3/2 * The increased sensitivity of LMR is mirror. (X100) mainly derived from operating at fre- In the apparatus shown in figure 1 fi = 3/2 qencies about 100 times higher than the detection volume is centered be- 2 Figure 2. Energy level diagram for the II the microwaves normally used in EPR. tween the pole faces of a 38 cm electro- + ground state of HBr . The small lambda and from the fact that absorption co- magnet. The magnetic field is modu- doubling in the fi = 3/2 state (case A efficients typically vary with the square lated at frequencies near 5 kHz, and notation) is multiplied by 100; that in the diamagnetic fi = 1/2 state is to scale. or cube of frequency in this region. phase-sensitive detection is employed, Detection limits for OH radicals resulting in a first-derivative line-shape. pendicular to the plane of the diagram. are roughly as follows: LMR: 1 X Until very recently, all of the spe- ,; " 12 :! In the new spectrometer used to ob- 10 cm and EPR: 2 X 10 cnr . cies observed by LMR had been pro- + flame has re- A schematic diagram of an optically- duced in various types of atomic and serve HBr , the been pumped FIR laser magnetic resonance molecular flames reacting inside the placed with the positive column of a glow discharge, and the magnetic field spectrometer is shown in figure 1 . In detection volume. These flames were this system, a tunable CCX laser oper- generated by using a concentric flow is instead produced by a 5-cm diameter ating single-mode with a power output tube mounted on the laser system per- by 33-cm-long liquid nitrogen cooled solenoid magnet, which car} generate FAR-INFRARED LASER fields up to 0.6 T (6 kG). This new^ 1.15m configuration provides for an increased Bolometer — detection volume, and the axial sym- Detector Fixed Moveable metry readily accommodates the elec- Mirror Mirror trical discharge. With this apparatus, WAVELENGTH RANGE we have succeeded in detecting several FAR-IR: 40-900 Mm short-lived species, including ground- state oxygen atoms and metastable ex-

Ins cited electronic states of 0 2 and CO, Diaphragm as well as the HBr+ ion. HBr+ was generated in a discharge cGrating through 1 percent HBr in helium at Micrometer 133 Pa total pressure and near-ambient Drive temperatures. Mirror We observed the two m - M i rror lowest rotational transitions in the Piezoelectric 2 Q spin-component of its II 2.60 m Drive = 3/2 ground electronic state, for which the CO LASER 2 enegy-levels are shown in figure 2. Bromine occurs in nature with nearly

Figure 1. Schematic diagram of an optically-pumped laser magnetic resonance spectrometer. equal abundances of the 79 and 81

12 CENTER FOR ABSOLUTE PHYSICAL QUANTITIES

Extension of Absolute Frequency Measurements to the Visible

Donald A. Jennings, F. Russell Peter- sen, and Kenneth M. Evenson, Time and Frequency Division

measure- T. he direct frequency H (Tesla) ment of visible light was achieved by .250 .350 .450 D. A. Jennings, F. R. Petersen, and K. M. Evenson of the National Bureau of Standards and K. M. Baird and G. Hanes of the National Research Council, Ottawa, Canada. This experiment paves the way for frequency- based spectroscopic investigations in the important visible spectral region which are potentially a thousand times more accurate than measurements of wavelength. In addition, this experiment may eventually lead to a new definition of the meter.

Visible frequency measurements were achieved by a laser frequency synthesis chain, shown in a very sim-

plified way in figure 1. This experiment links the cesium frequency stand- 79 ard directly with the frequencies of

m I the fre- u 10 hyperfine lines of 2 at quency of the second harmonic of the laser. measure- + 260 THz He-Ne The Figure 3. The laser magnetic resonance structure of HBr , as well as precise ments were made in seven steps from spectrum of the J = 3/2 -* 5/2 transition of values for the rotational constants and T HBr observed at 251.1 M m (CH..OH). This cesium to iodine by using only fre- ; lambda doubling parameters. These, spectrum was recorded on a single scan with quency counting techniques. Actually, in turn, can be used to deduce spin a 0.1 s time constant. The positions of the 14 lasers and 6 klystrons were used in densities, distributions, ac- eight doublets represent the hyperfine split- charge and the heterodyning experiments, finally ting in the Mj = —3/2 -> —1/2 Zeeman curate bond lengths for the molecule. culminating in the 520 THz visible transition. The doublet splitting is due to It will be very interesting and informa- frequency. The measurements of the :the lambda doubling, tive to observe the effects of the net first six steps were made at NBS and j positive charge on these properties by the final one at NRC. (The NBS 260 comparing them with properties of cor- mass isotopes, and the reduced masses THz laser was transported to Ottawa responding neutral molecules. More •of these two forms of HBr+ are nearly for comparison with NRC's 260/520 exciting, however, is the prospect of identical; thus we observe both isotopes THz composite laser in the final step.) detecting other similar ions by the same simultaneously. Since each Br isotope The schematic illustrating step 7 is method. In particular, HC1+ may be ^has a nuclear spin of 3/2, we find a shown in figure 2. of considerable astrophysical interest. pattern of two sets of four lines, addi- The NRC composite laser contained Measurement of LMR spectra of this tionally split into doublets by the small the iodine cell, the lithium niobate and other ionic species will provide the , ''lambda-doubling" interaction. The doubling crystal, and the 260 THz gain accurate rest frequencies required to additional hyperfine splitting from the tube. The 260 THz frequency was perform astronomical searches for proton is too small to resolve. This stabilized to hyperfine components of these molecules in interstellar clouds. 127 pattern is shown in figure 3. the P(62) 17-1 transition in T. by When completed, the analysis of using saturated absorption techniques, these will LMR spectra produce the 1 Department of Chemistry, University of and the 520 THz radiation generated first information on the hyperfine California, Berkeley, California 94720. from the 260 THz by the doubling

13 NATIONAL MEASUREMENT LABORATORY

Fine Structure Constant Determined to an Accuracy of 1 in 10 7

P. Thomas Olsen and Edwin R. Williams, Electrical Measurements and Standards Division

Frequency, THz Work is in progress to reduce the large uncertainty.

' 520.0 2 The redefinition of the meter in value for the fine-structure A 7 terms of the second is now a real pos- constant accurate to 1 in 10 has been X 2 sibility. In fact the CCDM (Consulta- obtained from a new determination of He -Ne 260.1 tive Committee for the Definition of the proton gyromagnetic ratio. This the Meter) at its meeting in June 1979 achievement has helped test one of + 63 THz has proposed that: modern physics' most fundamental

He -Ne 196.8 "The meter is the length equal to the theories: quantum electrodynamics

distance travelled in a time interval of (QED). QED is believed to provide an + 49 THz 1/299 792 458 of a second by plane exact description of all atomic He -Xe 147.9 electromagnetic waves in a vacuum." phenomena which do not directly The era of speed of light measure- involve nuclear forces, weak 60 THz ments may cease as the value for the interactions, or gravitation. He -Ne speed of light becomes fixed in the redefinition of the meter. X 3 The key to obtaining the new value The measurement of this iodine line was our development of a unique mag- co 29.4 with frequency measurement technol- 2 netic technique to measure accurately ogy some 50 years after this line was X 3 the relative positions of each of 1000 first measured by wavelength tech- of wire single-layered preci- H 0 10.7 turns on a 2 niques represents a truly exciting ad- sion solenoid. This new method allowed vancement in the science of measure- X 12 us to determine directly the proton ment. ' the prime HCN 0.890 gyromagnetic ratio, yp (where indicates that the protons are in a

X 84 spherical sample of pure H 2 0 at 25 °C). Cs We then obtained an order of magnitude improved value for the fine-struc- NRC 260 THz v Composite Laser ture constant, a, by combining our Figure 1. A block diagram of the laser fre- stabilized on \ ' 2 result for with the values of other synthesis chain connecting the yp quency Cs at 520 THz already well-known constants, such as frequency standard with the visible. the speed of light in vacuum, c, and the Rydberg constant for infinite mass, R. crystal. The visible frequency measure- NBS 260 THz The fine-structure constant is the funda- 20Ne Laser ^ ments were made simply mental expansion parameter of QED by combining Lamb-Dip stabilized the 260 THz beams from the two lasers and therefore must be accurately known on a high-speed photodiode. The beat in order to compare critically the cal- frequency was displayed on a spectrum culated values of various QED-depend- analyzer and measured with an adjust- ent quantities with their experimentally able marker oscillator and counter. The determined values. Indeed, with the estimated in t value a, is predict error the beat frequency counter new of QED able to spectrum is ., measurement ± 0.5 MHz, and the the value of the electron factor, ( analyser g g overall error is ± 60 MHz—a result of with an uncertainty of only about oscillator 10 the free running nature of two lasers in 3 in 10 . This theoretical result the frequency synthesis chain. The may then be compared with the experi- Figure 2. A block diagram showing how frequency of the O - hyperfine com- mental result obtained from the elec- the last step to the visible in the frequency ponent (nearest the center of the sec- tron . synthesis chain was accomplished. magnetic moment anomaly, a t = ond harmonic output at 520 THz) is (ge — 2)/2, as obtained in a recent f0 = 520,206,837 ± 60 MHz. experiment at the University of Wash-

14 ENTER FOR ABSOLUTE PHYSICAL QUANTITIES I

ngton. The agreement found is deemed .cceptable in view of the incomplete lature of the theoretical calculations.

' The quantity yp is an atomic constant vhich relates the angular frequency,

', ip with which the proton precesses vhen placed in a magnetic field, B, to

' he value of the field: y p = o> p'/B. Thus, the experimental determination

')f -/p' requires the measurement of both he proton precession frequency and he magnetic field which produces the recession. Determining the precession

requency is relatively straightforward vhen nuclear magnetic resonance N'MR) techniques are used. More

challenging is the critical problem of naking an accurate dimensional mea- Laser Interferometer surement of the precision solenoid so hat the magnetic field produced by a ,cnown current can be calculated. Figure 1. Solenoid dimensional measurement system used to determine the axial position and radial variations of the wires. The five coils A, A', B, B', and C are attached to a In our experiment, the solenoid was T silica tube T and can be pushed or pulled along the axis of the solenoid. Coils A and A' ingle layered and wound on a hand locate the axial position of the injected current and coils B, B', and C form a diameter-to-

. apped, fused silica form. Its critical voltage transducer. Mirrors M and M' are part of a linear interferometer. The Pyrex

limensions were measured by a mag- tube is evacuated. netic induction technique which located tained. solenoid to . he current in the windings. Coils A Coils B, B', and C formed a were measured an accuracy

ind A' in figure 1 formed a linear radius-to-voltage transducer which of 0.05 ^m. No turn deviated from that .differential transformer which located measured the variations in the radius of expected of a perfect solenoid by more

he axial current injected the injected current. The voltage in- than 2 yu.ni. j position of the Unto selected turns of the solenoid. duced in coil C was inversely propor- These improved dimensional meas- ijZoils A and A' were connected so that tional to the radius of the activated urements along with other refinements ..heir output voltages cancel when cen- turns of the solenoid. To detect small in the experiment provided a 0.21 ppm tered on the activated turns of the sole- changes we first "bucked out" most of (parts per million) value for y,/, which noid. A servo system locked the coil the voltage in coil C by using the two in turn yielded for the fine-structure assembly to the null point with a pre- coils, B and B'. At the same time that constant a" 1 = 137.035963(15) (0.11 -1 cision of better than 0.05 ^m. A mirror we were bucking out the voltage in ppm). This value of a is (0.33 ± corner cube, M') located in the center coil C, we were also increasing the 0.14) ppm less than the value of a~ x de- Ibf the coil assembly was part of a sensitivity of this three-coil system to rived from measurements of the anom-

I aser-interferometer system, with the changes in the radius R of the solenoid alous magnetic moment of the elec-

; reference mirror (M) connected to the because the voltage in coils B and B' tron and its current best QED estimate. \~nd of the solenoid. Thus, with the aid increased when the radius R increased. The 0.14 ppm uncertainty in this com- of a computer-automated system, the This three-coil radius-to-voltage trans- parison represents a test of QED which 'elative position of successive turns ducer was then calibrated by having a is a factor of ten more precise than all

' :ould be recorded. We chose to acti- few turns of wire on both ends of the previous tests. The small, 0.33 ppm vate 10 turns at one time, and then solenoid that were 25 p.m larger at one remaining discrepancy has the follow-

I move the current injector to successive end and 25 smaller at the other end. ing possible explanations: (1) it could 1 ) turns until information about all With this system, the axial position and be purely statistical, about a 2 per-

1000 turns of the solenoid was ob- the radius variations of the turns of the cent probability by chance; (2) the

15 NATIONAL MEASUREMENT LABORATORY'

Superconducting Thermometric Fixed Point Devices

Robert J. Soulen, Jr. and James F. Schooley, Temperature Measurements and Standards Division

he narrow and highly repro- OED numerical calculations of a,, are T tween 0.015 K and 7.2 K, five of not yet complete and in addition, are ducible superconductive transitions of which have been incorporated in a new sufficiently complex so that they could 10 materials have been shown to be provisional temperature scale, EPT-76, easily be in error; (3) one or more of useful as reference temperatures which extends the IPTS-68 down to the five fundamental constants required (fixed points) for definition of 0.5 K. -1 ', ' temperature scales which are being fixed to obtain a from yv yp itself, and/or The points are provided by of the experimental value of ae could have developed in the region 0.01 to 20 K. two devices, Standard Reference Ma- unknown systematic errors; or (4) there Two devices incorporating these 10 terial (SRM) 767 and (SRM) 768, could be something wrong with the superconductors are now available which are produced at NBS and dis- basic physics, i.e., quantum electro- through the Office of Standard tributed by the Office of Standard dynamics. We are now working to Reference Materials. Reference Materials. Each unit con- extend this test of QED to yet another tains five very pure, well characterized decimal point by carrying out a new materials for which the phase transi-

-/,,' experiment with an accuracy of a One of the cornerstones of the Inter- tions into the superconductive state s few parts in 10 . It is expected that by national Practical Temperature Scale occur over a temperature region of the time this new effort is finished, the (IPTS) —the lingua franca for ther- theoretical calculations of a as well as mometry—is a set of reference tem- e Temperature Superconductive its experimental determination will also peratures, or fixed points. These well (K) Transition be completed and that an unequivocal delineated and reproducible points are test of a OED prediction to an ac- defined by the change in physical state Nb Sn (18.00 K) s curacy of a few parts in 10 will be of selected materials. Their stability is 3 possible. used to hold the temperature scale V Ga (14.30 K) 3 10 and to check temperature calibrations. Nb (9.289 K) Notable examples are the melting Pb (7.201 K) point of gold, the triple point of water, in (3.146 K) and the boiling point of oxygen. When the IPTS was last formulated (1968), V SRM it was not extended downwards, partly 767 Al (1.175 K) because few conventional fixed points 1.0 Xn K) were available below 13.8 K. Tradi- (0.844

tional phase transitions (boiling points, Cd (0.515 K) melting points, triple points) of all materials except 3 He and 4 He take Auln (0.203 K) place at higher temperatures. There- 2 AuAl (0.150 K) fore, fixed points needed for this region 2

0 1 lr (0.112 K) must be found in phase transitions occurring within the solid state. A pro- ,SRM gram was started at the NBS in 1968 768 with assistance from the NBS Office of Standard Reference Materials 1 to de- Be (0.024 K)

velop cryogenic fixed points based on W (0.015 K) the phenomenon of superconductivity. - 0.01 There now exist ten such points be- Figure 1. Superconductive transition temperatures of several materials. The 10 materials included in the Standard Reference 1 Standard Reference Materials are well Material units are shown together with characterized materials or simple artifacts materials at higher temperatures which are with special chemical or physical properties being examined as candidates for fixed certified by NBS. points.

16 : CENTER FOR ABSOLUTE PHYSICAL QUANTITIES

Improved Laser Test of the Isotropy of Space

John L. Hall, Quantum Physics Division, for himself and Alain Brillet, Joint Institute for Laboratory Astrophysics

about 0.001 K and for which the figure 2, in this case for the element e report the application of reproducibility of the observed transi- indium which is used in SRM 767. w stabilized laser techniques to the tion temperature, designated T c , is Provided that the ambient magnetic precision test of one of the ± 0.001 K or less. The materials field as well as the measuring field are fundamental concepts of Relativity used in the two devices and the as- less than 1 ,uT (0.01 Gauss), some of Theory. These measurements signed T c 's are shown in figure 1. the transitions have been found to confirmed the isotropy of space to Observation of the phase change is reproduce to within ± 0.0001 K. 13 ± 2.5 parts in 10 , and represent a based on the fact that in the normal Referring to figure 1 , we see that the 4000-fold improvement over the best • state above T e the electrons exhibit a materials included in the SRM units do previous test. very weak magnetic response to an not cover the full temperature range

impressed alternating magnetic field, over which superconductivity is known

- while in the superconducting state the to exist. We are, therefore, conducting One of the most meaningful and - electrons respond with perfect diamag- research to produce satisfactory sam- significant human activities might be ; netism. This very strong effect is easily ples of Nb and Nb 3 Sn to extend the called "guessing with inspiration." In- registered as a change in mutual induc- region covered by superconducting fixed | deed, we daily are called upon to invent tance in two coils surrounding the ma- points to a full three decades in tem- an "outside reality" with only clues and terial. A typical transition is shown in perature. shards of evidence for input. But I am thinking here of the kind of scien-

tific guessing, or concept synthesis, that occurs in a field perhaps only a dozen times a century. This year we celebrate the 100th birthday of the genius that made three such contributions in a single year. Einstein's inspired guesses and their consequences must surely have been troublesome to conventionalists at the time. In explaining why Michelson and ndium Morley had failed to see a few tenths of T = 3.416 K c 7 a wavelength shift when their apparatus rotated, Einstein gave up such reasonable physical ideas as simul- taneity, absolute time, and absolute motion. New and unfamiliar ideas were introduced such as the equivalence (for

discussion of physical effects) of all intertial reference frames, the physical equivalence of mass and energy, a kind of symmetry between time and length, and certain group properties of frame transformations. Outrageous and deep-

-3.0 -2.0 -1.0 0 1.0 2.0 3.0 ly troubling paradoxes were identified, such as the "twins" paradox. Other T r (mK) "unphysical" effects were predicted, c Figure 2. Indium phase transition. The change in mutual inductance, AM, in a set of coils such as the shrinking of a physical bar i surrounding the In sample is plotted versus the output of an ac resistance bridge monitoring a along the direction of its motion. This calibrated germanium resistance thermometer. The transition temperature is chosen as the shocking physical is still mid-point of the transition, while the width W of the transition is defined as the theory of temperature interval, centered about T<, corresponding to 80 percent of the change AM. fundamental interest to us now 74 years

17 NATIONAL MEASUREMENT LABORATORY

later since these remarkable counter- tivity theory with lasers. It is probably did not even exist when the experiment intuitive predictions are found to be in fair to say that the first really mean- was first undertaken. Basically we had perfect agreement with every experi- ingful laser test of special relativity was developed new measurement tech- ment performed up to now. Indeed, a only recently completed. In this meas- niques that appeared useful for several major industry is based on the conver- urement, Alain Brillet (from the fundamental experiments, and the NBS sion of a bit of the binding energy of Laboratoire de l'Horloge Atomique, in management supported the work on the uranium nucleus into commercial Orsay) and I, using appropriate laser these completely adequate and reason- amounts of heat energy. Synchrotrons techniques, repeated the famous able grounds.) sweep their frequency as the acceler- Michelson-Morley "Isotropy of Space" Our experiment was designed to be ated particles gain mass. Time dilation experiment of 1887. Impressively highly sensitive, clear in its physical of the muon's internal decay clock enough, one still finds a null result, in interpretation, and free of problems gives a distance range in the CERN agreement with Einstein's postulate, involving spurious effects. Its basic laboratory some 29 times the distance even with a measurement sensitivity concept can be understood by consid- that light could travel in one (proper) 4000-fold improved over the best pre- ering figure 1. Fundamentally, we were decay time. vious experiment (the 1964 experiment comparing the laser frequency defined

It is not surprising that experimen- of Charles Townes et al.). by an etalon of length with the laser talists have found this special theory of Our experiment is timely in three frequency defined by an optical fre- relativity a powerful stimulus to the ways. First (and trivially), it is ap- quency standard. As depicted in figure invention of precision experiments propriate to do an improved relativity 1, a 3.39 /j.m laser was servostabilized against which to test the theory. In the experiment in an Einstein celebration to the standing wave condition in an 1950's and 60's, high resolution tech- year. Secondly, laser techniques have isolated and stable passive Fabry-Perot niques based on the Mossbauer effect been developed to the point that they etalon. The laser, Fabry-Perot inter- provided some major advances to con- permit a really major advance in sensi- ferometer, and associated electronics firm our knowledge of the validity of tivity. Thirdly, it is probably true that were mounted on a large granite slab special relativity (the so-called inter- all physical laws are approximate on a table which could be slowly ro- action of local clocks with distant mat- statements: As the measurements be- tated around a vertical axis. Informa- ter and "anisotropy of inertia" experi- come more refined and sensitive, a tion about the length of the interfer- ments). The other new experimental larger and larger fraction of "outside ometer was implicitly carried by the tool, lasers, provided a first glimpse in reality" needs to be taken into account wavelength (frequency) of the portion

1964 of the dramatic enhancement of in describing the system under study. of the laser beam which left the table measurement sensitivity that lay ahead. Ultimately then, one might expect some along its rotation axis and which was This first important experiment was image of the cosmic status in the micro- directed over to another table to be performed in a special isolated labora- scopic physics which we can probe. optically heterodyned with an "isola- tory near MIT by Charles Townes, Recently, several lines of investigation, tion" laser. This laser was frequency- Ali Javan, and their coworkers. They but most notably radio noise ther- stabilized relative to another 3.39 pointed out for the first time the gain mometry done from above the earth's laser which was in turn frequency- in sensitivity possible through the use atmosphere, lead to the conclusion that stabilized relative to the saturated ab- of frequency techniques to we are probably \ measure moving at speed ~ sorption peak in CH4 . In following small length changes. Unfortunately, 400 km/s relative to the 3 K blackbody Javan and Townes, we had thus pre- their laser experiment was troubled by background radiation remaining from pared a highly sensitive readout of the 1 systematic noise sources, probably in- the primordial "big bang". Thus it is optical length of the Fabry-Perot etalon ertial forces arising from the periodic interesting to see whether the supposed through the use of frequency measure- motion, and the attained sensitivity small lack of isotropy in physics "in the ment techniques. To see whether there was only somewhat better than that large" will have an ultimately observ- was a measurable spatial anisotropy in obtained in the 1930's by G. Joos using able effect on extremely sensitive lab- our laboratory, the laser and Fabry- Michelson's original techniques. Thus, oratory experiments which test some Perot etalon part of the apparatus was a large amount of technical develop- kind of local isotropy. (Honesty com- rotated slowly and we looked for a ment of laser techniques has been a pels me to note that the experimental change synchronized with this rotation. necessary preliminary to testing rela- basis for this third, philosophical point It is natural—perhaps inevitable when

18 ;hnter for absolute physical quantities

Phase Lock

'Figure 1. Schematic of isotropy-of-space experiment. A He-Ne laser (3.39 ,um) is servostabilized to a transmission fringe of an isolated and highly stable Fabry-Perot resonator, with provision being made to rotate this whole system. A small portion of the laser beam

:is diverted along the table rotation axis to read out the cavity length via optical heterodyne with an "isolation laser", which is stabilized

relative to a CHi-stabilized reference laser. The beat frequency is shifted and counted under minicomputer control, these frequency measurements being synchronized and stored relative to the table's angular position. After 30 minutes of signal averaging, the data

are Fourier-transformed and printed out, and the experiment is repeated.

lithe sensitivity is sufficiently high—that We observed that the spurious signal reduced or eliminated, a factor of at some : such synchronous perturbation at the double rotation frequency was least 1000 further improvement would jwill be observed. We saw a signal at essentially fixed in the laboratory co- be possible. In this domain, general jthe table rotation frequency of some ordinate system, and so averaging over relativistic effects (the earth's gravita- j200 Hz excursion, representing a 2 X a few days in 12 or 24 sidereal hour tion) would cause a vertical spatial 10~ 12 length change of our Fabry-Perot blocks reduced our spurious signal anisotropy of the type we can detect, : etalon. 2 10- 17 Its origin is now understood as essentially to zero. This null result Vi (GM/R) (1/c ) (1/R) ~ 3 X . a gravitational stretching of the inter- represented a fractional synchronous Unfortunately, such a vertical experi- ferometer which occurred because of length variation below 2.5 parts in 10 1 "', ment would be hopelessly contaminated residual asymmetry in the mounting of and this limit represents a 4000-fold by gravitational distortion of the Fabry- [the reference etalon when there was a improvement over the MIT laser experi- Perot etalon itself. A more promising

jslight misalignment of the rotation axis ment. Further, if it is supposed that the direction for growth is to redo the jfrom the vertical. A 10-fold smaller earth's inferred velocity of ~ 400 km/s Kennedy-Thorndike measurement that

signal was observed at the double fre- is the relevant one to be used in the shows the equivalence of the transfor-

quency of the table rotation, which is coordinate frame transformation, this mation of length and time. Extreme the spectral location expected for a "laserized" Michelson-Morley experi- care would be necessary to avoid drifts, ^genuine anisotropy signal. However, ment represents the most demanding since the only modulation in this experi- considering that the laboratory axes test yet of the Einstein-Lorentz trans- ment arises from the earth's motion. were rotated relative to the hypotheti- formation. For example, a temperature stability of

cal "preferred direction in space" by As mentioned above, the laser hetero- 50 pK for 1 month and creep rate noise ithe earth's rotation, a true ccsmic ani- dyne techniques provided a very high below 10~ u per day 2 would allow a sotropy signal would show a further sensitivity in these experiments. If the 20-fold improvement on the original rrodulation every 12 sidereal hours. spurious background signal could be experiment of Kennedy and Thorndike.

Center for Radiation Research

Chris E. Kuyatt Nuclear Radiation he Center for Radiation Research has a unique t ne A. Cassatt, Randall S. Caswell T mission. It provides high accuracy radiation measure- h. Robert L. Ayres Stephen M. Baloga ment techniques and quantitative data about atomic Donald Colle James W. Behrens 51mer H. Charles D. Bowman and nuclear radiation processes. These outputs are Patricia Burt** Eisenhower, Jr. required for the safe and effective use of radiation in rlenry T. Heaton, II Jacqueline M. lames S. O'Connell Calhoun key technologies. Center activities cover the electromag- Larry Cardman** and Plasma Atomic Allan D. Carlson netic spectrum starting at the infrared, through the Radiation Lucy M. Cavallo L. Wiese visible, the ultraviolet, and vacuum ultraviolet, to x-ray Wolfgang Bert M. Coursey Nicolo Acquista John J. Coyne and nuclear gamma rays. Particle radiations include lohn M. Bridges Hall Crannell** electrons, positions, protons, deuterons and neutrons David G. Cooper John B. Czirr Zharles H. Corliss Michael Danos below 140 million electron volts. Our facilities include a Jeffrey R. Fuhr William R. Dodge Ronald L. Greene* Kenneth C. Duvall 140 MeV electron linear accelerator, a 240 MeV electron Victor Kaufman Charles M. storage ring for synchrotron light, many lower energy ac- Daniel E. Kelleher Eisenhauer Arthur K. Kerman Larry Fagg** celerators, a polarized electron source, plasma devices, Jules Z. Klose Everett G. Fuller Andrea M. Henry M. and many sophisticated spectrometers with computer Malvezzi** Gerstenberg assisted data acquisition. Oscar Manley** David M. Gilliam Georgia A. Martin Daniel Golas** During the past year. CRR has directed special atten- William C. Martin, James A. Grundl tion to several areas of growing national concern. One Jr. Evans V. Hayward Tsu-Jve A. Nee Alan T. Hirshfeld of these areas, ionizing radiation health and safety, has William R. Ott Dale D. Hoppes Timothy L. Pittman John M. R. attracted a large amount of public and institutional con- Donald Priester** H. Hutchinson cern. At the request of the Senate Committee on Joseph Reader Kenneth G. W. Inn James R. Roberts Ronald G. Johnson Commerce, Science and Transportation, the Center is Larrv J. Roszman Arthur K. Kerman cooperating with the Conference of State Radiation Phiilip M. Stone** George P. Lamaze Jack Sugar John W. Liehtbody, Control Program Directors. Together, we are reviewing John Weiner Jr. Andrew W. Weiss Richard Lindgren** the need for an intermediate level of calibration services Jean-Francois Larry Lucas such as regional facilities to better couple NBS labora- Wyart** Marcia D. Maltese Stephen M. tories with State and industrial needs. This review will Younger* Romuald Zalubas ( Continued) contribute to an updated long-range plan and budget for ionizing radiation measurements.

CRR is working with the State of Illinois to set up a pilot regional calibration laboratory, with Florida to set up a calibration facility for low-level survey instruments, and with the University of Michigan on the testing of personnel dosimetry performance. Special measurements were made on dosimeters used near the Three-Mile Island Robert Placious adjusts equipment for char- nuclear power plant. Natural matrix radioactivity stand- acterization x-ray fluorescent intensifying of ards, such as human lung and liver, are being prepared screens. Measurements are made, on an for environmental monitoring. We are working on do- absolute basis, of the total light output and spectral distribution of the x-ray excited simetry measurements and standards for radiation steri- fluorescence from radiographic intensifying lization, a process which replaces the use of carcinogenic screens. This unique facility utilizes mono- ethylene oxide. Measurements of the response of x-ray energetic x-ray production together with highly sensitive radiometric detection equipment. screens are part of our work on the technology for reduc- These screens are universally used in almost all ing patient exposure. Because of the need for increased medical, and in many industrial radiographic cooperation between the various agencies concerned with procedures to maintain good image quality n hile reducing x-ray exposure. ionizing radiation, the Center is represented on two

21 *

NATIONAL MEASUREMENT LABORATORY

Wilfrid B. Mann David R. Penn recently formed interagency committees on ionizinj Xavier K. Maruyama Daniel T. Pierce Leonard C. Maximon Robert C. Placious radiation. Emmert D. McGarry Daniel Polansky Michael M. Meier John S. Pruitt CRR is working to improve the state-of-the-art in opti Sydney Meshkov Ronald W. Rendell* Vernon W. Myers JackC. Rife** cal radiation measurements. Better measurements an James R. Noyce Edward B. Saloman needed, for example, in solar and atmospheric monitoring James O'Brien** Stephen M. Seltzer Gerald Peterson** Christopher G. and by the electro-optics and lighting industries. To make, Mark Reeves** Soares Henry Rosenstock Julian H. Sparrow available to the lighting industry the new radiometric Francis J. Schima John Sutherland* scale, which will eventually replace the current scak John C. Schleter Nils Swanson 1 Roald A. Schrack Peter Takacs** based on the candela, 175 lamps are being calibrated or Ivan G. Schroder Bernard J. Waclawski both scales for simultaneous distribution. Progress towarc Robert B. Schwartz Harold H. Walker William E. Slater James T. Weaver, Jr. the 0. 1 percent radiometry needed in solar monitoring ii Daniel Sober** Radiometric Physics Valentine Spiegel, Jr. being made through quantum-based calibrations of sili Jack L. Tech William Stapor** con photodiodes, by improved characterization of black Richard D. Starr* Robert L. Booker, Michael P. III body sources, and through intercomparisons of NB5 Unterweger, Jr. Robert J. Bruening Oren A. Wasson Kenneth L. Eckerle radiometry standards. Improved measurements of UVI Joel B. Fowler Elisa Wolynec** radiation are made possible by new source standards anc Lungwen Yap** Jon C. Geist Warren K. Gladden Radiation Physics by accurate measurements of spectroradiometer response Jack J-G. Hsia Lewis V. Spencer Provision of characterized detector renta Henry J. Kostkowski packages on a Martin J. Berger Ernest Lewis, Jr. basis enables users to measure the spectral response o L. Randall Donald A. ' Canfield McSparron their detectors and to test the performance of their mea Robert Celotta Fred E. Nicodemus surement systems. measurement assurance program: Barrett E. Cole Charles H. Popenoe New Randolph Culp** Robert D. Saunders, are being developed for diffuse reflectance and for retro Joseph Dehmer** Jr. Charles E. Dick A. Russell Schaefer reflectance. Steve R. Domen John B. Shumaker, David L. Ederer Jr. Several of our activities contribute to the Nation': Margarete Ehrlich James H. Walker Raymond L. Falge, John F. Ward, III energy programs. Measurement and calculation of the, Jr. William R. Waters, spectra of highly-ionized heavy ions are essential to the; Annija D. Galejs Jr. John C. Hamilton* Victor R. Weidner diagnostics of fusion-plasma devices. Ionization rate mea, John H. Hubbell Charles R. Yokley surements in progress will provide benchmark date Lanny R. Hughey Edward F. Zalewski now Jimmy C. for analyzing heavy ion impurities in Tokamak fusioi Humphreys Radiation Source and Instrumentation Paul J. Lamperti devices. Vacuum ultraviolet radiometric calibrations anc Samuel Penner Donald R. Lehman** Harry Levine** John B. Broberg portable transfer source standards provide the measure Robert Loevinger Louis Costrell ment base for accurate measurements of radiation fron Thomas P. Loftus Stephen C. Ebner Toh-Ming Lu** David C. Green fusion plasmas. Radiation phenomena, such as plasm; Thomas B. Eric R. Lindstrom line broadening and dielectronic recombination, are beim Lucatorto Isador J. Livant Robert P. Madden Anthony B. Marella investigated as possible new plasma diagnostic tech William L. Joseph Parrington** McLaughlin George Rakowsky niques. Resonance neutron radiography is being devel Thomas J. Mcllrath John E. Rose oped as a non-destructive method for obtaining accurate Caroline Mehlman** Julian K. Whittaker Stanley R. Mark A. D. Wilson values for the total mass of fissionable isotopes. Thi: Mielczarek Post Doctoral Staff method is expected to be useful in calibrating measure Joseph W. Motz ** Research Associates John Noakes** and Guest Workers ment devices for nuclear fuel rods as well as for waste Hideo Onuki** Rathindra Pal** and scrap from nuclear fuel reprocessing plants. Accu rate measurements of neutron cross sections are beim made for both fission and fusion reactions. Integral anc

differential neutron fields continue to be developed fci

22 CENTER FOR RADIATION RESEARCH

: ie calibration of neutron detectors used in fusion re- nuclear constituents is well known from quantum elec- sarch, nuclear reactors, and personnel dosimetry. trodynamics, and the flexibility of electron scattering

. The Center has a dedicated synchrotron radiation fa- allows energy and momentum transfer to the nuclear

Jlity, SURF-II. This device produces electromagnetic system to be varied independently. Using a high-current .idiation in a narrow, intense, highly-polarized beam 140 MeV linear electron accelerator, CRR scientists have jHth a continuous and accurately known spectrum from investigated nuclear structure using electron scattering, fie infrared through the visible and into the extreme photon scattering, and electrodisintegration. Recent elas- 12 ftraviolet. This tunable, intense radiation is being ap- tic photon scattering measurements from C showed a plied to several new scientific and technical opportuni- surprisingly large electric quadrupole component. Simi- ie _es. Many of these activities involve collaboration with lar measurements on O will be of even greater theoreti- ther Centers of NBS or with laboratories outside of cal interest. Elastic electron scattering has yielded an 12 BS. Ultraviolet photodiode detectors are now being cali- improved value for the C nuclear radius which has be- rated for photon energies up to 250 eV. This capability come a world standard. Similar measurements on 4 He j was used in a collaborative program with Los Alamos to will provide a critical test of quantum electrodynamics. •leasure photoemission yields of several promising cath- In electron scattering, the cross section is measured as de materials. Several calibrations were carried out on a function of energy and momentum transfer. In electro- .oectrometers for space experiments and fusion plasma disintegration, however, outgoing particles or induced iagnostics. A new spectrometer calibration facility, un- radioactivities are integrated over all of the final states of ,ter construction with NASA funding, will give more the scattered electron. An entirely new type of experi- .ccurate and convenient calibrations. During the past ment has recently been pioneered by CRR scientists in .ear our new high-flux normal-incidence monochromator which the yield of particles is measured as a function of .arae into full operation, and has been used to measure the incident energy. From analysis of such measurements le angular distribution of photoemitted electrons from it is possible to locate and measure the strength of the idividual vibrational levels of highly excited molecules. El and E2 giant resonances that result in proton and « Hiis monochromator also has been used to study Stark particle emission. Results on 58Ni, 60Ni and r,2 Ni show nixing in high Rydberg autoionizing states of atoms and important E2 components in the « channel. Molecules. Further studies on excited state dynamics will

: ultimate experiment scat- tilize fluorescence measurements to investigate ion de- The would detect both the tered 'xcitation processes. Using another monochromator, we electron and outgoing particle in coincidence. Such

! experiments ave observed high pressure helium microbubbles in thin would require a 100 percent duty cycle electron 'luminum films by measuring the pressure-broadened (CW) accelerator and would be crucial in going beyond the classical description 'elium resonance in absorption. Scientists from the Na- of nuclear structure and

' reactions in terms of al Research Laboratory are studying insulators by using elementary neutrons and protons; : 2flection. photoemission and fluorescence. Scientists that is, in determining the effects of many-body correla- I'om Brookhaven National Laboratory are installing a tions, virtual pions, and excited nuclear states on nuclear structure. Circular dichroism apparatus for measurements of bio- These opportunities are so important that we have begun, with DOE funding, to develop a prototype holecules. Apparatus is being built in collaboration with 185 room-temperature electron accelerator. le Surface Science Division to study the photodesorp- MeV CW In addition to demonstrating the feasibility of building a 'ion of ions from absorbates at surfaces. We do not have 1 to 2 GeV machine using the same technology, the pro- pace to describe all of the activities at SURF-II. but totype accelerator will be a powerful tool in its own right 'nese examples illustrate the wide range of activities being for nuclear physics. 'arried out with synchrotron radiation. High energy photons and electrons are excellent These vignettes cover briefly only a fraction of the rjbes of nuclear structure. The coupling of photons and CRR programs. Some of these programs and others are lectrons to charge, current, and magnetic moments of described in more detail on the following pages.

23 :

NATIONAL MEASUREMENT LABORATORY

Radioactivity Standardization

Wilfrid B. Mann Nuclear Radiation Division

-A. adioactivity, like fire, has monitor of nuclear-fuel burnup). Meas- Such spectrometer systems are used in potential for both benefit and harm, uring the radionuclides released from monitoring at nuclear power plants, or and the increasing variety of its the burning of coal required standards wherever mixtures of radionuclides applications and public awareness of of several naturally occurring radionu- must be assayed, and special standards its hazards have brought challenges to clides. For example, a standard of are prepared for calibrating the effi-

r the NBS Radioactivity Standardization flyash from the burning of Eastern ciency of these systems as a function of | group. We are meeting those United States coal was developed in energy. A mixture of 9 radionuclides challenges in a variety of ways. which the sample is highly homogene- with well-established Py's was devel-

-' 2 ' ; Nearly 2000 samples of some 50 ous and is certified for Ra and oped for this purpose, and commercial different radionuclides were calibrated - 28Ra content. suppliers now submit samples of each or issued as standard reference For some applications, preparing the batch of their standards of this mixture materials (SRMs) in 1978. Co- SRM in the proper form and checking to NBS to establish explicit traceability. operative programs with nuclear- its uniformity offer more of a chal- Longer-lived radionuclides, such as 13S ir,2 medicine and environmental- lenge than the basic calibration. Stand- Ba and Eu, which emit gamma i monitoring quality-control groups ards which are prepared to check the rays at several energies, are also now S allowed the staff to assure correct yields from chemical procedures are of available from NBS as standards, and 12r ]54 155 instrument calibrations much beyond this type. Such natural matrix stand- a mixture of 'Sb, Eu, and Eu is their ability to supply direct services. ards of human lung and liver, soil being studied as a possible standard. from Rocky Flats, Irish Sea sediment, For some common counting conditions,

and river and lake sediments are being large errors can occur if the summing

Such cooperative measurements- developed. Homogenization of these of coincident radiations is not taken assurance programs work to establish materials is a significant challenge which into account, and computer programs

the traceability of radioactivity meas- is being met partly by borrowing tech- are now being used to calculate these i . urements to those of NBS. Why NBS? niques from powder-mixing technology. corrections for frequently encountered Because we are the legal promulgator Ultimately, not only the radioactivity radionuclides. of standards of the national radioac- of these materials will be measured, but Once a calibration has been estab- tivity-measurements system. The term the chemical forms will be examined. lished for any gamma-ray-emitting ra-

! traceability may not be completely apt, For example, we expect to tell users dionuclide, the response of NBS pres- i but it is being used more and more by approximately how much of the plu- sure ionization chambers to a known !: k regulatory agencies to connote meas- tonium in a sample is in the form of activity is measured relative to that of urements that are demonstrably in high-fired oxides. a 22,; Ra source producing a comparable agreement with those of the National Radioactivity calibrations are unique current. This preservation of what

Bureau of Standards. in that a value for a standard applies would otherwise be only a temporary , a

Each year NBS must decide on those only for an instant. The radioactive calibration, in the case of short-lived i radionuclides for which standards are decay law allows the standard to be radionuclides, eliminates the need for most urgently needed. Once those deci- used at a later time if accurate values frequent direct recalibrations. Doing so sions are made, our work begins. In the for the half life and the associated un- allows the group to offer an efficient last three years, an efficiency-extrapo- certainty are known. For a number of calibration service for many radionu- i

lation coincidence method employing a years, the half lives of most radionu- elides for which the demand does not : pressurized proportional counter and clides used as standards have been justify preparing SRMs. efficient gamma-ray spectrometers has measured at NBS. The probability per The preservation of calibrations, and been used to establish calibrations for decay for radiations is also required a comparison of them on an interna-

G7 f,!, 00m ul 123 1G0 Ga, Mo, Tc, In, I, Yb, if their emission rate is to be deduced tional scale, is also accomplished by i i 195 20Hp . ' AUj 201 Xl! _al J i and 5 rad onu from the activity given for a standard. the routine transmission of calibrated ti elides now used in nuclear medicine. Well-calibrated germanium spectrom- gamma-ray-emitting solution stand-

The same system was used to calibrate eter systems have been to meas- ards to the Bureau International des lj used !j 14 ~'Pm (used in a pressure measuring ure gamma-ray probability-per-decay Poids et Mesures (BIPM) and to the 14 140 device) and "Ba- La (a significant (Py) values for many recent standards. International Atomic Energy Agency !>

24 :enter for radiation research

Spectra of Highly- Ionized Atoms

Joseph Reader, Atomic and Plasma Radiation Division

c(IAEA) for measurment on their ion- ment. It describes in detail many of the T_JL he development of controlled chambers. calibration tech- concepts and techniques mentioned i zation NBS fusion reactors depends greatly on the niques are also checked by international here. Further, the NBS Radioactivity ability to observe and understand the

- ntercomparisons of the radioactivity Standardization group has become physical processes taking place in their concentration of split samples of the known for its willingness to help those very hot plasmas. One of the methods

:;ame master solution. Such intercom- with radioactivity-measurements prob- used to diagnose these plasmas is the .parisons, arranged by BIPM, have re- lems, through seminars or direct tech- observation of line spectra emitted by ;ently taken place for 139Ce, mCs, nical assistance. This attitude has the various ionic species present. Such

:;7 55 L Cs, and Fe. In the first two cases brought them considerable satisfaction, observations provide information about

(!:he agreement was quite good, with 22 and a very harried existence! the temperature of the plasma, impuri- l aboratories agreeing to within 0.7 perties, and spatial and temporal proper- cent for the m Cs, but preliminary re- ties. For lighter ionic species much of ports show a much greater spread for the basic atomic data needed to inter- 137 55 t he more difficult Cs and Fe. pret these observations are already on

I International cooperation on appli- hand. However, for heavier ions, such cations of radioactivity has been aided as those with Z^>20, relatively few

. n the past few years by the Interna- data are available. The problem is par- tional Committee for Radionuclide ticularly acute for Tokamaks, because Metrology (ICRM), with active par- of their use of aperture limiters to ticipation by NBS. The ICRM spec- prevent stray particles from striking trometry group has organized compari- the interior wall of the torus. Atoms

^ions of gamma-ray emission-rate meas- from the limiter, which is usually made urements with germanium spectrometer of a refractory metal such as molyb- iiystems for ir,2 Eu and 133Ba; collected denum or tungsten, are sputtered into nformation on impurity searching, per- the discharge, where they become tinent nuclear data, and germanium highly ionized and significantly affect detector stability; and initiated, through the plasma conditions. Knowledge of TAEA, a study of random pile-up ef- the spectra of the ions of these heavy fects in gamma-ray spectrometry. Sym- elements is therefore of great impor- posia were held by the low-level group tance. By studying spectra of laser- ho determine the need for specific nat- produced plasmas our group has made ural-matrix standards of radioactivity, progress during the past year toward ;and the group has organized interna- providing new data about these highly tional cooperative projects for their ionized atoms. The work was done in production and distribution, collaboration with G. Luther and u Any accurate radioactivity measure- N. Acquista. ment requires good technique in addi- jjtion to good standards, and several itmembers of the NBS group have con- Data about highly ionized atoms l tributed to written standards for the were obtained in two separate experi- ;use of ionization chambers, germanium ments. First we focused the light from f.and Nal (Tl) detectors and liquid scin- a 3 GW Nd/glass laser onto flat me- tillation counters. Much NBS effort tallic targets to produce plasmas of Iwent into the writing of the Handbook various elements. The spectra were re- of Radioactivity Measurement Proce- corded on a 10.7 m grazing incidence dures, prepared as Report 58 by a group spectrograph. Typical laser pulses were

of experts for the National Council on 15 I in 10 ns. By varying the energy Radiation Protection and Measure- and duration of the laser pulses we

25 NATIONAL MEASUREMENT LABORATORY

were able to distinguish spectra due to suits, figure 1 shows the energy of the Experimental different ionization stages. Analysis of 5s configuration in the Cu sequence, the data yielded the spectra, electronic along with the values obtained in fully

energy levels, and ionization energies relativistic calculations [1]. The cal-

of ions in the zinc and copper isoelec- culations predict the general variation 21+ 22 + ;: tronic sequences to Sb and Sb , of energy very well, although the ab- including the important ions Mo12+ solute values are somewhat displaced. 13+ and Mo . As an example of the re- As the scale is greatly expanded here the absolute differences between the

1 Theoretical Pd Figure 1. Plot of the energy (in cm- ) of two curves are actually not large; for 37 800 I- Se/ the 5s configuration in the Cu isoelectronic Mo the difference is only 0.5 percent.

sequence. Z . is the net charge of the atomic 3 5 9 11 13 15 17 19 ( In the second experiment we trans- core seen by the valence electron. Theoretical Z, - ported a 2.2 spectrograph and target values are from Cheng and Kim (Ref. [1]). m chamber to the Los Alamos Scientific Laboratory. There we used a very powerful (100 GW) Nd/glass laser to

generate still higher stages of ioniza- 1600 tion. Spectra were recorded with laser pulses of 30 J in 300 ps. With these observations we extended the Zn and Cu isoelectronic sequences to W44 + 4r, + and W . These results represent the longest isoelectronic sequences ever observed as well as the highest stages of 1200 ionization recorded with diffraction gratings. Figure 2 shows an isoelec- Laser-produced Plasma tronic sequence plot for the 4s-4p NBS/LASL) resonance lines of Zn-like ions. These experiments also produced e u significant results for molybdenum. Be- 12 o sides the resonance lines of Mo + and i::+ ? 800 Mo , two of the most important lines > used in Tokamak diagnoses are the (3 a resonance lines of the and Na-like UJ Mg Z o,0+ 31 + ions, . re- UJ Mo and Mo Until cently these lines had been observed only in Tokamaks. Confirmation of their proposed origins (ion and transi- 400 tion) were thus critically needed. Pre- vious observations of laser-produced Mo Laser-produced Plasma plasmas [2] gave spectra that could not (NBS) be interpreted unambiguously. In the present work we studied ions in the Mg 27 + 34 + and Na sequences from Sr to Pd .

Figure 2. Isoelectronic comparison of the 15 25 35 45 2 1 4s S0-4s4p !Pj transitions in the Zn isoelectronic sequence.

26 -CENTER FOR RADIATION RESEARCH

"igure 3 shows a plot of two transitions Tokamak lines. For other lines the checking theoretical calculations of the n the Na sequence that are nearly co- wavelengths do not agree so well, sug- properties of highly stripped ions in 31- incident in Mo . From the isoelec- gesting a possible problem in the Toka- high temperature plasmas (energy ronic regularities we could identify the mak measurements. levels, wavelengths, transition probabil- observed lines and correlate them with The results are also of interest for ities, cross sections for excitation and Tokamak observations. Wavelengths ionization by electron impact). In the

or Mo are compared in table I. For domain of highly stripped ions, where Figure 3. Isoelectronic comparison of 3s-3p vome lines the wavelengths agree well. relativistic effects are extremely impor- and 3p-3d transitions in the Na isoelectronic :onfirming the proposed origins of the sequence. tant, the wavelengths predicted by the most accurate fully relativistic calculations agree with our experimental values to within about one percent (see

fig. 4). As the predicted values are sys- tematically low by this amount, the experiment reveals a possibly important 800 effect not included in present theories. - 3p 2P 3/2 3d 2D 5/2 Table 1. Present values of Mo wavelengths compared with Tokamak measurements of E. Hinnov, Phys. Rev. 14, 600 1533 (1976). /Kr A(PRES- A(TOKA- /As Ion ENT)-NM MAK)-NM 400 Mo ,2 + 34.0909(5) 34.10(5) - 3s 2S 1/2 3p 2P 3/2 Mo 13 + 37.3647(5) 37.38(5) 42.3576(5) 42.35(5)

:; " + 200 Mo 11.596(1) 11.70(5) 31 + K Mo 12.7814(5) 12.9(1) 17.662(1) 17.7(1)

1 1 I i 10 15 20 25 20 25

References

[1] Chen, K. T. and Kim, Y.-K., Energy Levels, Wavelengths, and Transition + 0.2 Probabilities for Cu-like Ions, Atomic Data and Nuclear Tables Nd W Data 22, 547-563 ( 1978). 0.0 Mo 20 30 40 50 [2] Burkhalter, P. G., Reader, I., and Cowan, R. D., Spectra of Mo XXX, S -0.2 XXXI, and XXXII from a laser- produced plasma, J. Opt. Soc. Am.

Figure 4. Comparison between observed and calculated wavelengths in the Cu isoelectronic sequence. Calculated values are from -0.6 Cheng and Kim (Ref. [1]).

27 NATIONAL MEASUREMENT LABORATORY Polarized Low Energy Electron Diffraction (PLEED): A Technique to Study Surface Magnetism

Daniel T. Pierce and Robert J. Celotta, Radiation Physics Division

JLAJLt the surface of a magnetic spectroscopy), be- the polarization modulation must re- sion, LEED, Auger crystal, the lack of translational invar- cause their strong interaction with the sult from a spin dependent scattering. iance perpendicular to the surface and electrons in the material limits their The scattered intensity for incident the reduced number of neighbors of the probing to the outer few atomic layers. electron spin parallel to the net spin in

surface the I is i atoms cause magnetic prop- For example, in the Ni(llO) surface the ferro-magnetic surface j T in erties to differ from those in the bulk. which we studied, electrons with energy general different from the scattered in- We have developed a new way to meas- of about 100 eV can on the average tensity when the spins are antiparallel ure the magnetization of the outer few penetrate and return through only two I U . We measure a quantity S, which atomic layers of a single crystal surface atomic layers without losing energy and is this difference normalized to the spin based on polarized low energy electron being removed from the elastically scat- averaged scattered intensity, diffraction (PLEED). [1] tered beam. Polarized electron scatter- I ing provides a way to study the surface f t —I t i semi-infinite single Due to the exchange interaction, the of a well defined I 1 T T + t i diffracted intensities of electrons at the crystal. The polarized electron probe surface of a magnetic crystal depend is non-destructive and has potentially We also correct for the fact that the on the relative orientation of the inci- very high spatial resolution although we incident polarization is 43 percent dent electron spin and the magnetiza- have not exploited it thus far. One can rather than 100 percent. tion of the surface. The surface mag- envision a magnetometer for surface Theoretically, S depends on the ex- netization, its temperature dependence, microstructure which could display change potential J(E,#) and the Cou- and the transition temperature at the magnetic domains and bubbles. lomb potential V(E,0) seen by the elec- surface have been described by a num- The measurement of surface mag- tron scattering at an energy E and an scat- ber of theories which we now have a netization by polarized electron angle 6. It can be shown in the first straightforward way to test experimen- tering is made possible by the polarized Born approximation that S is propor- tally. The increasing interest in lower electron gun which we developed. The tional to the average magnetization M dimensional magnetism and surface electron gun, which is based on photo- of the outer few layers of the surface. affinity critical phenomena makes the develop- emission from negative electron If multiple scattering is included there ment of our "surface magnetometer" GaAs, provides a constant intensity may be higher order terms proportional

especially timely. electron beam like an ordinary LEED to M", where n is odd, but in Ni these We also expect the surface mag- gun, but the spin direction is modulated are very small and may be neglected.

netometer to have technological appli- at a rapid rate. Since the incident elec- The apparatus [2] consists of the cations. In the push to further minia- tron beam intensity on the target crystal polarized electron gun attached to a[

is constant, any ac component of the turize magnetic devices—by using conventional surface analysis chamber,

j thinner layers in magnetic storage scattered intensity synchronous with We cleaned a Ni(110) crystal by sue-! media, for example—the magnetic X cessive Ar ion bombardment and an-

properties of the outer few layers will nealing cycles. surface cleanliness I CO The become very important. Our technique - + .02 was checked by Auger spectroscopy provides a research tool to study, for using a cylindrical mirror analyzer.

instance, the magnetic changes at the Generally, it is not straightforward to

surface of a recording head resulting — 25 + 25 scatter low energy electrons from the - from corrosion. 1 ^— of Magnetic surfaces are -f-+- Hh- surface of a ferromagnet because also important in hydrogen storage and H (Oe) the deflection of the electrons in stray catalysis, such as in the coal gasification magnetic fields. The stray magnetic process. The surface magnetometer will fields have been minimized by making allow the study of the role of the mag- the Ni crystal the keeper of the small - - .02 netic electrons in chemisorption on well electromagnet that magnetically aligns characterized single crystal surfaces. the domains in the crystal. Figure 1. Measurement of S as a function Electrons are widely used in surface can measure S as a function of of H showing hysteresis in the surface mag- We analytical techniques (e.g., photoemis- netization. magnetic field to obtain a surface hys-

28 —

CENTER FOR RADIATION RESEARCH

X-Ray Image Information Theory

Joseph W. Motz and Michael Danos, Radiation Physics Division

>:eresis curve S(H). Such a curve is tion. There are definite predictions TM his study analyzes the depen- shown in figure 1 for specular scatter- from theoretical scaling laws relating dence of image information content on

7 ng at an angle of incidence of 12° the surface and bulk behavior in the x-ray exposure in medical diagnostic ind an energy of 125 eV. The maxi- critical region near the Curie tempera- x-ray examinations. The results indi-

mum S in figure 1 is less than 0.02; this ture Te , We were able to measure S(T) cate that with the employment of image

is a small signal but the noise level in from 0.5 to 0.8 Tc and found that the processing techniques and modified

:his sensitive measurement technique is temperature dependence of the sur- x-ray spectra, it is possible to signifi- :ess than 0.001. Reversal of the mag- face magnetization shows much less cantly reduce these exposures without netic field allows us to isolate any pos- curvature than that of the bulk; in fact, loss of the visual image information

sible nonmagnetic contribution to the it is approximately linear. Measure- that is available to the radiologist.

spin dependent scattering. We empha- ments were made by heating above Tc

size that figure 1 is a hysteresis curve and cooling in applied magnetic fields Medical x-ray examinations provide epresenting the magnetization of the of 18, 21, and 26 Oe as represented the major source of radiation exposure Duter two atomic layers. Such a curve by the different points in figure 2. The to the population in this country. s very sensitive to surface contamina- cooling was too fast to allow measure- Therefore, any reduction in this expo-

tion. In particular, segregation of sulfur ment near Tc . In future measurements sure would have a large impact in re- :o the surface decreases the magnetiza- we will have crystal temperature con- ducing the adverse effects of radiation tion to zero. Future measurements will trol and will measure up to T thereby to the general health of the population. e ,

study the effects of adsorbates on the providing a test of theories of surface Is it possible to reduce the exposure surface magnetization. We will inves- critical phenomena. We plan also to and still have the same diagnostic in- tigate the existence of magnetic dead extend the measurements to low tem- formation that appears in visual form

. ayers at the surface (considered un- perature in order to normalize S(T) to in present-day radiographs? We have ikely from these first measurements) the saturation magnetization. carried out a theoretical study which

.and the role of the magnetic d-electrons In summary, this very powerful new concludes that this is indeed possible. n the bonding of adsorbed atoms. technique may be compared to neutron This conclusion seems paradoxical,

We also scattering, inasmuch is , explored the temperature which has been so success- as an exposure reduction dependence of the surface magnetiza- ful in measuring bulk magnetic proper- always accompanied by a loss of image ties. We expect that analogous informa- information. However, our study in- Temperature, K tion will be obtained about surface dicates that only a fraction of the 400 500 600

i i i i i i i i -i i magnetic properties image information i using content in medical ————————— n r our new PLEED technique. radiographs is visually detected by the

radiologist. Accordingly, it is possible

.01 References to reduce the exposure without loss of the visible information obtained pre- [1] Celotta, R. J., Pierce, D. T., Wang, viously, providing methods can be G. -C, Bader, S. D., and Felcher, G. found to display all of the available n 0 P., Physical Review Letters 43, 728-731 (1979). image information. In fact, the exposure can be reduced to a theoretical [2] Wang. G. -C, Bunlap, B. I., Celotta, minimum or threshold value without -.01 R. J., and Pierce, D. T., Physical Re- loss of the visible information on film. view Letters 42, 1349-1352 (1979). Such a lower limit exists because of the statistical noise introduced by the

1 -.02 1 1 L quantum nature of x-rays. Temperature, (T/T have estimated the ex- c ) We minimum posures required for particular cases in Figure 2. Measurement of the temperature diagnostic radiology from a theoretical dependence of the surface magnetization showing a linear behavior as a function of relationship between the patient ex- :emperature. posure and the image information con-

29 NATIONAL MEASUREMENT LABORATORY

tent. The results for a few of these cases gens and from approximately 500 to 3 example the screen-film system. How- are shown in figures 1 and 2. Figure 1 milliroentgens. ever, any type of x-ray imaging system shows the minimum exposures at dif- The effect of reducing the exposure with adequate sensitivity, dynamic ferent x-ray energies that are theo- in this manner is to produce a very low- range, and linearity can be used to ob- retically required for the detection with density, low-contrast radiograph which tain the described exposure reductions. film of a tissue calcification sphere (0.5 provides a poor image on direct visual An example of such a system is the mm diameter) which simulates the inspection. However, with commer- x-ray image intensifier optically cou- structure involved in the radiographic cially available image processing tech- pled to a TV camera. The video signal detection of breast cancer (mammog- niques, it is expected that contrast and passes to an analog-to-digital converter,

raphy) . Present day mammography ex- density of the image can be suitably en- and is then stored on a disc. With such posures are in the region of 1 roentgen, hanced, essentially restoring image in- a system, the stored image can be dis- primarily because the high film densi- formation previously observed. This played on a TV monitor with any de-i ties produced by such an exposure are procedure assumes that a suitable sired type of image processing such as necessary to obtain optimum image screen-film combination can be found contrast or edge enhancement.

Visibility. However, the data in figure 1 to store all of the image information The application of these new systems indicate that it may be possible to ob- provided with a given x-ray exposure. to radiology requires data on the in- tain the same visible image information In this discussion, we used as an formation content and the correspond- with exposures that are one to two orders of magnitude smaller than 1 roentgen, provided that the x-ray spectrum is confined to an energy band (±10 keV) in the region of 30 keV. Figure 2 shows our estimates of the exposure reductions that are possible in chest and abdominal x-ray examinations. These data show that for chest and abdominal x-ray exams with respective 1 mm and 0.5 mm diameter spatial resolutions, the corresponding exposures may be reduced respectively from approximately 20 to 4 milliroent-

Figure 1. Minimum x-ray exposure theoretically required in mammography for the detection of a tissue calcification sphere (0.5 mm diameter) embedded in 5 cm tissue.

4 3 2 I0" I0" I0" 10"' 10°

Exposure, Roentgens

2 Figure 2. Dependence of the attainable area (cm ) of the spatial resolution element on the patient exposure for chest and abdominal (G.I. tract) x-ray examinations for body struc- 0 20 40 tures described in Table II of paper by I. W. Motz and M. Danos, Medical Physics 5, 8 Incident Photon Energy, keV (1978).

30 .

CENTER FOR RADIATION RESEARCH

New Studies of Electric Quadruple (EZ) Excitation in Nuclei Using Electromagnetic Probes

William R. Dodge, Nuclear Radiation Division

ing image storage requirements for The (e,p) and (e,a) cross sections normal oscillation or resonance to be 38 60 C2 x-ray pictures. As shown in figure 3, we for targets of Ni, Ni, and Ni have discovered, the giant electric dipole have determined the maximum number been measured in the electron energy resonance, GDR. Rather, the GDR of bits per cm2 required for the storage range 16 to 100 MeV. They have been was established and studied through of x-ray images as a function of the analyzed using the DVVBA El and E2 photonuclear reactions. (As a point x-ray exposure and the desired spatial virtual photon spectra. Protons are of interest many of the most impor- resolution at two x-ray energies (20 emitted primarily following El absorp- tant early experimental and theoretical and 50 keV) that are in the energy tion but a emission results from a com- studies of the GDR were done at NBS) region used in medical radiology. We bination of El and E2 absorption. However, during the past seven expect that these data will provide the years, normal oscillations of the nu- basis for developing improved radio- cleus other than dipole oscillations logical systems that will permit either Since the early days of Fermi, the have been experimentally observed. extracting more image information for idea that the effects of the electron's Unlike the GDR, these new giant a given x-ray exposure or reducing electromagnetic field in electron- resonances were discovered and stud- x-ray exposure for a given image in- nuclear scattering could be related to ied almost exclusively through inelas- formation content. the real photon cross section has in- tic scattering of medium energy ha- trigued nuclear physicists. As early as drons and electrons.

1934 Bethe and Peierls had computed Figure 1 shows a pictorial view of the electric-dipole electrodisintegration the monopole, dipole, and quadrupole 10 : cross section of the deuteron and had oscillations of the nucleus (A = 0, 1,

compared the result with the photodis- 2, respectively). Normal oscillations — * s 1(T / / integration cross section. In the next which do not involve a nucleon spin few years electrodisintegration cross flip are labeled AS = 0 while those

sections for magnetic dipole, electric which do are labeled AS = 1. Giant 10 c £ u quadrupole and octopole transitions resonances, which are internal collec- were calculated. These cross sections tive nuclear motions involving all 2 to -io were enhanced relative to the electric nucleons, are important because they (d) : dipole cross sections. The enhance- involve basic parameters of nuclear ment was understood on the basis of matter such as compressibility, polariza- 10 1 (e) \ the strong momentum dependence of bility, and the effects of the in-phase these cross sections and of the greater (isoscalar, AT = 0) or out-of-phase 10 l _1 1 linear momentum which an electron (isovector, AT 1 ) motions of pro- 10" 10 10° 10 10' = can deliver to the nucleus compared tons and neutrons. In addition to the Exposure, Milliroentgens to a photon. However, while all of normal oscillations shown in figure 1, Figure 3. Dependence on the number of these predictions were based on Dirac the giant isoscalar octopole resonance bits per cm2 on the x-ray exposure at the electron theory, they did not include has been reported in the literature. image plane for x-ray energies of 20 and Coulomb distortion effects (i.e., they resonances 50 keV. and for different spatial resolutions While the new multipole where the area a of the spatial resolution did not allow the incident or scattered (NMR) are commonly called giant

element is equal 10- 4 10-3, to (a) , (b) electron to change energy and momen- multipole resonances they are in fact 2 10-i. 2 'c) IO- 10° . , (d) and e ) cm ( tum by exchanging longitudinal pho- less intense than the GDR. Because of tons with the nuclear Coulomb field). the relatively small intensity of the Consequently, these early cross section NMR some method either to suppress estimates were grossly inadequate for the isovector GDR or enhance the

all but the lightest target nuclei. NMR is needed to facilitate the study Faced with these technical difficul- of the NMR. In hadron excitation of

ties, electrodisintegration experiments the isoscalar NMR, the method is to only played a minor role in mapping use an isoscalar projectile, thus sup-

out the properties of the first nuclear pressing excitation of the isovector

31 —

NATIONAL MEASUREMENT LABORATORY

GDR. In electron scattering or electro- field with no kinematic restraints such The virtual photon spectra are those

disintegration the method is to enhance as characterized earlier theories became functions involving only electron kine- the NMR excitation through the strong available. This theory, called a dis- matics which relate photo- and electro- momentum transfer dependence of the torted wave Born approximation disintegration cross sections. Figure 2 cross section in the manner alluded to DWBA theory in the trade—showed shows El and E2 DWBA virtual pho- earlier. that Coulomb distortion of the electron ton intensity spectra for Ni. The large Fortunately, almost simultaneously wave function greatly enhanced the difference between El and E2 virtual with the discovery of the NMR, a "virtual photon" intensity spectra, photon intensity spectra make it rela- theory which provided a solution of especially for E2 and higher electric tively easy to separate transitions in- the Dirac equation in a point Coulomb multipoles. volving El and E2 excitations in an electrodisintegration experiment. The upper half of figure 3 illustrates the conventional photoproduction ex-

periment, in which bremsstrahlung pro-

duced by electrons of energy E0 strikes a radiator of high atomic number. The real photons—with intensity spectrum, Ny (E0 ,o>), which of course does not depend on multipolarity of the reaction which the photons will induce—are absorbed by the target nucleus, which then emits particle x. The yield of x for A = 1 real photons is given by

Y(7,x) (Eo) — (1)

Eo — m e N \ 7 (£,„, L (w)do (7,x) thres

AS = 0 L AT = 0 where o- (y,x) is the photodisintegra- THOMSON tion cross section for the Lth multi-

pole and w is the excitation energy. A = 2 The lower half of figure 3 shows schematically an electrodisintegration experiment in which the electrons strike the target nuclei directly. The electro- yield, production Y u ., x) , is given by an

expression similar to Y,y iX ), except that the spectrum of virtual photons seen by the target nucleus now depends on AS = 0 AT = 0 the multipolarity L of the nuclear tran- E2 0 sition. In this case

Figure 1. The giant monopole, dipole, and quadrupole (\ = 0, 1, and 2, respectively) Y (e,x) (E 0 ) — (2) oscillations of the nucleus. Normal oscillations not involving nucleon spin flips are labeled L E„ m c N AS = 0 while those which do are labeled AS = 1. Normal oscillations which involve e (E0 , oj) in-phase motions of protons and neutrons are labeled AT = 0 while those which involve cr (tu)dc thres (7,x) out-of-phase motions are labeled AT = 1.

32 CENTER FOR RADIATION RESEARCH

E' target

E„

real photons

Bremsstrahlung radiator

product detector

E e- c virtual photons

10 20 30 40 50

E (MeV) product Figure 2. Virtual pho'on spectra for El and detector E2 transitions calculated in the distorted wave Born approximation for Ni (Z = 28). Figure 3. Schematic diagram of a typical photodisintegration experiment (top) and of a typical electrodisintegration experiment (bottom).

L Since Ne (E0 ,o>) rapidly increases as L increases for a given E0 and u electrodisintegration yields are very sensitive to the multipolarity of the transition. Thus armed with a good virtual photon theory the discover}- of the NMR by inelastic hadron scattering experiments, and the arrival of an energetic guest worker, E. Wolynec. -2 from the University of Sao Paulo, we - set about trying to exploit these devel- £ opments.

For our initial experiments, we selected targes of 58Ni, 60Ni, and 62Ni. Our measurements were made with electrons in the energy interval be- 10' tween 16 and 100 MeV provided by the NBS linac and with a magnetic spectrometer capable of separating the many charged particle groups (such as 24 32 protons, deuterons, tritons, and alphas)

E co (MeV) •vhich can be copiously produced in a 0 or nuclear disintegration. At each of Figure 4. The 58Ni (e,a) and (e 7, a) excitation functions. The open circles are the results several spectrometer-incident-electron- + for pure electroproduction and the filled circles are the results for electroproduction plus :>eam angles we measured proton and photoproduction. The histograms are El and E2 cross sections deduced from these alpha particle energy spectra as a func- excitation functions. E0 is the incident electron energy and u is the excitation energy.

33 NATIONAL MEASUREMENT LABORATORY

Glueballs and Quark States

Sydney Meslikov, Nuclear Radiation Division A tion of the incident electron energy, that alpha clusters might be continually JL jL systematic enumeration of the glueball spectrum of QCD and other E0 . These energy spectra were inte- forming and dissolving in the nuclear PC grated over all particle energies and surface. A quadrupole oscillation might SU(n) gauge theories yields J then the resulting angular distributions preferentially shake off these conjec- odd . even . and 0 states in integrated over all angles to obtain tured alphas. At the present time these addition to the usual quark model states. The properties of these new Y (e x) and Y (E ). The yields are only speculations and no theoreti- , ( y iP) 0 states are discussed. Y (Eo) and Y ,p)(E ) were both cal explanation of our results is avail- ( y, p) (e 0 accounted for when the previously able. measured cr (o>) was inserted in eqs. Quantum Chromodynamics (QCD), ( y, p) (1) and (2) using the electric-dipole our present theory of the strong inter- virtual photon spectrum for E0 =16 to actions, describes hadrons as being 50 MeV. These results confirmed the composed of colored quarks and anti-

expectation that photon emission from quarks. Color is a new degree of free- the Ni isotopes was a predominately El dom associated with the strong inter- process near 20 MeV. action. All single quarks have color. So On the other hand, the measurement far all observed hadrons are colorless. of the (y,a) and (e,a) yields showed The strong force between quarks and

that a sizable fraction of the a emission antiquarks is mediated by the exchange comes from the giant isoscalar quadru- of massless colored gluons, just as the ple resonance near 16 MeV. Figure electrical force between electrons and 58 4 shows our Ni(e,a) excitation func- between electrons and positrons is tion. The open circles are the results for mediated by the exchange of massless

pure electroproduction and the filled photons. Convincing manifestations of circles are the results for electroproduc- the existence of the gluon have been tion plus photoproduction when a displayed in recent e + e~ annihilation 2 0.217-g/cm Ta radiator is inserted in experiments at 27 and 31.6 GeV. QCD the electron beam. The histograms la- implies that glueballs (gluonia) should

' beled El and E2 are the cross sections exist [1]. These are mesons which are

which produce the best fit to both composed entirely of gluons; they con- the electroproduction and the electro- tain no quarks and have no color. A production plus photoproduction exci- current activity of the NBS Elementary

tation functions. Note that attempts to Particle Program is the study of glu-

fit the excitation data wih pure El or onium, i.e., to set criteria for its exis- pure E2 cross sections were not satis- tence, and to understand how to dis- factory. tinguish gluonic states from the usual The striking result of our measure- quark states experimentally.

ments is that the alpha decay mode of Understanding the formation, struc- the isoscalar E2 resonance exceeds ture, and decay of gluonic systems re-

statistical estimates by a factor of 4. quires a full knowledge of the spec-

Said another way the alpha decay chan- troscopy of "quarkful" mesons, i.e., nel accounts for 25 percent of the E2 those described in the usual way as isoscalar-energy-weighted sum rule quarks and antiquarks moving with a value. Similar conclusions regarding the relative orbital angular momentum L.

decay of the oxygen isoscalar E2 res- Fortunately, the latter is a major com- onance have been reached by the Hei- ponent of high energy activity of NBS delberg-Jiilich group on the basis of as manifested by work on charmonium 16 a 0(a,aa") coincidence measure- and other "onium" spectroscopy (2). ment. Theorists have long speculated Some of the simplest differences be-

34 "ENTER FOR RADIATION RESEARCH

Quantum Basis for Visible Region Radiometric Standards

Jon Geist, Edward Zalewski, and A. Russell Schaejer, Radiometric Physics Division

the spectroscopies are sim- s ilicon photodiodes are the de- flectance losses, was within 1 percent :ween two PC for radiation ple, i.e... in the quantum numbers J tectors most widely used of unity in the red region of the spec- (J=;total angular momentum. P=par- measurements in the visible region of trum. The range of internal quantum the electromagnetic spectrum. When efficiencies for five photodiodes along ity, C=charge conjugation) . The quantum number assignments depend on employed in this capacity, their cali- with their measured absorptance is whether the glueballs are made of two, bration is based on the thermal physics shown in figure 1. At wavelengths three, or more gluons, and whether the of either electrical substitution radiom- longer than 650 nm the internal quan- quark systems with which they are eters or the blackbodies. The possi- tum efficiency is 0.99 or greater. Be- compared are simple quark-antiquark bility of basing these calibrations di- cause it is so close to unity, there are systems, or are multiquark composites. rectly on the quantum physics of the less stringent requirements on the ac- PC =0~ _ internal photoelectric effect in silicon curacy of the theoretical model used to For example, states with J , 0~~, 2'~ occur in the 3 gluon system has been suggested frequently. Re- predict this small difference. but cannot be made with either 2 cently, we demonstrated a quantum- Photon energies in the wavelength gluons or the usual qq systems. Other based calibration that is accurate to region longer than 530 nm are less possible distinguishing factors such as within =0.3 percent. This uncertainty than twice the band gap. so that for modes of production and decay are compares favorably with that currently every photon absorbed only one elec- also currently being investigated. available from the best thermal based tron-hole pair is created. (Photons The charmonium quark structure calibrations, and, for silicon photo- absorbed by the free carriers in the work has continued with emphasis on diodes, the new procedure is more con- conduction band are a negligible frac- the existence and masses of the pseudo- venient as well. tion of the total absorption.) Photons that penetrate to the vicinity of the p-n scalar cc state n c and its radial excita- junction before absorption have a tion n'c . A reanalysis [3] of the _ _ In a study of the radiometric prop- German e e annihilation data which quantum efficiency of one because the erties of several high quality (high had led them to claim the existence of minority carrier thus generated is im- resistivity, shallow junction) silicon mediately swept the electric field the n, . at 2.83 GeV led to the conclu- by ; photodiodes. we found that the quan- sion that the claimed state was at most through the depletion region. It is only tum efficiency, after correcting for re- a 2-3 standard deviation effect. Recent in the regions away from the p-n junc- experimental searches at the Stanford tion that recombination or trapping colliding electron-positron ring have found no state at 2.83 GeV, but strong

1 1 1 10 1 1 indications of a state at 2.98 GeV F—

which may well be the n e . Work is

continuing on the properties of the n c.

and n'c .

References

[1] Bjorken, J. D. Electron-Positron An- nihilation. Talk given at European Physical Society High Energy Physics Conference, Geneva, Switzerland.

June 27-July 4, 1979. SLAC-PUB- 2366.

Fishbane, Paul M., Horn, David and [2] l i i i i i i 400 500 600 700 800 i i Meshkov, Sydney, Spin forces in char- 400 500 600 700 800 monium spectroscopy. Phys. Rev. Wavelength (nm) Wavelength (nm) D19, 288 (1979). Figure 1. The absorptance (one minus the Figure 2. Variation of the absorption coeffi-

[3] Samios. N". P. and Meshkov, Sydney reflectance ) and the internal quantum efficient of silicon in the visible region of the CBNL 25926) 1979. ciency for five silicon photodiodes. spectrum.

35 NATIONAL MEASUREMENT LABORATORY

can be expected to decrease the quan- of the function at wavelengths shorter values of « (A) at the appropriate wave- tum efficiency. than 400 nm. lengths can then be used to obtain K

Figure 2 shows the variation of the In the long wavelength region of the from eq. ( 1 ) . The types of measure- silicon absorption coefficient with spectrum the photons can penetrate be- ments involved in this procedure, ex-

wavelength. The absorbance is high in yond the junction into the bulk region cept for the current from the photo- the blue region, decreasing by about of the diode before being absorbed. diode, are all relative as opposed to

two orders of magnitude toward the Because the region between the junc- absolute measurements. That is, the re- red. Therefore, in the blue region the tion and the rear of the diode has a flectance, absorption coefficient and re-

majority of the photons are absorbed low concentration of impurities and lative spectral response all involve a near the front surface. Here the recom- defects, the lifetime of the electron-hole ratio of similar measurements—no ab-

bination probability, 1-P, where P is pair is long relative to the diffusion rate solute radiometric measurements are

the collection efficiency, is significant back to the junction. Therefore, recom- required. because of the large density of majority bination in the bulk region also has a This research has demonstrated the

carriers and possible surface effects. very small effect on the quantum effi- feasibility of adopting a new measure- One might expect that P(x) would be ciency. ment basis for absolute radiometry a complicated function of distance, x, The simplest mathematical model based on quantum physics instead of

into the diode; however, we have found which fits our data is of the form thermal physics. High quality silicon that a model which postulates a layer photodiodes can now be used for meas- of constant recombination probability K P+(l-P) exp(-«(A)J) urement of visible radiant power with followed by a region of zero an accuracy comparable that recombina- B to of tion probability fits our data quite well. (1) standards presently available. It is ex- a(A)L 2 An example of the quality of the fit of pected that improvements in the meas- this model to the data for one of the Here e is the internal quantum effi- ured values of the absorption coeffi- silicon photodiodes is shown in figure 3. ciency, P is the collection efficiency, cient of silicon plus a more complete

Other more physically significant mod- rt(A) is the spectral absorption coeffi- knowledge of the physics of these de-

els for the recombination probability cient of silicon, J is the junction depth, vices will yield more accurate and con-

distribution function were also fit to B is the thickness of the bulk region venient radiometric standards. The these data. They only affected the shape and L is the minority carrier diffusion long term goal of future work in this length. The coefficient K can be inter- area would be the development of preted as the calibration factor of the photodiodes whose radiometric properties could be accurately I I I determined 100 1 I I I I I I 1 I I I I I electrical substitution radiometer used I I J I

to determine the external quantum effi- from a knowledge of their structure ciency. From our data the values ob- and physical properties instead of the tained for K ranged from 0.9987 to relative radiometric measurements now 1.0033. These results are consistent used. Possible approaches are elec- with the estimated accuracy of the tronic test structures built into the electrical substitution radiometer we photodiode and calculations based on used and the assumption that absorp- device fabrication history. tion of a visible photon in silicon produces one electron-hole pair.

Equation 1 can now be used as the basis of an absolute calibration I ' of I I I I I I a Mil • ' ' ' 350 450 550 650 spectrally non-selective radiometer. In Wavelength (nm) order to do this, the internal quantum

efficiency of a silicon photodiode is Figure 3. The curve obtained from the measured in the blue and near infra- model of the internal quantum efficiency of silicon (equation 1) and the data for a red spectral regions relative to that in typical photodiode. the red. These data plus the published

36 1

Center for Thermodynamics and Molecular Science

\°eter L. M. Richard I. Martinez T,he Center for Thermodynamics and Molecular Heydemann Norman C. Peterson James A. Walker Science (CTMS) provides the scientific community with lalph Klein Francis Westley \lbert E. Ledford, the means to measure, interpret, and control accurately Robert E. Huie, III Jr. Chemical the behavior of complex chemical systems. Examples of ,»urfaee Science Thermodynamics such systems, whether manmade or natural, are ubiqui- Zedric J. Powell David Garvin engines, lasers, chemical reactors, distillation ames J. Carroll Stanley Abramowitz tous: heat R. iichard George T. Armstrong columns, catalytic processes, the oceans, even biological Cavanagh* Pierre J. Ausloos Villiam L.Clinton** Kenneth L. Churney systems. Villiam F. Egelhoff Jennifer C. Colbert Jr. Howard J. DeVoe** To researchers in this Center, these systems represent oseph Fine David A. Ditmars ohn W. Gadzuk Eugene S. Domalski mixtures of chemical substances in one or more phases, Iteven M. Girvin Robert N. Goldberg either in equilibrium or in the process of approaching David W. Goodman Thomas L. Jobe, Jr. jregory L. Griffin* Duane R. Kirklin equilibrium and interacting with both radiation and the David K. Hoffman** Sharon G. Lias ack E. Houston** Michael Mautner containment surface. "errence J. Jach George W. A. iichard D. Kelley Milne** The problems CTMS scientists address are often •am H. Kim** Ralph L. Nuttall solved generically, with result that "heodore E. Madey Vivian B. Parker the the solutions are Mian J. Melmed Edward J. Prosen Stockbauer applicable to a variety of specific practical questions. ioger L. Richard E. ohn T. Yates, Jr. Rebbert For example, the quantification of the behavior of fluids Kinetics Richard H. Schumm Tiemical can be applied to the development of standards for pres- Ving Tsang Frederick P. Walter Braun Schwarz sure measurement and to the provision of the values of Brown Lewis W. Sieck Robert L. those properties of steam needed in the design of electric Robert F. Hanapson, Bert R. Staples Donald D. Jr. Wagman power plants. Methods used to calculate the electronic ohn T. Herron Jin T. Wang** properties of transient species are Tadahiro Ishii** Stanley P. Wasik in the atmosphere William J. Webb vi. Odo Klais** applicable to transient species in chemical lasers. Proce- vlichael J. Kurylo, Harry D. III Youmans** dures for measuring of the rates of chemical reactions Mian H. Laufer (Continued) relate equally well to atmospheric chemistry or to combustion kinetics. Methods developed to measure the heat evolved in chemical or physical transformations can be applied to study biological cell metabolism or to determine the stability of oil in the simulated environment of an internal combustion engine. Accurate spectroscopic

selected the rmophysical properties of metals data can be used to calculate the concentration of atmos- are measured with millisecond time resolution pheric pollutants; such data can also serve as a basis for by Ared Cezairliyan (foreground) and Archie calibrating laser frequencies or identifying molecular "vliiller. The metal sample is heated electrically species in interstellar clouds. 'n the metal cylinder shown near the center of 'he photograph. Measurements are made either brief review of recent work within CTMS will illus- aider vacuum or in selected inert atmospheres. A The temperature of the sample is measured with trate the broad applicability of CTMS research and in optical pyrometer seen to the right of the development. sample container. The large lamp in the foreground is a tungsten filament lamp used for Researchers involved in surface science are concerned 'he calibration of the optical pyrometer. The with developing accurate methods to characterize the vyrometer, designed specifically for the project, properties of surfaces. Surface Is a unique instrument, capable of millisecond physical and chemical resolution. The laser pulse interferometer which science has been a rapidly developing field over the past n easures, with millisecond resolution, the 10 years, with important applications in corrosion, catal- thermal expansion of the sample is on the air pollution, integrated circuit technology. This fable around the sample container. ysis, and 39 NATIONAL MEASUREMENT LABORATOR

Thermophysics Molecular growth has resulted largely from the development of si Harold J. Raveche Spectroscopy Merrill M. Hessel face characterization techniques and instrumentation Francis W. Balfour** Anita Calcatelli** Stephen E. greatly improved specificity and sensitivity, from an Ared Cezairliyan Bialkowski* creased understanding of surface properties and pre Ren-Fang Chang Arthur J. Grimley* Jack H. Colwell Jon T. Hougen esses on a microscopic scale, and from the success! Theodore D. Doiron* Marilyn E. Jacox Nils E. Erickson Paul S. Julienne application of modern measurement methods and the Vincent J. Fratello* David S. King retical concepts to a wide variety of scientific and tea 1 John S. Gallagher Morris Krauss Robert W. Gammon Walter J. Lafferty nical problems. Lester Haar Frank J. Lovas John R. Hastings Arthur G. Maki, Jr. Yet, in spite of the growth in this area, serious me£ V" Richard F. Kayser, Frederick H. Mies urement-related problems and limitations remain. 1 Jr. David B. Neumann T John M. Kincaid Hideo Okabe Surface Science Division works to develop quantitati Max Klein William B. Olson James W. Little John C. Stephenson understanding of the details of the measurement proce Rosemary A. Walter J. Stevens as they relate to surface science and to apply this knou Mac Donald Lawrence L. Strow** John L. McClure Richard D. Suenram edge to the development of new and improved measur B. Mehl Alfons Weber James :. ment methods and standards. Surface properties Archie P. Midler Richard F. such Michael R. Moldover Wormsbecher* composition, atomic structure, electronic structure, Malcolm S. Morse Raymond D. atomic motions are being probed through a great varie Mountain of techniques, usually involving the interactions of ene E. Gordon Powell** Jan V. Sengers getic particles (ions, electrons, atoms) or the interactk Johanna M. Sengers of radiation with surfaces. Understanding of these pro G. Bruce Taggart** Donald H-N. Tsai esses on an atomic or molecular scale is critical to tl Meyer Waxman Arnold Wexler** development of reliable, quantitative measureme Harold Post Doctoral Staff W. Research Associates methods. Woolley** and Guest Workers Researchers are engaged in work on electron stimi lated desorption of molecules from surfaces, the surfa( characterization of certain catalytic processes, the appf cation of infrared spectroscopy to the characterizatiC of molecules absorbed on surfaces, field ion microscop

ion sputtering, the application of auger spectroscopy 1 the analysis of airborne particulates, and the develo] ment of theories for describing the interaction of photor or charged particles with atoms and molecules.

In chemical kinetics, our scientists develop methoc to measure the rates of chemical reactions and to dete mine the mechanisms of reacting systems. They also pre vide evaluated kinetics data and measure particular ral constants when data of sufficient accuracy are not avai

able elsewhere. In the chemistry of the atmosphere or i combustion, dozens of reactions often occur simultant ously, with intermediate products of a transient natur playing a significant role. Frequently, knowledge of th

nature of these intermediate products is purely specuk tive. The possible reactions they undergo must be cor sidered in any quantitative description of a chemic;

40 ^CENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE

process if the behavior of the system over time is to be and homogeneous decomposition of halocarbons and ni- dequately characterized, trous oxide (contributing to models of the ozone layer in the stratosphere) and developing a new mass spectro- ii Programs in the Chemical Kinetics Division are confined with the intermediates that occur in the atmos- metric technique for characterizing fossil fuels. A major pheric oxidation of olefins, the upper atmospheric effort of the Chemical Thermodynamics Division con- chemistry of fluorocarbons, the processes that occur in tinues to be evaluating and publishing of selected values of thermodynamic properties under the auspices of the :. lanetary atmospheres, and the high temperature decom- 1 osition of hydrocarbons. Important intermediates and National Standard Reference Data System. the Chemical Thermodynamics Division is i leir reactions are being characterized. New methods for Whereas -identifying species are being investigated and new meth- concerned primarily with the chemical properties of sub- ods for initiating chemical reactions, such as laser in- stances, the Thermophysics Division concentrates on the duced decomposition, are under study. Computer codes physical properties of substances. The Dynamic Measure- Ipr modeling these systems are also being developed. In- ments Program involves the measurement of such prop- deed, the linking of computer models and laboratory erties as thermal diffusivity, thermal expansion, heats of leasurements will characterize the activities of the fusion, and heat capacities under conditions requiring an Chemical Kinetics Division in the future. The evaluated extremely rapid response. At present, a millisecond sys-

ate constants and related kinetics data provided by this tem is operational and a microsecond facility is under ;|: division are used in large scale computer models of development. The capability now exists to measure both pper atmospheric chemistry. equilibrium and nonequilibrium properties at elevated | temperatures. In the area of chemical thermodynamics, we provide The equation of state program is concerned with: the accurate chemical thermodynamics data and standards, accurate measurement of the temperature, volume, and scientists develop state-of-the-art instruments for meas- pressure relationships in fluids and complex fluid mix- ring thermochemical properties. In support of the NBS tures, the measurement of such properties as dielectric Recycled Materials Program, researchers are designing constants near critical points, the development of quan- he world's largest combustion calorimeter. It will be titative theories to explain the behavior of these proper- j,apable of measuring the heat evolved during the com- ties in the vicinity of critical points, the development of bustion of multi-kilogram-sized samples. Although the new measurement techniques to measure these properties yrincipal purpose of this calorimeter will be to determine more accurately, and, the development of mathematical ,he calorific value of refuse-derived fuel, it will serve as f correlation models for the properties of a variety of fluids yart of a general research facility for studying combus- used in energy generation. tion processes. At the other extreme, microcalorimeters These correlation models have been used, for example, apable of measuring the heat evolved during metabolic to calculate accurately the properties of steam over wide •rocesses have already been designed and built. Labora- ranges of temperature and pressure. A spherical acoustic tories throughout the world use microcalorimeters based resonator has been designed to accurately measure of >n the NBS designs. c* the velocity of sound in gases, from which thermo- .. . J Other efforts include in determining equilibrium work physical properties can be calculated. A laser light scat- constants associated with ionic properties, high accuracy tering instrument has been constructed to measure >omb calorimetric measurements for certifying heat- , transport and equilibrium properties with the greatest the constant Standard Reference Materials (SRM's), accuracy currently achievable. In the future, properties I levelopment of SRM's to serve as standards for the calo- of complex mixtures of polar and nonpolar compounds imetric assay of radioactive plutonium, and the calcula- r will be measured and interpreted. Such compounds are tion of the high temperature thermodynamic properties

)f refractory compounds from molecular properties.

Also, Division scientists are studying the heterogeneous 1 See pages 99 to 114 for a discussion of the NSRD program.

41 NATIONAL MEASUREMENT LABORATORY

found, for example, in systems for tertiary oil recovery. length standards that will be two orders of magnitud The Thermophysics Division provides methods and more accurate than those presently available. This effoi standards for the measurement of pressure. Pressure is a response to requests from the instrumentation indu; measurements in industrial and government laboratories try, industrial and government laboratories, and stanc are made traceable to NBS standards through measure- ards groups. This group also uses nozzle-beam an ment assurance programs developed and maintained by single-moded laser techniques to determine the electroni division scientists. One current program consists of meas- structure of molecules and to improve the resolution anl urement assurance services (i.e., transfer standards sensitivity of molecular electronic spectra measurement: consultation, round robin intercomparisons) for aircraft The microwave spectroscopy group furnishes accurat altimetry measurements. Highly accurate pressure meas- microwave spectra to astronomers for use in studying th urements also find application in materials properties chemistry of interstellar clouds. Researchers determin research. For example, a new high pressure phase of the structure of small, unstable molecular intermediate CCli has recently been discovered which shows that a in atmospheric or combustion processes through analys substance can have dual melting curves. of the microwave spectra. NBS scientists have discovere

To meet the demands for static and dynamic prop- new compounds such as ethylenimine and dioxirane i erties data, a theoretical capability for prediction is these studies. required for systems not accessible with existing measure- The laser chemistry and photochemistry groups hav ment methods. In response to this need, the Thermo- as their objective the quantitative understanding of th physics Division has a statistical physics program in chemistry of molecules when they interact with hig which models based on molecular properties are being power laser radiation. The quantum chemistry grou developed to predict the behavior of bulk matter. develops and evaluates methods for calculating from fin

The work includes computer simulation of the be- principles the electronic, structural, and chemical prop havior of properties and development of analytical erties of molecules. Transient species in such systems ? theories. Both equilibrium phenomena and such non- lasers play a critical role in the functioning of the system but, of their very small concentration, equilibrium phenomena as metastability are being studied. because cannot b observed experimentally. chemistry grou Recent measurements of anomalously high values of solid The quantum calculates reliable values for the molecular properties heat capacities at high temperatures are being thoroughly q in applied ranging froi investigated by the computer simulation of the bulk species encountered problems atmospheric chemistry to laser fc properties of solids taking into account the effects of upper development defense energy purposes. large amplitude, anharmonic lattice vibrations. Theo- and retical work is also underway to predict the behavior of This brief description touches on the activities of th systems as they undergo phase transitions and to predict approximately 100 scientists who contribute to the pro behavior under conditions of intense pressure and high grams of the Center for Thermodynamics and Molecula temperature. Science. Each year, CTMS researchers publish approx 1 mately technical papers reports present In molecular spectroscopy, scientists are concerned 300 and and 25 papers at national and international meetings. They cc with the electronic properties of molecules and their in- extensively with scientists in teractions with electromagnetic radiation. The laser-high operate universities, indu: resolution spectroscopy group devises new measurement try, and other Government agencies and participate i techniques for generating highly accurate spectra and national and international standards activities. computer models for interpreting and extrapolating the The technical highlights that follow have been chose values of molecular properties obtained from the spectra. to illustrate in more detail the breadth of the Center

A laser diode spectrometer for measuring and analyz- activities. They represent work which is new or partici ing infrared spectra of atmospheric constituents is being larly exciting, but they by no means include all that ha developed. The group is also engaged in providing wave- been significant or new during the past year.

42 -

CENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE

Bond Energies and Chemical Reactivity

Wing Tsang Chemical Kinetics Division

B ond energies are the essential a role in predictive theories, accuracy values for the heat of formation of ingredients for understanding chem- requirements are extremely stringent. simple organic radicals presented in

ical reactivity. Recent studies at NBS For example, in the cases where the table 1. Also included for comparative

suggest that many long accepted values bond energy is directly reflected in the purposes are previously recommended I are in error. It is demonstrated that the rates for a particular process, an error values. It should be noted that the gen-

T new bond energies provide a sound of 6 kJs will result in an order of mag- eral trend of the new data is towards basis for understanding cis-trans iso- nitude deviation in rate at 298 K and increasing bond energies. In other merization and small ring decyclization a factor of 2 at 1000 K. v/ords, most aliphatic hydrocarbons are processes. The history of gas kinetics has been much more stable than previously as- ilf marked by repeated determinations of sumed. The course of any chemical reaction bond energies and, as new techniques An interesting consequence of these

: I is marked by breaking and forming and understandings have developed, by findings, which serves to illustrate the li chemical bonds. For a quantitative continual revisions of what had been use of bond energies as a predictive

; understanding of chemical kinetics, a thought to be well established numbers. tool, deals with the isomerization of

= knowledge of the appropriate bond Recent work at NBS has led to new in- small ring hydrocarbons. This is a sub- I energies is necessary. Chemical kinet- sights on such measurements. The va- ject of great current interest for which

icists usually define the bond dissocia- lidity of this work is suggested by the a considerable volume of data exists. tion energy as the enthalpy of reaction self-consistency of the results as de- The study of such highly strained sys-

' for the process AB A-fB at 298 K, rived from a wide variety of measure- tems pushes chemical theory to its r where A and B are molecular fragments. Specifically, the NBS studies limits. The simplest way of visualizing

I ments. Note that this is a thermody- have concentrated on the reactions such processes is to postulate a biradi-

namic quantity is and directly related kf cal intermediate formed from the to the enthalpies of the molecule and A2 <=> 2A

] fragments. However, since A and B kb Table 1. Comparison between old and I are frequently unstable, direct deter- (where A = ethyl, isopropyl and t- new values for carbon-hydrogen disso-

: minations through heat of combustion butyl radicals) , and have shown that ciation energies. measurements are not feasible. j Indi- high temperature determinations for k f I rect techniques must therefore be used, at 1100 K (carried out at NBS) and a Previous and for all except the smallest frag- reinterpretation of low ( temperature recom ments, this means kinetic j methods, measurements, are completely con- New NBS mended Thus, j one selects a reaction involving sistent with existing results on the re- values values the unstable species in question and verse reaction kb . These combined Compound (kJ/mol) (kJ/mol measures the rate of the reaction in studies cover a range of rate constants n Ethyl-H 420 412 both directions. Detailed balancing of 35 orders of magnitude. Thus, for

' Isopropyl-H 407 395 yields an equilibrium constant which these "over-determined" systems we N-Propyl-H 420 409 » can then be converted to heats of reac- arrive at a consistent set of thermal

tion if S-Butyl-H 407 398 I structural information for all properties for the unstable species A. species is I-Butyl-H 422 417 j known. Alternatively, a heat Furthermore, the calculated entropies T-Butyl-H 404 387 I of reaction can be calculated from the agree with present day ideas on the T-Amyl-H 405 temperature dependence of the equilib- molecular structure of the species in- Cyclohexyl-H 416 395 rium constant as measured across a volved. The enthalpies are, however, Allyl-H 377 369 range of temperatures. For all these significantly different from those de- Isobutenyl-H 373 measurements the instability of the termined by scattered measurements of : molecular 1-Methylallyl-H 375 355 fragment introduces severe the type described earlier. The NBS re- 1 Propargyl-H 384 393 uncertainties and the heats of forma- sults are thought to be more reliable 3-Methylpro- tion or bond energies so determined are due to the larger data base and con- pargyl-H 385 necessarily fraught with assumptions. sistency of determinations. Generaliz- Nevertheless, Benzyl-H 372 352 for bond energies to play ing from these results, we arrive at the

43 NATIONAL MEASUREMENT LABORATORY

Dynamic Thermophysical Measurements

A red Cezairliyan, Thermophysics Division

Table 2. Comparison between predicted and experimental activation energies JL. ^L»esearch has been carried out for ring opening reactions. to explore, by dynamical means, the thermophysical behavior of solids and Experimental From NBS bond Previous fluids under extreme conditions of Compound (kJ/mol) energies (kJ/ mol) results (kJ/mol) rapid heating. In addition to studying properties of materials of interest in the Ethylene (2-membered ring) 275 272 255 areas of high temperature sciences and Cyclopropane 268 255 238 technologies, we have attempted to: 1,2 Dimethylcyclopropane 259 263 250 1,1,2,2, Tetramethylcyclopropane 228 232 188 • clarify the anomalous behavior of 1,2 Dimethylcyclobutane 257 251 230 the heat capacity of metals at high

Vinycyclopropane 208 208 108 temperatures (fig. 1), • determine the effect of dynamic heating conditions on vacancy gen- breaking of the highly strained C-C perimentally measured. As a result, it eration and migration, bond, followed by appropriate rota- has been necessary to postulate a 20 to • study superheating in phase transitions and finally recyclization. The ac- 40 kJ activation barrier for recycliza- tions, and tivation energy for the isomerization is tion or rotation. Structural and quan- • investigate the non-linear contributions therefore: E = AHT (ring opening) -f- tum mechanical considerations are to the transport phenomena RT and AHT (ring opening) = 2DT however at variance with the assump- under large thermal and electrical (C-H)-DT (H-H) + aHt (hydrogenation of such a barrier, making this sub- gradients. tion) where DT is the bond dissociation ject quite controversial. With the new energy at temperature T and the heat bond energies, we obtain the activation Because of limitations in the conven- of hydrogenation refers to the conver- energy results summarized in table 2. It tional steady state techniques at high sion of the cyclane to the appropriate can be seen that the deficit has been temperatures (above 2000 K), we have aliphatic compound. Since the last two almost completely removed and in developed a dynamic technique for 1 terms are accurately known, the de- most cases is probably within the ex- conducting the research. For conven-" pendence of the activation energy on perimental uncertainty. Thus, it is now ience, our activities may be divided into 1 the bond dissociation energy is quite clear that the biradical mechanism does clear. Such calculations performed provide an accurate quantitative basis using the older bond dissociation ener- for predicting the activation energy for gies have always resulted in values ap- the isomerization of small ring com- proximately 20 to 40 kJ lower than ex- pounds.

1000 1500 2000 2500 300I Temperature, K

Figure 1. Heat capacity of molybdenum. The results indicate considerably higher values at high temperatures than the classi- 1 1 cal value of 3R (about 25 J mol- K- ).

44 CENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE

two categories, namely "millisecond" and "'microsecond" systems—terms indicative of the speed of the dynamic experimentation. In either case, the high-speed pyrometer

technique is based on rapid resistive self-heating of a specimen by an elec-

trical current pulse and on measuring the pertinent experimental quantities with appropriate time resolution. In the millisecond system, we under-

took two types of activities : ( 1 ) meas- the detector uring selected thermophysical prop- to erties of new materials with the facility developed earlier in this program, and

(2) developing new capabilities to enable the measurement of additional properties. In the microsecond system, we continued to develop a facility for extending accurate measurements to temperatures above the melting points of materials and up to 6000 K. Some of the highlights are given below. With the "millisecond" system, we Figure 2. Schematic diagram of the laser pulse interferometer for rapid measurement of thermal expansion. electrical resis- , measured heat capacity, tivity, and thermal emittance of a , graphite composite used as a high tem- in applica- . perature material aerospace tions. Similar measurements, in addi- experiment trigger tion to its melting point, were also timers and performed on palladium. It should be pulse generator noted that the melting point of palla-

dium is a reference point on the International Practical Temperature

Scale and that its value is based on measurements performed over 40 years specimen spark ago. New developments, as additions __-ri temperature gap trigger c O to the millisecond system, include a """"—LI measurement circuit capability to measure thermal expansion simultaneously with other properties. We designed, developed, and performed a feasibility study of a laser standard rim charging pulse interferometer for this purpose resistor circuit

(fig. 2) . Another addition was a tech- tttt

nique for measuring heat of fusion of capacitor metallic materials in experiments of bank sabsecond duration. To check the technique, we performed measurements Figure 3. Functional diagram of a dynamic (microsecond-resolution) thermophysical cn niobium. Other new areas that we measurements system.

45 NATIONAL MEASUREMENT LABORATORY

Pressure as a Probe of Molecular Interactions

Vern E. Bean, Thermophysics Division

JLhe. he use of pressure as a variable embarked on include (in conjunction T so that only la is formed, then la melts with S. Block's group, Center for parameter is a valuable tool for study- around 245 K. Thus, at atmospheric

Materials Science) developing a meas- ing molecular interactions. We have pressure, CC1 4 has two melting points urement technique under combined recently used differential thermocouple some 5 degrees apart. We have studied conditions of high temperatures and analysis to explore the phase transitions these phase changes as a function of high pressures. Research is presently in carbon tetrachloride (CCI 4 ). Our hydrostatic pressure up to 350 MPa underway to assess the feasibility of measurements demonstrate that ( C I, and over the temperature range of 210 pulse heating specimens in a diamond has dual melting curves up to a pres- to 330 K with the following results: sure of 100 MPa. anvil high pressure cell. The objective 1 . Between atmospheric pressure and is to determine the melting temperature 100 MPa, CCU has dual melting of materials as a function of pressure curves, one for la, and, a few degrees

(up to above 10 GPa). There are four phases in CC1 4 in the higher, one for lb, as shown in figure 1.

As an important step in developing temperature/pressure region investi- Above 100 MPa, there is only one an accurate microsecond resolution gated: liquid, a face-centered cubic melting curve, that of lb. These two thermophysical measurement system phase (denoted as la), a rhombohedral curves diverge with increasing pressure.

(fig. 3), we designed and constructed a phase (lb), and a monoclinic phase Above 273 K and 100 MPa, it was not prototype pyrometer which incorpo- (II). Both la and lb are plastic crys- possible to detect the formation of the rates a digital data acquisition system. tals. The interest is in the relationship la phase.

The pyrometer has the capability of between the liquid, la, and lb. When 2. The Ia-Ib phase boundary is meta- measuring the temperature of a rapidly CC1 4 is cooled at atmospheric pressure, stable. It is vibration sensitive and heating specimen every one micro- it solidifies into la at about 245 K and pressure rate dependent. The lb phase second with a resolution of about 0.1 with further cooling transforms to lb. will abruptly and spontaneously form percent. When heated, lb melts around 250 K from the la under a variety of condi- We have also advanced the state-of- without passing back through the la tions. Ib will also form from the liquid the-art in electrical power measure- phase. Should the cooling be limited without passing through the la phase. ments of high-voltage, high-current dis- charges. We plan to use this very fast 350 system to obtain thermophysical prop- + = Ross and Andersson erties data on solids and liquids up to X = Makita and Takagi

6000 K and to conduct studies on the 325 behavior of materials under rapid Liquid dynamic heating conditions. 300

2 3 Ib a JL-> ». 275 m a, Ia^> Liquid E * . A'

250

225 i*^

1 1 1 1 1 1 200 ' i i i i I i i I i i i ' i i i i i : i i i i i I i I I 0 50 100 150 200 250 300 350

Pressure, MPa

1. for . Figure Melting curves the la and Ib phases and the II-* Ib transition curve for CC1 4 46 ;enter for thermodynamics and molecular science

The Catalytic Methanation Reaction: A Source of Synthetic Natural Gas

John T. Yates, Jr., David W. Goodman, Richard D. Kelley, and Theodore E. Madey, Surface Science Division

5. There is no liquid-Ia-Ib triple point. ments in fluids which use the precision

of ultrasonic interferometry. The tech- ynthetic methane, CH 4 produced 4-. Liquid CC14 will superpress (which s

: nique involves making ultrasonic meas- from coal, will become a major com- s analogous to supercooling in an sobaric experiment) forming a band in urements on two samples of the same ponent of the Nation's natural gas Jihe T/P plane where the metastable fluid in the same pressure vessel. One supply after 1990. Heterogeneous

; catalysts will certainly iuid exists. The width of this meta- sample has a constant ultrasonic path play a signifi- length, the other a constant mass. cant role in achieving this objective. . table band increases with pressure. From the data, one obtains the speed We have been studying the funda- A number of other experiments of sound in the fluid as a function of mental mechanism of the key catalytic [where pressure is an independent vari- pressure and the isothermal compressi- reaction involved in the synthesis of able are either underway or are being bility without going through the usual methane. Using modern surface meas- 'developed. They include: adiabatic to isothermal conversion. urement methods, we have determined I. Light scattering measurements on This capability will be used to: (a) the chemical mechanism for the cata- ,iuids at high pressure through use of look for density anomalies in the meta- lytic hydrogenation of carbon mon- ?hoton correlation spectroscopy for oxide, stable liquid region of CC1 4 , (b) obtain CO, over nickel catalysts, and "righly viscous liquids, and Brillouin compressibility data on fluids having certain mechanisms of catalyst poison- .cattering for less viscous liquids Such nearly spherical molecules (to be used ing have been thoroughly investigated. measurements will yield: (a) the mean with change of the index of refraction This understanding of catalytic proc- ime for structural relaxation, (b) a with pressure data, obtained by K. esses should provide a conceptual basis measure of the distribution of relaxa- Vedam of Pennsylvania State Univer- for optimizing the efficient use of cata- ion times about the mean, and (c) a sity, in a critical evaluation of equations lysts in what promises to become a neasure of the elastic response of the of state of fluids), and (c) obtain com- multibillion dollar national enterprise. iquids to high frequency disturbances. pressibility data to be used in conjunc- 3ased on these data, researchers will tion with dielectric relaxation measure- nitiate a program to develop theories ments in an attempt to apply activation All proposed routes from coal to vhich relate the rearrangement dynam- volume concepts to metastability in high-BTU methane involve the initial !cs to the mechanisms which lead to fluids. gasification of coal using steam and air 'iscosity and connect these mechanisms 3. Capacitance measurements on fluids or oxygen and yielding coal gas, a mix- vith the ease with which metastable as a function of pressure, temperature, ture of CO, Ho, and CH 4 . While this itates can be realized. and frequency for the dielectric relaxa- mixture will burn, its BTU content is I. Isothermal compressibility measure- tion studies mentioned above. too low ( 150 to 500 BTU per standard cubic foot) for economical transport through the Nation's pipeline system. Therefore, the coal gas mixture must be catalytically converted to pure methane (1000 BTU per standard cubic foot). Although the catalytic synthesis

of methane from CO and H2 has been studied for the last 50 years [1,2], until recently the mechanism by which the reaction occurs had not been definitely determined. We have been successful in focusing modern surface measurement techniques on this reaction to

learn how it works and how to maintain efficient catalyst operation.

Carbon monoxide is the most strongly bound molecule known. The

47 NATIONAL MEASUREMENT LABORATORY

tion mechanism [3]. Use of the crystal position of the surface of the Ni crysta permits us to utilize modern surface following methane synthesis. We were spectroscopy techniques for studying able to show that a fraction of a mono-

the reaction. However, it was first layer of "carbidic" carbon always exists

necessary to show that the Ni crystal during the catalytic reaction. It origi-

behaved similarly to actual Ni catalysts nates from CO dissociation on the N: ? having about 10 6 cm2 surface area per surface: the rate of dissociation of the

gram. A spectacular result was ob- CO is closely related to the rate of CH,

tained, as shown in figure 1, where the production [5]. If either the H 2 /CC

rate of CH4 synthesis (turnover num- ratio or the temperature falls outside ar

ber) over the Ni(100) single crystal is optimal range, a graphite layer form;

plotted against 1/T as an Arrhenius on the surface; this graphite layer is £ plot. Two important conclusions are poison which deactivates the Ni cata-

apparent: (a) linear Arrhenius behav- lyst. Such poisoning is a principa -4 ' 1 1 1 1 1 1 1 1 L > 1Q 1 — ior is observed over a very wide dy- problem in methane synthesis ovei c 1.2 1.4 1.6 1.8 2.0 2.2 namic range (5 orders of magnitude), technical catalysts. It was also founc H 1/T X 103 (K-1) giving an activation energy of 103 that extremely small traces of gaseous

1 kJ/mol- (~10 percent of the C-0 Fe(CO) 5 present in many sources o: kJ/mol dissociation energy), and (b) unpurified CO produced by coal gasifi- Figure 1. Arrhenius plot of rate of CH, synthesis The agreement of the rate measured for cation will decompose to yield an Fe (NCH ) over different Ni catalysts in various temperature regimes. Ni(100) the single, crystal with rates and activa- covered Ni surface which immediately

present P,-,, work; PH 2 = 96 Torr, = 24 tion energies measured on Ak>On-sup- produces the graphite poison, shorten- Torr (0.85 cm2, E = 103 kJ/mol); (a) 25- a ported high-area Ni catalysts is excel- ing the catalyst lifetime. 50% Ni/Al 0 E = 84 kJ/mol; (b) 5% 2 3 , a lent. The single crystal may be used The mechanistic details which wt Ni/Al 0 E = 105 kJ/mol; (c) 8.8% 2 3 , a Ni/Al 0 E = 109 KJ mol. over a wider temperature range than have inferred are that both CO am 2 3 , a the technical catalysts, giving more hydrogen dissociate on the Ni surface 1

energy required to dissociate the car- confidence that true activation energies The carbon atoms form an active ! -1 bon-oxygen bond is 1070 kJ/mol . are being determined. "carbide" layer on the Ni; the carbide

Because of this basic fact, it is neces- We have directly shown that the species react with absorbed H to forn

sary to use a catalyst surface to achieve single crystal is a good model catalyst -CH X intermediates. Further hydro ,

a suitable chemical pathway leading to: [3, 4], and that results obtained on the genation of these intermediates result:

(a) to CO dissociation and (b) to single crystal are transferable to the in the formation of CH 4 product. A hydrogenation of the fragments to pro- real world. simplified view of the mechanism foi

duce CH 4 and H 2 0. The catalyst must Auger electron spectroscopy was the reaction of CO and H 2 over Nj

be able to carry out both reaction steps then used to study the chemical com- catalysts to yield CH 4 is schematically (a) and (b) with suitable efficiency and with the desired selectivity. In other

words, it must produce pure CH 4 efficiently without significant production of other undesired products. A number of transition metal catalysts are

suitable, but Ni is, by a large margin, the most inexpensive catalyst currently employed in laboratory and pilot plant

CH 4 synthesis studies. We have employed a Ni single crystal (1 cm 2 surface area) as a model catalyst for our investigations of the reac-

48 CENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE

Multiphoton Chemistry

John C. Stephenson, Molecular Spectroscopy Division

frustrated in figure 2. Efforts are JL..undamentalund studies of how in- been done with C0 TEA laser pulses F 2 rnderway to study directly the molecu- tense infrared laser light causes chemi- focused into bulbs of gas at high pres- lar nature of the postulated CH4 sur- cal reactions are described. Time- sures and have relied on chemical ace intermediates. We are using resolved in situ spectroscopy is used to analysis of the final stable end prod- leutron energy loss spectroscopy to detect reaction rates, product energy ucts. Such high pressure conditions

nvestigate these species [6]. Also, we states, and isotope enrichment factors, may be necessary for potential com- , plan to study the details of sulfur as functions of laser wavelength, mercially practical processes. How- poisoning of Ni single crystal catalysts; fluence, intensity, and gas pressure. ever, it may be difficult to relate the

t is known that this will be a major The data is compared to a comprehen- results of such experiments to the ^echnical problem in the industrial sive model, yielding new conclusions basic questions of how IRMR occurs,

orocess because sulfur is a major im- about the dynamics of excited This difficulty arises from complicat- purity of coal. molecules. ing factors related to collisions be- A general conclusion from this work tween highly excited molecules, radical

s that modern surface measurement Radiation from powerful and effi- chain reactions, reactions on cell walls, e ools may be applied successfully to cient infrared (IR) lasers may directly secondary photolysis of initially formed ; xnravel the secrets of catalytic chemis- induce chemical reactions such as mo- products, and undefined or inhomoge- ry and to provide new and useful lecular isomerizations and dissociations neous laser fluence (energy unit area) concepts applicable to practical cata- [1]. Since the energy required to break or intensity (fluence/unit time). ytic chemistry. a chemical bond (e.g., 20 000 to Our experimental approach [2-6] -1 40 000 cm ) is substantially greater has been to use a C0 2 laser pulse than the energy of an IR laser photon (about 10~ 7 second duration) to cause

•References (e.g., the energy of a C0 2 photon is an IRMR in a very low pressure gas, -1 1000 cm ), a molecule must absorb such as: Chemical and Engineering News, 57, many IR laser photons to react.' For '"CO., 20-28 (1979). "Synfuels: Uncertain this reason, such chemical reactions CFoHCl (v,J) and Costly Fuel Option," by Ward >CF2 Worthy. are called "infrared multiphoton reac- + HCl(v,J) - ET (1) laser = ; 2] Mills, G. A. and Steffgen, F. W., tions" or IRMR. This IR AH 209 kJ/mol Catalysis Reviews, Science and En- radiation selectively excites particular gineering, 8, 159 (1973). chemical bonds in a molecule, unlike '"CO., Goodman, D. W., Kelley, R. D., J3] ordinary thermal reactions in which CH.CFCl >CF 2 (vJ) Madey, T. E., and Yates, J. T. Jr., energy in the reactant molecules is dis- - CFCl(vJ) - I (2) Kinetics of hydrogenation of over CO tributed randomly. Chemists have AH = 419 kJ/mol a single crystal Ni catalyst, J. Catalysis hoped that particularly efficient or We then use a weak visible or uv probe (in press). unique reactions could be effected laser pulse (5 ns duration) to detect in '4] Goodman, D. W., Kelley, R. D., through laser induced "bond specific real time before collisions occur (10 ns Madey, T. E., and Yates. J. T. Jr.. in, Hydrocarbon Synthesis from Carbon chemistry." resolution) the rate [3. 4] at which Monoxide and Hydrogen, E. L. An enormous amount of research products are formed. This technique Kugler and F. W. Steffgen, eds., pp. on laser-induced chemistry is now lets us probe primary processes result- 1-6, Advances in Chemistry Series being done at universities, U.S. and ing from the coupling of isolated mol- 178 (Am. Chem. Soc, Washington, foreign national laboratories, and at ecules to the radiation field, before the B.C., 1979). several leading corporations. Potential complicating effects mentioned above 5] Goodman. D. W., Kelley. R. D., applications of IRMR include separa- have occurred. We probe a region of Madey, T. E., and White, J. M., tion of stable and fissionable isotopes. constant, known C02 laser fluence to Measurement of carbide buildup and Our research here at NBS has focused determine the vibrational (V), rota- removal kinetics on Ni(100), /. Cata- on developing the basic understanding tional (J) and translational energy lysis (in press).

of how IR laser excitation causes such (E ) of the products, 6] and the 6] Kelley, R. D., Rush, J. J., and Madey, T [2, 13 12 T. E., Chem. Phys. Letters, 66, 159- reactions. isotopic enrichment [5] (e.g., C/ C) 164 (1979). Most experiments in IRMR have in the products.

49 )

NATIONAL MEASUREMENT LABORATORY.

We have measured all these factors Any consistent theory of IRMR must factor of 6. This result shows that foi and determined their dependence on explain product yields, formation rates, CF2HC1, coherent excitation of the the intensity, fluence, and wavelength isotope enrichment, and the energy molecule is important, and that a sig of the IR laser. Another important content of the products as a function nificant barrier to laser excitation oc experiment is underway to measure of pressure and laser intensity. We are curs in the discrete lower vibration these parameters under conditions in making progress in developing a com- rotation levels of this molecule. It which the reactants are extremely dilute prehensive theory. figure 2, the absolute rate [3] at whicl 6 (e.g., 1 part in 10 ) in a high pressure Some specific results on reaction ( 1 CF2HC1 molecules decompose is showr bath of Ar atoms. Under these cir- in 1 2. figure as a function of pressure, for are shown figures and In Ar CO : cumstances, the reactants undergo 1, the collision-free reaction yield [4] laser pulses of two different intensities many collisions with bath molecules is shown as a function, of laser fluence The initial rise in reaction rate wit! during the laser excitation process. F for COo laser pulses of different peak increasing Ar pressure is due to col

1 intensity, I. One popular theoretical lision-induced rotational relaxation

model [7, 8] of IRMR, used to explain i.e., CF2HC1 collisions with Ar atom: many results on larger reactant mol- result in rotational equilibration withii ecules such as SF predicts that the the absorber pump mode, allowing tht 6 , reaction yield depends only on F and pumping of molecules which were orig

is independent of I. This is dramat- inally in rotational states outside th<

ically not true for CF L.HC1, where for power broadened line. The fact tha laser pulses of constant F, the yield there is no downturn in rate at high A 5 CD increased by a factor of 400 as the pressure shows that vibrational de

O average intensity I increased by only a activation of excited CF2 HC1 reactan 3 T5 molecules is not as efficient as theorie

[3, 8] had predicted.

By using a comprehensive model t< Peak intensity understand data like that in figures ] • 880 MW/cm2 and 2, we can make predictions o x140 MW/cm2 draw conclusions about molecular dy 29 MW/cm2 namics that are very difficult to meas 12 MW/cm2 ure directly. Figures 3 and 4 shov

such predictions [3, 8] based on thi

1 10 100 model which fits the data for figure 2 Fluence F (J/cm2) 100 200 300 In figure 3, the distribution of vibra Ar pressure (torr) Figure 1. Intensity dependence in the IRMR tional energy in the laser excited re of CF..HC1. Absolute product yields from Figure 2. Rates of production of XCF 2 actant (solid line) is shown at time the dissociation of CF ,HC1 (9.3 R(32) by L during the dissociation of CF..HC1 C02 (1, 2, and 50 ns) during the lase 1 COo laser transition of 1086 cm- ) for non- laser pulses, determined for CFoHCl very pulse when the average reactant ha< modelocked laser pulses for four different dilute in Ar at the indicated Ar pressures total energies. 2 absorbed 5, 10, and 15 photons, re The pulses all exhibited iden- (1 Torr = 133.3 N/m ). The pressure tical temporal envelopes and are labeled dependence was measured for two ex- spectively. For comparison, the dashei according to the peak intensity. The data laser oper- perimental conditions: (O), C0 2 lines show thermal distributions o given for each different pulse represent the transition at 1086 1 ating on the R(32) cm- the same average energy. Note that th yield of CF., observed at various times r , which is resonant with the CF ,HC1 one- D L laser-driven distribution always differ during the IR laser pulse plotted against the photon absorption, peak laser intensity of from a thermal distribution. Figure < 150 MW/cm-; (A), R(32) laser energy reduced by a factor of 10. The curved lines shows the absorption cross section fo fluence F(t d ) I(t)dt for that time through the data points are the theo- drawn a CF L.HC1 reactant to absorb CO retically calculated results, based on the radiation as a function of the amoun and laser pulse. The total pulse energies model described in ref. [3]. The isolated of energy the CF^HCl has absorbed were 2. 5, 25. and 150 mJ. respectively, for point shows the magnitude of the (± 35%) the 12. 29, 140, and 880 MW/cm? pulses. associated with each experimental datum. Note that the absorption cross sectici

50 Renter for thermodynamics and molecular science

ecreases rapidly as the energy in- and detailed understanding useful in J. Am. Chem. Soc. 100, 7151-7155

reases. This decrease is due to very all areas of gas phase chemistry and (1978). Stephenson, J. C, Bialkowski, S. E., =kpid intramolecular dephasing rates molecular dynamics. [6] and King, D. S., Energy Partitioning " short T2 times) caused by anharrnonic References in COo Laser Induced Multiphoton ouplings. result is The a "smearing [1] Cantrell, C. D., Freund, S. M.. and Dissociations: of > and L Energy XCF L X ut" of oscillator strength over a very Lyman, J. L., Laser Induced Chemi- CFC1 from CF.,CFC1, J. Chem. Phys. road spectral region. cal Reactions and Isotope Separation, 80, (1980). field of is extremely ac- in Laser Handbook, Vol. II (North- The IRMR [7] Bloembergen, N., Yablonoviach, E., Holland, Amsterdam, 1978). |ve. Questions related to coherent vs. Physics Today 31, 23-30 (May 1978). [2] Stephenson, J. C. and King, D. S.. icoherent excitation. T x and T 2 times. [8] Goodman, M. F., Stone, L, and J. Chem. Phys. 69, 1485-1492 (1978). ' nd the feasibility of "bond specific Thiele, E., Laser-Induced Decomposi- [3] Stephenson, J. C, King, D. S., Good- 'hemistry" are still controversial, as tion of Polyatomic Molecules: A man, M. F.. and Stone. L, /. Chem. ~ Comparison of Theory with Experi- re questions regarding the commer- Phys. 70, 4496-4508 (1979). ment, in Multiple Photon Excitation "ial practicality of IRMR processes, King, D. S. and Stephenson, J. C, [4] and Dissociation of Polyatomic Mole- however, these experimental and the- Chem. Phys. Letters 66, 33-38 (1979). cules, Topics in Current Physics, C. ' retical IRMR studies are creating new [5] King, D. S. and Stephenson, J. C. D. Cantrell, ed. ( Springer-Verlag, I New York 1979), and op. cit. therein. l

5 10 15

Average vibrational excitation

Figure 4. CF2HC1 absorbtion cross section calculated for various values of energy absorbed (v). These average absorptions were achieved by varying the time of irradiation at an intensity of 150 MW/cm2 and a pressure of 300 Torr Ar.

0 5 10 15

Vibrational excitation (v)

i ure 3. Laser excited distributions ( —) compared with thermal distributions ( ): a) average energy (v) = 5: (b) average energy (v) = 10; and (c) average (v) ! energy

= 15. The extra curve ( ) in (c) for (v) = 15 is the thermal distribution for a nclecule arbitrarily truncated at 20 levels. These model predictions are for a CO., laser ntsnsity of 150 MW and 300 Torr Ar.

51 '

NATIONAL MEASUREMENT LABORATQR'

Quantum Calculations of Excimer Molecules

Morris Krauss, Molecular Spectropy Division

Quantum chemical calculations All the molecules of a class are of states of the Group VI atom and thi are being performed to interest; thus the all-electron techniques ground state of the rare gas atom. Wi identify and analyze the potential of have been extended by introducing have determined that, with the excep candidate electronic laser systems for pseudo-potentials to replace the core tion of XeO, the transistions are dom the Department of Energy and Defense electrons. Both non-relativistic and inantly continuum-continuum. Calcu ! Advanced Research Projects Agency- relativistic effective core potentials lation of the transition line shape v sponsored programs for generating (ECP) have been generated and used found to require a multichannel coul ' high power laser beams. The net gain as necessary to treat high Z atoms. pling among all states arising from th of a candidate system must be known. Calculations of the electronic proper- valence asymptotes. The line-shape fo Experimental determination of this ties of ion-pair molecules using ECP the Ar-O^D) - Ar-0( ] S) transitioi property is difficult for exciplex or have compared favorably with all- has been calculated in reasonably! excimer systems. Calculation of the electron results, although more re- agreement with an experimental deter; spectroscopic behavior of excimer and search is still needed on the use of the mination. continuum-continuum systems has ECP. Using the pseudo potentials, we Pb vapor is now being explored as : proved to be an important source of have considered molecules containing Raman scattering media for producin; information. elements as heavy as Hg and Pb. The highpower backward scattered pulse ECP take the relativistic contraction of from an XeCl laser pulse. Rare ga We have calculated the spectroscopic the core into account, but at the present buffers are used to enhance the back properties of important molecules in time the spin-orbit interaction in the ward scattering cross section over thr the following classes of laser candi- valence shell is treated as a perturba- forward. Collision-induced absorptioi dates: Group Vl-rare gas, rare gas- tion.. at either the pump or Stokes frequency halide, Group lib-Group Ha and b The transition moments and energies would reduce the efficiency of the proc dimers, Group IV-rare gas, and rare have been calculated for the X-A ess. A calculation of Pb-Ffe and Pb-Xi gas-rare gas. bound to continuum transition in HgCl. energy curves and transition momenti :

Ab initio molecule calculation for The photodissociation cross section is reveals that collision-induced absorp j evaluating energies and spectroscopic illustrated in figure 1 as a function of tion is not significant at the frequencie: i properties can be done for small the transition energy and temperature. of interest. molecules containing low Z atoms using The laser energy for the B-X transition The rare gas halide diatomic anc all-electron multi-configuration varia- is close to the peak energy, and a signi- triatomic excimer spectroscopy owe i tional techniques. Considerable experi- ficant absorption will result from a great deal to theoretical calculation ence has been accumulated on dealing buildup in the concentration of X state Not all of the relevant states have beer with the calculation of the correlation HgCl in the course of exciting the laser observed directly by experiment anc energy involved in bonding interactions media. This absorption could explain assignments of some transitions an over the entire range of internuclear the low extraction efficiency observed still not definitive. The relative ener separations. Of particular importance for certain types of HgCl lasers. for excimer systems is the recent devel- Self-absorption in Cd 2 is also pre- 2 opment of techniques for calculating dicted by ab initio calculations for the van der Waals interactions into dis- excimer transition. From the energy tances where the overlap of the charge curves shown in figure 2, we see there distributions is no longer 3 -» 3 negligible. is strong bound-bound, IIg IIu ab- We have shown that the modification sorption near 470 nm. Absorption was of the usual inverse power series expan- also predicted earlier for the 335 nm sion by a damping function x(R)> emission from Hg 2 from the reservoir metastable excited states. -1 E(vdw) = Transition energy (1000 cm ) -2 C(/,L) v(/,L,R)R- 2 " + l +d The Group Vl-rare gas laser transi- Figure 1. Photodissociation cross section foi tion is a collision-induced electric 2 2 the X 2 — A II transition in HgCl as £ can be applied to all the systems men- dipole transition between electronic function of transition energy and tempe tioned above. states that correlate to the valence ature.

52 ENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE New Development in the Evaluation of Thermochemical Data

D. D. Wagman, David Garvin, Vivian B. Parker, Richard H. Schumm, and 1 J. Brian Pedley , Chemical Thermo- dynamics Division

;tes of the B and C states in the Theoretical calculations have proved A comprehensive, up-to-date iatomic system are of present interest, to be a fast, accurate way to obtain the compilation of evaluated thermochem- ve are now comparing our calculations spectroscopic data needed in modelling ical data has been completed by the ith the interpretations XBS Chemical Thermodynamics Data r of the experi- laser media. We are continuing our lental data. We agree with previous studies of the rare gas halide lasers by Center. The compilation is called leoretical all-electron calculations that examining photodissociation of bottle- Selected Values of Chemical Thermo- le B state is lower in energy than the necked states in XeF and the likelihood dynamic Properties. It covers the periodic table except for some of the f for XeF and XeCl. of positive gain for the Xe 2 Cl system. actinide elements and includes over 12,000 inorganic and low molecular weight organic compounds. These include pure substances in the gaseous, liquid, and solid phase as well as many binarv solutions.

The properties tabulated in the compilation are the enthalpy of formation, 3 A fH , Gibbs energy of formation,

A fG°, entropy, S°, and heat capacity,

°, Cp all at the temperature 298.15 K (25 °C). In addition, values are provided for the enthalpy of formation at 3 0 K. A f H (0). and the enthalpy

difference H° (298) - H° (0). The values are based almost entirely on experimental measurements and statistical calculations. Approximately 20,- 000 property values are included. Well over half of these are enthalpies of formation. The next two most numerous entries are for Gibbs energies of formation and entropies. These tables are designed for use in chemical engineering, chemical and biochemical research, rocketry, plasma studies, environmental control, energy conservation, and other related tech-

nological fields. Thermodynamic prop- 4 5 6 7 8 9 erties are essential for calculating the Internuclear separation (Bohr) energetics of chemical processes and

7 igure 2. Excited state energy curves for the Cd„ molecule calculated with multi-configura- optimizing of equilibrium yields. They | on self-consistent-field variational techniques. are widely used in interpreting physico- chemical research. The numerical values in the tables resulted from critical analysis and evaluation of many different types of

1 University of Sussex, England.

53 NATIONAL MEASUREMENT LABORATORY

physico-chemical measurement. Vari- values for the substances. The follow- ^=^LiAi 0 (c) ous calorimetric reaction enthalpies can 5 8 ing criteria are required for this com-

be combined to yield values for the II ponent: (1) The recommended values LiO(g) of of particular Life) = must reproduce well those measure- enthalpy formation substances. High temperature equilib- ments considered to be reliable. (2) rium studies may be combined with The values selected must be thermo- values of the thermal functions for the dynamically consistent throughout the. individual species in the reaction to tables and be in reasonable accord with yield desired values. Calculating ther- physico-chemical correlations (similai mal functions of gases by statistical- substances should have similar prop-

mechanical methods requires a knowl- erties) . edge of molecular structure parameters Recently, this second component ha; (bond lengths and angles) and vibra- been automated. More than 1,00(! tional frequencies; the calculation for a measurements involving 600 or mors

substance in the condensed phase re- substances may be involved in a partic-.

quires values for the enthalpy content J ular problem. Computer technique; LiOH(1 ,000H 2 O) LiOH(400H O) 2 as a function of temperature and en- are ideal for manipulating such large, thalpies of phase transitions. Low masses of data. In addition, a system],

° or less, con- fit temperature (298 to 0 K) Cp meas- Figure 1. Loops of 5 members atic of the data is obtained with ar

; each line between urements are used to obtain the third taining reaction (a) analysis of the network. This fit is usec> ; substances represents one reaction between | law S° (298 K). Similarly, data on as a guide to the strategy of date, lithium compounds; processes (b) and solubility, activity and osmotic coeffi- Q evaluation and to indicate the impor-. appear to contribute large errors. cients, vapor pressures, and enthalpies tance of various substances. Rapic

: of solution and dilution are required several different measurement paths. updating of the recommendation is pos- for the proper evaluation of measure- The interconnections among the paths sible when new and significant meas. ments on solutions. Enthalpies of com- urements become available because figure 1 . results may be seen in The | bustion, hydrogenation, and halogena- obtained from the various paths usually these evaluated measurements ar\ tion are used in the calculation of do not agree. Choices must be made in stored in reusable form. The network

thermodynamic properties for organic establishing the "best" values. solutions combining the measurement: ! molecules. are done in an iterative process ir There are two fundamental com- 3 linear To do the above evaluations, we ponents in this evaluation process. The which programming is used tc reference file of locate the most highly consistent set oj have amassed a numer- first is the heart of the process and must ical from measurements and weight accord- data abstracted more than be done by an expert. It is necessary them ingly, prior 40,000 articles describing over 200,000 to determine, for each measurement, in to a least squares solution measurements made since 1848 on the light of present knowledge, whether This two-step approach has been showr individual chemical substances. Each or not the interpretation of the chemis- to mimic the decisions made by date analysts year, an annotated bibliographic refer- try was correct and whether the tech- when they combine the meas- ence list of all material added to the nique was suitable. Reinterpretation, urements in the traditional "by-hand' Data Center files is published in the recalculation, and correction of the solution. It results in a solution fron Bulletin of Chemical Thermodynamics. data often are necessary. A consistent which the user may combine any set o list, by chemical formula values in the data to predict witl This arranged set of auxiliary data is used in reducing bank of confidence the the substance measured, indicates the measurements, e.g., physical con- thermochemistry of s the type of measurement made as well stants and properties of ubiquitous process. as the full bibliographic citatioi substances. Accuracies of the measure- Automation has been an importan Construction of these reference data ments must be assigned. factor in the development of the date

tables is an exercise in the art of data The second component is the com- bank. The selected values are stored

evaluation. Often the properties of a bination of the accepted measurements in a machine-readable file. Tables car substance can be determined by using to produce the final, recommended be printed in any system of units

54 CENTER FOR THERMODYNAMICS AND MOLECULAR SCIENCE

-20 10

Loop Error, Kcal

Figure 2. Errors in loops containing up to 5 members, and containing reaction (a) ; shaded portions are deleted when reactions (b) and (c) are discarded.

(Publication in SI and in "caloric units" rors of closure. An analysis of the loop

is planned.) A machine-readable cata- errors shows which measurements are

' log of the accepted measurements is highly inconsistent. As an example,

; being developed—both to document figure 1 shows a network involving the tables and for direct use in applica- enthalpy measurements for 16 lithium

tions. In a cooperative program with compounds. It shows all 5 membered the University of Sussex, England, au- (or fewer) closed paths involving proc- tomated quality control, network solu- ess a (there are 84 of these). The dis-

tion, and graphic display techniques tribution of errors of closure is shown

have been developed. These simplify in figure 2. The mean error is close to and speed up the routine parts of the zero, indicating that process a probably

task. These quality controls assure that is acceptable. But the distribution is

thermodynamic consistency is main- distinctly nonnormal. Processes b and

: tained. c account for all extreme values and

A recent development is "loop anal- therefore are of doubtful accuracy. ysis". In this analysis, measurements This technique holds the promise of 'are summed over every closed path in becoming a very rapid technique for hhe loosely connected network. Con- preliminary analysis of thermochemical 'servation laws require zero sums, but data and may play an important role in 'experimental inaccuracies produce er- future data evaluation strategies. 4 Center for Analytical Chemistry

~urt W. Reimann Richard M. Kristy L. Richie he Center for Analytical Chemistry maintains Lindstrom Robert Schaffer T .ucy B. Hagan activities George J. Lutz Lorna T. Snieaoski expertise in analytical measurement to support ohn K. Taylor Lawrence A. William J. in this Nation's industrial, clinical, environmental, and nstroment Machlan Sonnefeld** >eveIopment E. June Maienthal Pietr VanDerLijn** research laboratories. The results of analytical measure- ] ames R. DeVoe George Marinenko Ranee A. Velapoldi ments often form the basis for regulatory, medical, and Michael J. Welch** "homas C. O'Haver Radu Mavrodineanu Jerry D. Edward White. V 'illmer C. Ruegg Messman industrial decisions. It is essential that they be reliable. Klaus D. Mielenz Patricia A. White** .obert L. Sams Mitchell** Stephen A. Wise The Center serves as the reference laboratory for as- lonald W. Shideler J. John R. Moody William T. Yap .ichard K. Simon, suring this reliability. For example, methods of high Pedro A. Morales** Jr.* Gas and Particulate ohn C. Travis Thomas J. Murphy Science accuracy are developed for identifying substances and John A. Norris iregory C. Turk Harry L. Rook for quantifying the chemical composition of materials. Paul J. Paulsen Analytical Arnold M. Bass norganic Theodore C. Rains Materials standards of known composition and stability Research Robert L. Watters, Adville A. Bell 'rnest L. Garner Jr. John J. Blaha are prepared for aiding in quality control of measure- David S. Bright ohn Alienkoff** Tadashi Barry C. Cadoff ment procedures and for determining the reliability of ohn R. Baldwin Yamamoto** Rolf L. Zeisler R. Casciani** Lynus Barnes. Jr. data. Ilan S. Chabay lllyn S. Beary Organic Analytical Lloyd A. Currie The National Bureau of Standards uses several means ames M. Research William D. Dorko Blackman** Harry S. Hertz James W. Elkins to transfer to others the Center's analytical methods, Villiam A. Bowman. Robert E. Botto** Edgar S. Etz standards, and research results. These means include Ill Jeanice M. Brown Charles Fiori** Oavid E. Brown Stephen N. Chesler Robert A. Fletcher publications, seminars, consultation, cooperative pro- Lobert W. Burke Richard Alan Fried G. grams, and the direct distribution of materials standards . Stephen Carpenter Christensen Jimmie A. Hodgeson sabell Casar** Alex Cohen Ernest E. Hughes through the NBS Standard Reference Materials Program. Erie R. Deardorff Bruce Coxon Robert H. Kagann larry I. Lawrence M. George A. Klouda The diversity of materials standards produced in the Diamondstone Doane* Edwin C. Kuehner Center reflects the ubiquity of analytical measurements .ura P. Dunstan Richard A. Durst Ryna B. Marinenko •lichael S. Epstein Delmo P. Enagonio Raymond L. in this society. These standards include samples used in ohn D. Fassett Rafael Espinosa- McKenzie the metals industries (e.g., steel, copper), Lonald F. Fleming Leniz** W. R. Miller** aluminum, homas E. Gills Alexander J. Fatiadi Robert L. Myklebust chemical industries, and clinical and environmental labo- ohn W. Gramlich Mary L. Fultz Dale E. Newbury Lobert R. Franklin R. Richard Paur** ratories. To carry out the research underlying each of the Greenberg Guenther Peter A. Pella several hundred gas standards, the Center must maintain ose Guerrero** Diane Hancock William P. Schmidt higeki Hanamura Laurence R. Hilpert John R. Sieber a level of expertise not achieved in any other single ana- oe M. Harris** Barbara F. Howell David S. Simons lytical laboratory in world. ally H. Harrison Pamela L. Konash Gerald A. Sleater the Moreover, standards be- Emerson F. Heald William A. John A. Small come increasingly complex, and technical requirements Villiam R. Kelly MacCrehan Eric B. Steel loward M. Sam A. Margolis W. Thorn** for their development become more demanding. These Kingston, Jr. Willie E. May *Post Doctoral Staff factors create needs for new methods of measurement. Villiam F. Koch Reenie M. Parris ** Research Associates ames Koskoris Dennis J. Reeder and Guest Workers The Center works to achieve a balance between provision of currently needed standards and research in new ana- laman spectra from individual micrometer-size particles are •btained with the Raman microprobe, developed at NBS. The lytical methods. molecular identity of the constituent chemical species in the The technical activities of the Center focus on sample ample can be determined. Local environment and crystal- and compound types (analytes) —organic, inorganic, •nity of the microsample can also be inferred in many cases. r his rapid nondestructive technique requires no special sample gas, and particles. For each analyte, researchers in sev- reatment or special conditions such as vacuum. It compo- eral disciplines take part in an integrated measurements nents other microanalytical probe (electron- or ion-beam) standards program. echniques which furnish primarily elemental data. John Blaha The Center has developed more than 80 gas stand- rear) and Edgar Etz identify the composition of airborne 'usts and submicrometer atmospheric aerosols. The minerali- ards. Though initially stimulated by the need for reliable sation process of teeth and bone, the characterization of measurement of air pollutants, such standards are becom- environmental trace organic materials and the study of fluid ing increasingly important in occupational health and nd gaseous inclusions in minerals which are related to 'e ilogical processes are also studied. safety, technology applications (e.g., evaluation of en-

57 NATIONAL MEASUREMENT LABORATORY

gine performance), and atmospheric monitoring and techniques, with emphasis on improving the accuracy o research. Requirements for increasing accuracy, meas- multielement methods involving the use of lasers, neutro urements of trace amounts of constituents, and determi- radiation, electroanalytical techniques, and mass spec nation of a greater variety of gases and toxic vapors trometry. necessitate research and development in the area of Research in microanalysis is an important activity i| highly sensitive spectroscopic methods—those involving the Center. Work in this area has emphasized under

the use of lasers as well as mass spectroscopic techniques. standing the fundamentals of electron, photon, and io:<

Organic compound analysis is often called a new fron- interactions with matter so that more quantitative infor-

tier in analytical chemistry in recognition of two facts: mation may be derived from microprobes. Most existin' hundreds of thousands of organic compounds occur in electron microprobes now use one of a series of quanti nature or are produced commercially; yet methods of tative x-ray emission data reduction programs developer

measurement of complex mixtures are just now beginning in the Center. The most recent development, a compute

to emerge. Moreover, the basis for standardization of program called FRAME-«, has yielded promising im'

methods is in a very early stage of development. The provements in the ability to determine quantitativ importance of methods for the detection and quantitative elemental compositional information on micron size par

determination of organic molecules is evident in the tides. These basic studies have aided in research t potential applications: measurement of health-related develop improved techniques in particle analysis, incluc constituents, including therapeutic drugs, in body fluids: ing a major effort to standardize the analysis of asbestc

measurement of nutrients and contaminants in foods; in air and water.

measurement of carcinogenic compounds in water, in air, The field of analytical chemistry is undergoing a rapi on particles emitted with pollutant gases by factories and growth and an increase in scientific complexity. Th

in synthetic fuels. Though the Center has been involved "evolution" is stimulated by the growing application c

in organic analysis for a relatively short time, it has made analytical chemistry to areas of public concern. Alsc a number of significant contributions: development of advances m analytical chemistry are helping to promot

the first standard for trace organic analysis ( antiepilepsy scientific advancement in other fields. As analytical mea:, drugs in serum); development of quantitative methods urements become more sensitive, they are increasing of measurement of polynuclear aromatic hydrocarbons valuable in areas such as materials science, climate re in recycled oil, and development of definitive methods search, health research, and nutrition research. Thus, th for clinical chemistry. Research plans include the devel- Center for Analytical Chemistry faces a dual challenge

opment of sophisticated chromatographic separation to keep pace with the growing demand for standards t schemes and the linking of separation systems with novel support current analytical applications and to take pai spectroscopic, electrochemical, and mass spectrometric in the advance of analytical chemistry itself. quantitative measurement systems.

The largest research activity in the Center is in inorganic analysis. Studies in this area have led to hundreds of materials standards, ranging from trace metals in food

materials and blood to samples used at all stages in the production of metals—steels, cast iron, aluminum, and copper. In addition, high precision and high accuracy mass spectrometry on isotopic composition has led to the redetermination of atomic weights accepted as international standards. In the future, efforts will focus more on measurements of high technology materials and very complex matrices such as hazardous wastes. These en- deavors will require the development of very sensitive 3ENTER FOR ANALYTICAL CHEMISTRY

laser Enhanced Ionization n Flames

John C. Travis, Gregory C. Turk, 'ames R. DeVoe, and °eter K. Schenck 1 instrument Development Group and Chemical Stability and Corrosion Division (CMS)

energy population statistical weight

FJLrom its unception in the summer .f 1976 [1], until today, laser nhanced ionization (LEI) has held he promise of unprecedented

ame spectrometry . It is characterized n a recently awarded U.S. patent 92 No. 4-148,586) [2] as part of a eneral phenomenon called the opto-galvanic effect" (OGE). Jnderstanding the underly ing lechanisms of LEI has been a

Significant step in establishing a solid "1 9i oundation for the atomic spectro- netric method. optica processes

Figure 1. Energy level diagram showing the radiational and collisional processes used to model laser enhanced ionization. Laser enhanced ionization occurs

/hen a tunable dye laser is used to The flame rapidly evaporates the sol- levels. Recombination is insignificant electively populate a particular excited vent, while most trace metals present because the applied electric field lowers tate of an atomic species present in a in the solution are left as free atoms. the electron density and thus the ame. Since the laser-prepared excited Quantitation is achieved by comparing recombination rate. toms have a lower effective ionization spectrometric signals from unknown For example, consider the popula- otential than they did prior to excita- sample solutions to those from "cali- tion of level 2 (n L.) to be negligible in on, the rate of ionization of these brated" solutions. the absence of laser irradiation. Then toms by thermal collisions increases the LEI signal current is taken to be ignificantly. This ionization enhance- Calibration curves exist for a num- proportional to k2 in2 during the pulse, nent can be easily monitored by draw- ber of elements detected by LEI. They with level 2 being totally populated by ig a current through the flame and yield "limits of detection" (LOD) often the laser (see fig. 2). Assuming a y observing changes in the current less than 1 part per billion (ng/mL) square laser pulse which maintains a nat are synchronous with pulsed or for aqueous solutions, as shown in fraction R(< [g 2 /(gi + g2 )] of the hopped laser illumination. As shown table 1 [3]. These data show that the neutral atom density (n, + n L>) in level i figure 1, the laser-induced ac com- technique compares favorably with the 2 during the pulse, the LEI current onent of the flame current is separated others listed. Conspicuously absent should go as k Lj R (n, + n L.), where R 'om the dc background current by an from the table are any elements having is determined by the rate constants LC filter. It is then amplified and high (> 7.5 eV) ionization potentials. of figure 2 and the laser spectral ir- ignal averaged by gated integration. The role of the ionization potential, as radiance, Ev . Since the total number The analytical application of LEI well as other pertinent parameters, can

density, n T , is given by nT = n x + omes from the well established tech- now be explained on the basis of a

n . n , the final enhanced ionization L + ; ique of using a flame as an atom model using the simplified energy level/ rate becomes reservoir". It requires a liquid solu- rate constant scheme of figure 2. on of the sample to be "nebulized" Several processes with relatively slow (dn,/dT) f=a (n — n a signal L-i R T : ) nto the oxidant supply as a fine mist, rates are neglected, for simplicity, current. (1) arried through the fuel/oxidant mixing including the "recombination" rates number, and finally into the flame. from the ionization level to the neutral The effect of the ionization potential

59 NATIONAL MEASUREMENT LABORATORY

Table 1. Comparison of detection limits" (ng/ml). perimental sensitivities for 23 transitions within 17 elements using an air/ Laser acetelyne flame [4]. This expression, as Element LEI 1 ' FAA C FAE' 1 FAF e FAF f recently reported in Analytical Chem-

istry [4], can be used to predict sen- Act 1i 1 0 1 4 sitivities 1 S for observed transitions and LJCX 1 O

T. elements, Ri — 70000 J utilizing conventional spectroscopic cross sections. Ca 0.1 1 1.1 20 0.08 additional areas of study are in 1 Two Cr 2 2 2 5 j process: (1) the relative insensitivity ol Cu 100 1 0.1 0.5 1 Fe 2 4 5 8 30 high ionization potential atoms, and (2) Ga 0.07 50 10 10 0.9 the characterization of the mechanism In 0.006 30 0.4 100 0.2 of electron/ ion collection. The first problem is being approached throug K 1 3e 0.05 e sequential—or "stepwise"—excitation Li 0.001 1 0.02 0.5 to high energy levels using two tunable Mg 0.1 0.1 5 0.1 0.2 lasers, and satisfactory results have al- Mn 0.3 0.8 1 1 0.4 ready been published seconc Na 0.05 0.8 0.1 0.1 [5]. The is being botl Ni 8 5 20 3 2 area of study approached Pb 0.6 10 100 10 13 experimentally and theoretically. The Sn 2 20" 100« 50e experimental work involves measuring T) 0.09 20 20 8 4 a number of variables, including signa strengths, flame potentials, and flame a Detection limits for comparison taken from S. J. Weeks, H. Haraguchi, and J. D. Winefordner, background currents as a function o Anal. Chem. 50, 360 (1978), unless otherwise noted,

b Ref. 3, laser enhanced ionization in flames, other variables, such as electrode ge j c Flame atomic absorption, ometry, applied potential, flame com il Flame atomic emission. e Flame atomic flourescence, conventional light sources, position, and aspirated solution com f Laser induced flame atomic flourescence. position. The theoretical studies wili g Taken from J. D. Winefordner, J. J. Fitzgerald, and N. Omenetto, Appl. Spectrocs. 29, 369 (1975).

li Taken from V. A. Fassel and R. N. Knisely, Anal. Chem. 46, 1110A (1974). involve obtaining numerical solution:

of Poisson's equation in a flame sub ;

jected to an applied field, with realistii comes from the excited state rate at any time after the beginning of the electrode configurations and flame con constant pulse (for which n although the t 0), ditions.

LEI current is favored in k 2i for low k 2i ^«exp [— (Ei — E2 )/kT], (2) The outlook is promising for LE ionization potential elements, it is pe- where k is Boltzmann's constant, and in flames to become established as ; nalized by the factor (n T — ni) rep- T the flame temperature. Equations major method for quantitative spec resenting depletion of neutrals. Since (1) and (2) indicate an inverse ex- troscopic analysis. real laser pulses are not "square", but ponential dependence of signal current have a finite risetime, the population on ionization potential. However, References fraction R is building up in time as the semi-log plots have yielded slopes dif- [1] Green, R. B., Keller, R. A., Schenct neutral atom density nT — ni is dimin- fering from a realistic 1/kT value by P. K., Travis, J. C, and Luther, C ishing and thus, a complex solution is G, Opto-Galvanic detection of specie > 50 percent, and, therefore, an im- expected for the maximum value of in flames, J. Am. Chem. Soc. 9<{ provement of the model was required. dnydt. 8517-8518 (1976). Re-examining equation ( 1 ) reveals [21 Green, R. B., Keller, R. A., Luthe that a possible explanation to the prob- Such a solution has been obtained G. G., Schenck, P. K., and Travi lem might arise from the depletion of for a laser pulse shape approximating J. C, Apparatus for Galvanic Df the neutral atom pool (nT — n0 by that used in these experiments. The tection of Optical Absorptions, U>! ionization during the laser pulse. Thus, resulting expression agrees with ex- Patent No. 4,148,586, April 10, 197!

60 :ENTER FOR ANALYTICAL CHEMISTRY

Measurement of Trace Elements in Solar-Cell Silicon

Richard M. Lindstrom, Ronald F. Fleming, and Paul J. Paulsen, Inorganic Analytical Research Division

arge quantities of pure silicon ?] Turk, G. C, Travis, J. C, DeVoe, tion of atomic species in flames, L J. R., and O'Haver, T. C, Laser Anal. Chem. 51, 1516-1520 (1979). will be required in the 1980's as solar enhanced ionization spectrometry in Turk, G. C, Mallard, W. G., Schenck, photovotaic electricity becomes an j [5] analytical flames, Anal. Chem. 51, P. K., and Smyth, K. C., Improved important energy source. Accurate 1890-1896 (1979). sensitivity for laser enhanced ioniza- measurements of trace impurities help |t] Travis, J. C, Schenck, P. K., Turk, tion spectrometry in flames by step- to define the optimum balance between G. C, and Mallard, W. G., Effect of wise excitation, Anal. Chem. 51, 2408 purity and cost. selective laser excitation on the ioniza- (1979).

M The Low-Cost Solar Array project, energy managed by Jet Propulsion Laboratory probe mil (JPL) for the Department of Energy, flashlamp has as its goal the generation of elec- dye laser trical power from the sun at a cost comparable to that charged by electric util-

ities. Reaching this target requires a massive increase in the scale of opera- oscilloscope tions of the solar-cell industry. This in-

trigger crease is expected to cause a dramatic photodiode drop in the price of solar cells, their o structural supports, and the power-con- ditioning equipment required for a large distributed activity.

It now appears that the targets set trigger signa chart at the beginning of the project in 1975 out boxcar recorder be I averager can met by more than one combination of methods. The most mature

technology is that of flat-plate, single- crystal silicon arrays, which were .'-IGURE 2. Block diagram of the laser enhanced ionization apparatus. developed primarily for space applications. For such applications, the quantity of cells produced was small and the cost was relatively high. The cells were made by slicing disks, one by one, from large single-crystal ingots of semiconductor-grade silicon, the purest material produced in multi-ton quantities in commerce today. For generation of electrical power, much greater quantities of silicon will be required in the 1980's. Current estimates are that photovolatics will require tens of thousands to hundreds of thousands of tons of silicon to be produced per year by 1990. This quantity of silicon must be produced without an unacceptable degradation of cell quality, at a price no

61 NATIONAL MEASUREMENT LABORATORY

greater than one-fifth the current price activation analysis (NAA) and isotope the discrepancy is greater than a factor of semiconductor grade silicon. It is dilution spark-source mass spectrom- of two. The problem is most extreme the quality of this "solar grade" silicon etry (IDSSMS). In NAA, packaging for copper (see table 1 ) . Our measure that our work addresses through trace and handling methods have been estab- ments in some cases differ by three element characterization. lished and detection limits have been orders of magnitude from those re Workers at Westinghouse, Dow Cor- caluculated and verified by measure- ported by other analysts. Systematic ning, and Monsanto are studying the ment of blank silicon samples. Nearly errors in our NAA determinations are effects upon solar-cell performance all elements of interest can be deter- not likely since excellent agreement caused by adding specific impurities to mined quantitatively at concentrations was found in an analysis of the same silicon. These studies will define "solar where changes in electrical properties sample by both IDSSMS and NAA grade" silicon. Measuring cell effi- begin to be visible. Two notable excep- Measurements of concentrations of ciency as a function of concentration tions are titanium and vanadium, which other elements in samples also show requires that the concentration be we are still unable to measure at sub- large discrepancies when NBS meas known; however, for some impurity ppb concentrations in silicon. In urements were compared with others. elefnents, the concentration at which IDSSMS, the laboratory blank sample Experience in geochemical trace ele the cell degradation begins is less than and instrument sensitivity are roughly ment measurements has shown how

] 1 part in \0' (1 ppb). Such low con- equally important limits to the mini- discrepancies of this kind can be re- centrations cannot be measured by mum quantity of impurity that can be solved. The method used is an analysis most methods of trace analysis. In observed, which is generally in the of a common lot of materials in differ- those cases where sufficiently sensitive region of a few nanograms for the ele- ent laboratories by different techniques. analytical techniques are unavailable, ments of primary interest. If disagreement in results exists, the the concentration in a doped crystal The results of our measurements, process is repeated, with improvements must be inferred from that in the melt when compared with those made by in methods, until agreement is reached from which it is grown. This requires other investigators, show the difficulty To aid in such comparisons of anal- a knowledge of the liquid-solid distri- of the measurement problem. For yses, a silicon Research Material is in bution coefficient, which is difficult to forty percent of the samples compared, the process of being prepared. measure because of the great tendency of 15 1 crystallizing silicon to exclude cer- Table 1. Comparison of copper analyses (10 atoms/ cm ). tain impurities from its crystal lattice. Sample W-007-Cu W-017-Cu W-051-Cu/Ti W-056-Cu Related difficulties must be faced by chemical engineers who design large- g I Mass Spec 1.80 32 4 70 scale production processes for solar silicon. Since impurity concentrations NAA 1.65 25 NA $ { of interest are very low chemical analysis performed by most service labora- ' tories will have low accuracy. There- o fore, an engineer will have difficulty in 11.2 constructing an impurity mass balance NAA 0.050 \ | 0.002 ±0.002 45.3 ± 0.8

:0.01u - ul ±0.5 that would assure with high confidence m 0.065 }- 10.9,nn Z that some sensitive contaminant will IDSSMS — — 42.2 not concentrate in the final product. To this set of problems analytical chemists at NBS have applied two sensitive methods of trace analysis that have been shown to be reliable in the past by repeated agreement in the certification of Standard Reference Mate- rials. These two techniques are neutron

62 'Renter for analytical chemistry

Quantitative Analysis of Individual Trace-Level Organic Constituents in Alternate Fuels

Willie E. May and Stephen N. Chesler, Organic Analytical Research Division

I ncreasing energy demand in the Table 1. Interlaboratory comparison of shale oil analysis (mg/kg). Jnited States, coupled with increasing 3 Compound NBS 2a 3 4 5 6 7 8 9 •eliance on foreign sources of petro- eum, has caused . national concern that pyrene 107 155 360 150 168 212 141 620 185 las resulted in a program for develop- flouranthene 61 102 220 80 108 104 112 380 115 ing new sources of energy. The con- benzo(a) pyrene 22 117 3.3 — version of coal to gaseous or liquid benzo (e) pyrene 21 — 1.3 — uels and the utilization of oil shale and phenol 394 392 180 399 — 310 iar sands appears to be some of the o-cresol 338 350 150 381 2400 328 ( promising new energy sources. How- Numbers indicate another laboratory's determination. ever, a serious though largely unknown complication of developing these alternate fuels is their potentially delete- ious effect on man and the environ- Table 2. Shale analysis (mg/kg, 95 percent confidence level). ment. Of particular concern are the Acid/Base No lumerous compounds (for example, HPLC extraction with extraction with extraction jhose identified on EPA's priority pol- with quan- quantitation by quantitation by titation by lutant list) which result from process- Compound LC GC GC/MS GC GC/MS GC/MS ing alternate fuels, and which may

'. ventually be transported into the en- pyrene 108± 16 101 ± 4 102± 9 94±10 104±8 'ironment as gaseous or liquid effluents. fluoranthene 53± 6 55 ± 6 62 ± 6 75± 5 58±5 To evaluate the potential environmental benzo(a) pyrene 20± 3 21± 5 24 == 2 Iffects, the alternate fuels base source, benzo (a) pyrene 20± 6 22r= 5 ts industrial processing, and its result- phenol 383± 50 387 ±26 416± 28 334d= 63 int effluents must be analyzed. Results o-cresol 330± 34 334 ±86 350± 16 322d=45 »f these analyses cannot be assessed dequately without knowing the accuracy of the analytical techniques and aiata obtained. analyses did not exist because they mass spectrometric (GC/MS), and We have in j the past conducted require for certification a determination high performance liquid chromato- everal collaborative studies aimed at by a minimum of two totally independ- graphic (HPLC) methods for identify- ssessing the accuracy of data obtained ent methods. During the past year, ing f we and quantifying the compounds of om various trace organic analytical have developed the requisite methods interest in the various fractions ob- lethodologies. The results of a recent for certifying a Shale Oil SRM. tained from the isolation procedures. omparison of the determination of The methodology used in the certifica- Direct analysis by GC/MS was also ydrocarbons, and N- heterocylics by tion of shale oil was developed over an used when possible. veral laboratories under contract to 18 month period. Initially, we focused These data, given in table 2, show ;ae Department of Energy, are given our efforts on the evaluation of tech- that the overall agreement between the i Table 1. The variance among re- niques for extracting the compounds of various methods of extraction and ults reported by these laboratories interest from shale oil matrix. A classi- quantitation was good up to the 95 per- ocument the current state of measure- cal acid-base extraction scheme and a cent confidence level. This excellent ment in this area. One well-docu- high performance liquid chromatog- agreement between these results lented method of increasing the accu- raphic fractionation procedure, which compared to those obtained in the acy of analytical measurments involves was developed in this laboratory, served interlaboratory collaborative study pro- tie uses of quality assurance standards, as independent methods for isolating vided the basis for establishing state- r Standard Reference Materials the analytes from the sample. Next, we of-the-art methods for trace organic SRM's). However, until recently, investigated the use of gas chromato- analyses.

RM's for environmental trace organic graphic (GC), gas chromatographic- The specimen of shale oil being cer-

63 NATIONAL MEASUREMENT LABORATORY

Computer-Controlled Electrochemical System for Electrode Kinetics Research

Lawrence M. Doane and Richard A. Durst, Organic Analytical Research Division

tified is from a 150-ton retort for in-situ the end of the mercury drop lifetime, that polarographic problems simulated combustion operated by the M. odern electrochemical so several Laramie Energy Technology Center, research often requires measurements can be minimized, e.g., electrode charg-

Laramie, Wyoming. The shale is from of fast electrode kinetics and transient ing current and solution agitation. In the Mahogany Zone of the Colorado intermediates. Since commercial elec- addition, the potential is applied as a Green River formation. The stock of trochemical equipment generally do square wave while sampling the result- shale oil being certified was obtained not meet these rate requirements, a ing current at the end of each potential from Dr. Bruce R. Clark at Oak Ridge state-of-the-art, computer-controlled change and adding together the two National Laboratory, Oak Ridge, Ten- system capable of data acquisition at a sampled currents. Thus, this method nessee. It has undergone centrifugation rate of 100 kHz is in the final stages gives an increased current signal over to separate water (~ 40 percent) and of development. A specially designed that obtained from the linear ramp sludge from the oil prior to shipment to potentiostat, with dynamic and nearly experiment. This system is obviously NBS. complete compensation of the cell re- applicable for trace metal analysis, and This SRM, having the concentrations sistance, will be incorporated into this such an analysis will be used to evalu- of phenol, o-cresol, fluoranthene, py- system. Fourier transform experiments ate the CCES. rene, benzo(e)pyrene, and benzo(a)- can be implemented, since this system We have designed the CCES around pyrene certified, will be issued in early can sample two inputs simultaneously the relatively low-cost central process- 1980. Its use should greatly improve at a rate of 50 kHz. The system has ing units (CPU) that are now com- the accuracy of related environmental been designed for maximum data ac- mercially available. The CPU is a measurements. Accurate measure- quisition throughput and versatility, Zilog Z80A microprocessor, operated ments will assist scientists in correctly and simplified control. at a clock rate of 4 MHz, in which the relating health effects to levels of indi- fastest instruction cycle takes place in vidual pollutants, engineers in correctly 1 ju.s. Memory consists of 64 kilobytes assessing effectiveness of control tech- The recent emphasis on finding en- (K) of random-access memory (RAM) nologies, regulatory agencies in relating ergy-efficient processes has stimulated and 240 K dual-drive disk memory. A levels of pollution emissions with ambi- a resurgent interest in electrochemistry block diagram of CCES is shown in ent air and water quality, and the and electrochemically related fields. figure 1. The CCES acquires the elec-r Federal Government in correctly mak- Since maximizing the efficiency of electrochemical signals via two modes ing policy decisions on cost-benefit trochemical processes will involve the mode A uses fast commercial potentio- trade offs concerning environmental elucidation of electrode reaction mech- stats (an electronic instrument used to protection, energy conservation, and anisms it is necessary that fundamental control the potential of an electrochem economic health. investigations of these processes and ical cell) from which analog signals are This work represents efforts by the their associated elecrode kinetics be converted to digital format in the com staff of the NBS Organic Analytical undertaken. Commercially available Division who have expertise in gas electrochemical instruments are gen- chromatography, liquid chromatog- erally not versatile enough to be useful FLOPPY raphy, and mass spectrometry. in many of these types of investiga- disk tions. We are therefore constructing a computer-controlled electrochemical CPU system (CCES) that can generate complex waveforms, apply these waveforms

to an electrochemical cell, and then

acquire data at a high rate. A system POTENTIOSTAT similar to this is used in single-drop square wave polargraphy and will serve CELL as a model for evaluation of the CCES. In the single-drop experiment, the po- Figure 1. Block diagram of the compute- tential range of interest is scanned at controlled electrochemical system (CCES)

64 E ITER FOR ANALYTICAL CHEMISTRY

titer; mode B uses a laboratory- isters with the parameters of the elec- fication of eluting species; (c) signal Dnstructed potentiostat which digitizes trochemical signal to be outputted by enhancement by ensemble averaging ae electrochemical signals before en- the potentiostat. When this function techniques employing newly developed

Ting them into the computer. There is finished, the computer reverts to an pulse and relaxation methods, e.g., the

| a five-fold difference in signal acqui- input sequence that sends the digitized square wave experiment described tion rate between the two modes: 20 data to the computer upon signal from above. There also exists excellent op-

s for mode B versus 100 for mode the ADC's. One input sequence can portunities in data retrieval and pattern occur during the conversion of the recognition, two operations which have In mode A, the analog electrochemi- next analog signal; however, when two been underutilized in electrochemistry. al signals are digitized by two analog- signals are converted simultaneously, Such operations would be useful for s-digital converter boards (ADC) the second datum input to memory will the identification of specific compounds

esiding in the computer. The conver- add an additional 10 lis to the overall in complex mixtures (deconvolution)

ion can occur at a maximum rate of throughput, i.e., 20 /xs are required to and in the elucidation of electrode 0 kHz (i.e., one data point every 33 perform one complete two-data input. kinetics and mechanisms by compari- •s). however, execution of the pro- The registers can be reloaded during son of acquired data with theoretical ;ram to input and store the data an outputting sequence so that the out- curves generated by digital simulation

'equires an additional 30 to 60 tts, put can be dynamically altered in a of coupled chemical-eletrochemical epending on whether or not an analog manner dependent on the format of the J reactions. ignal is being sent to the potentiostat. program. "herefore, the overall throughput could Currently, we are developing soft-

: vhich two ADC's and a digital inter- normally occupy is now available for lace have been included to minimize data storage. However, a direct memory

he software required to operate the access (DMA) controller is being de-

ICES (fig. 2). In this scheme, the signed to directly input the data to :omputer loads a group of parallel reg- memory (bypassing the CPU). By using the DMA, the CCES speed will FROM depend only on the ADC conversion COMPUTER rate, freeing the CPU for other tasks. ii Following an evaluation of the CONTROL WORD CCES, its application will be broad- OUTPUT CONTROLLER ened to include: (a) electrode kinetics experiments via Fourier transform of 1 TIME LOW BYTE HIGH BYTE CONTROLLER AC voltammetry, a technique providing high-precision measurements of

VOLTAGE REGISTER faradaic admittance vs potential or frequency; (b) rapid scanning detectors TO for high performance liquid chroma- DAC tography, where the dynamic resistance compensation and rate of data acqui- F:gure 2. Block diagram of data flow through interface. sition will permit the real-time identi-

65 NATIONAL MEASUREMENT LABORATORY

High Resistance Junction Formation in House Wiring Circuits With Aluminum Conductors

Dale E. Newbury, Gas and Particulate Science Division

TJLhe.he mechanism of high resistance junction formation in aluminum-wire iron screw connections is shown to be the generation of iron-aluminum infermetallic compounds during arcing.

In the period 1968-1972, many new homes built in the United States were equipped with aluminum wiring and conventional brass-coated iron junc- ~Jk tions. In laboratory experiments, these aluminum-iron junctions have been observed to undergo a glow-arcing phenomenon accompanied by severe overheating which could lead to the initiation of a fire [1]. The origin of 0.1mm the overheating, which involves two normally satisfactory electrical conduc- Figure 1. SEM image of an arc-damage crater. tors, has not been previously explained. Electron probe microanalysis of the microstructures of circuit components which have undergone such failures has revealed the probable mechanism of the glow phenomenon [2]. Previous laboratory investigation of the glow phenomenon revealed that glowing began when the aluminum wire-iron screw junction became me- 1 m FeH chanically loose [1]. As a starting point for this study, conducted in cooperation with the NBS Center for Con- sumer Product Technology (CCPT) a loose connection between an aluminum wire and an iron screw was simulated in a circuit carrying 15 amperes of current. When the components were 1 separated by a small gap, an arc was

created that led to the glow phenome- j non. Scanning electron microscopy and Figure 2. X-ray spectrum of a region of a crater. x-ray microanalysis of the regions damaged by the arc revealed evidence of significant material transport across

the arc and of high temperatures at the

surface of the components (see figs.

1 and 2) . Iron and a minor amount of

66 1

:opper were transported to the surface

of the alurninurn wire, and aluminum eras transported to the iron screw ap- parently in the vapor phase. The zinc remponent of the brass was not de- ected and was probably lost by evap- oration. Temperatures in excess of 1500 C were suggested by the observation of solidification structures of learly pure iron at the periphery of •he crater on the screw. This combination of high tempera- xres and intimate mixing of the metals suggested the possible formation of ntermetallic compounds. A section Through the damage crater revealed the taresence of such a compound. FesAL £s determined by x-ray microanalysis. The existence of the intermetallic com-

i sound at the current-transmitting inter-

face is of special significance since its

-esistivity is generally a factor of 10 to 100 times greater than that of either of he pure metals. In order to confirm that this phenomenon occurs in actual junction boxes •sired with aluminum, a second series

:>: experiments was carried out in cooperation with CCPT. Typical household iunction boxes wired with aluminum vere subjected to high" current cycling

I ontil glow failure was induced. In the process large portions of the assembly -eached temperatures of 400 to 500 r C ll]- Samples for metallographic exam- nation were prepared from a section through the wire screw contact;

i typical section through a contact is phown in figure 3. A thick reaction ?cne spanning the entire region of con- iz: between the wire and screw can

r>e seen in the optical micrograph ( fig- Jre3). An electron microscope image 3f this region. Figure 4. reveals that

he zone is composed of two layers. X-ray microanalysis with the VBS Theoretical matrix correction procedure FRAME C indicates (1) that the layer '

NATIONAL MEASUREMENT LABORATORY

Discrimination of Natural From Anthropogenic Carbonaceous Pollutants Through Mini- Radiocarbon Measurements

Lloyd A . Currie and George A . Klouda, Gas and Particulate Science Division

adjacent to the aluminum corresponds References [1] Meese, W. J. and Beausoliel, R. W., to the intermetallic compound FeAl 3 adiocarbon dating rests on the Exploratory Study of Glowing Elec- and (2) that the outermost layer con- fact that living matter is in isotopic trical Connections, NBS Building Sci- sists of Fe2Al3 with 0.5 percent copper. equilibrium with cosmic ray-produced ence Series 103 (NTIS, Washington, Examination of samples taken at var- 14 C; death breaks the photosynthetic D.C. 1977). ious stages in the glow process suggests chain, leading to reduction of the [2] Newbury, D. E., Investigation of the 14 12 that initially intermetallic compounds Mechanism of the Formation of Re- isotopic ratio C/ C in accordance are either iron-rich or aluminum-rich sistive Junctions in Aluminum Wire, with the physical half-life of 5730 depending on the major constituent Report of Analysis 78-96, Center for years. We have taken advantage of B present. With continued reaction, inter- Analytical Chemistry, Gas and Par- this natural phenomenon to develop a Division, National metallic compounds containing roughly ticulate Science method for quantitatively discrim- Bureau of Standards (1978). the same quantities of each element are inating between biogenic ("living") [3] Hansen, M., Constitution of Binary formed [3]. All of the intermetallic and fossil ("dead") carbonaceous Alloys, (McGraw-Hill, New York, compounds containing aluminum and contaminants in the environment [1] 1958) pp. 90-95. iron have resistivities 20-200 times (see fig. 1). [4] Vol, A. E., Handbook of Binary greater than or iron aluminum [4]. Metallic Systems, Vol. 1 (U.S. Dept. The observation of the development of Commerce Washington, D.C, of thick intermetallic layers in the 1966) pp. 136-149. The importance of distinguishing current carrying region during the glow biogenic from fossil carbon arises from failure suggests the following chain of the effects of both natural and anthro- events: ( 1 ) an arc occurs in an initially pogenic carbonaceous species on loose connection, causing local heating health, stratospheric ozone, and cli- and material transport in the vapor mate. In order to plan responsible con- phase; (2) on both the wire and the trol strategies, it is therefore vital to screw, the metallic elements react in determine the relative contribution of the high temperature environment, man's activities to atmospheric hydro- forming intermetallic compounds; (3) carbons, halocarbons, carbon monox- the high resistivity of the intermetallic ide, and primary and secondary car-^ compounds leads to sharply increased bonaceous particles. Although natural! PR heating at the interface; and (4) emissions of hydrocarbons far exceed i- the heat produced leads to enhanced those originating from man s use ol

material transport, first as a vapor and fossil fuel [2], for example, there re- perhaps eventually as a liquid, which main very significant questions con-; then promotes rapid reaction to form cerning the sources of hydrocarbons even more intermetallic material, re- contributing to rural pollution episodes sulting in an unstable, runaway proc- as well as the influence of natural for-

ess. Knowledge of this basic mech- est emissions on urban air pollution [3], anism for glow failure of electrical Longer-lived tropospheric species, such junctions provides important informa- as methane, may lead to depletion ol

tion which could lead to the design of stratospheric ozone; and trace gases ;

safer circuit connections. and particles may influence the earth's i radiation balance directly through "greenhouse" and albedo effects or in-

directly through cloud nucleation [4],

It is worth noting that radiocarbon is unique biogenic/fossil dis- the only j criminator. Chemical composition and;

68 2ENTER FOR ANALYTICAL CHEMISTRY

13 12 1 .table isotope measurements ( C/ C) Dating. This requisite sensitivity was sional spectroscopy according to pulse are frequently helpful but the former achieved through the design of a high amplitude and risetime. Special fea-

's indirect and subject to interference purity, miniature gas proportional tures of the system, which includes an

md the latter exhibits large natural counter which is linked with a two- on-line computer, provide for rejection

• ariations. parameter low-level counting system, of spurious electronic pulses and long- Because of the small atmospheric described in the following section. term monitoring of stability. As a re- :oncentrations of many of the species sult of the attention given to materials'

of interest (table 1 ) , it has been neces- Measurement System radio-purity plus active (timing, spec- ;ary to advance the state of troscopy) passive re- the meas- Following collection and separation and background urement art by a factor of a thousand. duction, our residual background is according to size or chemical char- That is, measurements prac- comparable to the signal 0.1 must be acteristics, environmental samples (~ cpm). icable for samples containing This is equivalent a "noise" attenua- just a containing 5 to 15 mg carbon are con- to 'ew milligrams of carbon tion (with to —rather than verted quantitatively to CCX. for radio- reference the unshielded laving gram-size samples which are carbon measurement. Prior to count- counter) of more than a factor of two ypical of conventional Radiocarbon 14 hundred. ing of the C, the sample-CO L> must Figure is photograph of our 15 be rigorously purified—a few ppm 2 a fable 1. quartz Tropospheric concentrations of electronegative impurities interfere mL counter and the internal (mg-C/nr). shield of purity copper. with counting—and cryogenically trans- high This counter is used for samples containing Air ferred to a high-purity quartz counting up to 30 mg-carbon; a second counter quality tube. Measurement of the radiocarbon having an internal volume of 5 mL and Range standard takes place over a period of several operated days using a special low background up to 4 atmospheres C0 2 is used for COo 180. system incorporating massive shielding, samples having <10 mg- carbon. (The residual background rate :h4 0.76 0.03 — electronic anticoincidence cancellation for the :o* 0.06 20. 17. of cosmic ray muons, and two-dimen- 5 mL counter is remarkable: a -ic <10"3 2.00 0.14 recent measurement in the under- i a 3 10"3 ground counting laboratory at the Uni- xl, c,H (; i0" — 1 Articles <10"3 0.10 0.052 The required sample size is given by the versity of Bern yielded a mean interval fact that 10 mg of contemporary carbon "Dissolved of 25 minutes between counts.) A dis- 2 10(mg/l) — yields a net counting rate of about 0.15 cpm 'rain) which results in a 10 percent uncertainty play of the two dimensional spectrum, Organic (Poisson relative standard deviation) after for an air particulate sample collected about one day of counting. The correspond- 3 in Salt Lake City, is given in figure 3. larbon ing air sample size would be 100. m for a substance whose concentration is 0.1 mg- Valid pulses lie below the dashed line, 1 5 N-S asymmetry. C/nr [1]. and eight (rejected) spurious events RADIOCARBON are seen above. Further details concerning the sample preparation and purification system, and a brief descrip- 14 N Production: Cosmic Rays (p) —*. n 14/ CO, (Biosphere) tion of the counting system may be (atmosphere) found in References 5 and 6, respectively

= e~t/r Dating: A A [10 g, 1%] Experimental Evidence of Man's Activities Pollutant a Source: = f • (l-f) • research, A A 0 + 0 [10 mg, 15%] Our whose objective is to use radiocarbon to discern man's per- Figure 1. Radiocarbon production by cosmic rays, and its application to dating and turbations of environmental carbona- pollutant source identification, (f equals the fraction of contemporary or living carbon in m environmental sample; represents the equilibrium ceous matter, has been directed toward A0 "C/^'C isotope ratio in living natter. Typical sample sizes and Poisson Standard deviations are given in brackets.) three classes of samples: atmospheric

69 NATIONAL MEASUREMENT LABORATORY

gases, atmospheric particles, and organic matter in "recording" media, such as sediment and ice cores. Study

of the first two categories yields information on the significance of current

activities, while the third provides the historical record and a link to the global carbon cycle. Work with atmos-

pheric gases is still at an early stage; we have just begun sampling atmospheric methane, and suitable apparatus

for the collection of mg-quantities ofjj non-methane hydrocarbons and carbon

monoxide is under development. In cooperation with University groups, however, we have carried out a number of measurements of radiocarbon concentrations in atmospheric particles and in sediment. The results are summarized in Table 2.

The first point to be noted in Table 2 Figure 2. Photograph of the small quartz counter (15 mL volume) and associated shield. is that, with the exception of one sam- The shield, along with the inner wall of the counter, is constructed of OFHC copper. ple, all samples consisted of only about

10 mg of carbon. This is the amount daily Table 2. Radiocarbon in aerosols and sediment. which was available with normal, operation of a HiVol air sampler or Mass- Percent with the chemical processing of about Carbon contemporary Supporting 600 grams of surface sediment. The 14 12 only two previous studies of radiocar- (mg) [ C/ C] (±1 S.D.) data bon in urban aerosols involved a thou- Ambient particles sand times as much carbon and re Los Angeles 7.8 23 ± 15 Organic quired a week of sampling [7]. (Sim Salt Lake City 5.3 28 ± 13 (insoluble) ilarly, conventional radiocarbon meas Utah Desert 10.2 88 ± 16 composition urements of PAH in sediment would Impact particles have required extraction of a metric h «3fxm) ton of material.) Aside from the in-i - - Portland, OR. 5.1 10.7 62. 107. Inorganic creased ease of sampling and time res-:

composition olution, the ability to measure very 1 c Sediment small samples is essential in terms of

Puget Sound, WA. ( FA 60. 86 ± 21 Isotopic selectivity. That is, miniradiocarborr (13 /12 (PAH 14. 19 ± 21 C C ) counting makes it possible to derive composition reliable information on individual chemical or size fractions especially " — Ref. 5. » Ref. 8. fine particles which may be free from ' Ref. 9 (FA Fatty acid fraction, PAH = Polycyclic aromatic hydrocarbon fraction). local contaminants, such as resuspend- ed waste or spores. The ambient particles (table 2). studied in collaboration with the Uni

70 CENTER FOR ANALYTICAL CHEMISTRY

re 60 c . 0)

(Fossil)

u>- BCD (13 I I I Im O a a Q E © < 40

2 I I 0 0.25 0.50 0.75 1.00 1 t II ^112 1 2 I I 2 I 12 11 1 II o 12 3 I 2 2 2 I I I S 2 II - 3 3 2 11 Expected fraction of

1 I I I I I I X I 5 2 13 2 1 contemporary carbon _ *3S 2I3IS22 3 * (pure source) X . ? ?

box [3] indicates the median and l

Supporting studies on the composition between 27 and 51 percent of the fine land study was contemporary— i.e., of the insoluble carbonaceous compo- particulate mass, and up to 35 percent biogenic, whereas it was fossil in the nent (soot) also pointed to an anthro- of the total suspended particulate ma- Los Angeles and Salt Lake City study.) pogenic (fossil) source for the urban terial collected in the Portland and One of the most interesting areas of samples. Pyrolysis-gas chromatogra- Eugene, Oregon, airsheds. The radio- investigation is the record of man's past phy-mass spectrometry revealed sev- carbon analyses of filters selected for impacts on the environment. This was eral cyclic and substituted aromatic high impact from residential wood the stimulus for' the last study [9] in- compounds which have also been ob- combustion suggested that this source cluded in table 2, the measurement of served in vehicular exhaust. is currently significant and could be- the chemical and isotopic record in Collaborative work with the Oregon come one of the most significant sea- organic matter in a sediment core from Graduate Center was designed to assess sonal sources of respirable particulates Puget Sound, taken midway between the impact of vegetative burning (by in the future, unless steps are taken to urban North Seattle and the rural man) on urban (Portland) air particu- improve combustion efficiency and/or Olympic Peninsula. The initial results, late pollution [8]. The results of that emission control. Chemical Mass Bal- which apply to the top 1 0 cm sediment

71 NATIONAL MEASUREMENT LABORATORY

layer or material deposited during the Within the past few months, in co- machine availability and cost make it last two to three decades, show that operation with the University of Roch- currently somewhat less attractive than

different organic species are far from ester, we have begun also to explore a 11c if 5 to 10 mg of environmental isotopic equilibrium. The high radio- new measurement technique which carbon samples are available. How- carbon content (86±21 percent con- promises to extend our sensitivity by ever, its incredible sensitivity—literally 2 3 temporary) of the fatty acid fraction, an additional factor of 10 to 10 . a million-fold better than conventional indicating a biogenic origin, is con- Through the use of Accelerator Mass radiocarbon dating—promises to make sistent with the premise that this frac- Spectrometry (AMS), in which indi- practicable hitherto impossible experi- tion originated from recently-living vidual radioactive atoms (rather than ments. We can anticipate the use of planktonic matter. The Polycyclic their disintegrations) are counted in radiocarbon to determine the origin of Aromatic Hydrocarbon (PAH) frac- a nuclear accelerator, we have suc- the rarer atmospheric species (e.g., tion on the other hand, was found to ceeded in measuring the radiocarbon halocarbons), and we should be able

contain little or no radiocarbon, indi- concentration in just 40 of a con- to greatly increase our knowledge of

cating that it was primarily of fossil temporary radiocarbon sample [6, 11]. environmental sources and transforma- origin, presumably associated with The relationship between decay count- tions through the examination of indi-

man's use of fossil fuel. Stable isotope ing (11c) and atom counting (AMS) vidual organic compounds found in 13 12 measurements ( C/ C) of the PAH as well as the main features of our ex- ocean sediment, polar ice cores, and did not clearly discriminate between ploratory experiments are sketched in atmospheric particles.

coal and petroleum, but they did rule figure 5. Considerable work remains out waste oil as a primary source. to done, particularly in the areas be References These findings support the contention of ion source sample preparation and [1] Currie, L. A., Noakes, J., and [10] that PAH in the surface sedi- fast and reproducible switching, beam Breiter, D., Measurement of Small ment comes mostly fossil from fuel before AMS will be established as a Radiocarbon Samples: Power of combustion rather than from natural reliable and accurate technique; and Alternative Methods for Tracing forest fires. (PAH, which contains

known carcinogens and mutagens, is Accelerator Mass Spectrometry a common product of incomplete com-

bustion, and it is generally A = -dN/ XN = associated r N/ T with the particulate emissions.) Fur- ther studies, at increased in the depths : tic n = A'e'A t AMS : n' = A»f'«T sediments, should yield the historical record for both fossil and vegetative Exploratory Experiment combustion. (Rochester, July 1979) Future Directions OBJECTIVES Our immediate plans call for ex- • Environmental sample preparation/dilution panding the miniradiocarbon counting [particulate SRM] system to accommodate up to twenty or • Electromagnetic isotope enrichment/ implantation [C0 more simultaneous counting channels, 2 • Minimum sample size such that sample throughput need not [new radiocarbon dating SRM] be limited by the relatively long counting times associated with individual PRELIMINARY RESULT (dating SRM) measurements. Information on the global methane cycle and assessment 40 ,ag - C ^ 14C/12C = (1.41 ± 0.04) X 10-12 of the fossil contribution to urban and Figure 5. Accelerator Mass Spectrometry (AMS) Exploratory Experiment at the Microgram rural non-methane hydrocarbons will Level. (lie represents low-level counting; n, number of counts; A, disintegration rate; ef follow from our current work on at- detection efficiency; At, lie counting time; t = the mean life of radiocarbon; N, mospheric gas sampling. the number of atoms of radiocarbon.

72 CENTER FOR ANALYTICAL CHEMISTRY

Atmospheric Hydrocarbons, 9th Klouda, G. A., Application of car-

Internatl. Radiocarbon Conf., R. bon- 1 measurements to impact as- Berger and H. Suess, eds., Univ. of sessment of contemporary carbon Calif., Los Angeles and San Diego, sources on urban air quality, (to be June 1976; Currie, L. A. and Murphy, published in Environmental Science R. B., Origin and Residence Times & Technology, 1979). of Atmospheric Pollutants: Applica- [9] Swanson, L, Fairhall, A., and Currie, 14 tion of C , in Methods and Stand- L. A., Carbon Isotope Analysis of ards for Environmental Measure- Sedimentary Polycyclic Aromatic ment, W. H. Kirchhoff, ed., NBS Hydrocarbons (submitted to Nature, Special Pub. 464, 439-447 (U.S. 1979). Government Printing Office, Wash- [10] Hites, R. A., Laflamme, R. E. and

ington, D.C. 1977). Farrington, J. W., Sedimentary poly-

[2] Covert, D. A., Charlson, R. J., cyclic aromatic hydrocarbons: the Rasmussen, R., and Harrison, H, historical record, Science 198, 828- Atmospheric chemistry and air qual- 381 (1977).

ity, Review of Geophysics and Space [11] Currie, L. A., Klouda, G. A., Elmore, Physics, 13, 765-771 (1975). D., and Gove, H, Accelerator mass

[3] Maugh, T. H, II., Air pollution: spectrometry and electromagnetic where do hydrocarbons come from? isotope separation for the determi- Science, 189, 277-278 (1975). nation of natural radiocarbon at the

[4] Geophysics Study Committee, microgram level (to be submitted to Energy and Climate, NRC Geo- Science, 1979). physics Research Board, National Academy of Sciences. Washington, D.C. (1977).

[5] Currie, L. A., Kunen, S. M.,

Voorhees, K. J., Murphy, R. B., and Koch, W. F., Analysis of Carbona- ceous Particulates and Character- ization of Their Sources by Low- Level Radiocarbon Counting and Pyrolysis/Gas Chromatography/ Mass Spectrometry, in Proc. of the Conference on Carbonaceous Par- ticles in the Atmosphere, T. Nova- kov, ed., Lawrence Berkeley Labora- tory, Berkeley, California (1978).

[6] Currie, L. A., Environmental Radio- carbon Measurements, in Proceed- ings of First Conference on Radio- carbon Dating with Accelerators, H. Gove, ed., pp. 372-390 (University of Rochester, New York, 1978).

[7] Clayton, G. D., Arnold, J. R., and Patty, F. A., Determination of sources of particulate atmospheric carbon, Science 122, 751-753

(1955); Lodge, J. P., Jr., Bien, G. S., and Suess, H. E., Carbon-14 content of urban airborne particulate matter, Intl. J. Air Poll. 2 309-312 (1960).

[8] Cooper, J. A., Currie, L. A., and

Center for Materials Science

.hehe National Bureau of Standards has had a con- John B. Wachtman, Jo-Anne A. Jackson David T. Read TA Jr. Kenneth L. Jewett D. Ellis Roberts tinuing involvement in materials research and develop- Darrell H. Reneker Rolf B. Johannesen Raymond E. Jerome Kruger Schramm ment since its founding in 1901. Today, more than ever, Chia-Soon Ku R. Shives Thomas NBS is committed to a strong program in the materials John G. Abbott David S. Lashmore John H. Smith Jasper L. Mullen J. Tighe sciences, and has focused most of such research in the Harvey J. Anderson Nancv Fielding Ogburn Ralph L. Tobler David B. Blanchard Center for Materials Science. John W. Cahn Gregory J. Olson* Thomas A. Whipple Donald G. Fletcher Edwin J. Parks Sheldon M. This Center's program is geared towards providing Fred Goraick Patrick T-S. Pei Wiederhorn technical support to private organizations and govern- Melvin Linzer Ernest R. Plante William A. Willard John D. McKinley, Edison N. Pugh Polymer Science and ment agencies to enhance more effective use of materials. Jn Joseph J. Ritter Standards I Benjamin T. Further, we believe that a strong base of materials sci- IStephen J. Norton Ronald K. Eby 'Stephen L Parks Sanderson Joseph M. Antonucci ence at NBS is an essential foundation for U.S. industry Samuel J. Schneider, Peter K. Schenck Harold Argentar** Jr. Joseph H. in developing new products, increasing productivity and John D. Barnes Bruce W. Steiner Simmons** Jacob Smit Gary M. assuring future economic competitiveness. Accordingly, | Robb M. Thomson Beaudreau** Claudia F. Vosbeck Stephan J. Weeks the Center conducts major research activities in poly- 1 Barry Bernstein Melvin R. Wallace John P. Young Leonard H. Bolz mers, metals, ceramics, and glasses focus on areas where M. Katheryn Fracture and Wharton Rayphel L. Bowen** Deformation Gerhard Brauer measurement services and data are especially needed. Gerry N. Wixson M. Richard P. Reed Martin G. materials research program addresses major Chemical Stability Our John M. Arvidson Broadhurst and Corrosion Mark W. Austin Daniel W. Brown phases of the materials cycle, from processing through ! Thomas D. Coyle Walter E. Brown** Betty A. Beck design, manufacture, and assembly, to use and finally to Anthony J. Bur L'go Bertocci Elliot L. Brown James Cassel • Tames R. Bethin Bruce W. Christ M. recycle or discard. The Center's research program covers William R. Blair Leon Chuck Shu-Sing Chang three major aspects of materials science: transformation David W. Bonnell Roland deWit Tzeng Chen** Lawrence C. Chow** Frederick E. Ronald C. Dobbyn science, structure/ property science and durability science. Edward S. Clark Brinckman, Jr. James G. Early. Jr. James P. Colson is denrietta J. Brown John W. Elmer Transformation science the basic science which under- John M. Crissman Zhwan-Kang Richard J. Fields lies materials processing research, for it deals with the Chiang Stephen W. Freiman George T. Davis Ronald E. Dehl ' John J. Comeford Edwin R. Fuller, Jr. changes in state or microstructure of materials. It seeks Richard G. Crumley Arthur H. Heuer Aime S. Dereggi Brian Dickens to predict the microstructure which results from variables Arthur L. Cummings George E. Hicho James G. Dillon** Klan L. Dragoo Bernard J. Hockey of composition, temperature gradients, pressure, and Edward Escalante Charles G. Interrante Seymour Edelman Fanconi Klan D. Franklin Maurice B. Kasen Bruno M. mechanical stresses. Those same variables are of im- Barry L. Farmer Richard S. Gates Ralph F. Krause, Jr. portance in industrial processing. Lewis J. Fetters Structure/ property Villiam F. Gerhold Ronald D. Kriz* .udwig H. Grabner Hassell M. Ledbetter Roland E. Florin science addresses how to measure structures and how Joseph ohn W. Hastie Samuel R. Low, III H. Flynn L. chemical and physical properties depend upon structure. rVilliam S. Horton Clyde L. McDaniel Edmond Graminski tephen M. Hsu Harry L McHenry Durability science deals with the changes which cause iaioshi Ito** Terry A. Michalske* Warren H. Grant A arren P. Iverson Robert S. Polvani (Continued) deterioration in service, such as corrosion, wear and fracture. Center activities are formulated partly on the basis of materials' trends and problems, such as those identified

by the National Academy of Science. We find it espe- Siew of the counter arm of the five-counter diffractometer. cially effective, however, to plan our efforts through di- Antonio Santoro and Kay Hardman are investigating a r jterial at low temperature; the sample is enclosed in the rect contact with the users of our services, frequently by -ryostat shown standing in the central part of the picture. The working in cooperation with our many "clients" in indus- leutron beams diffracted by the powdered material are inter- try and government. Thus, our assessment of the needs ested and measured by the counters in the horizontal plane V the instrument. There are five such counters, placed 20° for materials' standards is aided by having members of ipart, so that different regions of the pattern can be measured our staff participate in more than 120 engineering stand- hiultaneously. Visible in the picture are also the catcher of ards committees concerned with measurements needed 'he primary beam (on the left), and shielding walls, necessary o "educe background. for effective use of materials. We are helped both in pro-

75 NATIONAL MEASUREMENT LABORATORY

Robert E. Green Lawrence A. William S. Brower, gram design and execution by 22 industry-supported Thomas M. Wood** Jr. staff. addition, Gregory** Wen-Li Wu Simon J. Carmel Research Associates who work with our In Stephen S. Hass** Hyuk Yu Mario J. Cellarosi we have attracted 54 guest workers from industry and Charles C-C. Han Louis J. Zapas Given W. Cleek** Gerald R. Harris** Martin I, Cohen universities. We also work closely with other government Metallurgy William P. Harris Lawrence P. Cook Robert Mehrabian agencies. In fact, the Center currently provides technical Harley L. Heaton** David J. Cronin Ture Heline** David B. Ballard Marilyn J. Dodge assistance to approximately 60 different organizations Elliot A. Kearsley Cletus J. Bechtoldt Louis P. James M. Kenney Lawrence H. Bennett Dominquez** within the Federal Government. About 40 percent of Francis S. Eloise H. Evans** Freddy A. Khoury our materials program is carried out to provide direcl David E. Kranbuehl Biancaniello Edward N. Samuel J. Kryder Peter J. Blau Farabaugh support to other government agencies. In addition, we William J. Albert Feldman William V. sponsor or co-sponsor national or international confer Loebenstein Boettinger Thomas A. Hahn Harry C. Burnett, Jr. David J. Lohse* Wolfgang K. Haller ences, symposia, and workshops. Robert E. Lowry Daniel B. Vicky L. Himes The technical activities of the Center are carried oui Curtis P. Mabie** Butrymowicz William R. Hosier Roger B. Clough Cyrus G. Camden R. Hubbard by six divisions: Chemical Stability and Corrosion; Frac Malmberg** Gabrielle G. Cohen Arnold H. Kahn Gordon M. Martin Sam R. Coriell Michael J. Lynch** ture and Deformation; Polymers Science and Standards John R. Cuthill Floyd A. Mauer Robert M. Marvin** Metallurgy; Ceramics, Glass and Solid State Science; and Arthur A. Maryott** Billy J. Evans Howard F. Mathai Mathew** Steven E. Fick McMurdie** Reactor Radiation. These divisions vary considerably ir John R. Maurey Anna C. Fraker Alan D. Mighell their balance between theory and experiment, betwee Jacob MazLir Stephen C. Hardy Marlene E. Morris** Frank L. McCrackin M. Ashraf Imam** Ronald G. Munro direct standards work and research, and in their orienta Donald Mclntyre Lewis K. Ives Taki Negas Gregory B. Daniel J. Kahan Craig D. Olson tion toward industrial needs and the needs of othe Masao Kuriyama Helen M. Ondik McKenna components of the scientific and technological co John E. McKinney Kenneth C. Ludema Boris Paretzkin** Daniel L. Menis** John R. Manning Harry S. Parker munity. Dennis B. Minor Thaddeus B. Alvin Perloff For instance, to meet the needs of the metals process Dwarika N. Misra** Massalski Gasper J. Frederick I. Mopsik George J. Mattamal Piermarini ing and manufacturing industries for data required to George C. Archie J. McAlister Carl R. Robbins alloys, Paffenbarger** Simon C. Moss Robert S. Roth develop durable, high performance the Cente Robert W. Penn Joanne L. Murray David M. Sanders carries out theoretical, experimental, and data-analysi Anton Peterlin Lucien A. Nedzi** Richard D. Spal Joseph C. Phillips Robert L. Parker Judith K. Stalick activities on alloy phase stability. The needs for accurat Michael E. Read Jon L. Waring Walter J. Pummer phase diagram information on new alloy systems hav Darrell H. Reneker Robert C. Reno Roy M. Waxier Steven C. Roth Steven T. Ridder John T. Wenzel been addressed frequently in recent years, and a join Arthur Ruff, Jr. S. Robert J. Rubin W. Grady White Society or Robert J. Schaefer effort with the American of Metals (ASM) Nelson W. Rupp** Reactor Radiation Isaac C. Sanchez Robert F. Sekerka Robert S. Carter phase diagrams is now underway. In this joint progran Charles P. Saylor** Robert D. Shull George A. Senich John A. Simmons Norman F. Berk the Center provides overall guidance to the program t Gregory B. Russell C. Casella Jack C. Smith assure reliability of the data evaluations, to maintai Leslie E. Smith Stephenson Yu-Tarng Cheng Jeffrey W. Stansbury Pei Sung** Donald A. Garrett coordination among participating phase diagram center Shozo Takagi** Lydon J. Charles J. Glinka (most of which are in the private sector), and to develo John A. Tesk Swartzendruber V K"' w-r-fiman* Elizabeth E. Toth Ann C. Van Orden Jeffrey W. Lynn techniques for the graphic presentation of phase diagrar Ming S. Tung** Richard E. Watson Bernard Mozer will direct interface with the David L. Morgan L. Williams John H. Nicklas data. ASM serve as a meta VanderHart Edward Prince Ceramics, Glass, industry so that needs and priorities can be properly dc Peter H. Verdier Tawfik M. Raby and Solid State Gerald L. Vogel** James J. Rhyne termined. An international council involving 70 men Science Herman L. Wagner J. Michael Rowe Hans P. R. bers from nine countries will advise us on this effort. Tw Francis W. John J. Rush Wang Frederikse Richard M. Antonio Santoro significant features of this joint effort are the publicatio Waterstrat** Michelle C. Austin James F. Torrence of a new bulletin for rapid information transfer and th James J. Weeks, Jr. Michael I. Bell Alexander Wlodawer Eric P. Whitenton** Rebecca A. Bird availability to industry of evaluated data directly froi Richard P. Douglas H. Whitlock** Blackburn Post Doctoral Staff ASM. * Research Associates William K. Wilson** Stanley Block and Guest Workers One unique feature of our structure/property scienc

work is the 10-megawatt research reactor located at oi

76 ;enter for materials science

•aithersburg site. The low energy thermal neutrons double bonds, methyl groups, and carbonyl groups result nitted by the reactor are used in the structural analysis from rupture of carbon-carbon backbone bonds and sub- f complex and sophisticated materials. Researchers in- sequent free radical reactions. These concentration olved in such analyses come from a variety of govern- changes can be detected in difference spectra obtained ment agencies, universities, and industrial organizations, through subtraction of the digital FTIR data. The de-

forking with our staff, these investigators have studied tectability of small concentration variations is enhanced tany materials and material transformations—among by the superior sensitivity of the FTIR method over con- lem cements, magnetic materials, and battery materials, ventional dispersive infrared spectroscopy. In addition, 'lost recently, they have carried out fundamental studies the FTIR method has potential application in the study

f the chemistry of ceramics by neutron diffraction profile of polymer matrix composites, a group of new materials nalysis, which promises to be a powerful new tool for of interest to Center scientists. These materials are find- "rystal structure analysis. ing increasing use as structural parts in the transportation We are constructing a special facility at the reactor industry because they can be fabricated into lightweight abr small angle neutron scattering, to be used, for ex- structural components. ample, in the study of polymer structures. Next year, The anticipated durability of materials and products

re will develop an additional facility at Brookhaven in service is a central consideration in their selection and aboratory for the application of synchrotron radiation use. Thus, a major aspect of the Center's materials pro-

p materials. This application, we feel, is to be the next gram concerns corrosion abatement as well as other as- eneration of x-ray techniques for structural studies of pects of materials durability including wear, which is the rystalline materials. primary cause of materials degradation in operating ma-

Similarly, we are broadening our research on the chinery; fracture, which is the cause of catastrophic :ructure of noncrystalline materials through new studies failure of structural components; and the migration of

n glassy materials which are formed by deposition from plastic stabilizers, which is one of the principal causes of jie vapor phase. Under construction is a film deposition failure in plastics. pparatus which allows the simultaneous codeposition One example of our research in measurement and cor-

rom separately controlled molecular beam sources. Dep- rosion science is localized corrosion wherein a metal is osition rates will be monitored during observation of opti- attacked at specific sites instead of being dissolved gen- cal properties of the forming films. In addition, the erally, leading to pitting, stress corrosion, and crevice cor-

eposition facility will be coupled to a diagnostic facility, rosion. When it happens, failure can result even when

llowing in vacuo transfer and analysis by several elec- most of the metal is relatively unattacked. In fact, local- ron and particle emissive techniques. When operative, ized corrosion accounts for the majority of corrosion

.te facility will be used for the production of reflective failures. This form of corrosion is the most difficult to

nd antireflective optical coatings and flat configuration detect, since the total amount of metal dissolved prior to aveguides, and in such fundamental studies as molecu- failure may be small. We consider localized corrosion to lar cluster formation in glassy films, the determination of be a major technical challenge and are working on sev- accommodation coefficients in multi-component amor- eral approaches to help understand and control it. In par- 'hous structures, and in the simulation of amorphous ticular, Center scientists are developing measurement >xide films found on corrosion-resistant metals. methods to determine the susceptibility of a variety of

Additional work on structure/ property relationships metals to localized corrosion. This information is essen- avolves investigations to define the molecular mecha- tial for developing standard methods of assessing the nisms in polymers associated with deformation and frac- suitability of a metal for service in a given environment ure. For polyethylene, these mechanisms are elucidated and for the developing of new, corrosion-resistant alloys. hrough comparisons between Fourier transform infrared Another and a very important aspect of materials

FTIR) spectra of undeformed and fractured specimens. durability is fracture control. Improved fracture control >mall variations in the concentrations of carbon-carbon methodologies depend upon predictive models. In fact,

77 predictive models coupled with nondestructive evaluation techniques are becoming increasingly important in developing materials performance codes and standards, including those for metals. For example, in 1976, Alyeska Pipeline Service Company requested waivers from the Department of Transportation (DOT) for a number of buried girth welds. The welds contained radiographically detected defects which appeared to be not in conformance with DOT regulation. The waiver was based on a fracture mechanics analysis. In anticipation of the waiver, DOT requested assistance from NBS in evaluating the fracture mechanics analysis and the nondestructive evaluation methods used to detect and determine the dimensions of specific girth weld defects. In large part because of the analysis provided by Center scientists, DOT did eventually conclude that under some circumstances, exemptions from existing

standards could be granted if warranted on the basis of fracture mechanics test results. Thus, more petitions for

exemptions can be anticipated. The Center is working with DOT now to establish a technical framework for ruling on these waiver requests—a framework which would minimize the need for subjective judgments and thus obviate a tendency toward overly conservative design. The breadth, scope, and accomplishments of the Cen- ters technical program are comprehensively reflected in the more than 420 articles by Center scientists published in 1979 in numerous professional journals, books and conference proceedings. The following brief reports high- light some current work of particular interest.

78 CENTER FOR MATERIALS SCIENCE

Erosion of Brittle Materials by Solid Particle Impact

Sheldon M. Wiederhorn, Bernard J. Hockey, A. William Ruff and Lewis K. Ives, Fracture and Deformation Division

products; and the erosive thinning of their transparency to electromagnetic n an effort to provide a better helicopter rotor tips by dust. radiation, but must also be able to

; understanding of the micro-mechanics Because of their outstanding erosion resist erosion to maintain their integ- of erosion and other impact-related behavior, ceramic materials are often rity. Similarly, in turbine nozzles and processes, the detailed nature of used in applications where erosion re- vanes, ceramics must withstand both damage produced in ceramics and j sistance is required. Thus, castable re- high temperatures and erosion from metals by impingement with i solid fractories have largely replaced metals combustion products and dust. particles has been investigated. in certain high wear areas of oil refin- The few examples given above Results obtained by scanning and eries and are being considered for use underline the need for improved ero- transmission electron microscopy for similar applications in coal gasifica- sion resistant ceramics, the develop- clearly specify the | elastic-plastic tion plants. In another application, ment of which could, in principle, be nature of the impact. This has led to ceramics have been usefully employed assisted through a better understand- the development of models j which as nozzle vanes in experimental radial- ing of the erosion process. Studies

I incorporate the concept of elastic- flow gas turbines, where they have ex- have been conducted primarily through plastic fracture in describing the hibited a pronounced superiority over the use of the scanning and transmis- erosive wear of brittle materials. the standard metallic vanes in resisting sion electron microscopes. Erosion rate erosion. Ceramics are also used in a measurements have also been made in

] number of applications where other a variety of chemical environments, Erosion is a major cause of material properties are important, but where and at temperatures ranging from room degradation in both civilian and mili- erosion is a factor in limiting reliability. temperature to 1000 °C. The combina- tary applications. It is caused by the In windscreens and radomes for ex- tion of techniques employed on this high velocity impact of solid particles ample, ceramics are used because of project has led to the development of or liquid droplets against solids. Al- though each impact causes only a minute amount of damage, the cumu- lative effect of millions of impacts can result in appreciable loss of material from structural components. Further- more, even the minute damage caused by the impact of a single particle often j

i can significantly reduce the strength of brittle materials. In many military applications, erosion shortens the lifetime and increases

1 the cost of replacement of mechanical components. The operation of gas turbines in dusty terrains, for example, 1 reduces turbine lifetime to one-tenth jl the normal value. As a result of this shortened lifetime, the cost of replac- ing turbines in helicopters used in Southeast Asia amounted to $150 million in 1970 alone. Other examples of component degradation include: the erosion of radomes by water, ice, and dust; the damage of optical components such as laser windows by rain and dust; the damage of turbine and rocket components by combustion NATIONAL MEASUREMENT LABORATOFT

a deeper understanding of the erosion brittle than MgO. For these materials, tangles of dislocations result in com- process, and to the development of new plastic deformation was confined to the plete scattering of the transmitted elec- elastic/plastic models of erosion that near vicinity of the impact site, while tron beam. At the edge of the impact are particularly applicable to brittle cracks propagated relatively large dis- site, the dislocation density drops materials. These models are based on tances into the crystal. The behavior abruptly to the density characteristic of the observation that all erosion of of SiC and AI0O3 was somewhat inter- the undamaged crystal. Ceramics and ceramic materials, no matter how brit- mediate, the extent of deformation metals differ in appearance because tle, is accompanied by plastic deforma- being approximately twice the size of metals are able to accommodate the tion. the residual impact impressions. Cracks mechanical strains resulting from the

The earliest indication of plastic de- in SiC and A1 2 0 3 propagated beyond impacting particle, whereas ceramic formation during erosion of hard brittle the deformation zone and were arrested materials crack when subjected to high materials, such as A1 2 0 3 , SiC and Si, in materials that was free of disloca- strains. As a consequence, crack for- came from examination of individual tions. mation is normally observed during the impact sites by light microscopy. Resid- A qualitative comparison of this erosion of ceramics. In fact, it is this ual plastic impressions were found to type of erosion damage with that oc- crack formation that normally accounts be associated with cracks that formed curring in metals indicates a great sim- for material loss during the erosion of as a result of the impact. The damage ilarity for both types of materials with ceramics. resulting from impact is, in fact, very regard to the density and extent of Investigations of impact damage by similar to that produced during hard- dislocation formation at the impact site light microscopy and by scanning j ^ ness testing using Vickers or Knoop ( figure 2 ) . For both types of materials, electron microscopy are useful for diamond indentors. For soft materials the density of dislocations at the con- revealing the relative extents of frac- such as LiF or MgO, etch-pit tech- tact site is so high that the crystal ture and plastic flow. Plastic deforma- i niques have been used to show that structure is severely strained; fruitful tion in ceramic materials is enhanced dislocation slip-bands are associated examination by transmission electron when particles strike target surfaces [. j with impact damage or hardness im- microscopy is not feasible. Dense at low angles of impingement. Scan- pressions. These observations naturally suggest that similar deformation accounts for impact impressions in hard materials such as Si, A1 2 03 and SiC. However, direct evidence for plastic flow in these materials was difficult to obtain, because the dislocations produced by impact are closely confined to the impact site. This close proximity of the dislocations to the plastic deformation was observed for all materials studied (Ge, Si, A1 20 3 , SiC, MgO), but the extent of deformation and the morphology of the cracks associated with the deformation was dependent on the crystal structure. MgO, an ionic crystal, exhibited the most extensive deformation at impact sites; cracks that formed (on 110 planes) in this crystal were always confined within the zone of plastic deformation. Quite the opposite behavior was observed for Si and Ge, which are covalent materials and more

80 :enter for materials science

Atomistic Theories of Crack Growth and Fracture

Edwin R. Fuller, Jr., Robb M. Thom- son, and Brian R. Lawn 1

structural component that ning electron micrographs of the dam- tion and material loss from brittle ma- A fails by a brittle fracture mode gen- age surface suggest extensive plowing terials during erosion. These theories erally involves the growth of an in- as a consequence of the particle im- are currently being tested in the Center trinsic flaw, or crack under sustained pact (figure 3). At low impact angles, for Materials Science in order to de- load until it attains some critical size. fracture during contact is greatly re- velop a body of knowledge that can be The desire to understand this crack duced. The damage sites are often used to model the lifetime of ceramic growth and fracture from a funda- free of cracks, and plastic deformation components in environments where mental level has provided considerable appears to be the primary mode of erosion is important. Studies suggest impetus to the atomistic modeling of material removal. These observations that the hardness, H, and the critical crack tips and crack-tip processes. are confirmed by transmission electron stress intensity factor, K c , are the two While these models are usually overly microscopy studies, which show that materials properties that control ero- simplistic in relation to real structures crack-free plastic indentations are fre- sion: a strong dependence of erosion and materials, they demonstrate the quently obtained in brittle materials rate on Kc , but a weak dependence on connection between microscopic proc- ^for small angles of impact. H, is observed. Although the available esses in the crack-tip region and 1 Based on scanning and transmission experimental data supports the view macroscopic fracture parameters; they electron microscopy data obtained on that K c and H (measured using dy- thereby provide a qualitative frame- ceramic materials, several models of namic loading) are independent of work for considering the complicated 'erosion have been developed recently temperature, additional research is phenomena of crack growth and frac- J which incorporate concepts of elastic- needed to confirm this conclusion. Con- ture. Only through this fundamental 'plastic fracture into erosive wear. The sequently, future research in the Center understanding can more realistic formation and growth of plastic zones for Materials Science will concentrate models be developed and more quan- during the impact process determine on the dependence of erosion on K c titative calculations be performed that the types of cracks that are formed and and H, and on the effect of temperature permit an assessment of crack growth hence, the amount of strength degrada- on these parameters. behavior, and accordingly, of component reliability from more fundamental principles.

Our studies in the Center for Ma- terials Science have been to develop atomistic models of a cracked solid, which minimize the mathematical com- plications that necessitate elaborate computer calculations, yet which contain the essential elements of the fracture process and do not preclude general conclusions. Reconsidering an

earlier model of Thomson et al [1], we were able to decouple effectively the nonlinear interatomic interactions in the crack-tip region from the linear elastic interactions in the bulk of the

solid [2,3]. This stratagem provides a mechanics model within which the influence of the crack-tip interaction can

Figure 3. Impact crater in dense A1 3-5 (im grain size, SiC particles 0 0,, caused by 150 fim 1 On study leave from University of New South

'i at a velocity of 90 m/s and an impingement angle of 15°. Wales, Australia.

81 :

NATIONAL MEASUREMENT LABORATORY

be explored. It was into this framework in the linear regime of their force- the variation of the system potential that we were able to incorporate, again displacement response with spring con- energy as the crack-tip bond ruptures. in a simplified manner, the molecular stant a. The one exception is the crack- We require, however, that for each interaction of a gaseous species with the tip interaction which, by the necessity crack-tip bond displacement the re- crack-tip atoms to obtain a theory of that it must rupture to advance the maining atoms of the solid are in an chemically assisted crack propaga- crack, must be allowed to be in the equilibrium configuration with respect tion [4,5]. nonlinear regime of its response. Trans- to this displacement. Necessary con- In the following section we describe verse opening forces P are applied to ditions for these equilibrium configura- the lattice model of a crack we have the end atoms of each chain to force tions are considered, and indicate how the atom- open the crack. [3U/?(2uj)] = 0, forj^n, (2) isticity of the problem leads to a num- The total potential energy of this ber of mechanically stable crack con- crack system is given by [3] which correspond to an infinite set of figurations ("lattice trapping"), from fourth-order difference equations. Ana- 2Uj U = — 2Pu„ + /JS°° [uj+ i — which the crack can advance by j=i lytical solutions of these equations are j 2 131 activation. The final section discusses + u^x] + nUan + U„(2u n ) obtainable in the functional form Vi Uj ]-, how a molecular chemical reaction can + [2 (1) Uj(P, n, u„). Substitution of these equi- j = n+l be incorporated into this mechanics librium-displacement solutions into Eq. where the first term on the right hand model to give a qualitative model of (1) accordingly reduces the system po- side is the potential energy of the load- chemically assisted bond rupture. tential energy to an explicit function of ; ing system; the second term is the Intrinsic Bond Rupture. Our atom- both the applied fracture mechanics strain energy stored in the flexural istic model of a crack in a quasi-one- parameters (load and crack length) and bonds; and the remaining terms are the dimensional solid is shown in figure 1. the crack-tip bond displacement contributions from the cohesive bonds It consists of two semi-infinite chains across the crack. These cohesive con- 8 2u„: of atoms that maintain their rigidity by n = tributions are, respectively, the total means of an interatomic interaction. U(P,n,8„) = nUBB - (PV6/J) bond energy from the n ruptured bonds, (2n :; 3n 2 t n) — KS This interaction is modeled as a flex- + + n the nonlinearly the bond energy of Vi (_L=J_) ural spring element with force constant + Ub (8n) + tt 8 2, (3)1 strained bond at the crack tip, and the /?. The chains of atoms are bonded to- linear-elastic strain energy of the re- gether by a series of cohesive interac- is an elastic coefficient defined maining cohesive bonds. From a frac- where tions between the chains, modeled as and (P/£> ture mechanics viewpoint, this poten- by 1) = (2/3/«), K = stretchable spring elements between (n-j-t) is an effective one dimensional tial provides both the driving force for the atom pairs. To form a crack, we "stress intensity factor". fracture and the resistance force to suppose that the first n stretchable ele- This expression is particularly useful fracture. The crack driving force is de- ments are ruptured, or stretched be- for the origin of the rived from the stored elastic strain demonstrating yond their assumed finite range of atomistic energy barrier to intrinsic energy which is released upon crack interaction. We assume further that the advance; whereas, the resistance to bond rupture and crack advance. Fo-r remaining cohesive spring elements are cusing our attention on the n th bond, fracture is contained in the cohesive P bond energy of the crack-tip bonds that the first two terms on the right hand must be ruptured to advance the crack. side of Eq. (3) are constant for a given In contrast to a Griffith continuum applied force and will not be con-j analysis, where the fracture resistance sidered further. The next term, — K8„, corresponds to the reduction in poten- is provided by the average cohesive energy required to form the new ele- tial energy that results from the applied ment of surface, the atomistic treat- fracture driving force as the crack-tip ment reveals a fine structure in this bond ruptures. The two remaining process through the discrete nature of terms represent the resistance of the the bond rupture event. To illustrate system to the rupture of the crack-tip Figure 1. Atomistic model of a crack in a quasi-one-dimensional solid. this atomistic influence, we examine bond. The first of these resistance

82 .

i, CENTER FOR MATERIALS SCIENCE

.terms. Ub (8n ), corresponds simply to displacement. We assume that the the interatomic interaction between the A—A atoms arrange themselves in the

: most energetically favorable position £ — 1 term, Vi * s more su e for each value of the B—B bond dis- "C^,—^ «' ^^ placement. Comparison of curve (1)

: crack-tip atoms; whereas the second with curves (2) or (3) show that the

• and corresponds to the harmonic re- presence of the A2 molecule makes it

- storing potential exerted on the crack- easier to dissociate the B atoms. Gen- tip atoms by the remainder of the erally this reaction occurs with an

(linear elastic) structure. It is the bal- activation energy barrier that could re-

ance between these last three terms that Figure 2. System potential energy as an ex- sult, for example, from the repulsive plicit function of the crack-tip bond dis- i determines the atomistic energy barrier, interaction of the polar component of

placement, 5 : (a) Individual contributions; las illustrated graphically in figure 2. n the AB molecule. If the reaction is en- and (b) Sum of these contributions, illus- -;The individual terms are plotted in dothermic, the final state of 2AB is trating the stable initial and final config- r figure their is shown in metastable with respect to that of 2a and sum urations (i) and (f) respectively, and the

-figure 2b. Clearly, the stress intensity activation state (/*) through which the sys- A 2 -f- BB, and the reverse is true for tem passes to rupture the crack-tip bond. , factor term, — K8U , biases the bond exothermic reactions. T is the activation energy barrier to this - rupture to make the final state the more AU The procedure for inserting this intrinsic : bond rupture, and aU- is the stable configuration in this example. chemical reaction, via a modified co- energy barrier to the activated healing of hesive force law, into the constraint of this bond. Chemically-Assisted Bond Rupture. a lattice crack is the same as that used

' An important part of the above con- in the construction of figure 2b. The struction was that the interatomic po- initial and final equilibrium configura- tential between the crack-tip atoms was tions now correspond to nonreacted effectively decoupled from the interac- and reacted (adsorption) states of the tions of the remainder of the solid. Ac- bond, respectively. The unstable con- cordingly, this construction provides a figuration relates to the activated com- mechanics model whereby an arbitrary plex of the reaction. The exact in-

I bond-rupture reaction can be studied in fluence of the chemical reaction on the rupture of the crack-tip bond depends jan equivalent linear elastic continuum Figure 3. Schematic of a bimolecular chem- I t_\ ical reaction to rupture the crack-tip bond. on the details of the chemical reaction of stiffness - ap- crack-tip a ( ). An A gaseous A., molecule reacts with the and its activation barrier. Various pos- crack-tip bond — B — B— to produce the sibilities have been considered in a re- I propriate crack-tip reaction to consider '"terminal" surface species A — B — cent manuscript [7]. The main effect iis a bimolecular chemical reaction. It of the chemical reaction is to further 1 provides a qualitative model for under- bias the intrinsic activation energy standing stress corrosion cracking in a barrier and to allow crack growth to gaseous environment. occur at a lower applied stress. An- H We begin by reviewing the energy other important conclusion is that the ' surface for a bimolecular chemical thermo-dynamic surface energy, which reaction of two free molecules (similar j determines the thermodynamic condi- I to Lawn and Wilshaw [6]. Consider tion for crack propagation, is also the reaction depicted in figure 3 where (exothermic) lowered to that of the chemically re- ; a gaseous molecular A interacts with 2 acted surface. ithe crack-tip bond —B—B— to form Figure 4. Possible potential energy surfaces ;the B for the bimolecular chemical reaction illus- product molecule A— — . Several trated in figure 3. The chemical reaction < possible potential energy surfaces for References reduces the bond-rupture energy or thermo- this . four-atom complex are depicted in dynamic surface [curves (2) or (3)] com- [1] Thomson, R. M.. Hsieh, C, and figure 4 as a function of the B—B bond pared to the nonreacted surface [curve (1)]. Rana, V., Lattice Trapping of Frac-

83 NATIONAL MEASUREMENT LABORATORY

Neutron Diffraction Profile Analysis and the Structures of Electronic Ceramics

Antonio Santoro and Robert S. Roth, Reactor Radiation Division

ture Cracks, J. Appl, Phys. 42, 3154- I n recent years it has become evi- the fluorite-type structure with oxygen 3160 (1971 ). dent that certain crystalline phases vacancies, and the conduction involves [2] Fuller, E. R., Jr. and Thomson, R. M., allow cations and anions to move some type of hopping mechanism Nonlinear Lattice Theory of Fracture, through the crystal structure of which through the vacancies. However, it has in Fracture 1977, D. M. R. Taplin, they are part much more quickly than been found that a new group of com- ed. Vol. 3, pp 387-394 (Univ. of had previously been believed possible. pounds based on the scheelite- and Waterloo Press, Canada, 1977). Such materials, which include, BaMnF -type structures [3] Fuller, E. R., Jr. and Thomson, R. M., for 4 may accom- Lattice Theories of Fracture, in Frac- example, sodium /^-alumina and cubic modate oxygen ions in interstitial posi- ture Mechanics of Ceramics, R. C. stabilized zirconia, are often called tions which then become the vehicle Bradt, D. P. H. Hasselman and F. F. "super-ionic conductors." In an effort for the ionic conduction. On the other Lange, eds. Vol. 4, pp 507-548 (Ple- to better understand the mechanisms hand, the Li+ containing materials num Publishing Corporation, 1978). involved in super-ionic conductivity, have structures with channels suffi- Fuller, [4] E. R., Jr. and Thomson, R. M., the phase equilibria, crystal chemistry, ciently large to permit the motion of Theory of Chemically Assisted Frac- and crystal structure of these materials the conducting ions. ture, in Proceedings of the Third are being examined in laboratories A problem associated with the study International Conference on Mechan- throughout the world. In fact, only a of compounds of the type mentioned ical Behavior of Materials, K. J. Mil- thorough understanding of the possible is related to the fact that heavy and ler and R. F. Smith, eds. Vol. 2, pp mechanisms of conduction in different light atoms are present simultaneously 485-494 (Pergamon Press). types of conductors will enable in the structures. It follows that x ray [5] Fuller, E. R., Jr. and Thomson, R. M., a more

Atmosphere Assisted Fracture: II. accurate prediction of properties and diffraction cannot lead to an accurate Atomic Models of Crack Growth", allow the researchers to tailor-make knowledge of the atomic distribution, (J. Matl. Sci., in press). solid ionic conductors for specific ap- and neutron methods are the only [6] Lawn, B. R. and Wilshaw, T. R., plications in batteries, fuel cells, etc. alternative. In addition, many of the Fracture of Brittle Solids, Section 7.5 materials of interest cannot be pre- (Cambridge University Press, Cam- pared as single crystals of sufficiently bridge, 1975). The two most promising classes of large size for single-crystal work and [7] Fuller, E. R., Jr., Lawn, B. R. and possible ionic conductors have 0~ 2 and have to be analyzed by powder dif- Thomson, R. M., Atomic Modelling Li+ ions as the mobile species. Con- fraction. Because of the random orien- of Crack-Tip Chemistry (to be sub- ventional oxygen ion conductors have tation of the crystallites in a powder, mitted to Acta. Met).

Figure 1. View from the top of the counter arm of the five-counter diffractometer. The counters are placed 20° apart and simultaneously measure different regions of the pattern. The sample is located at the geometrical center of the diffractometer.

84 CENTER FOR MATERIALS SCIENCE

the three-dimensionally diffracted inten- olution diffractometer for profile anal- the structure of H-LiTa 3 O s . As an sities are collapsed into a monodimen- ysis studies. example of the results obtained in these "sional distribution, which results in a With this facility, illustrated in figure studies, the calculated and observed

loss of information about the spatial 1, five counters are used simulta- profiles for the structure of LiNb 3 O s

properties of the crystal, and, if over- neously. The resulting gain in intensity are compared in figure 2.

- lapping of the diffracted lines occurs, may be used to reach an instrumental More recently, it has become ap- about the intensities associated with resolution such that the natural broad- parent that the structures of many ionic each set of crystallographic planes. The ening due to the physical conditions conductors are modulated. Modulation, Most information can be partially re- of the sample becomes observable and for example, has been clearly estab-

covered if the powder patterns are constitutes the predominant factor in lished in CeNbO.i.0K, and there are good

: measured with diffractometers of high determining the width of the diffracted reasons to believe that it is also present

resolution, and if, instead of using in- lines. in H-LiTa :1 O s . Little is known about tegrated intensities in structural refine- At the NBS reactor, studies of possi- modulated structures, and even their

ments, one fits the entire observed pro- ble ionic conductors are underway as definition is still being debated. We

'•file with one calculated in terms of a part of a collaborative effort between may define modulation from a strictly c plausible model of the structure ("Riet- scientists of the Reactor Radiation Di- crystallographic point of view in the veld Method," or "Profile Analysis"). vision and the Ceramics, Glass, and following way:

Not surprisingly, the application of Solid State Science Division. Com- We may say that a structure is "mod- -2 • the method of profile analysis in the pounds with both O and Li+ ions ulated" when it has one or more struc- 5 U.S. and in foreign laboratories has have been, or are being, examined with tural "features" having a periodicity

become increasingly important in the powder neutron diffraction. The struc- which is "inconsistent" with, or "in-

study of a wide class of polycrystalline tures of CeNb0 4 , CeTaO,, and NdTa0 4 commensurate" with respect to, the - materials, and numerous improvements have been well refined so that reason- periodicity of the bulk of the structure. of both the experimental techniques and able conjectures can be made about the When this condition occurs, the ratio of

data refinement methods have been relationship between the stiochiometric the two periods is not a rational num- •proposed or applied. In the past few and the oxygen-rich phases obtainable ber (perhaps, we should say that this

years, scientists and technicians at the by oxidation of some of these com- ratio is not "a simple" rational num- NBS reactor have established the fore- pounds. Similarly, the structures of ber). The "features" which cause the

' front capability in the U.S. in the LiNb O M-LiTa , and Ta modulation may be: 1) positional in 3 s , 308 2W08

| methods and applications of neutron have been accurately measured as part nature (i.e., shifts from the positions diffraction-profile analysis and have de- of an attempt to determine the still that one or more atomic species would

veloped a new multidetector high res- unknown positions of the Li+ ions in have if the structure were unmodul-

S 4

i? 3 "3 I 2 c

1 1 2 o Li 10 12 14 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64

Diffraction angle (degrees)

Figure 2. Calculated profile (continuous line) and observed data (dots) in the powder

pattern of . LiNb,O s The close agreement between theoretical prediction and experimental c'ata, shown, in the diagram, indicates that the refined structure is highly accurate.

85 NATIONAL MEASUREMENT LABORATORY

Recent NBS Advances in Physical Properties Measurements at Ultra High Pressures

Stanley Block, Ronald G. Munro, and Gasper J. Piermarini

ated), 2) substitutional (i.e., one or JL^ ressureres: is beginning to take its P Plate more sets of positions, which would rightful place beside temperature as an be crystallographically equivalent if the important research variable in the quest structure were unmodulated, have vari- to understand the nature and behavior able occupancies), or 3) a combination of matter. Two developments at NBS of the two. are chiefly responsible for the recent The diffraction patterns of such blossoming of high pressure science: structures contain reflections corre- (1) the capability to generate very large sponding to the periodicity of the bulk static pressure and (2) the capability to of the structure ("main" reflections) measure it. As a result of these two and reflections which are inconsistent milestones, the discovery of new phases with, or incommensurate with respect and properties of materials, new in- to. those of the first type ("satellite" sights into the physics of condensed Hemisphere reflections). These definitions are nec- matter systems, the synthesis of new essarily provisional and may well be materials, and the opportunity to test modified when more information is experimentally theories which pre- Figure 1. Schematic cross-section of the gained about modulated structures. viously could be considered little more Diamond-Anvil assembly of the NBS pres- The solution of structures of this sort than interesting conjectures have all sure cell. Two gem diamonds roughly 1/3 requires the development of new tech- become possible. A steady flow of re- of a carat in size act as opposed anvils. The sample is contained in a small hole (ap- niques for the interpretation and proc- finements and new applications are now proximately 0.5 mm in diameter) in the essing of the experimental data. Along being developed in our laboratory. gasket located between the anvil faces. The this line, we are at present developing Other developments can be found in diamonds also act as windows for 180° and methods for indexing powder patterns the current scientific literature. EPI optical access. in which the presence of satellite re- In the past, the direct generation of flections is suspected or ascertained. At

pressures by applying a known force been an increase in the number of ex- i the same time, least-squares fitting pro-

to a known area required the costly periments that can be performed at high 0 :.: grams are being, or have been, modi- construction of a large apparatus and pressure. Previously, work was res- fied so that the intensities of satellites

a considerable effort to operate it suc- tricted primarily to compressibility i can be taken into account, and so that cessfully. The development of the dia- studies, pressure-volume-temperature v. background and thermal motions of mond anvil pressure cell at relations, and electrical resistance (DAPC) i the atoms are properly treated. This the National Bureau of Standards sur- measurements. The DAPC adds micro- last development should be particularly mounted these difficulties. The scopic observations, infrared spectros- useful in the study of nonstoichiometric NBS pressure device obtains large pressures copy, x-ray studies, and, most recently, phases, in which one or more sets of by using small areas, The substance viscosity measurements. equivalent positions are partially occu- to be studied is placed Quantitative use of pressure in re- pied. between two diamonds (see figure 1). When the dia- search requires both a means to gen- We hope the NBS program to com- monds are squeezed together with a erate the pressure and a means to meas- bine advanced profile refinement

- moderate force, a large pressure is ure it. Pressure is force per unit area, . methods, high-resolution neutron dif- transmitted to the substance because and hence if a known force is applied - fraction, and new approaches to deal-

of the small area presented by the faces to a known area, the pressure is easily : ing with and predicting incommen- of the diamonds. The resulting appa- computed. This direct method of pres- surate features in complex diffraction ratus (see figure 2) is compact, small sure determination was the basis of patterns will lead to a significant im- enough to be held in one hand, rela- early pressure devices, but severe prob- provement of our understanding of tively inexpensive, easy to operate, and lems in accuracy develop above about structure and properties of ionic con- readily adaptable to existing measure- 3.0 GPa. Consequently, secondary ductors and broad classes of other crys- ment equipment. Furthermore, pressure standards are required for talline materials. the diamonds provide optical and x-ray access most of the work at high pressure. to the stressed sample. The result has Fortunately, a very useful, rapid.

86 ENTER FOR MATERIALS SCIENCE

ranslating Research Activity Diamond diamond with the Anvils Mount Diamond Anvil Cell late

[Siting ™ 4 Ruby Scale Diamond 01 Mount Calibrated at NBS lemisphere Ruby Pressure Sensor o Introduced by NBS < 2 - basket

Pressure 70 Plate 'oke Year

Pressure Plate Yoke Bearing Figure 3. The chart shows the intensity of

igl're 2. The mechanism for generating force consists of a spring loaded lever-arm research activity with the diamond cell rela- 'psembly shown in cross section above. Belleville spring washers are compressed by the tive to the year 1972—the year that the

: mple rotation of a screw. The force generated is transmitted through the lever arms to ruby pressure sensor was introduced by NBS. ie pressure plate yoke which loads the piston containing one of the anvils. The chart is based on results from a recent literature survey.

Convenient, and versatile pressure scale The impact of the ruby scale on high white radiation source, absorption in as developed at NBS using the fluores- pressure research with the diamond the diamonds, beam divergence and

;nce spectra of ruby. The famous red anvil cell is indicated in figure 3. The particle size of the sample, as well as

Dior of ruby is produced by chromium curve is the result of a survey we made collimator and slit system design. Im- npurities in the host material, alumi- of the scientific literature from 1960 provements in equipment and tech- num oxide, and the red fluorescence through 1978. After the introduction niques have led to obtaining better -:nes known as the sharp R-lines occur of the DAPC in 1959, activity in- data in less time, and thus have en-

t well defined wavelengths. The pres- creased slowly for a few years, but hanced NBS capabilities for high pres- ume dependence of the R^-line was without a quantitative knowledge of the sure crystallographic studies. alibrated at NBS as a pressure scale. pressure, interest was limited. However, As examples, we discuss our results 'Vim the original calibration, pressures after the introduction of the ruby pres- on CuCl, CdS, and ice VII. The be- p to 30 GPa can be determined ac- sure scale in 1972 and its calibration havior of CuCl under pressure has re- urately. With increased use of the in 1975, a dramatic increase occurred ceived much attention recently,

uby method, work to extend the scale in research activity which has not yet prompted by the reports that it under-

100 GPa has been reported by reached its peak. goes a transition to a metallic state at athers recently. The temperature de- Among the recent advances asso- about 4.0 GPa. Further, anomalies in pendence of the scale has also been ciated with the increasing activity have compressibility, resistivity and mag- xamined in other laboratories over the been the applications of x-ray energy netic susceptibility in the pressure range

ange in 4.2 K to 358 K; and it has dispersive and single crystal techniques. of 4.0 GPa and in the temperature een found that the rate at which the We have, for example, carried out with range 4.2 K to 300 K have been re- avelength of the R^-line changes with F. Mauer of NBS a detailed study of ported, and the existence of super- bressure is nearly constant over the factors affecting intensity and resolu- conductivity has been postulated for

ntire range. As a result of its con- tion in energy dispersive powder dif- the anomalous diamagnetic behavior.

enience and its large range of utility, fraction in order to optimize the ex- The pressure dependence of the

ie ruby fluorescence technique is the perimental conditions for obtaining x- electrical resistance of CuCl to pres- tost widely used pressure scale pres- ray diffraction data with the diamond sures up 14.0 GPa was measured using

ntly available for pressures above 1 anvil cell. The factors studied included our recently developed two-probe PPa. size and intensity distribution of the method adapted to the diamond anvil

87 NATIONAL MEASUREMENT LABORATOf:

high pressure cell. The results of our with 0-0 distances. We carried out x- electrical measurements, which were ray powder diffraction studies using made on four samples of CuCl pre- wavelength dispersive film and energy pared by different methods, do not dispersive counter techniques. We

1 show any evidence for a "metallic' Ila found that the cell dimensions changed Phase even after repeated cycling. We from 33.55 nm at 2.36 GPa to 30.35 nm m also considered the possibility of a at 20.2 GPa indicating a AV/V, of 26 I photo-conduction effect in the CuCl percent and a shortening of the O-O samples. The electrical measurements bond distance from 29.06 nm to were performed in the absence of light, 26.38 nm. There is a sharp break in in dim red light, and in white light, the compressibility at approximately and no significant photo-effect was ob- 15.86 GPa with very little change in served. In additional experiments done A„ (30.52 to 30.35 nm). This probably in collaboration with A. Jayaraman, indicates that the repulsive forces be- 1.5 2.0 2.5 3.0 3.5 4.0 4.5 et al., at Bell Telephone Laboratories come highly significant. Pressure, GPa with a piston cylinder device, no large At the request of Watervleit Arsenal Figure 4. Viscosity of n-butyl chloride as resistance drop was observed in the we are studying the electrical and function pressure at 20 °C. The equilibriun 0.1 to is in of as a func- range of 5.0 GPa, which structural properties CdS freezing pressure is 2.3 GPa: above thii agreement with our diamond cell re- tion of pressure and temperature. The pressure, the liquid is metastable (super sults. semiconductor-conductor transition has pressed ). We have also performed x-ray meas- been measured electrically, by energy urements on CuCl. We used both film dispersive x-ray diffraction, and ob- techniques with ungasketed CuCl sam- served microscopically. The transition new type of diamond anvil pressun cell designed for single-crystal x-ra^ ples and gasketed samples with our is dependent on particle size purity. newly developed energy dispersive Single crystals transform at 2.65 GPa, studies. A single crystal of S 4 N 4 wa: method mentioned previously. Our re- while powders transform at 3.25 GPa. first mounted on a glass fiber and tht unit cell determined by centering 15 re<. sults confirm the transitions from zinc- Interestingly, the slope AP/AP tl . a iis is diffractometeri blende to tetragonal structures at 4.4 negative. At 100°C, the transition oc- flections on the Syntex of the monoclinio GPa and from tetragonal to the NaCl curs at 2.49 GPa and at — 44°C, the The cell parameters unit cell (a b 71.73! type structure at about 8.0 GPa. Our transition is at 3.24 GPa. All measure- = 87.91(1), = 92.47(1)° x-ray pattern of the black opaque ments were performed in a 4:1 meth- (13), c = 87.06(1) nm;/? = phase taken in an ungasketed diamond anol-ethanol solution and were there- numbers in parentheses represent un anvil consists of NaCl-type lines. We fore hydrostatic. certainty in the values) agree witl therefore believe that in the NaCl We also made x-ray diffraction meas- those obtained from a single crysta phase of CuCl, a disproportionation re- urements above and below the CdS placed within the pressure cell with n( applied. Unit eel action sets in, in the presence of large transition. The transition is from the external pressure shear stress. In summary, we found no zincblende structure where each atom data were also determined at pressure 4'J evidence for a dielectric-transition in is tetrahedrally coordinated to the so- of 1.3, 3.0, 3.7, and 4.2 GPa. At CuCl. dium chloride structure where each GPa, the cell parameters decrease to = In collaboration with G. E. Wal- atom is octahedrally coordinated. Pre- a = 81.56(21), b = 66.67(5), c rafen of Howard University, we have liminary results indicate that the low 81.62(8) nm; « = 89.93(7), p = carried out high pressure Raman and pressure phase has a volume compres- 91.74(16), y = 89.90(15)°. The vol x-ray diffraction experiments on ice sion of — 3.1 percent to the transi- ume of the unit cell decreases fron 5 3 VII. The group at Howard University tion and an additional — 19.3 percent 5.481(2)X10 nm to 4.437(13) X 10 3 whif. did Raman studies in a diamond anvil upon transforming to the high pressure nm , a reduction of 19 percent, cell. They determined that the O-H form. the a and c axial lengths become equa

vibrational frequencies of ice VII shifts The compressibility of S,N 4 was within the accuracy of the measure

downward with pressure and they ex- studied with J. K. Stalick of NBS to a ments. The monoclinic metric symme pected that the shift would correlate pressure of 4.2 GPa through use of a try is retained to 4.2 GPa.

88 : • 'ENTER FOR MATERIALS SCIENCE

Piezoelectricity and Pyroelectricity in Poly(vinylidene fluoride)

Martin G. Broadhurst and George T. Davis, Polymer Science and Standards Division

Since the diamond anvil cell places ;vere restrictions on the acquisition l n the past ten years polymer trans-

f data, especially with respect to ducers made of po!y(vinylidene

ngle, a theoretical investigation is fluoride) (PVDF) have been discovered eing conducted to examine the con- and accepted as reliable, practical and equences of these limitations for radial unique measurement devices. NBS istribution function analysis. The was involved in early developments in

nethod of investigation is to form a this field and demonstrated a variety omparison of the reliability of various of applications such as intrusion and

omputational procedures and to con- fire detectors, laser beam profile

SO I i i i i i i i i— truct the comparison as a function of 0 2.0 4.0 6.0 8.0 measurement devices, absolute radiom- he degree of data limitation. An exact Pressure, GPa eters, hydrophones, etc. NBS work et of data having an exactly known has stimulated commercial development Figure 5. Line width of the ruby Rj fluores- adial distribution function is used so their cence relative to the one atmosphere value, of transducer devices and hat the absolute error can be deter- as a function of pressure at 20 °C when application to measurement problems nined in each case. n-butyl chloride is used as the pressure throughout government. transiting medium. Glass transition pres- The recently developed diamond an- During this period, we were particu- sure = 4.2 GPa. il cell falling-sphere viscometer has larly involved in establishing a broad >een utilized to measure the pressure description of the basis of piezoelec- determined for n-butyl chloride, 4:1 lependence of the viscosity of stable tricity and related phenomena in methanol-ethanol, and isopropyl alco- .nd metastable n-butyl chloride (see PVDF. This knowledge was needed to hol, and the results have been used to igure 4). This substance was chosen as guide optimization of the effects and confirm the ruby line-broadening meth- " uitable for a cooperative intra-NBS define reasonable limits on how much od of determining glass transition pres- investigation of metastable states of could be expected from these devices. sures. For the noted liquids, the model -natter. The equilibrium freezing pres- In the following, we present a general yields, respectively, 4.6, 10.2, and 5.1 { ure (2.3 GPa) and the glass transition description of the molecular structure GPa for Pg , and the ruby method yields Pressure (4.2 GPa) were determined at of PVDF, show how a simple piezo- 4.5 - 4.2, 10.4, and GPa. oom temperature by separate measure electric model was developed and In addition to confirming the line- nents to identify the region of meta- successfully tested, and describe the broadening method of finding P r , the tability (2.3 GPa < p > 4.2 GPa) important consequences of this viscosity model might prove to be use- • see figure 5). The viscosity was then development. ful for describing the properties of obtained by starting at a pressure of lubricants. Since lubricants are used Microscopic Details of PVDF ill. 94 GPa. This transport property was under conditions of widely varying PVDF crystallizes from the melt into i ound to vary smoothly through the stresses, the value of the exponent in spherulitic structures. Figure 1 is a transition from the stable to the meta- the viscosity model could be important photomicrograph of a melt-crystallized stable state. for evaluating the ability of a lubricant PVDF film taken between crossed > The viscosity results for n-butyl to respond to the changes in the stress. polaroids. The volume fraction of crys- Chloride were also used in conjunction The unusual magnitude of the critical talline material is typically about vith the critical point viscosity model, exponent also presents a challenge to 50 percent, depending on thermal his- )-" 7 = A(1-P/Pe theoreticians in this field. tory. Most of the uncrystallized molecules are in a metastable supercooled (A-here -q is the viscosity, PK is the glass The temperature dependence of the transition pressure, and A and v are glass transition in 4:1 methanol-ethanol liquid phase. The glass transition pther parameters. Use of the model was determined by the ruby line broad- temperature for this liquid phase is

r akes it possible to obtain a reasonable ening method. At — 48 °C the glass around —50 °C.

estimate of the glass transition pres- transition is 5.65 GPa yielding APg/ We assume, as shown in figure 2, sure if the viscosity as a function of AT = 0.68 GPa per degrees. This that the spherulites consist of stacks of

oressure is known for sufficiently high value agrees with values obtained at lamellae which grow outward from the _ alcohols. p essures. Values of P g have thus been lower pressures for other center of the spherulite. These lamellae

89 NATIONAL MEASUREMENT LABORATORi

are typically 10-20 nm thick, depending liquid phase, the molecules continually carbon covalent bonds, or by rotatioi on crystallization conditions. The change shape due to rotations about of molecular segments about their lorn molecular chains are approximately carbon-carbon bonds. axes, or both. The driving force is tht normal to the large lamellae surfaces The net dipole moment of a group decrease in the applied field-molecula and to the radii of the spherulites. of molecules in a liquid region will be dipole interaction energy which is ex

Much of the liquid material is probably zero in the absence of an applied tremely large. A typical moleculai located between the crystalline lamellae. electric field because of the random segment in the crystal has a dipole 28 That is, a typical region (figure 2) orientations of individual dipoles. moment of roughly 2 X 10~ Cm (70 consists of parallel layers of alternating In the crystal phase, certain regular Debye). The energy of interaction be 8 _1 crystal and liquid material, each layer molecular conformations are energet- tween it and a field of 10 Vm i of the order of 10-20 nm thick. Of ically favored for crystal packing. several times the thermal energy course, on a larger scale one expects Viewed down the molecular chain, four room temperature and comparable t( the usual stacking faults, grain bound- such conformations are shown in figure the energies needed to rotate molecule aries, and other course defects typically 4. All four crystal structures are long in the crystal phase. Thus PVDF i found in a polycrystalline solid. Since lived (whether stable or metastable) at truely ferroelectric, but with the pola the molecular lengths are of the order room temperature, and three have alignment caused by VanderWaal of 100 times the lamellae thickness, polar unit cells. That is, a crystal of forces rather than dipole-dipole force each molecule must pass many times polar a, (3, or y phase has a net dipole and with a Curie temperature wel through one or more of the crystal moment. Crystal phase transformations above the melting range (170-200 °C layers and is free to assume flexible can occur through mechanical stretch- of the various crystal phases. and irregular configurations in the ing, high temperature annealing or Model for Piezoelectricity liquid layers. application of large electric fields. In The preceding established the exist PVDF is inherently polar (figure 3). addition to phase changes, changes in ence in PVDF of polar crystals pref The hydrogen atoms are positively orientation of the crystal moment can erentially aligned by an applied fielc charged and the fluorine atoms nega- occur in large electric fields. The field so that a PVDF film has a net electric tively charged with respect to the car- induced changes in orientation and moment. The charge on the film dut bon atoms in the polymer. In the phase occur by rotations about carbon- to aligned dipoles is the product of the polarization (dipole moment per uni volume) and area.

Q = N/x0 (ec + 2) < C os0 o >/3L, (1)

Crystalline Lamella

Non-crystalline Component

Figure 2. A schematic diagram of a sphere Figure 1. Photomicrograph of a rapidly cooled melt of commercial unoriented PVDF lite and a detail of a section emphasizing the between crossed polaroids. lamellar structure of the radiating branches 90 ,

"center for materials science

-/here N is the number and /j. 0 the 2 dp = Pofr J_(, _1) _i_J_<£ y fluorine ; acuum dipole moment of the dipoles, the angle between the dipoles and (2) -he normal to the film, /s the film thick- d( lnv c.)

less and £C the permittivity of the 2 Irystal. All of these quantities are Py — —Po«C 4-(«c-D +4- [y j Z o neasurable. (Ink) (2T« )- 1 As a beginning, we asked what the + C ] (3) (lnvc) j fleet of temperature and pressure will

>e. assuming the dipoles remain orien- where ationally frozen—the simplest possible J„ (<*» Po = *(€c+?)Nf | 0 0 ) ase. This question does not preclude

/3v (. (4) •>ther possible effects, but we will at phase ast take into account a very basic Effect before looking for more subtle nechanisms. Most workers in this field

: iave begun with special mechanisms

: uch as special space charge distribu- vc, /?c and «t. are the crystal volume and -ions, changes in the molecular con- its pressure and temperature coefficients ormations or orientations, etc., without of expansion, and y are the ampli- irst eliminating thermal expansion and tude of thermally induced dipole libra-

impressibility effects. This neglect is tions and associated Gruneisen con- 'iue in part to the relative unimportance stant, $ the crystal volume fraction and

)f volume expansion in inorganic J0 ( <>) a zeroth order Bessel function polar « phase Mezoelectrics. of the first kind. Taking the temperature and pressure Independent experimental values derivatives of the surface charge on were used to evaluate the right hand he PVDF film and assuming no change side of eq. (2) and eq. (3) without

in fio or the mean dipole orientation, we adjustable parameters and the results )btain the hydrostatic piezoelectric are shown in figures 5 and 6 together con- o itant . at dp and pyroelectric coefficient p v with data obtained NBS.

-y phase

v Jf o

fi phase Figure 3. Molecular model of PVDF in the all-trans conformation found in the p crystal phase. The backbone is composed of carbon atoms, the spheres represent fluorine atoms and Figure 4. Projection of four crystal struc-

the hydrogen atoms, if shown, would appear at the ends of the "sticks" which represent tures of PVDF viewed along the molecular zl emical bonds. axes. Dipole moments are shown by arrows.

91 NATIONAL MEASUREMENT LABORATORY

Electrochemical Noise Measurements: A New Measurement Technique for Diagnosis and Study of Localized Corrosion

Ugo Bertocci, Chemical Stability and Corrosion Division

These results show that without versible crystallization, molecular rota- changing the number of dipoles, vac- tions, etc., appear unjustified by the uum dipole moment or dipole orienta- data. FJLluctluct uations in the electron potential tion, essentially all of the observed and/or current of corroding electrodes Significance of this Development piezoelectric and pyroelectric response are giving information on the break- of PVDF is accounted for. Additional Based on the above mechanism of down and repair of protective films mechanisms connected with transport piezoelectricity in PVDF, we can make and are being studied as a possible of space charge, stress dependent re- some very strong statements about the means of detecting the output of optimization and limitations of this localized corrosion. Iron and aluminum material. in solutions that can cause pitting were

( 1 ) Present PVDF has about half the examined. A large increase of noise theoretical maximum polarization, so signal was detected when pitting that improvements in performance are occurred. expected to be small. (2) To find a "better" polymer than PVDF, one needs to have a dipole The detection and analysis of the with a smaller volume and larger fluctuations in current and potential of

moment than the compact (CF2-CH2 ) electrodes, which are generally called group. electrochemical noise, are receiving in- (3) Increasing the crystal to amor- creasing attention in electrochemistry

phous volume fraction can increase P0 , as a means for the study of some

but will decrease sample compressi- aspects of electrode kinetics [1] and 6 3 4 5 bility and expansion coefficient with are also being applied to the field of P ( MC/cm2) 0 small net gain. Some increase in com- corrosion, where the method has the

Figure 5. Piezoelectric coefficient of PVDF pressibility can result from incorporat- potential to be a useful tool, particularly film for a change in hydrostatic pressure ing small voids in the specimen. for the detection of localized corrosion versus polarization. Circles represent meas- (4) Use of the polymer is limited to [2]. ured data points; the solid line represents temperatures below the melting point. In the area of electrode kinetics, values to be expected using eq. (2) with the appropriate values of material properties. Increased stability can be expected one of the advantages is that there is from increasing the crystal lamellar no need to apply an external signal to' thickness and thus reducing the ratio the system under study. The analysis of fold surface area to crystal volume. of the experimental results, however,

(5) Moisture and moderate amounts still requires further development of of ionic impurities have negligible effect the mathematical models necessary for on the performance of PVDF. interpreting the results. As far as local-

(6) Piezoelectricity and pyroelectricity ized corrosion is concerned, the exami- in PVDF are directly related to the nation of the characteristics of the compressibility and expansion coeffi- deviations from the average values of cient so that both effects will exist the electrode potential and current can together. give insight in the processes of forma- 01 23456789 10 (7) The phase having the largest dipole tion of protective films and their break- moment will be the most active phase. down and might be able to signal the P 0 (/X/cm2) These conclusions give confidence in onset of pitting when the average cor

Figure 6. Pyroelectric coefficient of PVDF the dependability of PVDF transducers rosion rate is still low. film versus polarization. Open circles are used within the stated limitations and A number of techniques can be em- NBS data points, solid circles are from the provide firm guidance in the optimiza- ployed to characterize the electrochem- literature; the solid line represents values to tion of transducer processing and ical noise. main effort has been be expected using eq. (3) with the appro- My priate values of material properties. polymer modification efforts. directed toward looking at the signals

92 JtENTER FOR MATERIALS SCIENCE

trode characteristics, can be detected oscillations which were probably re- above the instrumental noise. lated to the formation of pits.

Noise measurements are now being Another system I examined was Al made on various electrodes in order to in a boric acid: sodium metaborate gather phenomenological information mixture with and without addition of about the noise levels generated by dif- a small amount of NaCl. In the

ferent electrochemical processes and chloride-containing solution, which is

to test the range of usefulness of the known to cause pitting if the potential

apparatus we have developed. For in- is raised approximately above —700 Frequency, Hz stance, I studied, as examples of low- mV vs SCE (Saturated Calomel Elec- ! noise processes [3], the deposition of trode) [4], the spectra obtained below -igure 1. Current spectra recorded in poten- iostatic conditions on different resistive copper and nickel. and at the pitting potential are shown ,

oads. Because it is an electrode system of in figure 3, together with the voltage

interest in corrosion science, I ex- spectrum. amined the Fe electrode in a neutral By dividing the voltage by the cur- n the frequency domain, through use borate solution with a small amount of rent at each frequency, a plot of the

if a spectrum analyzer. In order to cholride added. The voltage noise absolute impedance \L\ of the electrode ninimize interference, some instru- under galvanostatic conditions was as a function of the frequency can be ments are battery operated, and electro- much larger when the electrode was constructed. Figure 4 shows such a plot.

nagnetically shielded as well as pro- polarized anodically. As shown by the Here again, the onset of pitting is

' :ected from mechanical vibrations. The spectra in figure 2, the noise levels in- marked by a fundamental change of evel of instrumental noise often limits creased by more than a factor of ten the impedance at low frequency. When

^.:he electrochemical systems that can be when the electrode was above the pit- pitting does not occur, the random examined, thus, reduction of such noise ting potential. Correspondingly, obser- fluctuations in the current density are

. s of paramount importance in these vation of the voltage signal as a func- the system response to the voltage

> studies. tion of time revealed sudden bursts of noise input, and the electrode behaves

! 1 For the purpose of studying corro- 1 10-4 | , 1 sion processes, it is often convenient to control the potential of the electrode

jnder study. This is accomplished by neans of a potentiostat. Commercial f i nstruments, however, tend to intro- duce noise of unacceptable levels into ,:he circuits, preventing the study of the

signals of interest. For this reason, I developed a special low-noise poten- Ltiostat with the cooperation of the Spe- :ial Analytical Instrumentation Group, Center for Analytical Chemistry. The

{.instrument is designed for the detection bf fluctuations in the frequency range

0 1 2 3 4 5 1 103 between 0.1 and 2000 Hz. The internal 10 102 'noise is of the order of 3-10" 8 V/ V Hz. Frequency, Hz Frequency, Hz Figure 1 shows the noise current pro- Figure 2. Voltage spectra iron in borate Figure 3. Voltage and current spectra for educed by the potentiostat on different solution with chloride added, fa) anodic aluminum in borate solution with chloride + 1050 mV versus SCE; (b) anodic +750 added below and at the pitting potential. I resistive loads. With such an instal- mV versus SCE; (c) open circuit —280 mV Current density (a) —650 mV versus SCE; lment, current fluctuations in the nA versus SCE. Spectra are an average of 16 (b) 800 mV versus SCE. Dashed line is range, caused by changes in the elec- spectra. voltage.

93 .

NATIONAL MEASUREMENT LABORATORY

Migration Models For Polymer Additives

Isaac C. Sanchez, Shu Sing Chang, and Leslie E. Smith, Polymer Science and Standards Division

noise is due more to the faradaic or to n important problem in food 105 A the capacitative current. However, the packaging is the absorption into food increase in impedance at higher fre- of chemicals present in the packaging 104 quency indicates that the capacitative material. For the past two years, the a "a term is probably negligible above about Polymer Science and Standards £ 103 100 Hz. These results suggest that a Division has been studying the migra- u dynamic breaking and repairing of pro- tion of additives from polymeric c: tective films the metal are packaging materials to food simulating NT 102 on surface responsible for the signals observed. solvents. This study has been carried This work should thus provide new out under the aegis of an interagency 10 insights into the basic processes that agreement with the Bureau of Foods

control corrosion and, in so doing, also of the U.S. Food and Drug Admin- offer a new and more effective tool for istration. Our primary goal in this 10-1 1 10 102 103 detecting, characterizing, research program has been to develop Frequency, Hz and measuring various forms of localized corrosion a migration model which can be used Figure 4. Plot of the absolute impedance |Z| such as pitting. as a guide for the food packaging for aluminum in borate solution. — — 330 industry and serve as a sound scientific mV versus SCE (no chloride); = —800 A References basis for FDA regulatory decisions. mV versus SCE (0.05 M NaCl); = -650 1 Tyagai V. A., Elektrokhimiya 3, 133 1 mV versus SCE (0.05 m NaCl). 1 J (1967); Blanc G., Gabrielli C, Ked- Initially our research has focused on a class of called polyolefins. 2 dam M., Electrochim. Acta 20, 687 polymers like a capacitor of about 0.3 ^F/cm . (1975); Blanc G., Epelboin I., Gab- These polymers, of which polyethylene Above the pitting potential, the noise rielli C, and Keddam M., J. Elec- is the prototype, represent the majority current increases more than two orders troanal. Chem. 62, 59 (1975). of plastics used in food packaging ap- the plot of magnitude, and impedance [2] Iverson W. P., J. Electrochem. Soc. plications. Most polyolefins are semi- cannot be explained except as being 115, 617 (1968); Okamoto G., Su- crystalline, with the glass transition caused by fluctuations of the electrode gita T., Nishiyama S., and Tachibana temperature of the amorphous com- current, independent of the applied K., Boshoku Gijutsu 23, 439, 445 ponent below room temperature. For voltage signal. (1974). Bertocci such polymers migrations of additives The detection of the current noise [3] U., 7th International Con- gress on Metallic Corrosion (Rio De at room temperature often follow is a good indication of the beginning Janeiro 1978). p. 2010 Abraco, Rio simple Fickian kinetics. of pitting, and the large fluctuations de Janeiro (1979). Bertocci, U., /. We have found a "migration deci- observed are probably due to hydrogen Electrochem. Soc. (submitted for sion tree" (MDT), as shown below, evolution from the pits [5]. I found publication) to be a useful device in focusing atten- the onset of the large current noise to Pourbaix Corrosion [4] M., 26, 431 tion on the specific technical questions be reproducible within 20 mV. (1970); Galvele J. R., and de De- that must be answered when evaluating The noise current, which can be ac- Micheli S. M., Corr. Sci. 10, 795 migration. The basis for the MDT is counted for by oscillations of the order (1970); Nisancioglu K., and Holtan that migration can be limited by any of a few percent of the average dc cur- H., Corr. Sci. 18. 835 (1978). of four physical factors: [5] Bergeron C. B., and Givens R. B., rent, is the sum of several contribu- 1 initial /. Electrchem. Soc. 124, 1845 (1977); ) loading of additive tions; however, it can be shown that Corrosion Res. Conf., NACE, At- 2) solubility of additive in food (simu- the two major ones are caused by fluc- lanta (1979). lating solvent) tuations of the electrode resistance 3) equilibrium partioning of the addi- from the breaking of the protective film tive between food and package and the fluctuations of the double layer 4) diffusion. capacitance caused either by changes

in surface area or in the thickness of Non-Diffusion Limited Migration

the insulating film. At the moment, it The maximum permissible amount

is not possible to decide whether the of migration of a particular additive is

94 Migration Decision Tree

plastic (mass Given: Packaging M p ) containing M 0 gram of additive.

Compare: Migration of additive into a food

simulating solvent of mass M s to the maximum permissible dose or concentration C*:

C* = M*/M s

1 M < M* M M* 0 I 0 > i 1 Acceptable Migration Determine Solubility (S) of (load limited) additive in simulating solvent

S < C I S > C

Determine: Is migration limited by partition- Acceptable Migration ing of additive between solvent and plastic 7 (solubility limited) Residual concentration of additive in plastic must be greater than = - M*)/M C (M0 p Since the partition coefficient K is greater

than S, the ratio C*/C is compared with S

S < C*/C I s > c*/c

Estimate or determine: InK =ln S+X°° + 1 Additive polymer interac- 00 tion parameter X I Equilibrium extraction.

M 1 + /KM M p S

M M' I oo 1 Estimate: Maximum migration for Estimate or determine: Diffisuion Acceptable Migration a package thickness, £,,and a coefficient (D) of additive in (partition limited) shelf Jife, t: polymeric film M = 2 Vz 4M (DI/tt£ ) t 0

I M, > M* £ 1 Acceptable Migration Unacceptable Migration (diffusion limited) NATIONAL MEASUREMENT LABORATORY

determined from its toxicology. This this example the ratio of solvent to density polyethylene to various etha-

amount is presumed known and is usu- polymer is about 100.) nol/PFO mixtures. ally expressed as a maximum allowable At equilibrium, the ratio of additive Direct experimental determinations

concentration of the additive in the concentration in the solvent to its con- of K or x 00 are difficult and time con- food (C*). centration in the plastic is called the suming. However, K can be easily esti-

It is obvious that any level of migra- partition coefficient, K. We have shown mated by eq. ( 1 ) from a known solu- x tion is acceptable if the initial loading [1] that if the additive has only limited bility and a theoretically estimated x - of additive to the plastic package is solubility in the food simulating sol- Theoretical methods of estimating x x

small enough so even if all the addi- vent, then the relationship between the are available [2], but even this calcula-

tive migrates to the food, its concentra- partition coefficient and the solubility tion can be avoided by assuming that

00 tion level will never exceed C*. This is given by: X has a worst case value of 2 (or

is the first branch of the MDT. larger) . Using the definition of the In K = In S + x ^ + 1 (1) It is also obvious that if the solubility partition coefficient, the amount of ad-

(S) of the additive in the food simu- where x°° is a parameter (a reduced ditive that will migrate at infinite time lating solvent is limited and S

acceptable limits. This is the second finite dilution of the additive with the Equation ( 1 ) is based on the as- branch of the MDT. plastic packaging material. The param- sumption that the food simulating sol-

If the migration is neither load nor eter is also related to the volume vent is not absorbed by the plastic

solubility limited, then it may be lim- fraction activity coefficient (y°°) at package. If the solvent is absorbed, ited by partitioning. At very long infinite dilution by [2] then an appropriate correction can be

times, the additive will equilibrate or made [2]. Fortunately, in most cases 00 In y°° = x + 1 (2) partition itself between the food simu- where partitioning is an important con- lating solvent and plastic package. A Combining eqs. (1) and (2), we ob- sideration, solvent absorption is not.

residual amount of additive is always tain the simple result For example, polyolefins do not ap- retained by the package. An example preciably absorb water or alcohol and K/S (3) of partitioning is shown in figure 1 for it is usually aqueous and alcoholic food

the migration of radiolabeled (carbon- Experimental values of x 00 are con- stuffs for which partitioning is im- 14) n-octadecane from low density veniently measured by inverse gas portant.

polyethylene to a 50/50 ethanol/water chromatography [3]. For non-polar Diffusion Limited Migration mixture. Notice that more than 80 per- additives and non-polar polymers, x°° If migration is not load, solubility, cent of the original amount of additive is invariably positive and usually falls or partition limited and the maximum is retained by the polyethylene. (In in the range: 101 0 M 2 100 < x <

Positive x°° values are indicative of an unfavorable thermodynamic interaction 10-1 It between additive and polymer. Values of x 00 larger than 2 can occur when:

10-2 1 ) the additive/polymer combination

is polar/non-polar and/or

2) the additive is large in size.

10-2 10-1 100 101 1Q2 1Q3 1Q4 Positive x 00 values imply through eq. (1) that 70 80 90 100 Time, hours K > S. Ethanol Wt. % Figure 1. Migration of n-CVH, s from LDPE Figure 2. Partition coefficients of n-GsHas into 50/50 ethanol/water mixtures at 60 °C. In figure 2 measured values of K and between PDPE and ethanol/water mixtures S are compared for the migration of and solubilities of n-GsHss in ethanol/water radiolabeled n-octadecane from high at 60 °C.

96 .

CENTER FOR MATERIALS SCIENCE

-helf-life of the packaged food is much cient in low density PE that is about .horter than the time required to 30 times larger in heptane than in tchieve equilibrium partitioning, migra- ethanol. tion may be diffusion limited. For Use of eq. (4) for migration assumes simple Fickian diffusion from a plastic that diffusion in the plastic package is controlling. This is not always the ;heet of thickness / to an infinite reser- rate voir of solvent, the amount of migration case. Our experiments have shown that solubility additive is it time t can be estimated from the when the of the

ollowing equation [4]: very low in the food simulating solvent 10° 101 (K<<1), migration often follows a t 2 )'- t /M 0 = 4(Dt/7r/ (4) M Time, hours rather than a VA law. This behavior

can be seen in figure 1 where the initial Ahere D is the diffusion coefficient of Figure 3. Migration of n-C-Hes from LDPE curve is approxi- he additive in the plastic sheet. For slope of the migration into n-heptane and ethanol at 60 °C. Dashed of low is linear in mately unity. Under conditions -ickian diffusion, M t /M 0 lines have a slope of one half. A V2. solubility and solvent stirring, diffusion 'i for M0 /M t values up to about additive the relatively For longer times eq. (4) overestimates of the through layer surrounding :he amount of migration. As shown in stagnant boundary rate he fourth and final branch of the the polymer sheet can become controlling. have shown that a linear VlDT, if the maximum shelf-life of We

t law is consistent with boundary layer he packaged food is known, then the naximum amount of diffusion limited limited diffusion. We are currently nigration can be calculated. (For a working on a more general theoretical model that will bridge the gap between inite solvent reservoir, M0 is replaced polymer limited and boundary layer oy M*> in eq. (4) ) 480 720 960 1200 limited migration. Most of our migration experiments t1/2, s1/2 :rom polyethylene (PE) have exhibited ;mall departures from ideal Fickian Figure 4. Migration of labeled n-GsHse from HDPE into unlabeled n-CisHM at 30 °C. References oehavior. The most common deviation s illustrated in figure 3 which shows [1] Smith, L. E., Sanchez, I. C. Chang, :he migration of the normal alkane are the completely absent because PE S. S., McCrackin, F. L., and Senich, n-C 3 oH 0 c from PE to heptane and to is already fully swollen. This experi- G. A., Models for the Migration of ethanol. The migration curves tend to ment was compared with one in which Additives in Polyolefins, NBSIR-79- be slightly sigmodal in character rather the PE was only partially swollen with 1779. than completely concave to the time labeled C 1S (1 percent by weight). The [2] Sanchez, I. C, Equilibrium Distribu- axis as is characteristic of Fickian dif- results are shown in figure 4. As can tion of a Minor Constituent between fusion. This positive deviation, which be clearly seen, simple Fickian kinetics a Polymer and Its Environment, ACS is an unexpected acceleration of migra- are observed for the fully swollen Symposium Series Vol. 95, R. K. Eby, tion, is thought to be caused by the sample whereas a positive deviation is ed. (American Chemical Society, absorption swelling action of the and observed for the partially swollen Washington, DC, 1979). food simulating solvent. To test this sample. conjecture, migration experiments were Although solvent absorption would [3] Senich, G. A. and Sanchez, I. C, performed on high density PE contain- appear to cause migration to deviate ACS Organic Coatings and Plastics Chemistry, Vol. 345-349. ing radiolabeled n-octadecane in con- from ideal Fickian behavior, this de- 41% (Ameri- can Chemical Society, Washington, tact with unlabeled n-octadecane as viation is relatively unimportant com- DC, 1979). solvent. In one experiment the PE was pared to the effect that solvent absorp- saturated with the labeled C 1S (about tion has on the absolute value of the [4] Crank, J. and Park, G. S., Diffusion

5 percent by weight) . In this kind of diffusion coefficient. In figure 3 n-C32 in Polymers, Capt. 1. (Academic experiment solvent absorption effects H G6 has an apparent diffusion coeffi- Press, New York, NY, 1968).

pffice of the Deputy Director or Programs

C. Gravatt, Jr. Robert C. Thompson s even program offices have been established within Howard J. White, Jr. ndestructive NML to focus diverse technical activities on specific | Measurements for iluation Nuclear Technology problems and goals. These offices fund, coordinate, and 'old Berger H. Thomas Yolken direct efforts in all NML and several NEL (National es B. Beal** Recycled Materials rge Birnbaum Harvey Yakowitz Engineering Laboratory) technical centers. Thus, they nard Mordfin Donald A. Becker bert G. Sutter draw on expertise in many scientific disciplines to meet i Joseph G. Berke

I . ironmental national needs for measurement services and data. The asurements Directorate for I Measurement Services offices can be readily separated into two distinct cate- I Ham H. Kirchhoff Arthur O. McCoubrey tie T. McClendon gories: 1 ) those based on traditional outputs fundamental I Weights and les R. McNesby [ Measures to the NBS mission and related to general, ongoing Liter L. Albert D. Tholen £ielinski, Jr. national needs; and 2) those that deal with topical areas ndard Reference Carroll S. I Brickenkamp and direct NBS expertise in measurement science toward terials Stephen Hasko \orge A. Uriano solving specific current problems. Joe H. Kim

: bert L. Alvarez Henry V. Into the first category fall offices managing the tradi- semond K. Bell** Oppermann ^aul Cali Richard N. Smith tional NBS outputs: Measurement Services, Standard 'Keith Kirby O'to K. Warnlof Reference Materials, and Standard Reference Data. .onora T. Klein Harold F. Wollin bert E. Measurement They provide the means for tracing measurements to the vlichaelis. Jr. Services nley D. Rasberry national reference standards. Further, they provide Brian C. Belanger llliam P. Reed quality assurance for the application of NBS measure- Shhard W. Seward Kenneth W. Edinger Ines I. Shultz** Lee J. Kieffer ment technology in the field. Dating back to the founding Kathryn O. Leedy Jack L. Vogt of NBS, these programs have promoted the growth of 5 ndard Reference Domestic and organizations which improve measurement, data, and \vid R. hide, Jr. International standardization services throughout the United States |:rman P. Measurement ivozinsky Standards and abroad. ms H. Gevantman David E. Edgerly rithia A. Goldman Helmut M. The Measurement Service activities—weights and rla G. Messina Altschuler measures, calibration coordination, and coordination of tijoyce B. Molino Samuel E. Chappell rtrude B. * 'Research Associates voluntary standards—are specifically concerned with the Sherwood and Guest Workers practical application of fundamental standards in the real world of measurement. The weights and measures effort, which predates NBS, provides the fundamental ward White is calibrating the high services in olution organic mass spectrometer, standards for the exchange of goods and the is instrument is capable of achieving marketplace. Although NBS does not have statutory tic resolutions of better than 70.000 authority to regulate weights and measures within a State, AM). The gas chromatographic it does provide State weights and measures offices with et feature of this instrument permits titifying individual organic compounds the necessary technical support to insure fair and equita- shale complex mixtures such as in oil ble measurements throughout the Nation. It also works t e elemental compositions of the to assure that State weights and measurement activities •lecules and fragment ions are obtained, e field desorption capability of this are traceable to national standards and are capable of iss spectrometer is used in the analysis dealing with the complex measurement problems of the peptides with molecular weights of modern marketplace. In the area of calibration co- 'ater than 1,000. The clinically im-

rtant peptide, angiotension I, is being ordination, Measurement Services provide leadership and ihiated as a potential Standard services to laboratories in industry, State and local ference Material. This instrument has governments, and other Federal agencies through NBS id applications in studying samples of erest to several of NML's programs. calibration and measurement assurance programs. A new

99 ;

NATIONAL MEASUREMENT LABORATORY

program has been established to improve the manage- data evaluation efforts in high resolution infrared spectra. ment and use of precision test equipment in the Federal A highly successful vehicle for disseminating reference Government. The Measurement Services staff also co- data has been the Journal of Chemical and Physical^ ordinate the NML participation in domestic voluntary Reference Data. standards and plays a major role in managing key inter- The "topical" program offices of NML presently focus j national standards responsibilities. This office is specifi- the scientific and technical resources of NBS on national : cally charged with coordinating NBS participation in problems in the recycling of waste materials, nuclear the International Organization for Legal Metrology, the technology, nondestructive testing, and the environment.

Treaty of the Metre organizations, the International The Office of Recycled Materials is presently engaged

Organization for Standardization, and the International in activities related to urban solid waste and reycled oil. Electrotechnical Commission and in preparing policy Due primarily to environmental considerations, the prob- positions for NBS delegates to these organizations. lem of the disposal of urban waste is becoming acute for The Office of Standard Reference Materials produces many cities across the Nation. The best solution to the and certifies materials standards for the quality assurance disposal problem is to use waste as a resource, both for of most of the physical and chemical measurements of the raw materials and for fuel. To be a resource standards United States—and the world. This program presently are needed. Standards, however, are particularly difficult provides more than a thousand different Standard Refer- to set for a substance as heterogeneous as urban waste. ence Materials (SRM's), ranging from over 100 different A current major focus for this office's Resource Recovery steel reference materials, through such biomatrix SRM's Program is the development of standards for refuse- as trace elements in spinach and bovine liver, to such derived fuel. physical SRM's as Mossbauer Spectroscopy standards In the Office of Measurements for Nuclear Technology and superconductive fixed point devices. The range of current activities focuses on the need for accountability jii

SRM's provided serves traditional measurement activity throughout the nuclear fuel cycle. This program is con-i|iio

as well as measurement needs related to new technologies. cerned with all aspects of the handling and measurements jlif

Considerable effort is now going into the production of of nuclear fuels, from the determination of volume and

new environmental and clinical SRM's— a reflection of flow of nuclear material through the determination of the- : national concerns in these areas. Presently, the Environ- precise isotopic composition of nuclear fuels by nonde-

mental Protection Agency requires that the calibration structive evaluation, to the measurement of more accu-jilli standards used in atmospheric measurements be traceable rate values of the life-times of radionuclides. Future to NBS SRM's, whenever possible. plans include standards for nuclear waste disposal.

has: j The third traditional output program is the Office of The Office of Nondestructive Evaluation (NDE) of the reliability and Standard Reference Data (OSRD) . The need for coordi- as its objective the improvement nation and support at NBS of an international effort to durability of materials and structures by means, of im-#i compile and evaluate scientific and technical data has proved NDE procedures and standards. The program's

been specifically recognized by Congress in the Standard role is to help industry develop and bring to use methods Reference Data Act of 1968. One-third of the Federal for accurate and reproducible NDE measurements. This

data collection effort is funded through OSRD, and includes technical investigations of standards (both phys-

another third is directly managed or coordinated through ical calibration standards and procedural documents), \% development of im- this office. In addition, OSRD and its constituent data characterization of instruments, p centers in this country cooperate with other data centers proved techniques and the assessment of the meaning of 1 throughout the world to avoid duplication of effort and the NDE on material performance.

1 to aid the promulgation of standard reference data. Data The Office of Environmental Measurements manages ; efforts in NML range from the large, ongoing data centers the oldest topical program in NML, the air pollution it in chemical kinetics and thermodynamics to shorter term program. It has evolved since 1972 from a single effort

100 E OF DEPUTY DIRECTOR FOR PROGRAMS

Particulates

Safeguards

^ Water Resource + Recovery

Climate

Toxic Substances

research & maintenance seed growth development

Time

,he 1. Schematic development of nationally focussed NML programs.

the problems of standardization in homogeneous air

llution, to its present major thrust in the development standards for water pollution and particulates. Areas i intense effort involve the measurement and character- ation of environmental asbestos and the development of :lM's for trace organic compounds in water. A recent roject resulted in a significant revision in the mechanism

I ozone-alkene reactions used in photochemical smog :3dels.

5 The topical programs, unlike the traditional "output" rograms, are planned for limited duration. Figure 1 .es an idealized time development chart for topical Lograms and shows the approximate status of various jograms. The time axis can be quite different for the

|fferent programs. For example, Recycled Oil is ex- cted to have a lifetime of 5 to 10 years, while NDE will •obably last considerably longer. Nevertheless, the chart strates the philosophy of the topical programs as they ply NBS expertise to national problems. When the

BS contribution is complete, the program is terminated, le ongoing effort represents the transfer of the products the topical program to output programs and, possibly, me ongoing activity carried out within a technical nter. The types of products transferred to output ograms would include SRM's, data compilations, and libration services.

101 NATIONAL MEASUREMENT LABORATORY

Load Cell Mass Comparator ill ;

Randall M. Schoonover, Length and Mass Measurements and Standards Division for Office of Weights and Measures i a

TJL he calibration of mass standards vent the release of more than 10 percent stop runs against the bed plate. This upwards from 25 kg in value is pres- of the spring tension. stop is adjusted to provide about 90 ently performed on mechanical balances Figure 1 shows the device in detail. percent of full load to the cell when used as mass comparators. The slow The lifting eye carries the rigid frame the mass load is removed. i J balance response time and the handling and is itself suspended from a building Universal flexure joints provide for precautions required to protect the structural member, "A" frame, etc., a repeatable load axis. Without these, knife-edge bearings make this a time through a hydraulic cylinder. the minor mechanical misalignment • if consuming process. In addition, a large A coil spring, with a predetermined that occurs during the load/unload amount of laboratory space must be spring constant and length, rests on the cycle would degrade cell performance. dedicated to a permanent installation. bed plate centered about a passageway. In addition to the components shown i HI Economies of time and laboratory On the upper end of the spring rests a in the schematic drawing, a hydraulic space would be effected if the mass floating plate with its passageway simi- lifting cylinder is required, as is a comparisons could be accomplished larly centered. Horizontal motion of weight transport system. Dollies and using a strain gauge load cell, provided the floating plate is restricted by verti- tracks provide excellent weight han- i that its precision could be improved at cal rods that are part of the frame. dling (up to 1000 kg) for the existing least tenfold over its inherent repeat- The rods, however, do not impede device. Also, an electronic device is ability. I have constructed a simple vertical motion of the spring and plate. required to indicate load cell output comparator based on the use of a load Resting on top of the floating plate signal. However, a single such device cell, which demonstrates that the ten- are a thrust bearing and ball-and-socket will suffice for loads from 25 kg to f fold improvement is readily attainable. assembly. The ball-and-socket permits 25 000 kg.

crude vertical alignment of the cell I have constructed a realization of * during initial assembly. The bearing this device at the 225-kg (500-lb) ft allows the spring to rotate during com- level [1]. Using the device, calibration tin The cell pression and extension and also re- of 500-lb State mass standards was problem in using a load to Ell moves torque from the cell during the accomplished to a precision of 0.0002 compare a mass standard with an un- St loading cycle. percent. In addition, the calibration known object of the same nominal pi

is Adjustable shunt stops are set to effort required only a few hours as value that hysteresis in the cell and i|i loading errors are too large interdict vertical motion of the floating compared to a few days for the same to allow ! ilil one to exchange loads in the most plate just before full loading is reached, measurements using an equal-arm bal- direct way, i.e., completely unloading thus providing a rigid suspension of the ance. the cell and reloading with a different cell. When the load is removed, slight I believe that the same approach mass. Use of a load cell in this manner vertical motion of the spring and cell will be applicable for load cells operat- is limited to a precision of 0.005 per- is permitted before the constant load ing in the range between 25 kg and cent. However, the hysteresis and load- 5 000 kg—and possibly higher. In paring errors would be greatly reduced if ticular, at the 25 kg level such a device most of the -Lifting Eye operating at a precision of 0.001 per- tension on the load cell Ball & Socket^ could be maintained during the inter- Frame cent or better would be an important contribution to industrial metrology. change of loads. I was able to do this Thrust Bearing in a very straightforward way, as Floating Reference described below. Shunt Stop Plate [1] Schoonover, R. M., A high precision About 90 percent of the tension on Flexure load cell mass comparator, /. Res. the load cell is maintained during Universal the 347 ' NBS, 84, '(1979). load interchanges by use of a coil Load Cell-- - Spring spring. Essentially, this is accomplished by suspending the cell coaxially Bed Plate inside the coil spring, allowing the load Constant Load to compress the spring, and then adjust- Stop Figure 1. Schematic of load cell com- Point ing the surrounding rigid frame to pre- OVLoad parative.

102 OFFICE OF DEPUTY DIRECTOR FOR PROGRAMS

Interagency Cooperation in Mass Spectral Data Base Dissemination

Lewis H. Gevantman, Office of Standard Reference Data

: F^M^ or over three years the Office of scientific and technical audience which the molecular formula which is ordered Standard Reference Data (OSRD) has is involved in mass spectral measure- by the Hill convention [4] and derived >een involved in a joint data dissemi- ment but does not require the degree of by computer from the CAS-supplied tation project with the Environmental use implied in an automated system. connection table record. The molecular Protection Agency (EPA) and the Each entry in the tape file contains weight is derived by computer from the National Institutes of Health (NIH). a spectrum, quality index, compound same records using the integral atomic

This project deals with the collection, name, molecular weight and formula, masses C = 12, H = 1, O = 16, etc. evaluation, and dissemination of a and the Chemical Abstracts Service The structure diagrams, a unique fea-

nass spectral data file containing over (CAS) Registry Number. The entire ture of the book, are also computer : H,000 chemical compounds. The data base is written on a single 2400- derived from the connection table •ompounds profile relate largely to foot reel of magnetic tape in IBM record for each substance. wealth and environmental concerns, unformatted Fortran(G) and is avail- The "Quality Index" (01) employed out the file is sufficiently diverse to able in 9-track, 800 or 1600 bpi. Up- in selecting the data is based upon a nclude information on chemical sub- dates of data are provided on a regular set of criteria established by F. W. stances of interest to other commercial (annual) basis. McLafferty [5]. Originally the overall

ind academic areas. The book form is published as four index was made up of seven individual volumes plus an index volume. The factors dealing with such items as Prior to OSRD involvement, the spectra are indexed in four ways: by evaluation of the spectrum by an ex-

IH EPA Mass Spectral Data file was chemical substance name and syno- perienced mass spectroscopist, ioniza- accessible to users via an on-line com- nyms, formula, molecular weight, and tion conditions, impurity peaks ob- puter system as a commercially avail- CAS Registry Number. Supplements served, illogical neutral losses, isotopic able service [1]. The entry of the OSRD are being produced periodically, and a abundance, number of peaks, and the

: nto this project has broadened the complete revised set of indexes is pub- comprehensiveness of the data. Addi- 'nethods for disseminating the data and lished as each supplement appears. tional factors such as compound purity olaced a greater emphasis upon the Every entry is formatted as shown and instrument calibration procedures quality of the spectra, both contained in figure 1. Each spectrum is accom- have been added to the OI to select the

-n and being added to the file. The panied by the CAS Registry Number, spectra in the book and tape.

'iltimate intent is for the activity to -provide a depository of the highest Molecular Weight Molecular Formula CAS Registry Number juality spectra and to make the high Quality data available in a number of

' ormats so as to achieve the widest dissemination tailored to user require- 173 C H 86-57-7 10 7 NO 2 Substance Name - Naphthalene, 1-nitro- ments. Finally, it seeks to influence NO, 'Jhe community of mass spectral data 'isers and producers to continue to im- Structural Formula. prove the quality of mass spectral data eported 100 in the open literature. 80 The OSRD, in collaboration with 60 40 he EPA and the NIH, has made the * 20

- .J,IU ,l„ll ,.J., , llj 0 , lata base available in two forms: mag- .r n f '35 10 30 50 70 90 110 130 150 netic tape and book. The magnetic § 100 Mass Spectrum c 80 ape is available under a lease agree- ~ 60 nent [2] and is intended for use by 40 20 organizations who engage in a large 0 limber of spectral analyses in the 170 190 210 230 250 270 290 :ourse of their normal operation. The Mass/Charge iata book [3] is intended to reach the Figure 1. Typical mass spectrum.

103 NATIONAL MEASUREMENT LABORATORY

Standard Reference Material For Antiepilepsy Drugs in Serum

Robert Schaffer and Dennis J. Reeder, Organic Analytical Research Division

Further efforts to refine and extend tral Data Centre (MSDC), now based PJL hysicianshV! have depended on the the use of quality criteria have been at the University of Nottingham. Origi- laboratory testing of blood from epi- pursued in the OSRD. Several technical nally this group, based at Aldermaston, lepsy patients to establish that their contributed large discussions have been held under the a portion of the serum levels of anticonvulsant drugs auspices of the Joint Committee for spectra contained in the file. After are within the medically beneficial con- Atomic and Molecular Physical Data some hiatus, this group is again active centration ranges. Patients might other- seeks collect, (JCAMP) concerning the identification and to evaluate, and sub- wise suffer seizures or toxic effects, and promotion of measurement speci- mit for inclusion in the data file the depending on whether the drug levels fications which would lead to the better spectra now found in the litera- are below or above the therapeutic improvement of mass spectra in the ture. The Centre also plans to solicit ranges. A large-scale comparison of open literature. A preliminary manu- spectra from industrial, academic, and laboratory results from testing identical script based upon the McLafferty cri- miscellaneous sources for ultimate in- serum samples containing the com- teria but extending and clarifying the sertion into the data file, provided the monly prescribed drugs revealed unac- required steps is now being prepared. data meet the quality requirements ceptably large differences among labo- Further refinement of the discussed above. It is believed that the paper based ratories and methods [1], For each of critiques on by acknowledged experts renewed involvement of the MSDC the drugs tested for, the laboratories will should lead to a final document which is lead to the achievement of im- reported results ranging from not pres- planned for proved publication in 1980. Ac- spectra and a system which ent to present in toxic amounts. ceptance of these criteria by the Inter- conforms fully to the OSRD goals of help overcome this problem, national Union for Pure and Applied improving the dissemination of scien- To NBS was asked to provide a Standard Chemistry (IUPAC) is also being con- tifically reliable data. templated. Reference Material (SRM) consisting of three serum samples, each to contain In employing the two new mecha- four antiepilepsy drugs at different nisms for disseminating the mass spectral concentration levels. The SRM would data the OSRD has attempted to effect Reference fill a basic role for the achievement of the widest distribution possible. The [1] Heller, S. R., Milne, G. W. A., and accurate analysis, and help ensure the leasing of the tapes has resulted in a Feldman, R. J., A computer-based reliability of analyses for these drugs. number of arrangements, some related chemical information system, Science, Table 1 lists the drugs and actual certi- to primary (in-house) use of the tape 195, 253-259 (1977). fied concentrations which typify medi- and others to more extended use. The Heller, S. R., and Milne, G. W. A., cally designated levels. The SRM was latter has involved lease by large cor- The NIHI EPA Chemical Information System, M. M. Milne and W. J. Howe, to be supplied as freeze-dried material porations for distribution within their eds., ACS Symposium No. 84, Chapt. which would be reconstituted by the subsidiaries, lease by data services for 10. (American Chemical Society, user by adding an exactly specified dissemination via on-line service, lease Washington, D.C., 1978). volume of water. by foreign nations for distribution of [2] For further information regarding the the data within the specified borders of leasing of this magnetic tape, please In order to certify an antiepilepsy the country, and finally lease by instru- contact: Office of Standard Reference drug SRM, a series of problems had ment manufacturers for incorporation Data. National Bureau of Standards, to be considered: (1) analytical criteria of the data in an automated search file A323 Physics Building, Washington, and specifications had to be developed associated D.C., with experimental mass spec- 20234. for selecting a special supply of each trometer [3] Heller, S. R., and Milne G. W. A., systems. In this way, a wide of the four drugs. These supplies would and divergent group EPA/NIH Mass Spectral Data Base, of users has been be employed in preparing the SRM NSRDS-NBS 63, 5 volumes, 4685 pp. embraced. and, of greater importance, they would 1978. PB290661 NTIS, Springfield). final step A taken by the OSRD to serve as the primary standards for the [4] Hill, E. A., /. Amer. Chem. Soc, 22, improve the data quality and guarantee 478-494 (1900). certifying analyses that would be per- the acquisition of mass spectral data [5] Speck, D. D., Venkataraghavan, R., formed; (2) the serum used as the is a more recent agreement reached and McLafferty, F. W., J. Org. Mass matrix for the SRM had to be com- with the United Kingdom Mass Spec- Spectrometry 13, 209-213 (1978). patible with the many methods in cur- 104 )=FICE OF DEPUTY DIRECTOR FOR PROGRAMS

Table 1. Concentration Level, jug/mL is illustrated in figure 1. All of the serum as detected by reversed-phase LC Drug Toxic Therapeutic Subtherapeutic Blank then divided. A portion would serve as the serum blank and the remainder, Phenytoin in three separate portions, as the matrix Diphenylhydantoin) 60.7 ± 0.9 a 16.7 ±0.3 4.2 ± 0.1 for the drugs. Appropriate quantities Ethosuximide 174.7 ± 0.6 75.9 ±0.5 11.8 ± 0.4 of the four drugs were added to the Phenobarbital 103.6 ± 0.3 21.6 ± 0.2 5.3 ±0.2 three portions. The four pools were 'Primidone 18.6 ±0.7 8.1 ±0.2 3.6 ±0.1 then dispensed into vials with a mech- anized pipettor which was precali- The uncertainties represent one standard error for the above certified values. The imprecisions observed I ^oth within and between the liquid and gas chromatographic anlyses used for this certificate are included brated for accurate delivery. The vials ~n the standard errors. of serum were freeze-dried in a pro- longed single operation to minimize any irregularity in residual moisture level,

-ent use and. to the degree possible, source, for example, might exhibit and then, when under nitrogen, they vith prospective methods. This re- cross-reacting antibodies to bovine com- were stoppered. quired serum that would be free from ponents in radioimmunoassay reagents.) A random selection of vials contain-

:hemical and biological (e.g., imrauno- The large volume of serum needed ing blank serum and three levels of .ogically active) interferences; and required the pooling of sera. Since drug-spiked serum were reconstituted :3) two, independent, highly accurate serum from different donors could be with weighed quantities of water. The

: analytical methods had to be developed expected to contain different interfer- homogeneity of each kind was deter- i or determining the drug levels in the ing substances due to individual dietary mined by differential refractometry. SRM preparations, because NBS cer- habits and drug utilization, we em- With the refractive index of serum iification would not be based on the ployed several treatments to obtain an about 0.01 greater than that of water, ;veighed-in concentrations. The two interference-free serum. First, hepatitis- detection of variations of 0.1 percent methods would have to yield sufficiently free plasma from several donors was required refractive index measurements

accurate and precise data that the clotted by adding calcium chloride, and better than 0.000 0 1 . This was accom- ssignment of values could be unam- the serum obtained by centrifuging. plished by comparing the serum samples biguous. This required use of methods The lipoproteins were removed and the to a reference aqueous solution con- or achieving accuracy and precision serum was heated to deactivate some taining about 100 g/L of potassium lot ordinarily utilized in clinical labo- of the enzymes. The serum was again chloride, using a differential refractom-

; atory situations. centrifuged and then dialyzed to remove eter with a sensitivity about 0.000 001. For selecting supplies of each of the electrolytes. After re-equilibriation with The vial-to-vial variations in the recon- hugs, a variety of chromatographic 0.15 mol/L sodium chloride at pH 6.5 stituted sera were less than the variance md spectroscopic techniques were ap- to 7.0. the serum was passed through found by chemical analysis. plied to samples whose properties gen- a bed of charcoal and a 0.22-^m pore- Quantitative analysis of drugs in

erally conformed with those given in size membrane filter into a sterile serum was performed by reversed- he literature for the drugs. The purity container. phase LC, with a fixed-wavelength bf the selected supply of each drug was We tested the processed serum by detector at 254 nm and, for primidone, ietermined by liquid chromatography observing the clarity of the product a variable-wavelength detector set at LC), the compositions of fractions after portions were freeze-dried and 210 nm. Peak areas were determined )eing characterized by supplemental reconstituted. Unprocessed and proc- by use of an electronic integrator. :hromatography and spectroscopy. essed serum was also tested by reversed- Standard samples containing each of The specifications we set for com- phase LC and gas chromatography (as the four drugs in 20 percent methanol- patibility of the serum matrix with a described below) for evidence of inter- 80 percent water were run alternately /ariety of antiepilepsy drug assay fering substances. The difference be- with serum samples. Calibration data nethods precluded the use of any but tween the original and the processed were thus obtained concurrently with Timan serum. (Serum from a bovine serum as detected by reversed-phase LC the sample analyses and were applied

105 NATIONAL MEASUREMENT LABORATORY

ethosuximide

internal standard (i.s.)

phenobarbital

t -» Human Serum t -» Processed human serum matrix diphenylhydanto Figure 1. Reversed-phase liquid chromat- ograms of unprocessed, pooled human i.s. serum (left panel) and processed serum matrix (right panel).

2 3 4 5 6 7 3 to give the quantitative results. Figure 2 Time, minutes illustrates a typical chromatogram, re- Figure 3. Gas chromatogram of antiepi- corded at 254 nm. lepsy drugs. _L J_ Our gas chromatographic results 10 20 30 40 were obtained with a special column packing, GP 2 percent SP-2510 DA Time, minutes 15 mL conical centrifuge tube and evap- on 100/120 Supelcoport, to separate orated at 20 °C under vacuum. The Figure 2. Reversed-phase liquid chromat- the anticonvulsant drugs in their unde- ogram of antiepilepsy drugs in processed concentrate was dissolved in 50 yuL of rivatized forms. Flame ionization detec- human matrix (detection at 254 nm). chloroform and vortex-mixed. One mi- tion was used, and the output was croliter of this solution was then in- interfaced with a computing integrator. the SRM sample. Then 1 .0 mL volumes jected into the gas chromatograph. The compounds used as internal stand- of the serum blank was pipetted into Figure 3 illustrates the separations ards in the analyses for the four drugs the tubes containing the mixtures of achieved by gas chromatography. are listed in table 2. Uniform quanti- primary and internal standards. Usu- Standard reference material (SRM) ties of of the internal standards were com- ally four SRM samples and two the 900 is now being used in clinical bined with 1 mL of the reconstituted bracketing standard mixtures were laboratories throughout the nation. A SRM sample and, in separate contain- analyzed as a group, as follows: new comparison of laboratory results ers, 1 combined with the primary stand- Each was acidified to pH to 2 and is now in progress, and preliminary ard drugs of high- taken in quantities about then extracted with lOmL findings indicate that this SRM is hav- 2 percent lower chloroform. After centrifugation, and about 2 percent purity ing a dramatic effect on the quality of . higher than the expected (from the the aqueous layer was discarded and the clinical tests for antiepilepsy drugs weighed-in) quantities of the drugs in the chloroform was decanted into a in human serum.

Table 2. Drugs and internal standards used in GC analysis Reference [1] Pippinger, K. F., Penry, J. K., White, B. G., Daly D. D., and Buddington, Drug Internal standard R., Interlaboratory Variability in De- termination of Plasma Antiepileptic Phenytoin 4-(p-Methylphenyl)-5-phenylhydantoin Drug Concentrations, Arch. Neurol. (Diphenylhydantoin) 33, 351-355 (1976). Ethosuximide a-Methyl-«-propylsuccinimide Phenobarbital 5-Ethyl-5-(p-methylphenyl)-barbituric acid Primidone 4-Methylprimidone

106 Development of Test Procedures to Determine Calorific Value of Refuse and Refuse-Derived Fuels

Eugene S. Domalski, Chemical Ther- modynamics Division and Joseph G. Berke, Resource Recovery Program

V«X ver the past o two years, we ministrators, and private owners could greater impact than any other material have been working on a joint Depart- evaluate with confidence whether their property on market development be- ment of Energy/Environmental Protec- large incinerators or refuse-fired boilers cause burning will be an important :ion Agency project to develop infor- are in compliance with contract per- mode of reclaiming energy from proc-

I nation for a product specification formance specifications. Such a test essed solid waste for at least the next m refuse-derived-fuels (RDF). More procedure can be validated from meas- decade. Thus, a test procedure for

! -ecently, we have been working with urements carried out in a multi-kilo- determining the calorific value will :he Office of Recycled Materials to gram capacity flow calorimeter on large better equip commercial laboratories to ;ontinue the RDF work in direct re- samples of refuse which have under- certify accurately the energy content sponse to the mandate of the Re- gone little or no processing. Both the of RDF of various compositions. In source Recovery and Conservation Act business and industrial communities addition, the determination of the calo- "RRCA) PL94-580. have expressed the need for a test rific value of RDF at various stages of The research problems stem from procedure. In addition both American processing will identify any stages :he fact that the calorific value of mu- Society of Mechanical Engineers and which impart a different chemical char- nicipal solid waste (MSW) or refuse- American Society for Testing and Ma- acter to the solid waste. derived-fuels (RDF) is currently deter- terials have also lent support to the Since many of the current incinera- nined in a bomb calorimeter by using NBS program. tors for waste-to-energy systems burn •gram-size samples which have been unprocessed garbage as received, both NBS Activities subjected to a high degree of processing. ASME and ASTM expressed the need The processing includes grinding, shred- During the initial phases of the pro- to determine the calorific value of the ding, milling, riffling, and pelletizing of gram at NBS, the tests were carried unprocessed garbage. To do so requires he waste sample. Engineers who are out in a 2.5 gram bomb calorimeter. using large samples of garbage to en- 'experts in incinerator technology and However, due to the concerns of proc- sure some sort of representative "grab" solid waste processing believe that after essing the material cited above, a larger sample. NBS, in cooperation with

;uch treatment the waste sample no sample size calorimeter was required. DOE, is designing and plans to build longer resembles or represents the orig- NBS met the challenge by building a and operate a multi-kilogram sample inal waste. They claim that significant 25-gram sample size bomb calorimeter. size flow calorimeter to handle unpro- :hemical changes take place as a result Processing of samples, although still cessed or minimally processed refuse. Df the size reduction (from gross sizes required, was less severe than for the To date, pilot units of the flow down to below one millimeter) which 2.5-gram unit; in addition, samples of system using 25-gram samples have is necessary in the preparation of a as received-pellets of commercially been built and tested. Currently, tests lomogeneous waste sample for conven- produced RDF could also be com- are being conducted to determine ional calorific value determination. busted. The specifics of this work are proper operating parameters of the Similarly, such engineers have stated available in NBS publications: NBS small units that can be scaled up. :hat the calorific value determined for IR 78-1494 and NBSIR 80-1968. Calibration work with known materials

highly processed waste sample may Much of this work is being coordi- is going on to correlate between the the lave little or no correlation with the nated with the ASTM committee E-38 25-gram bomb and the 25-gram flow original heterogenerous waste. Because on Resource Recovery, which is cur- units. Over the next two years, NBS, pf the enormous heterogeneity of rently working to establish standard test with assistance from DOE will con-

: ~efuse, a suitable procedure has not methods to characterize refuse-derived- struct and operate the multi-kilogram emerged as yet which may be used fuels and enhance their potential for calorimeter and conduct a variety with confidence to select and extract becoming an article of commerce. The of experiments on fuels and various p truly representative sample. A suit- development of markets and sales of fuel mixes as related to resource recov- able test procedure for the calorific refuse-derived-fuels will be linked ery and waste-to-energy systems. The v'alue of solid waste from which a strongly to appropriate characterization data will include heating values, mois- "eliable thermal efficiency could be procedures. In particular, the calorific ture, ash, and elemental analysis. Al- j :alculated would supply a mechanism value, which provides a measure of the though the principal thrust of this

1 1 which engineers, public works ad- energy content of RDF, will have research is on the combustion calorime- L 107 NATIONAL MEASUREMENT LABORATORY

Nuclear Fuel Assay Using Resonance Neutrons

Roald A . Schrack, James W. Behrens, and Charles D. Bowman, Nuclear Radiation Division

try of large samples, a variety of other The Center for Radiation Research rods and to measure nuclear material in problem areas will be examined which is developing a technique that will scrap and waste from a nuclear plant.

have application to the conversion of produce images similar to the familiar The physical phenomenon that is

waste into energy and the processing of x-ray image, that will show the dis- utilized in the technique is the selective municipal solid waste into refuse- tribution of any isotope desired. Using absorption and scattering of neutrons derived-fuels. The problem areas are neutrons generated by the NBS linear by the nucleus. Every isotope has a

analysis and characterization of stack accelerator, this technique—resonance different set of energies for which it effluents, temperature measurement and neutron radiography—provides a interacts with neutrons. These energies materials durability in hostile environ- unique tool for determining both have been carefully cataloged so that ments, sample composition and char- distribution and content of any sample. the observations of a characteristic acterization, combustion and pyrolysis The technique is being developed as a absorption pattern can be used to iden- kinetics, corrosion, and statistical stud- reference standard measurement tify the presence of an element—and ies in waste stream sampling. Various system for the NBS Nuclear in particular the amount and isotope of measurement techniques will be used Technology Program. the element. Previous research at NBS in the research, including calorimetry, has determined to better than 1 percent gas-liquid chromatography, spectropho- accuracy the values of these interaction tometry, infrared spectrometry, and cross-sections for a wide variety of thermometry. This program seeks to provide accu- materials. The values previously deter- rate and timely techniques to keep mined at the Bureau were submitted to track of nuclear materials. The reso- the national clearing house maintained

nance neutron radiography technique is at the Brookhaven National Laboratory used to measure the concentration and on Long Island. We drew on these uniformity of distribution of fresh fuel files in the analysis of our data. Resonance neutron radiography thus Collimator Detector represents a new application and use for information that has already been

accumulated. The procedure used is Source shown in figure 1. Neutrons generated by the NBS linear accelerator are collimated down to a thin fan-shaped beam. A sample of material to be

analyzed is placed in this beam. The number and location of neutrons that pass through the sample are determined with a position-sensitive neutron detector placed behind the sample. This

information is accumulated in a com-

puter as the sample is scanned. When

the sample is completely scanned, the accumulated data are analyzed to produce a picture of the distribution of the desired isotope in the sample. The maging total amount of the isotope in the System sample can then also be determined. The ability to discern details in the Figure 1. Schematic diagram of Resonance Neutron Radiography system. Neutrons are distribution of the desired isotope de- generated in a source by using the electron beam of the NBS linac. The neutrons are then pends on the resolution of the detector collimated into a fan-shaped beam. The sample to be examined is passed through the beam to produce an image in a computer controlled system. system. Commercially available detec-

108 CFFICE OF DEPUTY DIRECTOR FOR PROGRAMS

r'tors have a resolution limi t of about 5 mm; details smaller than this could : not be observed. Fuel pellets used in } nuclear reactors are 10 mm in diame-

Pter, so the available resolution in commercial detectors does not permit a meaningful measurement of nuclear fuel. To overcome this deficiency, a high-resolution detector system was •developed by NBS in collaboration with

7 ('the Oak Ridge National Laboratory .

This detector, shown in figure 2, has

: a resolution of about 1 mm and has I'been used in the analysis of nuclear ffuel.

In addition to its value in the

[ Nuclear Technology Program, resonance neutron radiography promises to

['be of value in other fields, such as Figure 2. High-resolution position sensitive neutron detector. The detector region is the i-nondestructive evaluation. The exami- center small-diameter tube. Special amplifiers are enclosed in the large-diameter cylinders at each end. The valve allows the gas used in the ionization chamber to be changed to nation of manufactured items for hid- determine the effects of different mixtures and pressures. The detector is presently operating rden defects is frequently done with with a mixture of 3.1 X 105 pascals of 3He, 7.9 X 10 3 pascals of xenon and 0.4 pascals of I low-energy neutrons as well as x-rays. 5 carbon dioxide. (10 pascal == 1 atmosphere). Hidden defects in welds can be de- fected before the equipment is used, ['preventing costly equipment failure, an example of this type of use, a ample was made with an intentionally 'defective joint. Figure 3 shows the 'result of a resonance neutron radiography of this sample when the distribution of silver in the brazing material was made. No defect was observable with normal x-ray imaging techniques. The defect was only marginally detectable using low-energy neutron imaging techniques.

Research is continuing on the resonance neutron radiographic technique to improve accuracy and resolution and to reduce the time required to produce

an image. Although the technique is Figure 3. Computer-generated distribution low restricted to laboratories with high- of silver in a defective silver solder joint as energy linear accelerators, methods are determined by a resonance neutron radio- being worked on to allow the technique graphic scan. to be used on a broader basis and thus become more available as a tool for technology.

109 NATIONAL MEASUREMENT LABORATORY

Improved Uses of X-rays for Nondestructive Evaluation

Masao Kuriyama, William J. Boettinger and Harold E. Burdette, Metallurgy Division

TJL. he current need in industry for x-ray beam containing structure infor- outgoing beams, respectively. Obvi- quality control and early detection of mation (radiography or topographic ously, high magnifications are obtained

flaws in materials cannot easily be images) before it reaches a detector. when <9 in is very small. The "quantum"

satisfied by simple refinements of This magnification capability can be efficiency of an x-ray magnifier is de- existing NDE technology. Fundamental used either to improve the overall reso- termined by reflectivity under Bragg principles underlying the existing lution of existing x-ray radiographic conditions. The ratio of the diffracted traditional techniques should be techniques which are limited by the total flux (photon/second) to the inci-

re-investigated from a fresh point of detector or to permit the use of less dent total flux is called the reflectivity

view. Here we will report the results complex (and high gain) imaging sys- and is a function of the deviation from of such efforts in two different aspects tems to attain resolutions already pos- the Bragg condition. For a thick per- for NDE techniques: one is the sible. In our demonstration an x-ray fect crystal with negligible absorption,

resolution improvement of radiography magnifier displayed a detailed image this ratio is unity for a range of angles and the other is the measurement of of 1000 mesh gold grid (25-p.m spac- centered about the Bragg condition and residual stresses within materials. ings) with good contrast on a real-time falls to zero rapidly for larger devia- basis on an ordinary image intensifier tion from the Bragg condition. Because

of 100-/xm resolution. In order to mag- the reflectivity is unity regardless of the A. X-ray Image Magnification nify an x-ray beam which contains diffracting plane, the intensity (photon/ 2 Industrial applications of radio- structure information (such as one second cm ) of a parallel beam magni- graphy presently demand the early real- containing radiographic images), the fied in one dimension by asymmetric time detection of small flaws such as beam is successively diffracted from diffraction from a perfect crystal is tight cracks in materials. The live two silicon crystals. The first diffrac- decreased by a factor (1/m) only be- images of flaws or structures are re- tion magnifies the beam horizontally cause of the magnification of the beam corded (or viewed) usually on a de- and the second in a perpendicular area. Since refractive indices for any tector (an electro-optical imaging sys- direction. materials cannot be larger than one for tem) with similar size as the image This magnification technique makes the ordinary x-ray energy range, the

itself. In order to detect small objects, use of Bragg diffraction from a unity reflectivity is the highest effi- an improvement in the resolution of nearly perfect silicon crystal free from ciency—that is, no loss—which one ! real-time imaging systems is required, dislocations. This type of Bragg dif- can expect in x-ray optics. As shown hopefully without loss of image bright- fraction, often called dynamical diffrac- i ness. However,, attempts to improve tion, has unique features which are A # 0 ut the resolution usually result in a signifi- essential for making x-ray image mag- cant loss of intensity. Thus, the de- nification possible. When an x-ray tectability (or intensity sensitivity) of beam diffracts from the surface of a I'll small signal changes caused by flaws crystal and the diffracting planes are din will be lost because of lack of bright- not parallel to the crystal surface, the l / L ness. Any approach in the improve- diffraction is termed asymmetric. There sm ment of resolution for real-time imaging are three aspects of asymmetric dynam- 0 out d = — out = md in ; m j- always encounters two opposing prob- ical diffraction which are important for 0>S lems: If one improves the resolution, the x-ray magnifier: beam magnifica- j A0 O ut= A

d , is magnified by a factor, ration and the fiber diameter in fiber allel monochromatic beam of x-rays coming beam, in

m, to produce the outgoing beam d . The optic plates. They may be about 50 is magnified in one dimension by a out extremely narrow divergence of the outgoing /xm, down to 10 /mi at best. factor m as shown in figure 1, where diffracted beam created by dynamical diffrac- We have developed [1, 2] a tech- 0, n and 0Out are the angles between the tion guarantees the faithful magnification of nique which enables us to magnify an crystal surface and the incoming and x-ray images present in the incoming beam.

110 CFFICE OF DEPUTY DIRECTOR FOR PROGRAMS

inversely proportional to -in figure 1, the maximum divergence capability of detecting strains (or proximately interior of bulk ma- the third power of the energy of x rays. :of the outgoing beam, A0ont , is related stresses) in the The use of an energy dispersive sys- to the quantity w s , which is determined terials. evalua- by the crystal property of silicon, and Currently, improved quality control tem for possible residual stress tion tested in 1973 by Leonard ;iis of the order of 6 arc second. This of industrial components demands was small value of 0 guarantees the one- quantitative information concerning the [3], who concluded, however, that this j ollt .to-one correspondence of the details stress distribution near cracks and re- system did not have sufficient accuracy have proven that ! .between the unmagnified images and sidual stress distributions after different for this purpose. We conclusion was premature, and that c the magnified images. Since the reci- types of cold working and heat treating. his Lprocity law holds for x-ray magnifiers, These demands naturally lead to the it is still too soon to dismiss the pos- .demagnification needed for lithography necessity of measuring residual strains sible industrial use of this technique ban, in principle, be achieved by these in the interior of materials. Ordinary for the evaluation of residual stresses. magnifiers. diffractometry will not be suf- In so doing, we have proposed [4, 5] ; Bragg ficient to respond to these demands. a method for determining a strain ten- the There is an entirely different approach sor in predetermined volumes in B, X-ray Residual Stress Evaluation in to the x-ray evaluation of residual interior of a material, as shown in the Interior of Materials spectros- strains although it uses the principles figure 2. In energy dispersive copy, one is concerned with the energy | One of the nondestructive methods of x-ray diffraction. This approach in- currently used for measurements of volves the use of energy dispersive solid spectral profiles and peak positions at residual stresses in industrial materials state detectors with high energy x-ray different energy values, instead of ..nvolves the application of x-ray anal- photons. The use of solid state de- Bragg angles which play a key role in ysis for stress measurement. X-ray tectors has been introduced very ef- Bragg diffractometry. Lattice constant .diffraction phenomena are used to de- fectively in the fields of scanning elec- will be measured in energy dispersive termine macroscopic residual stresses tron microscopy (microanalysis) and spectroscopy by the energies of diffracted photons: I n engineering components. The meth- x-ray fluorescence analysis of materials

odology of analysis and its basic prin- in the past ten years. It is called J/A , / 6.199 \ 1 M x ciples are well known. Briefly, when energy-dispersive spectroscopy. Unlike d(A) =1 ^—-I -Fry. r- d) v smO > E(kev) -naterials are under stress, x rays are these successful applications, this tech- diffracted with a Biagg angle which nique has not been applied to the re- Here, d is an atomic interplanar spac-

s slightly different from the Bragg sidual stress (strain) measurements of ing related to lattice constants, E is the ingle expected in unstressed materials. materials. Extremely high resolution energy of diffracted photons measured

The change in the Bragg angle is re- and accuracy are required for strain from peak positions of the spectra, and ated to the alteration of atomic inter- measurements in comparison with the the coefficient including half the scat- :>lanar spacings when the crystal lattice other applications. If one could im- tering angle 9 is constant for a given

jf the materials is strained. As a meas- prove the resolution sufficiently and scattering geometry. For our purposes, irement system, the essential part of overcome other important technical the scattering angle should be smaller

ihis x-ray technique is the accurate problems, this new system will be ideal than 30° to ensure complete trans- n:asurement of lattice constants (or for residual stress measurement in in- mission geometry. heir changes) in materials by use of dustry; the equipment will be compact One may claim that the energy reso-

veil refined Bragg diffractometry. The and the results will be displayed visu- lution of current solid state detectors is ;urface residual stress can be evaluated ally and almost instantaneously, and not adequate for the evaluation of

rom the strains measured by x-ray analysis of the data will be made on residual strains. Fortunately, it has iiffractometry using a set of assump- site by a computer. Also this system, been known that each diffraction pro-

icn regarding material elasticity and in general, works better for high energy file in the energy spectrum obtained

ic mogeniety. Although the accuracy photons. High energy photons also from a solid state detector is very close

'btained by this technique is, in prin- make practical the use of the trans- to a Gaussian. If the mathematical

ciple, superior to any other methods, mission geometry, since the x-ray ab- shape is known for a spectral profile,

n obvious shortcoming lies in its in- sorption coefficient of materials is ap- then the claim mentioned above can be

111 NATIONAL MEASUREMENT LABORATORY

Detector Sample ideal for an industrial use on site, particularly when data are handled by a mini-computer. By no means has an Rotation energy dispersive system such as used 11 here reached the ultimate resolution; i further improvements of x-ray optics, and detectors should be made, in particular, with high energy photons Translation

800 820 840 860 880 900 8 Channel Number Reference [1] Kuriyama, M., Boettinger, W. J., and Figure 2. An example of the geometrical Figure 3. An example of the observed Burdette, H. E., Proceedings of Sym arrangements for the evaluation of residual diffraction profile fitted with a Gaussian posium on Real-Time Radiologic' strains in a predetermined volume inside curve with linear background. Imaging, National Bureau of Stand- III materials. The sample is rotated three times ards, May (1978) i around an x-ray beam to give six independ- Boettinger, ent sets of measurements with two detectors. [2] W. J., Burdette, H. E., The mapping of strains throughout the and Kuriyama, M., Rev. Sci. Instr., i

volume of the sample is achieved by trans- 50 26-30 (1979). lations. [3] Leonard, L., Franklin Institute Re search Lab (Philadelphia) Report F-C3454 luly (1973). circumvented: Mathematically fitting accuracy in the determination of dif- [4] Kuriyama, M., Boettinger, W. J., and the observed peaks with Gaussian func- fraction peak positions. Our result Burdette, H. E., American Society for Nondestructive Testing, National Fall tions, one can determine the centroids demonstrates that the strains of 5 X 5 Conference, Denver, CO, 49 October (or peaks) of the profiles far more 10~ can be detected in the predeter- (1978). accurately than the AE/E value of de- mined volume of 0.5 X 0.5 X 0.5 mm :! [5] Kuriyama, M., Boettinger, W. J., and tectors would indicate. As shown in with reliable reproducibility, if suffi- Burdette, H. E., Symposium on Ac figure 3, one can improve by a factor cient counting statistics are established, curacy in Power Diffraction, Nationa 10~ 4 of 100 the accuracy of determining the while strains of 3 X can easily be Bureau of Standards, lune (1979). centroid positions. detected even with moderate counting Figure 3 shows an example of the statistics. observed spectral profile which has In conclusion, an energy dispersive been fitted by a Gaussian curve with diffraction system has demonstrated its a linear background. The accuracy of capability as a useful tool for the de- determining the centroid (or peak) po- termination of residual stresses inside sition has been found to depend simply materials, when the curve fitting tech- upon counting statistics. Reproduci- nique is used simultaneously. The pres- bility of those positions has also been ent demonstration has been carried out tested; as long as the same predeter- with low energy photons. The use of mined volume is viewed, the centroid high energy photons certainly makes position remains identical within the more practical the use of these energy statistical error. As a reference to pos- dispersive systems for residual stress sible instrumental instability, Ag Ka evaluation. More penetration and in- line spectrum in each run has been creased energy resolution of detectors fitted with a Gaussian curve. (White will improve the resultant accuracy and radiation was obtained using a Ag tar- detectability of strains. Since the en- get tube.) The Ag Ka line result indi- ergy dispersive system can be made cates that the instability, if any, is not simple and compact, without any deli-

significant enough to jeopardize the cate moving parts, the system is quite

112 FF1CE OF DEPUTY DIRECTOR FOR PROGRAMS

ligh Accuracy Particle Size and •article-Fluid Interaction Measurements Using the Particle )oppler Shift Spectrometer

favid S. Bright and Robert A. Fletcher, }as and Particulate Science Division

In addition to the particle diameter information contained in the frequency T T e have increased the accuracy of the scattered light, we are able to nd precision of the Particle Doppler also obtain particle size information hift Spectrometer (PDSS) measure- from the amplitude of the scattered jent technique through improvement light. The intensity of light scattered f its light optics and its data reduction from a spherical particle is a complex nd analysis systems. With the PDSS, function of the particle size, angle of e can now determine the diameter of observation relative to the incident pherical aerosol particles in the 5- to light, the indices of refraction of the 5-fj.m range with an accuracy of 0 Doppler frequency F, Hz 2000 particle and the suspending media, and .05 ixm. This high accuracy made it the wavelength of the light. The in- 0 5 10 16 18 ossible for us to study particle/fluid Diameter, tensity of scattered light as a function iteractions that have been undetectable of particle diameter may be calculated Figure 1. Relative scattered intensity as i y other techniques, though such from Lorentz-Mie theory. The calcula- a function of diameter calculated from nteractions have been theoretically I Lorentz-Mie theory (lower axis) and of tion for the parameters of our experi- iferred. Doppler frequency obtained experimentally ment (6° angle of observation, known (upper axis). ... is calculated data, — is wavelength, and indices of refraction) experimental data.

• The basis of particle diameter meas- give the curve shown in figure 1, which

irement by the PDSS is a modification contains many maxima and minima small as a few ^m/s, or equivalent to

if the laser velocimetry technique. The features. The calculated curve is ex- diameter differences of only 0.05 /xm.

•ample cell, a vertical tube, is inter- perimentally equivalent to summing the This sensitivity allowed us to verify acted by a horizontal laser beam. scattering intensities from a uniform the Stoke's equation for particles in

3 ight is scattered by both the stationary distribution of particles having diame- the 5- to 15-um diameter range. In

indows of the sample chamber and ters covering the range of the Mie order to accomplish this verification, it

y any particles falling through the calculations. Such a summation of was necessary to demonstrate that we

iser beam. Scattered light at 6° above intensities is also shown in figure 1 were not perturbing the settling veloc-

orizontal is detected and measured and clearly demonstrates the features ities in some way. Thermal convection ith a photomultiplier having well col- of the calculated Mie curves. The adds large velocity increments to the mated collection optics. The light duplication of the calculated Mie curve settling velocity and removes most or

:attered by the settling particles is features in the experimentally obtained all size dependency, which may easily •oppler shifted with respect to the curve allows us to use the features in be seen in the loss of the features of icident light; the magnitude of the the experiment curve as accurate size the scattering curves. Though we find

equency shift is proportional to the calibration markers. For a given fea- no evidence of thermal convection, we sttling velocity of the particle. The ture, the Mie calculation gives us an do find under certain sampling condi-

'oppler shifted frequency is measured accurate value for the particle diameter, tions a particle-induced fluid flow m mixing the light from the particle from which we can calculate the parti- phenomena. nd from the windows (unshifted light) cle settling velocity on the basis of When a particle gravitationally set- id analyzing the resulting signals, Stoke's law. The settling velocity can tles through a fluid (air in our experi- /ith a Fast Fourier Transform ana- also be determined experimentally from ments) the drag force of the fluid r zer, the amplitudes of the scattered the Doppler shift frequency for that results in a small momentum transfer ght signals are stored according to feature of the curve. Through sophisti- to the fluid causing a small fluid flow equency. From the frequency of the cated data analysis, we are able to in the direction of the settling particle. :attered light, we determine the set- compare the settling velocities calcu- This flow then serves to increase the ing velocity of the particle; from the lated from the Mie theory with the settling velocity of neighboring particles

ittling velocity, we can calculate the settling velocities experimentally deter- because the fluid is no longer station-

article's diameter by using Stokes's mined from the Doppler shift such that ary, but is moving with the settling LW. we can detect velocity differences as particles.

113 We demonstrated this effect by com- flow is uniform over the entire volume paring the following two experiments: and effectively cancels the drag effect

In the first experiment, we measured on the particles. the settling velocity of particles when The induced fluid flows we observed they uniformly filled the settling cham- are on the order of 10 to 100 /xm/s. ber and compared it to the calculated An order of magnitude calculation, bal- velocity to verify that no velocity shift ancing the fluid viscous diffusion force occurred. We use this sample geometry and the Stoke's drag force, gives fluid for particle size measurements. In the velocities which are in good agreement second experiment we measured the with these values. settling velocity of a 4-mm diameter The sensitivity of the particle settling column of settling particles in a 9-mm velocity to the induced flows of such diameter settling chamber. The air low magnitude can be measured by the sheath around the column of particles PDSS due to the accuracy we can was void of particles. With this experi- obtain by using the Mie scattering mental condition we observe an in- curve features as size calibration mark- crease in the settling velocity, which ers. We know of no other technique

is particle concentration dependent, as that can determine particle diameter in shown in figure 2. The velocity increase the 5- to 15-^m diameter range to an represents the velocity of the induced uncertainty of 0.05 or that has the air flow, which adds to the settling sensitivity to detect particle-fluid inter- velocity. actions of the low magnitudes we have

The induced fluid flow is manifested described above. in the second experiment because a return flow path exists—the particle- void region of the settling chamber. In the first experiment, no velocity

shift is observed because the return

.02

.00 500 1000

Particle Concentration (cm -3)

Figure 2. Fluid velocity as a function of particle concentration.

114 U.S. GOVERNMENT PRINTING OFICE : 1980 O-311-046 (133)