University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange

Doctoral Dissertations Graduate School

3-1960

Gas Chromatography of Hydrogen-Deuterium Mixtures

Paul Payson Hunt University of Tennessee - Knoxville

Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss

Part of the Chemistry Commons

Recommended Citation Hunt, Paul Payson, "Gas Chromatography of Hydrogen-Deuterium Mixtures. " PhD diss., University of Tennessee, 1960. https://trace.tennessee.edu/utk_graddiss/2921

This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council:

I am submitting herewith a dissertation written by Paul Payson Hunt entitled "Gas Chromatography of Hydrogen-Deuterium Mixtures." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, with a major in Chemistry.

Hilton A. Smith, Major Professor

We have read this dissertation and recommend its acceptance:

John W. Prados, Jerome F. Eastham, William E. Bull, John A. Dean

Accepted for the Council:

Carolyn R. Hodges

Vice Provost and Dean of the Graduate School

(Original signatures are on file with official studentecor r ds.) January 16, 1960

To the Graduate Council:

I am submitting herewith a dissertation wr itten by Paul Payson Hunt entitled "Gas Chromatography of Hydrogen­ Deuterium Mixtures." I recommend that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy , wi th a ma jor in Chemistry.

We have read this dissertation and reccmmend its acceptance:

Accepted for the Council:

_ tl�uJ. __ Dean,i&l/ of th ){aue G� ool GAS CHROMATOGRAPHY OF HYDR OGEN-DEUTERIUM MIXTURES

A Disser tation

Pre sented to

the Graduate Counc il of

The University of Tenne ssee

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

by

Paul Payson Hunt

March 1960 ACKNOWLEDGMENT

The author wishes to express hi s most sincere appre­

ciation to Dr. Hilton A. Smi th for his fri end ship and counsel during the cour se of this research. The work was made pos-

sible through a research grant from the Un ited States Atomic

Energy Commission for the period from September 1957 to

Decembe r 1959 .

The flint quartz employed in this work was donated by

the Whittaker , Clark , and Daniels Company , Inc., and the

Raney nickel and Rane y cobalt were contributed by the Raney

Catalyst Company.

The author also wi shes to thank his wife , Helen , for her patience , understanding , and encouragement toward the

completion of the work . TABLE OF CONTENTS

CHAPTER PAGE

I. INTRODUCTION • • ...... 1

A. Scope ...... 1

B. His torical ...... 2

c. • . . Chromatographic Techn iques • • . . 5

1. Elution Development . . • • • .•• 5

2. Frontal Analysis . . • • ...... 6

3. Displacement Development •..•.... 6

D. Dis tribution Isotherms • . . . • . . 7

E. Hydrogen and Deuterium • ...... 7

II. APPARATUS AND MATERIALS • ...... 12

A. General ...... 12

B. Carrier Gas System . . . . • . . . . • . • . 12

1. Carrier Gases 12

2. Flow Ra te Control ...... 16

3. Flow Rate Measurement ...... 17

4. Circulatory System . . . . . 17

c. Sample System ...... 19

• • • . D. Column Assembly . . . . . • 21

1. Column Preparation .•.••.. 21

2. Cooling System .••.• 31

3. Heating System . • . • • . • • . . . 32 iv

CHAPTER PAGE

II. (continued)

E. Detection System ...... 33

1. General • • . • • . . • ...... 33

2. Katharometer . . . . • . • . . . . . 34

3. Measuring and Record ing Circuit . . . . 36

III. EXPERIMENTAL PROCEDURES AND RESULTS . • . . 38

A. Introduction...... · ..••. 38

B. Elution Experiments . . . . . • . . 39

1. Silica Gel Columns . • ...... 39

2. Molecular Sieve-5A Columns . . . . . 52

3. Charcoal Column . . . . • • . . . . 52

4. Raney Nickel Columns ...... 52

5. Raney Cobalt Columns ...... 60

6. Chromia-Alumina Columns ....•. . . 62

7. Chromia-Flint Quartz Column .•. . . . 70

IV. DISCUSS ION OF THE RESULTS . . . . . • . . 75 A. Retention Volumes 75

B. Separation Factor ...... 77

C. Column Efficiency .•...... 81 . D. Analytical Application ...... 83

E. Comparison with Previous

Chromatographic Results ...... 86 v. SUMMARY ...... 88

BIBLIOGRAPHY ...... 91 LIST OF TABLES

TABLE PAGE

I. Percentage Composition of the Hydrogen

Isotopes ...... 48

II . Retention Time for Hydrogen Gas on Raney

Cobalt Column ...... 63

III . Isotopic Composition for Equilibrium '.

Mixtures of Hydrogen Isotopes • • . • • . . . . 73 IV. Experimental and Theoretical Composition

of the Isotope s ...... 8? LIST OF FIGURES

FIGURE PAGE

1. Distr ibution Isotherms and Corresponding

Typical Elution Peaks • ...... 8

2. Func tional Diagram of Gas Chromatography

Apparatus Employed with Argon, Helium, and

Hydrogen Carriers • . . • • . . • • . • • 13

3. Photograph of Gas Chromatography Apparatus . . • 14 4. Functional Diagram of Gas Chromatography

Apparatus Employed with Neon Carrier . . • 15

5. Circulatory Pump ...... 18

6. Pyrex Glass Column Packed with Raney Nickel • . . 24

7. Katharome ter Block . . . . • ...... 35

8. Complete Circuit for Detection System . • . . 37

9. Effect of Sample Magnitude on the Shape of

Hydrogen Peak ...... 40

10. Typical Elution Peak for Deuterium with

Hyd rogen Carrier on Silica Gel Column . . 42 11. Recorder Trace for Hydrogen-Deuterium Sample Passed Through Palladium and Silica Gel

Columns • . • ...... 43

12. Typical Elution Curve for Hydrogen , Deuterium,

and Hydrogen Deuteride Sample on Silica Gel

Column • 0 • • • • • • • • • • • • • • 44 vii

FIGURE PAGE

13 . Calibration Curve for the Determination of

Deuterium . • • . . • • ...... 46

14 . Typical Elution Curve for Hydrogen and

Deuterium at the Temperature of Boiling

MethBile . . . • . . . • . . • . . . . . 49

15. Typical Elution Curve for Hydrogen, Deute­

rium, and Hydrogen Deuteride at the

Temperature of Boiling Methane . . . • . . 50

16. Typical Elution Peak for Hydrogen and Deute-

rium Sample with Neon Carrier at -161° •. • • 51

17 . Typical Elution Trace with Hydrogen, Deute­

rium, and Hydrogen Deuteride Sample on

Charcoal Column . • • . • • . . . • 53

18 . Typical Recorder Trace for the Elution of

Hydrogen or Deuterium from Raney Nickel

Column ...... 55

19 . Elution Peak for Hydrogen-Deuterium Sample

after Elevating the Raney Nickel Column

Temperature • • • ...... • . . • • • . 56

20. Elution Trace Obtained after Cooling and

Reheating the Raney Nickel Column . • . • • . • 58

21. Elution Peak Obtained with Deuterium Sample

on 6-ft. Raney Nickel Column . . ' ...... 59 viii

FIGURE PAGE

22. Typical Elution Curve for Hydrogen or Deute-

rium Sample on Raney Cobalt Column 61

23. Effect of Time between Runs upon the Reten-

tion Time of Hydrogen on Raney Cobalt . o • • • 64

24. Typical Elution Curve for Hydrogen Sample

on Activated Chromia-Alumina Column ...... 65

25. Typical Elution Curve for Hydrogen and

Deuterium Sample on Partially Poisoned

Chromi a-Alumina Column . . . . • . 67

26. Typical Elution Curve for Hydrogen , Hydrogen

Deuteride , and Deuterium Sample on Par­

tially Po isoned Ch romia-Alumina Column 68

27. Elution Trace Obtained with Highly Deacti-

vated Chr omia-Alumina Column •.•. 69

28 . Typical Elution Curve for Hydrogen , Hydrogen

Deuteride , and Deuterium Samp le on Par­

tially Poisoned 8-ft . Chromia-Alumina

Column 71

29. Calibration Curves for the Determination of

Hydrogen and Deuterium ...... 72 CHAPTER I

INTRODUCTION

A. Scope

In recent years the gas chromatographic method has found widespread applications, chiefly for analytical pur­ poses. This method is principally employed for the separa­ tion, identification, and quantitative determination of vola­ tile compounds. It is potentially of great value for the quantitative determination of mixtures of the hydrogen iso­ topes. The development of a rapid, convenient, and inexpen­ sive method for the analysis of the hydrogen isotopes would be a valuable addition to many laboratories, especially those where a mass spectrograph is unavailable. This problem was undertaken in an attempt to develop further the chromato­ graphic method for the separation and quantitative determina­ tion of hydrogen and deuterium mixtures.

The term "gas chromatography" describes all chromato­ graphic methods in which the moving phase is a gas. The sub­ ject may be subdivided into gas-liquid chromatography and gas-solid chromatography. The former is employed to describe all gas chromatographic methods in which the fixed phase is a liquid, and the latter refers to those methods in which the fixed phase is a solid. 2

The ma jor difference between liquid chromatography and gas chromatography is the nature of the mobile phase . The former employs an incompressible liquid as the mobile phase , but in the latter the mobile phase is a compressible gas. The chromatographic me thods depend on the distribution of the samp le between tw o phases and the subsequent se paration of these two phases. Gas-solid chromatography consists of a solid with a large surface area as a stationary bed , and a gas which percolates through the stationary bed.

B o Historical

Credit for the introduction of the chromatographic me thod is given to Tswe tt. 1 This investigator employed liquid-solid chromatography for separating components of plant pigments . During the years which followed the original papers, this method went practically unnoticed until Kuhn and

Lederer2,3 resolved the carotenes on a pr eparative scale . Li quid-liquid chromatography was introduced by Martin and Synge4 in 194lo Wa ter-saturated silica gel was employed as the stationary phase for separating some amino acids. In the same paper these investigators proposed the possibility of using a gaseous mobile phase with a supported stationary phase . Consden , Gord on , and Martin5 showed that fi lter paper sheets and strips can also be used as supports for the sta­ tionary phase in partition chromatography . This type of analysis is designated as paper chromatography. 3 The use of se lective adsorption processes for the separation of gaseous materials is a me thod of long standing .

Adsorption chromatography was adapted to mixtures in the form of vapors by Turner6, 7 and Claesson. 8 James and Mar tin9 introduced gas-liquid chromatography in 1952. These authors employed a fixed phase of silicone grease and stearic acid supported on kiese lguhr to se parate some volatile fatty acids.

Since the introduction of the method, gas chromatog­ raphy has been of interest for possible separation of iso­ topes. Glueckauf, Barker, and KittlO reported the enri chment but not complete separation of neon isotopes on charcoal .

The first separation of hydrogen isotopes by the gas chromatography method was repor ted by Glueckauf and Ki tt . 11

These authors separated pure deuterium from hydrogen-deuterium mixtures by displacement chromatography at room temperature in a column containing palladium black supported on asbestos as the fixed phase . Thomas and Smithl2 obtained partial resolut ion of hydrogen-deuterium mixtures by gas-elution chromatography in a column containing pallad ium black sup­ ported on flint quartz . Arg on was employed as the carrier gas, and the separation of the two components was less than that which mi ght be expec ted from a comparison of the elution peaks for pure hydrogen and for pure deuterium . Moore and

Ward13 reported the se paration of orthohydrogen fr om para­ hydrogen and the partial se paration of orthodeuterium fr om paradeuterium by means of an activated alumina column . These 4 authors employed helium as the carrier gas at 77°K. To obtain signal amplification from the katharometer, the iso­ topes were oxidized by passing the gas stream over copper oxide . The alumina column has been further investigated at

-195° by Van Hook and Emmett. l4 These investigators obtained good separation of ortho- and parahydrogen; ortho- and para­ hydrogen and deuterium; and of parahydrogen, deuterium, and orthohydrogen and hydrogen deuteride. Percentage composition of the isotopic mixtures was calculated from area measure­ ments of peaks obtained after the isotopic forms were oxi­ dized . Ohkoshi , Fujita, and Kwanl? empl oyed a Linde molec­ ular sieve column and hydrogen as the carrier to obtain se paration of hydrogen deuteride and deuterium at 77°K .

Chadwickl6 has described a me thod for separating tritium from 2-1 . batches of tritium and hydrogen by dis­ placement chr omatography on palladium. A' recent patentl7 has been granted for separating hydrogen from tritium by con­ tacting a gaseous mixture with finely divided palladium . The evolved fractions were collected as the pallad ium was gradu­ ally heated. Gant and Youngl8 employed a molecular sieve column and helium as the carrier gas to analyze hydrogen, tritium hydride, and tritium mixtures. The column was kept at -160°, and an ionization chamber was employed in series with the thermal conduc tivity cell for detection . 5 C. Chromatographic Technique s

The chromatographic processes may be subdivided into technique s by which the process is carried outo These tech­ niques are termed :

(1) elution development;

(2) frontal analysis;

(3) displac ement development.

1. Elution Develogment

In elution analysis, a small sample is introduced at the top of the column . This is usually accomplished in gas chromatography by placing the sample in the carrier gas stream. After the gas has been ads orbed on the column , the carrier gas , which is not as strongly ad sorbed as the sample by the stationary phase, is continually passed thr ough the column. Each sample component distributes itself in a char­ acteristic manner between the gaseous phase and the fixed phase , and that portion in the gaseous phase moves with the carrier . The zone s occupied by the ad sorbed substances travel down the column at different speeds, and that sub­ stance which is ad sorbed least strongly appears at the end of the column first. The appearance of band s will be from the least to the most strongly ad sorbed . Of the various chromatographic techniques, the elution me thod is the only one which normally may be expected to separate completely the components of the sample . 6

2. ErQntal Anal�sis

The mixture to be separated is fed into the column

during the entire course of the process. An adsorbate of the most stron gly adsor bed component will then be progressively built

up in the column until the adsorptive capacity of the latter

is reached. The components appear at the end of the column

in the order of their relative affinities for the adsorbent.

The components are not separated into bands, and only the

least adsorbed substance is separated in pure form.

3. Displacement DeveloQment

In displacement analysis, the sample is first carried

onto the column in a stream of the mobile phase. It is then steadily pushed or displaced along the column by causing the

mobile phase to carry with it a constant concentration of the

displacer vapor. The displacer substance is more strongly adsorbed than any of the sample components. The bands ob­ tained are not separated by bands of relatively pure carrier

as in elution development. The more weakly adsorbed sub­ stance is actually displaced by the more strongly adsorbed substance. The plot of some property of the components against the emerged volume usually produces a series of steps

in a displacement analysis. If the property observed does

not differ in value from one component to another,-·a step will not be obtained. 7

D. Distr ibution Isotherms

The ratio of the concentrations of a substance dis- tributed betwe en two equilibrated phases under specified con­ ditions is termed the distr ibution or partition coefficient.

The distribution coefficient , k, is defined by equation 1:

k = concentration of solute in fixed phase ( l ) concentration of solute in mobile phase

For a linear distribution isotherm, k is a constant equal to the slope of the isotherm. If the distribution coefficient is not a constant, the isotherms may be concave toward either axiso Ideal elution chromatography under the latter conditions produces peaks which are asymmetrical. The three types of distribution isotherms and the corresponding typical elution peaks are shown in Figure 1.

Isotherms for the ad sorption of mo st substance s are not linear but resemble the center example of Figure 1. The elution peaks obtained from ads orbents usually have sharp

"fronts" and long· "tails."

E. Hydrogen and Deuterium

An ordinary mixture of hydrogen and deuterium contains the three molecular species: hydrogen, hydrogen deuteride, and deuteriumo The conversion of hydrogen deuteride to hydro­ gen and deuterium is catalyzed by many surfaces , notably those of me tals and of oxides. Any mixture of hydrogen and deute­ rium or hydrogen , deuterium, and hydrogen deuteride, which is 8

Q) Q) (I) (I) ctl ctl ..c ..d 0.. 0.. Q) '0 ...... Q) .... >< ..0 •.-I 0 � s J:: J:: •.-I •.-I J:: J:: 0 0 •.-I •.-I .j.l .j.l ctl ctl '"' '"' .j.l .j.l J:: J:: Q) Q) u u J:: J:: 0 0 u u

Concentrat ion in Time -+ mobile phase

Figure 1. Distribution isotherms and correspond ing typical elut ion peaks . 9

in contact with these surfaces for a sufficient length of

time , will reach an equilibrium condition.

If hydrogen and deuterium are separated from each

other in a chromatographic column which catalyzes the con­ version of hydr ogen deuteride to hydrogen and deuterium, the hydrogen deuteride concentration will be continually reduced as the sample passes through the column . The hydrogen deuter­

ide concentration may eventually reach zero if hydrogen and deuterium can be completely se parated . For substances on which the exchange reaction does not occur, the separation of the isotopic species by adsorption and desorption as the sample goes through the column should produce pure samples of hydrogen , hydrogen deuteride, and deuterium at the column exit.

Quantum theory predicted in 1927 that the hyd rogen molec ule could exist in two distinct and stable forms, para­ hydrogen and orthohydrogen. The two forms have different optical and thermal properties , and normal hydrogen is a mix­ ture of the two . The two nuclei have parallel spins in the

ortho state; and in the para state , the nuclear spins are anti-parallel .

The more intense lines, obtained from the spectrum of molecular hydrogen und er ordinary conditions , are produced by ortho molecules. These lines correspond to odd rotational

levels in the ground state. The less intense lines, which represent even rotational levels, are due to para molecules. 10

The ratio of orthohydrogen to parahydrogen is approximately

three to one under equilibr ium conditions at room temperature .

The energy content of the gas decreases as the temperature of

the gas is lowered. At low temperatures, the molecules tend

to occupy the lower levels, and the molecules of hydrogen would be principally in the para state. At the tempera ture

of liquid nitrogen, the equilibr ium ratio of orth ohydrogen to

parahydrogen is one to one.

The surface catalysis of the conversion and exchange

of the hydrogen modifications may be by one of two mechanisms .

In the first type of mechanism, the spin isomeriza tion of hydrogen is catalyzed by inhomogeneous magnetic fields which

exist near paramagne tic ions or molecule s. This type of mechanism is likely to operate only at low temperatures,

because at higher temperatures the Van der Waals ad s orbed hydrogen at the surface is not large enough to allow a feasi­

ble rate. The second mechanism involve s the chemisorpti on of

hydr ogen atoms on the surfac e. The chemical mechanism is

capable of catalyzing both conver sion and exchange, whereas

the paramagnetic mechani sm is capable of catalyzing only

conve rsions.

Deuterium al so exists in two forms. However, the even

rotational levels constitute the ortho state and the odd

levels the para form. For this reason, deuterium exists

almost entirely in the ortho state at low temperatures. If 11 a molecule has two nucle i which are not identical , the ortho

and para state s cannot exist.

A chromatographic column packed with a diamagnetic

substance , which does not catalyze the ortho-para conversion, would separate the ortho and para forms , as well as hydro­

gen, deuterium, and hydrogen deuteride, if the separation factors for the components are different. A substance which

catalyzes the ortho-para conversion should , however, separate

only the three isotopic species. CHAPTER II

APPARATUS AND MATER IALS

A. General

A functional diagram of the apparatus employed for the experiments in which argon , helium , and hydrogen were used as the carrier gases is given in Figure 2. Figure 3 is a photo­ graph of this apparatus. The flow system was changed consid­ erably for the investigations in which neon was employed as the carrier . A functional diagram of this apparatus is shown in Figure 4. The carrier gas system, the sample system, the column assembly , and the detection system are described in the following sections .

B. Carrier Gas System

1. Carrier Gases

Argon , helium, ne on, and hydrogen were employed as carrier gases. Ar gon represents the be st choice from the standpoint of detection requirements. Since the thermal con­ ductivity of a gas is related to the molecular weight, the carrier gas should have a molecular weight quite different from that of the sample if a thermal conductivity device is employed for detection . For hydrogen and deuterium detection, argon fulfills thi s requirement . Argon in cylinders was obtained from the We lding Gas Products Company. 13

To atmosphere

Di�>phragm pressure Capillary regulator I \

Carrier "'-- Needle tank valve

Sample preparation cell

�Pressure manometer Vacuum -? manometer

�--��To hydrogen tank

To helium tank To vacuum pump

Figure 2. Functional diagram of gas chromatography apparatus employed with argon, helium, and hydrogen carriers. 14

A -Recorder H - Bri dge Current Meter

� - Hydrogen Tank ;r - Katharometer C - Deuterium Tank J - Soap Film Flow Meter D - Column Heater K - Sample Manipulation System

. E - High Pressure Manometer L - Input Pressure Gauge F - Galvanometer M - Associated Electrical Circuit

G - Argon Tank N - Sensitivity Controls

Figure 3

Phot ograph of Gas Chromatography Apparatus 15

Circulatory Charcoal trap pump Silica gel 1 trap \

Oxidation furnace

( Katharometer

<. Column

i Pressure r------By-pass cell Neon ._.manometer tank �Sample exchange cell

+- Pressure manometer Vacuum ....-::, manometer To hydrogen tank

To deuterium To helium tank tank To vacuum pump

Figure 4. Functional diagram of gas chromatography apparatus employed with neon carrier. 16

The molecular we ight of helium is very close to that of deuterium; therefore , helium is a very poor choice for the detection of hydrogen and deuterium by thermal conductivity .

At liquid nitrogen and boiling me thane temperatures , argon is either condensed or ads orbed very strongly and is not a suit­ able carrier" For this reason , helium or neon was employed at the lower temperatures. Helium wa s obtained from the

National Cylinder Gas Company and was used without further purification .

Neon is not as good a choice as argon for the detec­ tion of hydrogen and deuterium by thermal conductivity meas­ urements. Neon does have the advantage , however , that it can be employed at lower temperatures and produces much greater sensitivity than helium . Neon was obtained from the Matheson

Company.

Hydrogen and deuterium are ad sorbed strongly on silica gel at 77°K. Hydrogen was found to elute the isotopes from the silica gel column at this temperature , and the detection of hydrogen deuteride and deuterium peaks was obtained . Hydrogen was obtained from the National Cylinder Gas Company and was used without further purification.

2. FlQ! Rate Control

The flow rate was regulated by the diaphragm pres sure regulator on the carrier gas cylinders. For the neon carrier system, the flow rate depended upon the pressure differential 17

obtained with a gas circulating pump and the flow resistance

offered by the column .

3. Flow Rate Me asurement

A soap-film meter19 was employed for flow rate measure­

ment. The soap-film meter is simple and accurate, it requires

no calibration, its reading is ind epend ent of the nature of

the gas be ing measured , and it offers essentially no back

pressure . A dilute solution of a commercial liquid detergent wa s used in the meter .

4. Circulatory System

Since neon is a relatively expensive gas , it wa s neces­

sary to conserve it as much as possible . For this purpose a

circulatory system was construc ted in which the hydrogen and deuterium were removed by oxidation and subsequent adsorption

of the water . The complete functional diagram for the neon

circulatory system is given in Figure 4. The double-acting

piston pump employed in the gas flow system is shown in Figure 5.

The pis ton barrel was approximately 5 in. long and was constructed from 12-mm . Pyrex glass tubing . A Teflon-covered magne tic plunger, 1! in. in length, was employed in side the

piston chamber .

The outlet and inlet connections to the pump were made

of Pyrex tubing with an out side diame ter of 6 mm. Each valve

consisted of a section of 9-mm. glass tubing approximately 18

To pulley and motor '

------Intake for gas

Outlet

Figure 5. Circulatory pump. 19

! in. in length , fitted over a i-in. section of 6-mm . capil­ lary tubingo The capillary and the 9-mm . tubing were sealed to the 6-mm. tubing. The capillary tubing was ground on the end , and a pie ce of microscope cover glass was seated in the narrow space between the capillary end and the 6-mm. tubing connection to the 9-mm. tubing. Indentations were mad e in the 6-mm. tubing just above the cover glass to prevent the cover from turning edgewise and sticking during the pumping process.

The plunger was drawn by two circular Indox magnets obtained from the Indiana Steel Products Company of Valparaiso,

Indiana. The se were clamped in a brass carrier . The magnets and carrier were placed around the pis ton barrel before the valves were sealed to one end of the compression chamber.

The magnet car rier wa s connected by means of a string across a pulley to the motor to furnish vertical movement of the plunger. A two-speed motor (45 and 78 r.p.m.) was employed to move the magnet. A slot cut in a disc on the turntable of the motor allowed the stroke length to be properly adjusted .

The motor was usually employed at the 45 r.p. m. speed.

C. Sample System

The sample system was employed to prepare hydrogen and deuterium mixtures of known composition, and also hydrogen, deuterium, and hydrogen deuteride mixtures for inje ction into 20 the column . A schematic diagram of the sample system is given in Figure 2.

In order to prepare a sample of hydrogen and deuterium, the by-pass cell was ini tially evacuated . Hy drogen or deute­ rium was then ad mitted and the pres sure measured. The by-pass cell was closed , and the remaining part of the sample system was again evacuated . The second gas was admitted to the sample system at a higher pressure than the first. The stop­ cock to the by-pass cell was opened momentarily to allow the second gas to enter . At a designated time , the carrier gas was rerouted to sweep the sample into the column.

The following apparatus was employed for the prepara­ tion of a mixture of hydrogen deuteride , deuterium, and hydro­ gen . Two tungsten leads, brazed to a coiled Nichrome wire , were sealed into a 100-ml ., round-bottomed flask. The flask was connected to a capillary manome ter through a three-way stopcock. The third arm of the three-way stopcock was con­ nected to the sample system through a straight stopcock.

In order to prepare hydrogen , deuterium, and hydrogen deuteride mixtures, the sample bulb was evacuated and filled with hydrogen and deuterium. The Nichrome wire was brought to a faint glow for a designated leng th of time to catalyze the isotopic exchange. The temperature of the wire was con­ trolled by a Powerstat . Samples were then introduced into the by-pass cell and swe pt into the column as previously discussed . 21 D. Column As sembly

1. Column Preparatigg

a. Ge�l· Pyrex glass and copper tubing were employed as columns. Copper tubing column s were prepared from tubing with 5/16- or 1/4-in . outside diameterso To facilitate ease of handling and cooling, the packed columns were coiled in a spiral of approximately 4 in. diameter . In order to cool a column, the coiled column was fitted into a

Dewar flask and covered with a cooling liquid or solid .

Columns used at temperatures above room temperature were folded in sections to an over-all leng th of about 3 ft . This allowed these columns to be conveniently fitted into the column heater. Pyrex glass with an inside diame ter of 4 mm . was employed to prepare short Raney cobalt or nic kel columns .

b. Silica gel columns . Refrigerator grade silica gel obtained from the Davison Chemical Company was dried in an oven at 140° , crushed, and screened through a series of stand­ ard mesh screens. The silica gel was again dried before the column was packed.

Two columns were prepared from copper tubing with an outside diameter of 5/16 in . and an in side diame ter of 7/32 in .

Columns having leng ths of 6 and 10 ft . were packed with 40-60 and 60-80 mesh silica gel, respectively. Before packing , the columns were folded once and the packing was poured in to both ends . The packing was compacted by vibration of the column . 22

After the columns were packed , they were coiled as previously

indicated.

A 25-ft. section of copper tubing with an outside

diame ter of 0.25 in. and an inside diameter of 0.19 in. was

packed with 40-60 me sh silica gel. The column was coiled in

such a manner that one section of the spiral fit conveniently

inside the other.

c. Molecular sieve 5-A QQlumn. Zeolitic molecular

sieves are marke ted by Linde Air Products Company, a division

of Union Carbide and Carbon Corporation , in 4A , 5A , and 13X sizes. Also available for experimental purposes are 3A and

lOX sizes. The size designations correspond appr oximately to

the size of the pores in ang stroms .

A 10-ft. section of copper tubing with an outside

diameter of 5/16 in. was packed with 40-60 mesh Linde Molec­

ular Sieve-5A. The packed column was coiled to facili tate

ease of handling. The molecular sieve- 5A column wa s acti­ vated at 200-220° while passing nitrogen through the system. d. Charcoal column . A 5-ft. section of copper tubing with an outside diameter of 5/16 in. was employed to prepare

an ac tivated charcoal column. The tubing was packed with

40-60 me sh activated wood charcoal. This column was coiled

to facilitate ease of handling .

e. Raney nickel columns . Raney nickel was obtained

from the Raney Catalyst Company� Chattanooga, Tennessee. The

Raney nickel alloy consisted of 42 per cent nickel and 58 per 23 cent aluminumo Rane y nickel catalysts are prepared by leach­ ing the aluminum from the aluminum-nickel alloy with sodium hydroxide . Large volume s of hydrogen are liberated dur ing the leaching. In the preparation of the packing for the columns , short reaction times and low concentrations of the sodium hydroxide were employed in an attempt to le ach the aluminum from the surface only .

Several Raney nickel columns were prepared for pre­ liminary chromatographic inve stigations with 4-ft . sections of Pyrex glass tubing . The reaction time between the aluminum and the sod ium hydroxide was var ied with each column in an attempt to prepare a superior column . Only those columns which were employed in the final investigations are described below . These columns were shaped as indicated in Figure 6.

Thirty grams of 60-80 me sh Raney nickel alloy was placed in a 1-1. Erlenmeyer flask equipped with an all-glass stirring rod attached to a me chanical stirrer. The alloy was covered with 250 ml. of distilled wa ter and was stirred for several minutes to induce wetting of the surface. A water bath was employed for temperature control . The initial temperature of the suspension was 20°. Fifty milliliters of distilled water containing 5 g. of sod ium hydroxide was added to the contents of the flask in the time of one minute. The temperature of the reac tion mixture was gradually increased by adding warm water to the bath . The reaction of the sodium hydroxide with the aluminum 24

------Liquid level

------To aspirator �------

------Glass wool plugs

Figure 6. Pyrex glass column packed with Raney nickel. 25 also generated considerable heato The contents of the flask were agitated for seven minutes after the initial addition of the sodium hydroxide solution. The temperature of the mixture reached 45° , and considerable evolution of hydrogen was evident.

At the end of the seven-minute period , the Raney nickel was washed several times with distilled water by decantation in order to remove the alkali. Fine particles of alloy were also removed in this manner. The wash water was tested with litmus to determine if the alkali was completely removed .

The Raney nickel alloy was transferred to a small beaker by means of a stream of water fr om a wash bottle . Air contact was avoided as much as possible when transferring the Raney alloy.

One end of the column was equipped with a funnel, and the column was completely filled with water. The Raney nickel was transferred to the funnel by means of a spatula and a stream of water from the wash bottle. The alloy gradually settled from the funnel into one side of the column. The second side of the column was filled by pulling the packing into it by means of a water aspirator . Settling of the pack­ ing wa s facilitated by gently tapping the column with a glass rod. After the column was filled to the desired height, the ends were stoppered with glass wool.

The argon flow system was connected to the column , and the water was removed by slowly heating the column to 110°. 26

After the packing wa s dry, the heater was removed , and the

column was tapped gently with a glass rod to settle the Raney

nickel into the visible air spaces. Before any runs were

performed, argon was permitted to flow through the column for

twelve hours .

Two 30-g . samples of 60-80 me sh Raney nicke l alloy

were employed to prepare the packing for a 6-ft . copper tub­

ing column . Each 30-g . sample was treated with 5 g. of

sodium hydroxide in the same manner as previously indicated.

The initial temperature of the reaction mixture wa s approxi­

mately 35°, and the final temperature reached 45°. The time

consumed between the first addition of sodium hydroxide and

the initial washing of the Raney nickel was three minutes.

After the alkali was completely removed , the two porti ons of alloy were thoroughly mixed in a small beaker. The alloy was

washed three times with ethanol .

The 6-ft . secti on of copper tubing with an outside

diame ter of 5/16 in. and an inside diameter of 7/32 in. was folded once . Each end of the column , which was filled with

ethanol , was equipped with a funnel, and the Raney nickel alloy was added to both ends of the column. After the 60 g. of Raney alloy was placed in the column, the remaining space

was filled with flint quartz. The ethanol wa s removed by

allowing argon to flow through the column as the temperature

wa s slowly raised to 1100. The column was cooled and gently

tapped to settle the packing . 27

Two 44-g. samples of 60-80 mesh Raney nickel alloy were employed to pack a 10-ft. copper tubing column which had an outside diameter of 0.25 in. and an inside diameter of

0.19 in. The samples of Raney nickel were treated with 5 g. of sodium hydroxide as previously indicated, except that the temperature of the reaction mixture was kept below 15°. The reaction was allowed to proceed for three minutes before washing the alloy.

The two 44-g. samples of 60-80 mesh Raney nickel alloy were placed in the 10-ft. section of copper tubing. The ends of the column were filled with flint quartz, and the ethanol was removed by heating the column to 220° while passing argon through the system.

f. Raney cobalt columns. Raney cobalt was obtained from the Raney Catalyst Company, Chattanooga, Tennessee. The

Raney cobalt alloy consisted of 40 per cent cobalt and 60 per cent aluminum.

Several 4-ft. Pyrex glass columns were packed with

Raney cobalt alloy in order to perform preliminary investiga­ tions concerning the elution of hydrogen and deuterium samples through the column. After the preliminary experiments, a

12-ft. section of copper tubing with an outside diameter of

5/16 in. and an inside diameter of 7/32 in. was packed with

40-60 mesh Raney cobalt.

Five 42-g. samples of Raney cobalt allo·y were treated individually as indicated below. The alloy samples were 28 covered with 350 ml. of distilled water in a 1-1. Erlenmeyer flask, and the contents of the flask were mechanically stirred with a glass stirring rod. A water bath was employed for temperature control, and the temp erature of the bath was maintained at 20°o Fifty milliliters of distilled water con­ taining 7 g. of sodium hydroxide was added to the contents of the flask in the space of thirty second s. The reaction mix­ ture was stirred for tw o and one=half minutes, and the con­ tents were immediately washed several times with distilled water to remove the alkali. Litmus paper was employed to test for the presence of alkali.

After the five samples had been treated with alkali and subsequently washed , they were thoroughly mixed. The entire sample was again washed several times with distilled water. The final three washings were with 95 per cent ethanol. The Raney cobalt was transferred to a beaker and was completely covered with ethanol.

A 12-ft. section of copper tubing with an outside diameter of 5/16 in. and an inside diameter of 7/32 in. was folded once. Each end of the column was equipped with a funnel, and the column was completely filled with 95 per cent ethanol. The Raney cobalt was ad ded to both funnels , and the column was gently vibrated to facilitate settling. A short section at the top of each end of the column was filled with flint quartz , and the packing was held in position with glass wool plugs. 29

The Raney cobalt column was connected to the gas flow system, and argon was allowed to flow through the column for eight hour s at room temperature to remove the ethanol. The column wa s disconnected, and the ends were sealed to prevent air from coming in contact with the packing. The packing was resettled by gently vibrating the column. Two ad ditional bend s were made to shorten the working length of the column to 3 ft.

g. Chromia-alumina columns. For the preparation of the chromia-alumina columns� Grade F-1, 8-14 mesh, activated alumina obtained from the Aluminum Company of America was crushed and screened through a series of standard screens .

Two hundred and twenty-five grams of the 20-40 mesh material was placed in a 1-1. Erlenmeyer flask. The alumina wa s covered with 2 50 ml. of water. Sixteen grams of chromium trioxide was dissolved in 100 ml. of water, and the solution was added to the contents of the Erlenmeyer flask. The con­ tents were thoroughly stirred immediately, and the stirring process was repeated intermittently during a three-hour period. At the end of this period, the excess water was removed by filtration, and the residue was placed in a glass tube for subsequent drying and reduction. The glass reduc­ tion tube consisted of a section of glass tubing 16 in. long and 28 mm. in outside diameter. This section was sealed to short sections of 11-mm. tubing on the ends. The glass tube and contents were placed in an electric furnace. The 30 reduc tion tube was connected to the hydrogen tank by means of a short piece of rubber tubing. The hydrogen flow was started, and the furnace temperature was gradually elevated.

The furnace temperature was controlled by means of a variable transformer . Considerable water was desorbed from the con­ tents before reduction oc curred . The maximum furnace tempera­ ture reached 360°, and the yel low material turned gr een upon reduc tion .

After the reduction was complete, the material was again screened to remove small dust particles. A 12-ft . sec­ tion of copper tubing with an outside diame ter of 5/16 in . was packed with the 20-40 mesh chromia-alumina. The column was coiled for ease of manipulation . This column is referred to as the activated chromia-alumina column in future references.

The activated chromia-alumina column was deactivated with distilled water by drawing the water into the column .

Considerable evolution of heat was noticeable as the water was added . The column was placed in an oven at 140-150° for three hours to reactivate the column . Nitrogen was employed to remove the desorbed water from the column .

One hund red and twenty-five grams of the 40-60 mesh cut of activated alumina was employed to prepare a chromia­ alumina column from an 8-ft. section of copper tubing with an outside diameter of 1/4 in . The alumina to be treated was placed in a 1-1. Erlenmeyer flask and was covered with 150 ml . of water. The chromia-alumina was prepared by trea ting the 31 activated alumina with 10 g. of chromium trioxide which had been dissolved in 100 ml . of water. The mixture was allowed to stand , with occasional shaking , for three hours. The material was reduced as previously described�

Before the column was packed , the chromia-alumina packing was covered with water . The excess water was decanted , and the chromia-alumina was dried at 150° for two hours. The column was then packed and coiled . The chromia-alumina was reactivated by heating at 150° for three hours while passing nitrogen through the column.

h. Chromia-flint guartz columg . One hundred and twenty grams of 40-60 mesh flint quartz was placed in a 1-1.

Erlenmeyer flask with 150 ml� of water. Ten grams of chromium trioxide dissolved in 100 ml . of water was added to the flask .

The contents were shaken intermittently during a three-hour period . The contents were filtered, and the residue was reduced by passing hydrogen over the mixture at 360°. This material was emp loyed as the packing for a 6-ft . section of copper tubing with an outside diameter of 1/4 in .

i. Alumina column . A 6-ft . section of 1/4-in . copper tubing was packed with 20-40 mesh , Grade F-1, alumina . The column was coiled to facilitate ease of handling .

2. Cool ing Syst.ru!'!

The cooling system consisted of a 2-1. Dewar flask into which the coiled column was placed . The column was cov­ ered with the cool ing agent to obtain the temperature desired . 32

Liquid methane for use as a coolant was prepared by condensing natural gas , containing approximately 98 per cent methane , in a liquid nitrogen cold trap . The liquid nitrogen

trap was preceded by a "Dry Ice" trap to remove substances condensing at higher temperatures. The liquid nitrogen trap consisted of a 2-1. Dewar flask surrounding a 300-ml . Kjeldahl flask . The Kje ldahl flask was fitted with a one -hole rubber stopper containing a short piece of glass tub ing , which was connected to the lead from the gas line . The Kjeldahl flask was stoppered before immersion in liquid nitrogen.

The condensed methane was transferred to a small Dewar which served as a storage container. The liquid was later transferred to the cool ing bath . A small amount of liquid nitrogen was initially added to the cooling bath in order to conserve the prepared liquid methane . The tempera ture of the liquid methane bath was checked with an iron-constantan thermocouple and a Whelco portable potentiometer.

3. Heating System A 4-ft . section of Pyrex glass tubing was employed for preparing the heater for the copper tubing columns . Two heating elements o·f Nichrome ribbon were wound on to the tube .

The heating elements were operated independently by two variable transformers. The glass tube furnace was insulated with a l-in. thickness of prefabricated pipe insulation .

For the glass columns , the vapor jacket for a Victor­

Meyer gas density apparatus was wound with 13 ft . of Nichrome 33 wire with a resistance of 2.06 ohms per footo The heater was insulated with asbestos paper.

E. Detection System

1. �lnttal

Detector requirements for gas chromatography have been discussed in detail in previous publications.20,21,22 The two basic types of detectors are differential detectors and integral detectors. The thermal conductivity cell is a dif­ ferential detector. The integral type instrument provides a signal which is a function of the total amount of sample which has passed through the detector. A signal which is a function of the sample concentration in the detector at a specific time is produced by the differential type of instrument.

A detector which is satisfactory for gas chromatography should exhibit as many of the following characteristics as possible:

(1) sensitivity to small amounts of sample;

(2) rapid response;

(3) proportionality between the signal and the concen-

tration of the component;

(4) low susceptibility to fluctuations in flow rate;

c;) stability and reproducibility;

(6) simplicity; 34

(7) versatility;

(8) adaptability to automatic recording.

2. �.fllgarome ter

Detection instruments based on the principle that heat is conduc ted away from a hot body situated in a gas, at a rate depending on the nature of the gas, are known as thermal conductivity cells, or katharometers. The theory concerning the design of hot-wire katharometers has been thoroughly dis­ cussed . 20, 21 The katharometer employed for this work has previously been described. 23, 24 A cross-sectional view of the symmetrical, double-cell katharometer is given in Figure 7.

A matched pair of mounted thermis tors obtained from the Victory

Engineering Corporation was installed in the reference and sample compartments of the katharometer. The thermi stors were held in place with Sauereisen cement.

A thermistor is a more sensitive detection element than is a hot-wire element, since a thermistor has a large negative coefficient of resistance.

The thermi stors were heated by a 32-volt battery . The temperature of the katharometer was controlled by a copper cooling coil soldered to the exterior of the brass block .

Tap water or a liquid from a thermostat in a flow system wa s employed, depending upon the desired temperature . The kath­ arometer was usually operated at a temperature of 200 to 25° . 35

B W - Direction of gas flow

X - Sauereisen cement plug

Y - Approximate position of thermistor bead

Z - Silver soldered seams

A A' Approximately 2/3 full size

Top view BV z

- / , , ,•-- ---·" I I I I I I I I I I I I : I I I I I I I t I y w I I J..-l,' ---�X •- �--- Ct --.1 I 1 �.:__ ',,_.., I I I I I I I I l I i I I I I I I I I I w I I I I I I I I I I I I I I l I �- � �

Section AA' Section BB'

Figure 7. Katharometer block. 36

3. Measuring and Recording Circuit

The complete circuit for the detection system is shown in Figure 8. Two 1500-ohm precision resistors and two ma tched thermistors were employed in the Wheatstone Bridge .

The resistance of the thermi stors at the temperature employed was much le ss than that of the fixed resistances in either arm; therefore, the current through each arm of the bridge was nearly constant, regardle ss of the var iations in the thermistor resistance. For this reason, the voltage signal from the bridge was almost a linear func tion of the resistance change in the sensing thermistor.

A Brown potentiometer recorder, obtained from the

Minneapolis -Honeywell Company, was used to record the output signal from the katharometer circuit. The recorder had a full-s cale sensitivity of 5 mv ., a full-scale pen spe ed of

4.5 sec. , and a chart speed of 6 in. per hour . 37

t-1

R-4

To Brown potentiometer t-2 recorder

R-7

R-8 -7 I R-9

R- R-10

t-1, 2: A-33 matched thermistors B.-131 100 ohms (Victory Engineering Corp.) R-14: 47 ohms R-1, 2: 1500 ohms R-15: 10 ohms R-3: 125 ohms, wire-wound R-16: 7500 ohms R-4: 10 ohms R-17: 1500 ohms R-5: 75 ohms, wire-wound S-1: Bridge current ON-OFF R-6: 25,000 ohms switch R-7: 1000 ohms S-2: Fine sensitivity selector R-8: 5 ohms switch R-9: 10 ohms S-3: Polarity reversing switch R-10: 500 ohms M-1: Bridge current milli­ R-11: 4700 ohms ammeter R-12: 1500 ohms

Figure 8. Complete circuit for detection system. CHAPTER III

EXPERIMENTAL PROCEDURES AND RESULTS

Ao Introduction

For all the elution experiments, the sample was in­ jected into the column from the by-pass ce llo The sample cell had a constant volume of 4.72 ml ., and the magnitude of the sample was determined by the measured pressure . Wi th hydrogen, helium , and argon carriers, the reference exit was vented directly to the atmosphere. The carrier and sample gases proceeded through the flow-rate measuring apparatus to the atmosphere.

Hyd rogen and deuterium were removed from the neon circulatory system by passing the samples through a trap filled with copper oxide at a temperature of approximately

500° . The water formed was removed in a silica gel trap sur­ round ed by solid carbon dioxide . The flow direction could be diverted around the. copper oxide trap through the flow rate measurement dev ice and an activated charcoal trap . Consider­ able neon was adsorbed on columns wh ile in use , and the desorbed gas exerted excess pressure when the liquid nitrogen bath was removed . When inserting a different column , or dur­ ing periods when the col umn was not in use , a large portion of the neon in the system was adsorbed in the charcoal trap by immersing the trap in liquid nitrogen . During the periods 39

of inactivity, the charc oal trap outlets were tightly closed

with pinch clamps in order to store the neon . This allowed

the neon employed for one series of runs to be conserved for

future use.

B. Elution Experiments

When helium is employed as the carrier gas and hydro­

gen samples are passed through a silica gel column at room

temperature or at -78° , positive and negative deflections may

be obtained for a single sample . The nature of the deflec­

tion for a particular temperature depends upon the size of

the sample , the flow rate, and the type of column. This 25 behavior is illustrated in Figure 9 . Mad ison has indicated that this type of behavior is obtained because hydrogen and

helium mixtures exhibit a minimum thermal conduc tivity, and

at certain gas concentrations the deflection will be in one

direction, while at other concentrati ons the deflection will

be in the opposite direction.

Experiments were performed employing hydrogen, helium,

and neon as the carrier gases at the temperature of liquid nitrogen. Helium apparently failed to elute hydrogen from

the silica gel columns.

Hydrogen was employed as the carrier to elute deute­

rium through the column at 77°K. Considerable retention of 2.0 1- Temperature -78° Temperature -78° Hel ium carrier Hel ium carrier Hydrogen at 2 psig Hydrogen sample at 10 psig

""' � 1.5 1- .._, 1::: 0 •.-I � u Ql r-1 <1-1 Ql "0 1.0 1- "" (I) "0"" I \ �

g 0.5 t

o.o �------�------�------L----�------�------�------�--� 0 10 20 30 0 10 20 30

Time in minutes

Figure 9. Effect of sample magnitude on the shape of hydrogen peak .

� 0 41 deuterium and .wide separation of helium and deuterium were obtained . A typical deuterium peak is shown in Figure 10 .

A 2-ft. section of a column containing pallad ium metal deposited on flint quartz was connected directly to the 6-ft . silica gel column . The palladium column , which Thomas23 had prepared by reducing palladium chloride in a stream of hydro­ gen , contained 0.163 g. of pallad ium per foot . Hydrogen was employed as the carrier gas, and samples containing helium and deuterium were passed through the columns . The pallad ium­ flint quartz section was heated to 190- 200° , and the silica gel column was immersed in liquid nitrogen . The palladium column catalyzed the formation of hydrogen deuteride, and a mixture of deuterium, hydrogen , and hydrogen deuteride emerged from the palladium column . This mixture pass ed directly into the si lica gel col umn , and the hydrogen deuteride and deute ­ rium were detected by the thermal conductivity unit at the exit . Complete separation of the hydrogen deuteride and deuterium peaks was not obtained. A typical recorder trace is shown in Figure 11 . Mixtures of hydrogen deuteride, deuterium, and hydro­ gen were prepared in the apparatus described in Chapter II .

The 10-ft . silica ge l column was inserted into the system, and mixtures of the hyd rogen isotopes were swe pt directly into the column from the sample chamber. Complete separation of deuterium and hydrogen deuteride was obtained . This sepa­ ration is illustrated in Figure 12. The retention time for Temperature 77°K. Hydrogen carrier Fl ow rate 60 ml ./min .

1.5

......

'oJt 1:: 0 ·� +.1 () Q) 1.0 ...... 14-1 Q) '0 � Q) '0 � 0 () � 0.5

o.o 0 10 20 30 40 50 60

Time in minutes

Figure 10. Typical elution peak for deuterium with hydrogen carrier on silica gel column .

� 1\) 2. 0 Temperature 77°K. Hydrogen carrier Flow rate 68 ml ./min .

- . � 1.5 ._, 5 ...... HD u He

0.5

o.o 0 10 20 30 40 50 60

Time in minute s

Figure 11. Recorder trace for hydrogen- deuterium sample passed through palladium and silica gel columns.

.Jl. w 2. 01- Temperature 77 °K. Hydroge n carrier Flow rate 144 ml ./mi n.

1 .51 """. > e .._, s:: 0 •.-l � u D Cl) 2 ..... HD 4-1 O Cl) L r "0 A 1.< Cl) "0 1.< 0 u Cl) � o.sL ) \j

0. 0 �------�------�------�------�------��------�------� 0 1 0 20 30 40 50 60

Time in minutes

Figure 12. Typical elution curve for hydrogen , deuterium , and hydrogen deuteride sample on silica ge l column . � � 45 pure deuterium samples establishes the first peak as hydrogen deuteride and the se cond as the deuterium peak. The identity of the two peaks may also be established by adding a sample of pure deuterium to a mixture of the isotopic forms. The peak that is enlarged in relation to the other peak is desig­ nated as the deuterium peak . A stream of hydrogen was passed through the 6-ft. column, and the column was heated to 165° for eighteen hours to activate the column. The column was immersed in a liquid nitrogen bath, and samples of pure deuterium and mixtures of hydrogen deuteride , hydrogen , and deuterium were injec ted into the column . Complete se paration of deuterium and hydrogen deuteride was obtained for samples with total pressures up to

700 mm . of mercury.

The area under the pure deuterium peaks was employed for calibration plots. The area was obtained from the prod­ uct of the peak height and the wid th at half height . A deuterium calibration plot is shown in Figure 13 . Several samples, with a constant ratio of hydrogen, deuterium , and hydrogen deuteride, were eluted through the column. The partial pressure of deuterium was dete rmined from the calibration plot. From the initial and final pres­ sures of deuterium, the partial pressure of hydrogen deuter­ ide in the mixture was calculated . To obtain the hydrogen pressure in the mixture , the decrease in deuterium pressure was subtracted from the original hydrogen pressure. The 46

Temperature 77 °K. Hydrogen carrier Fl ow rate 64 ml .Jmin . 1.2

1.0

0.8

-

• s u H

• t ...... , �0 . 6 cc Q) �

0.4

0.2

o.o 0 100 200 300 400

Pressure of sample in millimeters of mercury

Figure 13. Calibration curve for the determination of deuterium . 47 results for a series of runs are given in Table I. The original sample contained 65.2 per cent deuterium.

The 25-ft . silica gel column was placed in the 2-1.

Dewar flask, and liquid methane was employed to cool the column . A small quantity of liquid nitrogen was initially added to the cooling bath . The initial temperature of the column fell considerably below that of liquid methane but began to rise as more methane was ad ded . He lium was allowed to flow through the column during the cooling process. Sam­ ples containing hydrogen and deuterium; hydrogen, deuterium, and hydrogen deuteride; as well as samples of pure hydrogen and pure deuterium were eluted through the column . Typical recorder traces for the hydrogen-deuterium and the hydrogen, deuterium, and hydrogen deuteride samples are shown in

Figures 14 and 15.

The 6-ft. silica gel column was placed in the system, and neon was employed as the carrier gas at -195° and at

-161° . At the temperature of liquid nitrogen , the neon car­ rier apparently failed to elute the isotopes from the column .

The isotopes were eluted from the column at -161° , but the overlap obtained wa s considerable . The sensitivity was much greater than that obtained with the helium carrier. A typi­ cal elution curve for hydrogen and deuterium is shown in

Figure 16 . Complete overlap was obtained for samples of hydrogen, hydrogen deuteride, and deuterium. 48

TABLE I

PERCENTAGE COMPOSITION OF THE HYDROGEN ISOTOPES

Total Pressure Partial Pressure of Sample mm. of Hg Per Cent mm . of Hg H2 D2 HD H2 D2 HD

1225 224 597 404 18.3 48 .7 33 .0

1090 164 496 430 15.0 45.5 39.4

973 150 445 378 15.4 45.7 38.8

871 144 409 318 16 . 5 47 .0 36.5 2.0 Temperature -161 ° 25-ft. silica gel column Hel ium carrier Flow rate 58 ml ./min .

1.5

...... � ._, � 0 •....l .j.J � 1.0 ..... "-! Q) '0 k Q) 'E 0 CJ � 0.5

o . o L------L------�------�------��----�=------�------20 30 40 50 60 70 80

Time in minutes

Figure 14. Typical elution curve for hyd rogen and deuterium at the temperature of boil ing methane .

� '-() 2.0 Temperature -161 ° 25-ft. silica gel column He l ium carrier Flow rate 60 ml ./min.

1.5 HD

,-... . � .._, s •.-I t! 1.0 Ql � IJ.I Ql "C ,.. Ql 'E 0 � 0.5 �

o.o �------�f------�------�------�------L------4------J 10 20 30 40 50 60 70

Time in minutes

Figvre 15. Typ ical elution curve for hydrogen , deuterium , and hyd rogen deuteride at the temperature of boiling methane .

\J\ 0 Temperature -161 ° Neon carrier Silica gel column

1.5

,...... H2 .._,� 5 ',j 1.0 () Q) r-1 II.! Q) '0 "" Q) 'e 0 () 0.5 �

0.0 �------._------�------�------_.------��------._� 20 30 40 50 60 70 80

Time in minutes

Figure 16. Typical elution peak for hydrogen and deuterium sampl e with neon carrier at -161°.

\J\ 1-' 52

2. Molecular Sieve-5A Columns

The molecular sieve-5A column was employed at room

temperature and at -78° for some preliminary experiments.

This column was also employed at the temperature of liquid nitrogen , with helium and neon as the carrier gases. Both

carrier gases apparently failed to elute the isotopes from

the column .

3. Charcoal Column

He lium failed to elute the isotopes from the charc·oal

column at -195° . When hydrogen was emp loyed as the carrier,

partial se paration of hydrogen deuteride and deuterium was

obtained . A typical recorder trace is shown in Figure 17 .

4. Raney Nickel Columns

Argon was employed as the carrier gas in all the elu­

tion experiments with Raney nickel columns. At room tempera­

ture the hydrogen or deuterium peaks were characterized by

sharp fronts and considerable tailing. The first samples passed through the column after the argon was allowed to flow overnight gave considerably smaller peaks than those obtained

with identical samples, which were injected after the recorder

had again leveled at the base line . Apparently a portion of

the hydrogen sample was very strongly adsorbed on the column.

Elution times were very much dependent upon the size of the

samples and the time betwe en runs . The retention times Temperature 77 °K. Hyd rogen Carrier Flow Rate 75 ml ./Min .

1.5

- . � -

6 1.0 ...... , (J G) ....c � G) '0 � G) '0 J.f 0.5 0 (J G) �

o.j___J__�----t ���- 0 0 10 20 30 40 50 - ----;-60

Time in minutes

Figure 17. Typical elution trace with hydrogen , deuterium , and hyd rogen deuteride sample on charcoal column .

\J1. w 54 indicated that no se paration of the isotopes was possible at room temperature .

The temperature of the column was gradually increased , and hyd rogen-deuterium samples were injected into the column at different intervals. The first elution traces, obtained for samples above 1 50° , were characterized by an initial, almost immediate , sharp peak which had slight tailing at the base . A second peak with a sharp front and tailing similar to that obtained at room temperature also appeared . The sharp initial peak became smaller and gradually almost dis­ appeared as more samples were inj ected thr ough the column .

Figure 18 is typical of a recorder trace obtained at 179° by passing the sample through the 4-ft . column .

The temperature of the column was elevated to 200° .

The first hydrogen-deuterium sample passed through the column at this temperature gave the recorder deflection as ind icated in Figure 19 . This trace was characterized by the initial sharp peak which was followed by the second tailing peak with a definite offset in the front . A sample of hydrogen and deuterium introduced forty minutes later produced only a slight deflection in the front of the second peak . At

223-228°, a slight break in the peak front was also obtained .

After the column was cooled to room temperature, hyd rogen was permitted to flow through the column for fifteen minutes. The argon flow through the column was resumed , and the column was slowly heated . The first hydrogen-deuterium sample introduc ed ;;

2.S Temperature 179° Argon Carrier

2.0

,......

•

.._,� = 0 '" +I 1.5 CJ Q) .-1 "-! Q) '0 '"' Q) '0 '"' 0 CJ 1.0 �

o.s

o . o 0 10 20 30 40 Time in minutes

Figure 18. Typical recorder trace for the elut ion of hydrogen or deuterium from Raney nickel column. 5'6

Temperature 200- 203° Argon carr ier

2 . 5

2 . 0

,...... � .,_, � 0 •.-l +l tJ Q) 1.5 ...... 4-1 Q) "0 1-1 Q) "0 1-1 0 tJ � 1.0

0 • .5

0.0 L------�------�------L------�------� 0 10 20 30 40

Time in minutes

Figure 19. Elution peak for hydrogen-deuterium sample af ter elevating the Raney nickel column temperature . 57 into the column at 233-238° produced a definite ledge on the peak front . This is illustrated in Figure 20. Here , it is noted that the initial sharp peak has disappeared . The third sample injected into the column at the elevated temperature produc ed no break in the peak . The samples were injected into the column forty minutes apart.

Samples eluted through the column at 233-238° , after allowing hydrogen to flow through the column for thirty minutes at this temperature , gave no indication of separation .

Recorder traces, similar to the one shown in Figure 20, could be repeated by cooling the column below 80° and allowing hydrogen or deuterium to flow through the column. After the column was reheated to a temperature between 200° and 240° , the same type peak was obtained for pure hydrogen or deuterium samples, as well as for mixtures of the isotopes .

The results obtained with the 6-ft . copper tubing column were similar to those for the shorter columns . A typical peak obtained for a deuterium sample passed through the column after several heating and cooling processes is shown in Figure 21 . This type of peak was obtained only for the initial sample added after the column had been reheated .

The column temperature was elevated to 255° , and deute­ rium was allowed to flow through the column for twenty minutes.

The column was cooled to 70° while deuterium was passed through the column . After reheating the column to 235-240° , the 2.0 Temperature 235-241 ° Argon carrier Flow rate 25 ml ./min.

""". 1.5 � '-J c: 0

.... .J.J u Q) r-1 � 1.0 '1:1 � Q) '1:1 � 0 u �

0.5

---L----!::--- L __ o.o __ ...J_ 7n--�;;---� 0 10 20 30 40 50 60

Time in minutes

Figure 20. Elut ion trace ob tained af ter cool ing and reheat ing the Raney nickel column . \..1\ co Temperature 226-231° Argon carrier Flow rate 48 ml ./min.

...... 1.5 .

.....,� ,:;: 0 ..... � () G) ;;::: 1.0 G) '0 "" G) '0 "" 0 (J G) r;x: 0.5

0.0 �------�------�------�------�------�------�� 0 10 20 30 40 50 60

Time in minutes

Figure 21 . Elution peak obtained with deuterium sample on 6-ft. Raney nickel column .

'-11. '-() 60

results obtained for the initial deuterium sample were similar

to that shown in Figure 20.

Samples of hydrogen and deuterium eluted through the

10-ft . copper tubing column at various temperatures between

25° and 250° gave little indication of possible separation of hydrogen and deuterium. The peaks obtained tai.led consider­

ably , and the retention volume of the subsequent sample was highly depend ent upon the time elapsed between samples . As

the temperature of the column was increased, the samples appeared to be ads orbed more strongly than at room temperature .

5. Rane_y Cobalt Col.J:!.!!!!!.§.

Equivalent gas samples were introduced into the 12-ft.

Raney cobalt column at several different temperatures. Each

sample contained helium with a pressure of 6 mm. of me rcury

to serve as a marker . A typical elut ion peak is shown in

Figure 22 . The hydrogen or deuterium peak was characterized by a sharp front with considerable tailing . Gas samples

initially introduced into the column, after the column tem­ perature was elevated, gave larger retention time s and lower peak heights than successive samples of the same magnitude.

This phenomenon indicated that some of the hydrogen or deute­ rium is held strongly by the cobalt column, and the gas is very difficult to desorb .

The retention volumes for hydrogen and deuterium sam­ ples were determined for gas samples of the same magnitude . 61

Temperature 54- 56° Argon carrier Flow rate 45. 3 ml ./min . 2.5

2.0

,-, . > 8 ._, n H2 or 2 d 0 •..l .j.l 1.5 () (I) ..-4 11-1 (I) '0 � {I) '0 '"' 0 () I.0 �

0.5

o.o 0 10 20 30 40

Time in minutes

Figure 22 . Typical elution curve f o r hydrogen or deuterium sample on Raney cobalt column . 62

Some data obtained at 54-56° are recorded in Table II. Alter­ nate runs started thirty minutes apart indicate that hydr ogen is held more strongly than the deuterium. Attemp ts to obtain partial separation by the elution method fai led to show any break in the eluted peaks.

In order to determine the effect of the time be tween runs upon the retention time of the succeed ing sample , hydro� gen samples were introduced into the column at different time intervals after a deuterium sample was injected . Each deute­ rium sample was swept into the column thirty minutes after the preceding hydrogen sample . A plot of the retention volume of hyd rogen samples versus. the time elapsed between the intro­ duction of the hydrogen and deuterium samples is given in

Figure 23 .

6. Qhromia-Al umina Columns

Neon was used as the carrier gas for all of the inves­ tigations wi th the chromia-alumina columns . The 12-ft . chromia-alumina column was first employed in a highly acti­ vated state after the reduction of the chromic acid at 360° . This column comp letely separated hydrogen , hydrogen deuteride, and deuterium at 77°K. , but the retention times were extremely long. The peaks were charac terized by sharp fronts and extreme tailing. A typical hyd rogen peak is shown in Figur e 24 .

The deactivated, or poisoned , chromia=alumina column , which was reactivated at 150° , gave much shorter retention 63

TABLE II

RETENTION TIME FOR HYDROGEN GAS ON RANEY COBA LT COLUMN

F p tHe ts ( ts - tHe ) F Run Sample ml . min . -1 mm. of Hg sec. sec . ml .

4 D2 46 .2 262 107 134 20.8 5 H2 45 ·3 262 108 155 3 5.5

6 D2 46 .2 262 108 134 20.0

7 H2 45.3 263 108 156 36.3

8 D2 45.3 261 108 13 5 20 .4 9 H2 45 ·3 266 107 157 37-7

12 D2 45.3 196 110 174 48 .4

13 H2 45.3 196 110 224 86.1

14 D2 45 ·3 196 110 175 49.1 15 H2 44 .4 196 110 221 82.1

18 D2 45.3 228 109 148 29. 5

19 H2 45.3 228 109 180 53 .6 20 D2 45.3 229 109 146 27.9

21 H2 46.2 229 109 178 53 .. 2

The flow rate is given by F, the pressure of the sam- ple by P, the retention time of helium by tHe ' and the reten- tion time of the sample by ts . 64

4.0 -

- --- 3. 0 ------

Temperature 54-56° Argon carrier at constant flow rate -- Deuterium retention times for sampl es added thirty minutes apart 1.0 � Hydrogen retention times

0.0�------�------��------��--� 0 10 20 30

Time in minutes

Figure 23. Ef fect of time between runs upon the retention time of hydrogen on Raney cobalt. 65

Temperature 77•K. Neon carrier 2.0

- . 1.5 � ...., d 0 •.-I +ol C)

0.5

0.0�------�------�------�------�------� 100 110 120 130 140

Time in minutes

Figure 24. Typical elution curve for hyd rogen sample on act ivated chromia-alumina column. 66 times. The peaks were sharper and the tailing was less evi­ dent. Almo st complete separation of the hydrogen and hydrogen deuteride peaks , and complete separation of the deuterium and hydrogen deuteride peaks were obtained . The separation ob­ tained for a typical hydrogen and deuterium sample is shown in Figure 25. Sing le peaks were obtained for samples of pure hydrogen. A recorder trace for a typical sample of hydrogen , hydrogen deuteride , and deuterium is shown in Figure 2 6 . The

12-ft . chromia-alumina column was completely filled with water.

After heating for one hour at 125° in the nitrogen flow system, samples were injected into the column at 77°K . The retention time s were reduced cons id erably, and the separation was much less than that obtained when the column was activated at 150° .

A typical elution curve is shown in Figure 27 .

The column was reactivated again at 150° for two and one-half hours. Separation was again obtained at -195° , but at -161° the peaks completely merged for hydrogen , deuterium, and hydrogen deuteride samples. The 8-ft., partially poisoned, chromia-alumina column was connected directly to the alumina column . This arrange­ ment should separate orthohydrogen and parahydrogen , as well as hydrogen deuteride and deuterium. Sample s eluted through the above column s at 77°K . gave extremely long retention times. The elution curves were very low and tailed consider­ ably . Temperature 77 °K. 2.0 Neon carrier

1.5 ,......

.....,� c 0 •rl ... t.J ! 1.0 11-1 Q) ., � Q) ., � 0 t.J �

� 0 . 0 �------�------+------�------�------�------.______80 90 100 110 120 130 140

Time in minutes

Figure 25. Typ ical elut ion curve for hydroge n and deuterium sample on partial ly

poisoned chromia-alumina column . 0' -....:1 Temperature 77°K. 2.0 Neon carrier

,.....,. 1.5 � '-" c: H 0 I 2 •...1 .j..l t.l r..._ � I 1 \ f \ o2 .-I � 1.0 "0 ""' G) "0 1-1 0 t.l Cl) �

0.5

�------L------�------�------�------� 0 . 0 �------.______80 90 100 110 120 130 140

Time in minutes

Figure 26. Typical elution curve for hydrogen , hydrogen deuteride , and deute- rium sample on partially poisoned chromia-alumina column . 0' co 2.0 Temperature 77°K. Neon carrier

1.5

,...... � I HD ._, c 0 ...... jj u Q) 1.o r- ..... H2 11-4 (I) '0 1-4 (I) '0 1-4 0 u � 0.5

o. o�------._------�------�------��------._------�� 0 10 20 30 40 50 60

Time in minutes

Figure 27. Elution trace -obtained with highly deactivated chromia-alumina column .

0" "' 70

The separation obtained for hydrogen , deuterium , and hydrogen deuteride samples with the 8-ft. chromia-alumina column is shown in Figure 28. The peaks are sharper, but the overlap is greater than that observed with the 12-ft . column.

Calibration plots, which were determined for pure samples of hydrogen and deuterium , are illustrated in Fig­ ure 29. The areas were obtained from the product of the peak width at half-height and the height of the peak . The percentage compositi ons were calculated for several samples in which the ratio of hydrogen , hydrogen deuteride , and deuterium was the same . The partial pressures of hydrogen and deuterium were obtained from the calibration plots.

The se partial pres sures were subtracted from the pressure of the original sample in order to obtain the hydrogen deuteride fraction. The results are given in Table III . The first four samples are for one series of runs , and the remainder are for a second group .

7 . Chromia-Flint Quar tz Column

The chromia-flint quartz 8-ft. column wa s employed to determine if the chromia alone were instrumental in the sepa­ ration of the isotopes. The column was immersed in liquid nitrogen , and several samples of hydrogen and deuterium , as well as samples of mixtures, were injected into the column.

The retention times for the pure samples indicated that 2. 0 Temperature 77°K. Neon carrier

1.5

,...,. . � '-' � 0 I \HD ·� +l I H2 CJ Gl 1.0 ,...., "-! Gl "0 1-1 Gl "0 1-1 0 (J � 0. 5

o.o �------�------�------4------4------��------�------� 20 30 40 50 60 70 80

Time in minutes

Figure 28. Typical elution curve for hydrogen, hydrogen deuteride , and deuterium sample on part ially poisoned 8-f t. chromia-alumina column .

...... ;) I-' 72

Chromia-alumina column

600 Neon carrier Temperature 77°K.

500

e- :;:J (,I 1-f 400 � 4-1 0 Cll 1-f Q,) � Q,) s 300 ...,j ...... •.-I s l=l •.-I Q,) 1-f :;:J 200 Cll Cll Q,) 1-f p..

100

o �------�------�------._------�------� 0 1.0 2.0 3.0 4.0 5.0

Area (mv. x em .)

Figure 29. Calibrations curves for the determination of hydrogen and deuterium . 73

TABLE III

ISOTOPIC COMPOSITION FOR EQUILIBRIUM MIXTURES OF HYDR OGEN ISOTOPES

Partial Pressures Percentage mm . of Hg P Sample ___ Composition___ P P P mmo of Hg H2 n2 HD H2 D2 HD

331 59 107 165 17.8 32.3 49 .8

170 32 57 81 18 .8 33 .6 47.6 496 96 161 239 19 .4 32.5 48 .2

140 23 49 68 16 . 4 35.0 48.6 588 90 227 271 15.3 38.6 46.1

593 85 237 271 14 .3 40 .0 45.7

438 59 177 202 13 .5 40 .4 46.1

641 89 251 301 13 .9 39. 1 47 .0

350 46 139 165 13 .1 39-7 47 .1 74 separation could not be obtained . The retention times for hydrogen and for deuterium were slightly less than four minutes. CHAPTER IV

DISCUSSION OF THE RESULTS

A. Retention Volume s

The retention volume or retention time may be employed for the qualitative identification of the components eluted from a gas chromatographic column. The retention volume , VR , is given by

VR = Ft ( 2) where F is the flow rate of the carrier gas measured at the outlet, and t is the time of emergence of the peak maximum after the injection of the sample . Equation 2 is applicable to a system in which the mobile phase is incompressible . For a compressible carrier, the flow rate is no longer constant throughout the column, and a correction factor must be applied . The corrected retention volume is given by,

V� = fVR (3) where the pres sure gradient factor, 9,21 , 26,27 f, has been shown to be equal to

(4 ) where P i is the inlet carrier gas pressure , and P0 is the outlet pressure . If F is measured at a temperature differ­ ent from that at which the column is operated , V� may be further corrected to the column temperature . 76

A further correction, introduced by Littlewood , Phil­ lips, and Price, 28 may be applied for the dead space in the column . The retent ion time , ti , is determined for a gas wh ich is not retarded by the column pac king . The fully cor­ rected retention volume is then given by ,

Pi/Po � v� = .3. ( t - t ) F � ( 5) 2 s i � 3 l � Pi/Po ) - � where t s is the retention time of the sample . The retention volumes determined on the Raney nickel and Raney cobalt columns were very much dependent upon the size of the sample and the elapsed time between the injec tion of the samples. The retention times for these column s were determined for the initial break from the base line , since the peaks had sharp fronts and long tails . Figure 23 clearly illustrate s the effect of the time elapsed between samples upon the retention volume of the sample . The retention time s also indicate that the hydrogen-deuterium sample would not lend itself to separation because of the overlap of the peaks . Possibly some adsorption sites are occupied by the sample introduced first, and the second sample does not have access to these sites. For this reason , the second sample wi ll proceed through the column at a more rapid rate .

The retention volumes obtained on the silica gel and chromia-alumina column s were affected very little by the length of time elapsed between samples and the size of the samples. The peaks were nearly symmetrical for these columns . 77

B. Separation Factor

The separation of two components in a chromatographic column depend s upon one of the components remaining at the fixed phase for a longer period of time than the other . In terms of a chromatogram of two components, the separation fac tor29 is given by ,

(6) where t1 and t2 are the retention times of the two com­ ponents, and t0 is the retention time of the marker . Melkonian and Reps30 have previously studied the ad - sorption of hydrogen and deuterium on silica gel at low temperatures and very low adsorption pres sures. These authors found that upon ad sorption , a shift occurs in favor of the lighter isotope in the gas phase. The separation factor for deuterium and hydrogen deuteride was calculated from the results obtained on the silica gel columns . The retention times were determined from the time of introduction of the sample to the peak maxima . Helium was employed as a marker to determine t . The separation factor ,�HD , obtained 0 , n2 at 77°K. , was 1.16. The ratio of the rates of diffusion is inversely proportional to the ratio of the square root of the masses . This ratio for deuterium and hydrogen deuteride is lol5, which agrees very well with the numerical separation factor obtained for hydrogen deuteride and deuterium on the silica gel column . 78

The relative separation factors for the three isotopic species were also determined for the chromia-alumina column .

Helium was employed as a marker� and the retention times were determined at the peak maxima . The gas flow rate was 30 ml . per minute at the column exit. The calculated separation factors are: lol5 for cx H HD , lo46 for o< H n and 1.27 2, 2, 2, for It should be noted that the product of o<. cx.Hn , n2· H2, HD and C:.:: is equal to c<. , · The ratio of the square Hn , n2 H2 n2 root of the ma sses for hydrogen and deuterium is 1.41 . This factor compares favorably with the value obtained for the hydrogen-deuterium separation . The ratio of the diffusion rates for hydrogen deuteride and hydrogen 1. s 1.22 , whereas the calculated separation factor is 1.15. For deuterium and hydrogen=deuteride, the separation expected by diffusion is

1 . 15, but the separation factor obtained on the chromia­ alumina column was 1.27. The separation factors for hydrogen deuteride and hydrogen and for deuterium and hydrogen deuter­ ide have numerical value s which are almost the reverse of the ratios of the square roots of the ma sses. Since other factors influence the separation obtained in a chromatographic column , the se value s would not be expected to be identical . The helium marker was possibly retarded to some extent by the column at this temperature . The appearance of the marker peak was approximate ly ten minutes after the injection of the sample o Orthohydrogen and parahydrogen were not separated on the chromia-alumina column . The chromium oxide catalyst has 79 previ ously been shown to be a very effective catalyst for

the ortho-para conversion at low temperatures . 31, 32 Grilly33

employed a chromia�alumina catalyst for the preparation of

85 per cent parahydrogen by catalyzing the conversion in the

feed gas at 76°K . and in the liquid phase within the lique­

fier . Apparently the orthohydrogen and parahydrogen inter­

conversion occurs rapidly enough on the column to prevent

separation .

Al though the conversion of hydrogen and deuterium to hydrogen deuteride occurs at higher temperatures on chromia

catalysts, no evidence of the exchange reac tion wa s observed

in this work . Voltz and Weller34 have indicated that the

hydrogen-deuterium conversion is completely inhibi ted at -78 °

by water ad sorption equivalent to 15 per cent coverage . Even

at the lower temperatures� this phenomenon would be less pro­ nounced without the poisoning of the chromia-alumina column with water . The recorder pen returned completely to the base

line between the deuterium and hydrogen deuteride samples. This indicates that the hydrogen=deuterium conversion was not occurring . The chr omia=flint quartz column did not produce sepa­ ration of the isotopes. The surface area of the chromia deposited on the flint quartz was apparently much smaller

than the area of the chromia�alumina surface . In preparing the alumina column , the water solution of chromic acid was ad sorbed into the many small pores available . Thi s left a 80

much larger surface area of chromium oxide after the reduc­

tion with hydrogen than was obtained on the flint quartz, which probably had no small pores available .

Although a separation factor other than one wa s indi­

cated by the lead ing edge of the samples eluted from the

Raney cobalt column , the extreme tailing exhibited by these

samples failed to produce any notable separation with the

mixtures . The tailing� which occurred , possibly could be

eliminated to a greater degree by a thinner surface layer"

In preparing the Raney cobalt and Raney nickel, attempts were

mad e to remove only the surface aluminum .

Differences in the rate of adsorption of hydrogen and

deuterium on nickel catalysts have been previously shown . 3 5,36

Eluted samples failed to show any separation of the isotopes.

The break in the peak front , which wa s obtained for the sam­

ples eluted through the Raney nickel columns , may have re­

sulted from some species previously adsorbed on the column .

The ease of the desorption of these species may have been increased by the heat treatment process. The occurrence of the isobaric desorption-read sorption of hydrogen on reduced nickel catalysts has previously been discussed by Sadek and

�'7 Taylor . ....J 1 81

C. Column Efficiency

The over-all separation achieved with a column depends upon two things, the separation fac tor and the number of theoretical plates. The column efficiency is usually ex­ pressed in terms of the number of theoretical plates. The following procedure is recommended38 �4l for calculating the number of theoretical plates. Tangents are drawn to the peak at the points of inflection. The length of the base , 6T, cut by these tangents can then be measured . The retention time , tR , is determined from the start of the run to the center of the base line section. The number of theoretical plates is then given by equation 7,

(t ·)2 n = 16 \A� (7) The theoretical plate is an abstract term with no physical significance other than the measure of the relative

var' iance of the peak width and retention time . The theoret- ical plate number may vary with the compound as well as the column . The number of theoretical plates is also influenced by the magnitude of the sample , and this quantity usually increases with longer retention times. The number of theo­ retical plates was calculated for the 10-ft. silica gel column and the 12-ft. chromia-alumina column . The values obtained for the silica gel column were 1930 for hydrogen deuteride and

930 for deuterium. For the 12-ft . chromia-alumina columns , the 82 value s were 1740 for hydrogen, 1150 for hydrogen deuteride, and 1160 for deuterium .

There are some serious limitations for calculating the number of theoretical plates. Since most bands are asymmet­ rical in gas-adsorption chromatography, the peak maximum is no longer a true measure of the retention volume . The peaks observed were slightly asymmetrical on the chromia-alumina columns . Even though the number of theoretical plates may not be truly represented by the determined values, a large value of tR/� is de�irable . A second limitation is that, in practice , it is not possible to introduce the total sample into the first plate of the column. Small sample volumes tend to eliminate this effect.

The hei ght equivalent to a theoretical plate may be obtained by dividing the column length by the number of theo­ retical plates. The plate theory is a very useful me thod for evaluating column efficiency, but it does not help the inves­ tigator to decide how to operate the column so as to obtain the optimum efficiency. Deempter, Zuiderweg, and Klinken­ berg42 have discussed the effects of several column and operational parame ters upon the height equivalent to a theoretical plate . 83

D. Analytical Application

The separation of small samples of hydrogen , hydrogen deuteride , and deuterium mixtures into separate bands permits a simple analytical analysis of the isotopic mixtures . Cali­ bration plots for pure hydrogen and pure deuterium were deter­ mined und er the same operational conditions employed for the analysis . These plots are shown in Figure 29. The results obtained on the chromia-alumina columns , for mixtures con- taining the same ratio of the isotopes, are given in Table III .

The equilibrium constant for the isotopic reaction may be calculated from partition functions for the temperature of the equilibration . The temperature of the Nichrome wire em­ ployed for preparing the mixtures was estimated to be between lOOoOK. and 1200"K.by comparing the color of the wire wi th a color chart. The equilibrium constants calculated at 1000°K . and 1200DK. were 3.89 and 3.94, respectively .

The equilibrium constant for the isotopic reaction is given by,

( 8)

Equation 8 may be transformed into,

2x)2 K = ( 9 ) (PH2 - x) (Pn2 - x) where x is the decrease in the hydrogen or deuterium pres- sure . Equation 10 is obtained by rear ranging equat ion 9:

(10) 84

The average experimentally determined composition and the calculated compositions for K equal to 3.8, 3.9, and

4 .0 are given in Table IV. The variation of the theoretical isotopic percentages over the range 1000°-1200�. is within experimental error for the assumed equilibrium constants.

For this reason, it was unnecessary to determine the exact temperature of the wire employed for equilibration . The standard deviations for the experimentally determined per­ centage compositions are also shown in Table IV.

The separation of deuterium and hydrogen deuteride on a silica gel column allows the determination of the isotopic composition. This method is hindered by the smallness of the recorder peaks . It was also impossible to determine directly the amount of hydrogen in the sample . An analysis is espe­ cially limited by this method for mixtures of unknown composi­ tion. Calibration plots would have to be determined for known mixtures of deuterium and hydrogen deuteride . The amount of deuterium and hydrogen deuteride would then be determined for the sample . Assuming an equilibr ium mixture , the hydrogen content can be calculated from the equilibrium constant .

Other method s emp loyed for the determination of small gaseous mixtures of hydrogen and deuterium include ma ss 4 44 spectrometry, 3 , , 45 thermal conductivity, 46 effusiometry,47,48 and optical spectroscopy . 49 The total hydrogen and deuter ium content of a sample may be determined by the latter three methods. However, these procedures fail to indicate the 85

TABLE IV

EXPERIMENTAL AND THEORETICAL COMPOSITION OF THE ISOTOPES

--

Per Cent ComQosition Ex peri-

Experi- Theoretical_ mental Isotope mental K=4 .0 K=3 ·9 K=3:8 C)

Sample 1:

H2 18.1 17.5 17 .7 17.8 1.14 D2 33 .4 33.7 34 .0 34.1 1.07 HD 48 .6 48 .7 48 .4 48 .0 0.81

Sample 2:

H2 14 .0 12.3 12 .4 12 .6 0.67 D2 39 .6 42.1 42 .3 42 .4 0.64 HD 46 .4 45.6 45.3 45.0 0.55

NOTE : cr is the calculated standard deviation for the results presented in Table III. 86 hydrogen deuteride content of the sample . The fraction of the sample which is hydrogen deuteride may be determined by mass spectrome try . The results obtained by this method are quite often difficult to interpret, especially for samples in which more complex gases are present . The presence of extraneous gases should produce no difficulty with the gas chromatographic method . The eluted gases should either be widely separated or completely adsorbed on the column . If he lium were present in a large proportion in the sample , some difficulty might be encountered , since this species is con­ tinually circulated , and peaks would be observed with each passage through the katharometer. The helium sample would gradually diffuse throughout the column .

E. Comparison with Previous Chromatographic Results

The results obtained by Ohkoshi , Fujita , and Kwanl ? on a Lind e molecular sieve-?A column suffer from the same limitations discus sed for the silica gel columns in the previous section. In the separation obtained by Van Hook and Emmett, l4 complete overlap of the hydrogen deuteride and the orthohydrogen peaks was obtained . The area measurements allowed the determination of the isotopic composition from a knowledge of the orthohydrogen and parahydrogen ratio in equilibrium. The porti on of the orthohydrogen-hydrogen deuteride species attributed to each sample was then deter­ mined from a calibration plot for parahydrogen. By eluting 87 the samples of the same isotopic composition with hydrogen and neon carriers, the composition of a sample in which the orthohydrogen-parahydrogen ratio was unknown could be determined .

Greene50 has employed neon as the carrier in separa­ tion attempts on silica gel, charcoal, and aluminum wi th the column immersed in liquid oxygen. The results ind icated that or tho- and parahydrogen were well separated, but hydrogen deuteride could not be completely separated from deuterium.

These results are in considerable disagre ement with the sepa­ ration of hydrogen deuteride and deuterium obtained in this work. The hydrogen peak wa s slightly overlapped with the hydrogen deuteride peak, but complete separation of hydrogen deuteride and deuterium was obtained . The separation of or thohydrogen and parahydrogen was not ob tained on the chromia-alumina columns. The chromia effectively catalyzed the conversion of the spin forms as the isotopes passed through the column ; therefore, at any instant an equilibrium mixture of the spin isomers should exist.

Although Thomas and Smithl2 obtained partial resolution of hydrogen and deuterium on a pallad ium col umn , the results are of little analytical value , since conditioning of the column was necessary before each introduction of the isotopes.

The peaks were characteri zed by overlapping and considerable tailing. CHAPTER V

SUMMARY

The feasibility of adsorption columns containing silica gel, Raney nickel, Raney cobalt, charcoal , molecular sieves-5A, and chromia-alumina has been investi gated for the separation of hydrogen � deuterium, and hydrogen deuteride by gas-elution chromatography. Complete separation of hydrogen deuteride and deuterium was obtained on the silica gel column with hydrogen as the carrier at 77°K. Partial resolut ions of hydrogen and deuterium and of hydrogen , deuterium, and hydro­ gen deut eride were obtained at -161° with helium as the car­ rier gas. The peaks were small, and a large sample was neces­ sary in order to produce a sufficient deflection with helium.

Neon failed to elute the isotopes from the column at 77°K. , and very little separation was obtained at -161°.

Argon was employed as the carrier gas with the Raney cobalt and nickel columns . The retention times obtained on Raney nickel columns were greatly influenced by the size of the sample and the time between runs. Apparently some hydro­ gen was irreversibly ad s orbed on the nickel column . Raney cobalt column s gave different retention times for hydrogen and deuterium passed through the column at the same interval between runs. The sample tailing wa s considerable, and the overlapping of peaks did not permit the separation of the isotopes. 89

Helium failed to elute the isotopes from the charcoal and molecular sieve- 5A column at -195° . Partial separation of deuterium and hydrogen deuteride was obtained on the charcoal column at -161° .

Complete separation of hydrogen � deuterium, and hydro­ gen deuteride was obtained on a 12-ft . chromia-alumina column at 77°K . Neon was employed as the carrier gas in a circula­ tory system. Tailing wa s very pronounced on the column when it had been activated at 360° . The tailing was reduced , and the sharpness of the peaks was increased by poisoning the column wi th water and reactivating at 150° for three hours.

The chromia catalyzed the interconversion of orthohydrogen and parahydrogen and thus prevented the separation of these species.

An analytical method was developed for determining the composition of the isotopes present in unknown mixtures .

Calibration plots were obtained for pure samples of hydrogen and deuterium. Isotopic compositions were determined for equilibr ium mixtures . The experimental results agreed very well with the composition calculated from theoretical considerations .

The peaks which were obtained for the isotopes on the chromia-alumina column were completely overlapped at -161° .

Further deactivation of the column wi th water decreased the separation obtained at -195°. 90 The separation factors obtained on the 12-ft. chromia­

alumina column are :

hydrogen deuteride . The separation factor determined for

deuterium and hydrogen deuteride on the silica gel column was

= 1.16. Q(HD ,D2 BIBLIOGRAPHY BIBLIOGRAPHY

1. M. Tswett, Ber. deu�. botan . Ges ., 24 , 316 , 384 (1906) .

2. R. Kuhn and E. Lederer , Naturwissenschaften , 12, 306 (1931) .

3. R. Kuhn and E. Lederer , Ber. , 64 , 1349 (1931) .

4. A. J. P. Martin and R. L. M. Synge , Biochem. �. , 32, 1358 (1941) .

5. R. Consden t A. H . Gordon, and A. J . P. Mar tin , �i ochem. �. , la, 224 (1944) .

6. N. C. Turner , Natl . Petroleum Ne� , 32, R-234 (1943) .

7 . N. C. Turner, Petrol. Refiner , 22, 140 (1943) . 8. s. Claesson, Arkiv . Kem1 . Min&£al Geol., 23A , No . I (1946) .

A. T. James and A. J. P. Mar tin , Biochem. �. , 2Q, 679 (1952) .

10 . E. Glueckauf , K. H . Barker < and G. P. Kitt, Discussions Faraday Soc. , 2, 199 (1949J .

11. E. Glue ckauf and G. P. Kitt , in D. H. Desty, "Vapor Phase Chromatography," Butterworths Scientific Publica­ tion, London , 1957, pp. 422-427. E. Glueckauf and G. P. Ki tt, in the "Proceedings of the International Symposium on Isotope Separation," Interscience Pub­ lishers, Inc., New York, N. Y., 1958 , pp. 210-226 .

12 . C. o. Thomas and H. A. Smith, l· Phys . Chem. , Q}, 427 (1959) .

13 . W. R. Mo ore and H. R. Wa rd , J. Am . Chern . So£. , 80 , 2909 (1958) .

14 . W. A. Van Hook and H. Emmett , private communicati on of work to be published .

15. S. Ohkoshi, Y. Fuj ita, and T. Kwan , Bull . Chem. Soc. �apan , 31, 770 (1958) .

16. J. Chadwick, "A Palladium Column for Concentrating Tritium from 2-Litre Mi xtures of Tritium and Hydrogen ,n AERE I/M-47 , Berks , England , 1958 . 93

17. United States Atomic Energy Commission, U. s. Patent 2,863 ,526 (Dec . 9, 1958) . 18. P. L. Gant and K. Young , §Ql§n£§ , 129, 1548 (1959 ) .

19. A. I. M. Keulemans , "Gas Chromatography ," Reinhold Publishing Company , New York, N. Y. , 1957, pp . 56-57.

20. D. H. Desty, "Vapor Phase Chromatography," Butterworths Scientific Publication , London , 1957.

21. A. I. M. Ke ulemans , "Gas Chromatography ," Reinhold Pub­ lishing Company, New York, N. Y. , 1957.

22. R. L. Pecsok, "Principles and Practice s of Gas Chroma­ tography ," John Wi ley and Sons, Inc., New York, N. Y. , 1959 .

23. c. 0. Thomas , Doctoral Dissertation, The University of Tennessee , Knoxville , Tennessee , 1958.

24. C. o. Thomas and H. A. Smith , �· Chern . Ed ., lQ, 527 (1959 ) .

25. J. J. Madison , Anal . Ch�. , lQ, 1859 (1958 ) .

26 . J. F. Young in V. J. Coates, H. J. Noebels , and I. s. Fagerson, "Gas Chromatography," Academic Press, Inc., Publishers, New York, N. Y. , 1958 , pp . 15-23.

27. C. Phillips, "Gas Chromatography ,. .. Academic Press, Inc., Publishers, New York, N. Y. , 1956 , p. 97.

28. A. B. Littlewood , c. S. G. Phillips, and D. T. Price, l· �. So�., 1480 (1955) .

29. V. J. Coates, H. J. Noebels, and I. S. Fager son, "Gas Chromatography ," Academic Press, Inc., Publishers , New York, N. Y., 1958 , p. 315.

30. G. A. Me lkonian and B. Reps , �. ElectrQchem. , �' 616 (1954 ) .

31. D. s. Chapin and H. L. Johns ton, �. Am . �. So£., 22, 2406 ( 19 57) •

32. Mme . F. Botter and G. Dirian, Com-m. energie at m� .U:Z:sm£.§.1 , Rappt . No . �.2 (1956). Q . A_. , 3!, 1 6201 ( 1957) .

33 - E. R. Grilly , Rev . §9! · Instr ., 24 , 899 (1953 ) . 94

S. E. Voltz and s. Weller, l· Am. Chem. Soc., 22, 5231 ( 1953 ) .

35. R. Klar , Naturwissenschaften , 22, 822 (1934).

36. S. Iizima, Rev . PhlS · Chern., Japan , 12 , 83 (1938) .

37. H. Sadek and H. s. Taylor, l· Am. Ch�. Soc., zg� 1168 (1950) .

38. V. J. Coates, H. J. Noebels � and I. S. Fagerson� "Gas Chromatography," Ac ademic Press, Inc., Publishers , New York, N. Y., 1958, p . 316.

39. D. H. Desty, "Gas Chr omatography" " Academic Press , Inc., Publishers, New York, N. Y. , 195� , p. xi.

40 . A. I. M. Ke ulemans , ''Gas Chromatography , tt Reinhold Pub­ lishing Corporation , New York, N. Y. , 1957, p. 113 .

4L H. W. Johnson and F. H. Stross, Anal . Chern. , lQ, 1586 ( 1958) .

42 . J. J. Van Deempter, F. J. Zuiderweg, and A. Klinkenberg, Chern. Eng . Sci. , 2, 271 (1956).

43 . W. Bleakley and A. J. Gould , Phys . Rev. , 44 , 265 (1933). 44 . I. Kirshenbaum in "Physical Properties and Analysis of Heavy Wa ter ," Vol . 4A of nNational Nuclear Energy Series III," edited by H. C. Urey and G. M. Murphy , McGraw-Hill Book Company , Inc., New York, N. Y., 1951 .

R. B. Alfin-Slater , S. M. Rock, and M. Swislocki, Anal . Chern. , 22, 421 (1950). 46 . N. R. Trenner , �. Ch�. Phys ., 2, 382 (1937) .

47 . 1. E. Line, Jr ., B. Wyatt, and H. A. Smith , �. Am . �· Soc., 21, 1808 (1952) .

48 . D. A. Lee, Anal . Chern. , lQ, 1296 (1958).

49 . H. P. Broida and G. H. Morgan , Anal. Chern. , 24 , 799 (1952) .

50 . S. A. Greene in R. 1. Pecsok , "Principles and Practices of Gas Chromatography ," John Wiley and Sons , Inc., New York , N. Y. , 1959 . VITA

Paul Payson Hunt was born in Elkview, We st Virginia, on December 17, 1930. He attended Aarons Fork Elementary

School and was graduated from Elk District High School in

May 1949.

The following September , he enrolled at Glenville

State College in Glenville, We st Virginia, where he received the Bache lor of Ar ts and Bachelor of Science degrees in June

1954 . From January 1954 until May 19 55, he taught chemistry and science at Montgomery High School , Montgomery, We st

Virginia.

He entered the Graduate School of The University of

Tenne ssee in June 1955, holding a graduate assistan tship.

In August 1958 , he received the degree of Master of Science in Chemistry . Since September 1958 , he ha s been a research assistant on a contract between the United States Atomic

Energy Commission and The Un iversity of Tennessee.

He is co-author of a publication, ''The Kinetics of the

Esterification of the Cyclohexanedicarboxylic Ac id s with

Diphenyldiazomethane ,11 l· Am . Chern. Soc., 81 , 590 (1959) .

He is a member of the American Chemical Society and the Society of Sigma Xi .

He is marr ied to the former He len Mar ie Brammer and has two sons, Daniel Payson and Larry Steven .