[COSTRIBUl'IiJS FROM TI113
I ~ V I S I O XOSF
1'HYSICS A X D ~I(JP11YSICSA S D OF STERnIn ~ I O C H E M I S T R Y , SLOAX-KETTERING 1XSTITI'I.E FOR CASCER l?ESB;.ARClf
1
Equilibrium between Protium, Deuterium and Tritium in the System : HydrogenAcetic Acid ; Isotopic Fractionation Factors in a Catalytic Hydrogenation nu
i r A X I v E L I , LEIGII
EIDIXOFF, JOSEPH
E.
KNOLL, I>.\VID
K . ~ r - ~ r s n Ar N~D . T. - ~F. G:ILL.\GHER'
RECEIVED MAY 12, 1952
+
Iwtopic distrilmtion coeflicients i l l the equilibria at 33" of: 11: + AcOIT' ( l i -+ 1-113' AcOI-I (1) have been measured. where 11' refers to deuterium : L I I ~ tritiuni. Equilihrinm constants were calculated for the corresponding gas reactions using these isotopic forms of acetic acid. This isotopically equilibrated hydrogenation system wah used iii the presence of platinum catalyst t o saturate the double bond in methyl 3u-acetoxy-~1L-cholenate. The isotopic composition of the hydrogen atoms incorporated during the hydrogenation was measured relative t o the isotopic composition of the carboxyl hydrogen of the acetic acid medium. These restilts indicate that the exchange between dissolved and carboxyl hydrogen a t the catalyst surface is rapid relative to the hydrogenation steps. Relative to the hydrogen gas composition, the heavier isotopes are preferentially incorporated into the conipound undergoing hydrogenation by a factor of 1.10 and 1.26 for deuterium and tritium, respectively. Relative to the carboxyl hydrogen composit ioii, protiuin is preferentially incorporated hy a factor of 3.31 and 5.12 with rcqpect to cleuterium and tritiuni, respectivelv.
Introduction The catalytic hydrogenation of an unsaturntd conipound is ;1 frecjue~itly used method for the preparation of compounds containing tleuteriuni 2nd tritium. The hydrogenating systrni : hvtlrogen gas-acetic acid-platinum catalyst is '111 e y cellent medium for such syntheses. The relationship between the isotopic composition of the hydrogenating medium and that of the hydrogen :itoms incorporated into the coinpound that was reclticed is of interest from the synthetic viewpoint as well as from the viewpoint of the isotopic irnctionation factors involved. In order to provide iiiformation on these relationships, the hydrogen isotopic equilibrium has been studied at 25" i n the system hydrogen-acetic acid using the isotopes of misses 1, 2 and 3 and the isotope distribution coefficients measured. The isotopic fractionation factors for these isotopes were also measured in the hydrogenation of a steroid, methyl 3cu-aceto.ul;2,11-cholennte(I), to form the correymntling cholaii,Lte (TI).
I
C following the isotopic atom fraction refers to the hydrogen o f o m s itmrporcited during the hydrogenation of the unsaturated compound ( I ) . Analytical Methods.-Relative tritium activities were measured by internal gas counting in which radioactive hydrogen gas ( a t approximately 5 cm. pressure) was admixed with 65-cm. methane gas. This gas mixture, inside a counter tube of about 18 mm. inside diameter with silvered cathode and 2 mil tungsten anode wire, was counted in the upper portion of the proportional region using a pulse amplifier and scaling circuit, the discriminator of which \vas set to accept all pulses greater than one millivolt. Csing this counting gas mixture, the characteristic curve had a plateau starting a t 2900 volts, and extending for approximately 1000 volts with a slope of less than 0.3% per huntiretl volts. The counting rate is directly proportional to the partid pressure of the radioactive hydrogen gas and is independent of the methane gas pressure. At methane partial pressures of 20 a n . , the plateau is shortened to 400 volts. The use of internal gas counting insures reproducible geometry. Resolving time corrections were made using the method described by Reid, h'eil arid Dunning.2 A half-life of 12.1 years was used to correct all tritium activities for natural decay. 3 The meawrement of relative tritium activities h y this procedure is precise to within 1%. Water containing tieutrrium or tritium was first converted to hytlrogcsn g:ii over zinc a t 420" in a good vacuum. The act i v i t y of the hydrogen gas wits then measured as described xbove. The isotopic analysis of the hydrogenated organic compound w a s made after complete combustion in a stream of dry oxygen with the resulting water collected in a tr'ip cooled t o -70". The water was then converted to hydrogen / , over zinc. Memory effects were rliminated b y discarding the product of the first conibustioii--converi i t r r i procedure i i i i t i checking the results of a secoiid a i d third run. The latter inexsureineiits were equal 11 within 2%. Deuterium measurements weremade using a dual collector S i e r type hydrogen mass spectrometer. Hydrogen gas for the latter wus obtained from heavy water and the deuteriutn-contuiniiig compound i i i :t manner similar to t h a t stated above. The relin1)ility of the deuterium measurements is l-2yo of tlie measureti deuterium content. The tritiiun countiiig rates, corrected to a staritlard partial pressure inside the Geiger--hfiiller counter tube used in all the iiieiisurements are directly proportional to the (T,") ratio iii the hydrogen gas or organic compound. The mass spectrometer data yield ( D / H ) rntios directly since the (mass .?/mass 2 ) current ratio is compared with ratios for hydrogen samples made from heavy water samples of known deuterium content. Sorm:,l tlruterium ahunrlancr was taken as 0.020%. ,/
I
Experimental Expression of Isotopic Ratios.-The isotopic concentrations of cleuteriurn ;tnd tritium are expressed as atom ratio.: ( D / I T ) and ( l / H ) . Thus, (I),") refers t o the ratio of the riuinber of atoms of mass 2 to mass 1. The subscripts (>ancl hf following these atoms ratios refer to the gas phase over the solution and to the hydrogen atoni i n the ctirbosyl g r o u p of the acetic acid, respectively. Only this hytlrogeii atom is of interest in these studies because the methyl group hydrogen atoms do not exchange with the gas under the experimental conditions. The concentrations ( D / I I ) c and (T!Il)o were measured clirectly in the equilibrium studies while (D,.")M anfl (T 'I$)>, were krioivn. The \Libscript ..
( I ) This investigation was jointly supported b y tlie Office of Naval Research contract Are. SC-ori-99, T . 0 . 1, the Atomic Energy Commission, a n d grants f r o m the Ziulionul Cnricrr I n s t i t u l r , I;niterl S1nte.i l'ublic Health Svr\,irv.
(21 A. I?. Keiil, A . S. n'eil a n d J . R . niinriing, A m l . Chcni., 19, 821 il947j. (31 A. Novick. Piivc. Re;,.,72, 972 (1!)47)
Nov. 5 , 1952
PROTIUM,
DEUTERIUM AND
TRrrru&f IN
Reagents.-Methyl 3a-acetoxy-Ail-cholenate ( I ) , was further purified by treatment with Raney nickel i n ethyl acetate solution and filtration through Celite. The product melted 118-119" (reported4 116-117"), [ a I Z 5 +53.5" ~ (acetone) (reported5 52.2"). Platinum oxide catalyst (Baker and Co., Newark) was used without further purification. Heavy water (99.8%) and deuterium gas (99.6%) were obtained from the Stuart Oxygen Co. Oxygen was rcmoved from tank hydrogen (Matheson Co.) by passage through a tube containing active platinum catalyst, trapping thc water formed a t Dry Ice temperature. Radioactive hydrogen gas was obtained from the Isotopes Division of the U . S. Atomic Energy Commission. Gas of lower activity was prepared by serial dilution of this material with oxygen-free tank hydrogen. T h e radioactive water used to prepare the acetic acid medium was prepared from hydrogen by reaction with copper oxide a t 350". The acetic anhydride (Eastman Kodak Co., 99-100%) was distilled in a n all-glass 20-plate column packed with glass helices, and a middle fraction boiling at 140" at one atmosphere was used. Titration with sodium methylate and sodium hydroxide following the procedure of Smith and Bryant6 showed t h a t the acetic anhydride was free of acetic acid within 0.2%, the experimental error in this procedure. The abbreviated formulas AcOD aud AcOT refer to acetic acid containing deuterium and tritium in the hydrogen of the carboxyl group while the methyl hydrogens are of normal isotopic abundance. (D," )M was approximately 0.03 and 1.0 for the equilibrium studies aiid hydrogenation experiments, respectively. (T/H)Mwas approximately and respectively. The acetic acid medium was prepared by adding water to acetic anhydride in equimolar proportions. The isotopic composition of the carboxyl hydrogen of the acetic acid was thus equal to that of the water used in its preparation. Measurement of Isotopic Equilibrium (Table I).-The glass-stoppered flask containing acetic acid, hydrogen gas and platinum catalyst (0.35, 8 X IOM3,and 1.63 X moles, respectively) was shaken vigorously in a thermostatic bath for approximately three hours. The hydrogen gas in isotopic equilibrium with the liquid was freed from acetic acid vapor by passage through a trap cooled in liquid nitrogen and the deuterium or tritium content measured as described above. Thc isotopic composition of the carboxyl hydrogen of the acetic acid was practically constant throughout the shaking period, since the acid is present in large molar excess. I t s composition a t equilibrium was calculated from the known initial composition by correcting for the uptake of deuterium or tritium by thc gas phase using the final isotopic composition of the gas and its volume a t the measured temperature and pressurc. This correction is of the order of 0.2%. An additional correction was made for the dilution of the isotopic content of the medium caused by the ordinary water, associated with the small quantity of platinum oxide catalyst used, the formula of which was considered to be PtO,. 1320. This correction is of the order of 0.2%. Measurement of Isotopic Fractionation Factor (Table 11). -Approximately 21, 0.5 and 0.04 g. of the aceticacid, steroid I and Adanis platinum oxide catalyst, respectively, were shaken for 2.5 hours. This corresponds t o molar proportions of acid, compound and catalyst equal to 3 0 0 , l and 0.14, respectively. Approximately 0.006 mole of hydrogen gas wcrc used in each experinicnt corresponding t o a n approximatel>. 4.5-fold cxcess. Thc hydrogenation reaction was cornplcted in approxiIiiatcly 13 niinutcs, corresponding to the theoretical uptake of hytlrogcn gas within the experimental error of one per cent. The uptake was uniform for the first seven minutes, corresponding to 75% completion and was slower in the remaining six minutes. At the end of the hydrogenation, the contents of the reaction flask were brought t o Dry Ice temperature, the hydrogen gas was removed and the acetic acid containing deuterium or tritium was removed b y distillation in Vacuo. Residual quantities of the latter were removed by successive additions of ordinary acetic acid followed b y vacuum distillation. The product I1 was freed of catalyst by filtration
+
(1) J. Press and T. Reichstein, Hclu. Chim. A c f o . '26, 878 (1942). ( 5 ) E.Seebeck and T. Reichstein, i b i d . , 16, 536 (1943). ( 6 ) D.M .Smith and W. M. D. Bryant, THISJOURNAL, 68, 2452 (lcJ3G).
HYDROGEN-ACETIC A C I D SYSTEM
5ZSl
of the solution in acetone and purified by two fractional rccrystallizations from petroleum ether. The melting point of the purified I1 was 134-135". The tetranitromethane test for unsaturation gave uniformly negative results; [a]2 5 +48.1" (acetone), (reported6 1n.p. 134"), [ a ] 1 5+48.4". ~ I n preparations A, B, E and F of Table 11, the reaction gas was first isotopically equilibrated with a large excess of acetic acid of isotopic composition equal to that used in the subsequent hydrogenation. The purpose of the prior equilibration was to minimize the effect caused by the changing isotopic composition of the gas phase during the hydrogenation. For preparations E, F and G, the (D/H)M was approximately 0.5 in order to furnish D / H ratios for the hydrogel1 from the combustion of I1 of approximately 0.01, a preferred concentration range for the mass spectrometer measurements. The acetic acid medium for preparations A to I) contained approximately atom fraction of tritium. In preparation C and G , the gas used had been equilibrated for a shorter time than would correspond t o isotopic equilibrium. In preparation C, the gas used for the hydrogenation had attained slightly less than of the equilibrium T / H ratio. In preparation G, the gas compositioii was not measuretl. I n preparation D, the hydrogen was of normal isotopic abundance a t the start. The isotopic composition of the carboxyl hydrogen of the acetic acid medium was calculated as described above. The isotopic compositioii of the hydrogen atoms incorporated into I1 during the hydrogenation was calculated from the measured ( D / H ) and ( T / H ) coniposition of the hydrogen gas obtained from the combustion of I1 using its molecular formula and the assumption that twq hydrogen atom positions were enriched in D or T while the other positions have the normal abundance of deuterium and no tritium. In order to determine whether the methyl group of the acetic acid exchanged hydrogen during the reaction, silver acetate was prepared by treating a portion of the acetic acid medium after the hydrogenation. Tritium analyses of the hydrogen from the combustion of the silver acetate showed that such exchange did not take place t o within thc experimental uncertainty of 0.2 %.
Results
A. Isotopic Equilibrium in the System Hr Acetic Acid.-The approach to isotopic equilibrium in the reaction: H2 AcOT(1) HT AcOH(1) is shown in Fig. 1. The lower curve refers to an experiment in which tank hydrogen was shaken with AcOT containing a large molar excess of acetic acid relative to hydrogen gas. After two hours, the radioactivity of the hydrogen gas reached a constant value. The equilibrium position was approached in the opposite direction by starting
+
+
0.
z
\
0
START WITH TANK H2+ EXCESS AcOH. A$
0
START WITH HT, H2 t
'I
I1
ll
\
a:
\ \
TIME IN HOURS.
Fig. 1.-Isotopic
composition of hydrogeu gas ( k T J H ) as function of shaking time.
~
with hydrogen gas containing more tritium than the equilibrium concentration (upper curve). The approach to isotopic equilibrium iii the rwctions
+ A~oL)(I) I& + AcOII(1)
ii2
+ ACOI-I(I~ FIT + AcOH(1j
€11)
(ai
(bj
The isotopic cxchange coeflicient ( a k ~ , for ~ ) the reaction (b) is similarly defined as: CYH,T= (T,'")hf /(T/H)Gwhere (T/H)hI and (T/H)G are the atom ratios in the carboxyl hydrogen and the hydrogcii gas, respectively. The average of the three experiments in Table I U yields a value of 6.83; (3.~1. i O . O " j ) at 2'4.3 + 0.5O.
\vas sludied simultaneously by starting with a
system of tank hydrogen and acetic acid containing atom fractions of deuterium and tritium equal to 0.03 and respectively. The gas isotopic composition relative to equilibrium is a measure of the approach to isotopic equilibrium and is plotted as ordinate in Fig. 2 . The fractional approach to isotopic equilibrium as a function of time is not significantly different for the hydrogentritiuin and hydrogcn-tle~~teriilllicscllangc.
TABLE I1 ISOTOPIC F R A C T 1 m U " I U N 1'ACTUK HYDKOGESATIOK AT 25 i 0.5"
&IEASURBMENT O F Preparation
E 1'
G
C I)
25" c I
z
'
H2 t A c O T e H T t AcOH H2
+ A c O D e H D t AcOH
PHASE ACOH, AcOT, AcOU (Appror 3 Atom%)
:n
0
in
3
I
05
I 15
io
I
1
20
25
Factor
Equilibrated Equilibrated Partial cquilibrittioii
3.33 i 0 . 0 7 3.28 i . o i 3.46 i . O i
1'
/I3.315 av.
lI1,I
13
095
1)UK-
f13, D
.I
p
ING Isotopic composition of gas
Equilibrated Equilibrated Partial equilibratioii Tank gas
5.41 i 0 . 1 , ;,,43 , CJ.42a v
*
1
U.01 i . 1 6.03 i . I
The isotopic exchange coefficient Q H , T (Table I ) is constant over a wide range of tritium concentrations where the latter atom fraction is small relative to the concentration of light hydrogen. This is also the case during the hydrogenation experiments listed in Table I1 where the atom fraction of trjtiuni is approximately lop9. This constancy of a l ~ , ' rmay be shown by deriving a relationship between N I I , , ~ and the equilibrium constant ICF( I i , T ) for the reaction € 1 2 (9) 4- AcOT (g) 11T ( g ) AcOH ( g ) . i t iiwtlerate pressures
+
1-IME lk HOURS.
I'ig. 2 , ~Ap1)roachto isotopic cquilil)riuiii.
'I'he isotope exchange coefficient a i i , r , for the .\ssuliliiig that the p u t i a l pressures of .IC()El reaction (aj is defined as: C Y ~ ? . I )=* ( L X : ' H ) h i > ' c t ~ .~k dO T follow idedl solution laws, it is calculated (~)/H)G where (D/H)hl and (D,/H)c are the atom tlldt ratios in the carboxyl hydrogen and hydrogen gas, respectively, at equilibrium. The average of cxperiineiits 2--4 in Table 1-1yields a value of :j.hL); /I ' . i C o l ~aiid $J'A~ kn > k11. Polanyi'j reached a similar conclusion after consideration of zero point energies for initial and activated state. These considerations will be applicable to the results of these experiments if it is assumed that the reacting hydrogen atoms are relatively loosely bound a t the catalyst ~ u r f a c e , 'so ~ that isotopic zero point energy differences here are small. The isotopic fractionation factor would also be affected by the isotopic exchange coefficient between hydrogen in the gas phase and a t the catalyst surface. Acknowledgments. - U7c wish t~ express our appreciatioii to Robert l V . Jsiler for carrying out the deuteriuiti a~i,tlj.~cs a i i t l to Dr. Harold Bcyer and 11oiics I3erniCtiiwho luriiislic(1 \-aluable xssistance i r i coiiiicctioii witli tlic tlcutcriuiii iiiass sprctronietcr. NElT' Y O K K ,
T.
(10) D. Rittenberg, S.l i a t n c r a n d H. I). Iloberman, T H t s J O ~ T R N A ~ . . 62, 2219 (1940). (11) T. W. X e w t o n a n d G. K.Rollefson. J. Chcm. Phys., 17, 718 (1949). (12) B. J. T h o r n , Biochewz. J.. 49, 603 (1951). ( 1 1 0 J. Bigeleisen, J . Chem. P h y s . , 17, 675 (1949). (1-1) h f , Polanyi, S n l i r r e , 26, 133 (1934). ( I s > ) A . Sliermsii ;iod H . Eyring, T H I SJ O U R N A L , 64, 2661, 4191 (l!IW).
