GAS-PHASE DECOMPOSITION OF FORMIC ACID IN CARBON COATED VESSELS Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY SANORA WOLLERI’ WAISOII 1967 LIBRARY Michigan State Umwr ity THESIS *— ABSTRACT GAS-PHASE DECOMPOSITION OF FORMIC ACID IN CARBON-COATED VESSELS by Sandra Wollert Watson Formic acid decomposes thermally to form two sets of products, H2 + C02, and H20 + CO: HCOOH'-+ The decompositions (1) and (2) are subject to surface > C02 + H2 (1) > co + H20 (2) catalysis by a large number of substances, the exact pro- duct distribution depending on the relative selectivity of the catalyst for the two paths. Recent work has indi- cated that the surface decomposition proceeds through formate ion, rather than by a molecular or radical mechanism. The photolytic decomposition, on the other hand, proceeds by a radical mechanism. The kinetics of the decomposition occurring in carbon coated vessels have been examined at 451, 478, and 505°C and good agreement with earlier work by Blake and Hinshel— wood (BH) was obtained. On the basis of the kinetics and the invariability of the rate to an 8-fold increase in sur- face/volume ratio and to addition of modest amounts of radical inhibitors, BH prOposed that the decomposition under these conditions was homogeneous and that reactions (1) and (2) are molecular. Sandra Wollert Watson Considerable interest attaches to molecular processes, and in order to examine the molecular character of the de- composition, the product ratios CO/COZ and Hz/COZ, and the isotopic content of hydrogen produced from the decomposi- tion of DCOOH and HCOOD, have been determined at 505°C. The results indicate that slightly less hydrogen is produced than C02 which is inconsistent with a molecular Split mech- anism; however, hydrogen may be removed by side reactions with the coating or other materials under the conditions of the experiment, and this result taken alone is inconclusive. The Hz/COZ product ratio was found to be insensitive to the initial formic acid pressure, whereas the CO/C02 product ratio increases significantly with increasing initial formic acid pressure. When DCOOH and HCOOD were decomposed under the same conditions, all three isotopic hydrogens Hg, HD, D2 were produced as shown by both gas chromatography and mass spectrometry. Under identical conditions, a mixture of H2 and D2 was found to be stable toward exchange, and the re- sult indicates that processes other than the molecular de- composition (1) and (2) must be Operative in the system. In all cases, the HD/Hz and Dz/Hz ratios were nearly twice as large in the case of DCOOH than for HCOOD, whereas the decomposition rates were comparable at all pressures. The results with DCOOH are very similar to those observed in the photolytic decomposition of formic acid, suggesting similar mechanisms as operative. Sandra Wollert Watson The present results are consistent with a radical mechanism in which the chain is carried mainly by the processes H- + HCOOH > H2 + ‘COOH H- + HCOOH > H20 + HCO- -COOH -——————> co2 + H- HCO' > C0 + H' Molecular processes may also be occurring in the system. The results are discussed with respect to the previous ex- planation. It is suggested that the experiments of BH are inconclusive but not inconsistent with a radical mechanism. GAS-PHASE DECOMPOSITION OF FORMIC ACID IN CARBON COATED VESSELS BY Sandra Wollert Watson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1967 .9 2/ H 9 .7 ,},;fj I To my mother, father and Jim ii ACKNOWLEDGMENT The author is indebted to Dr. L. B. Sims for his direction, assistance and encouragement in making this in- vestigation possible. The author would also like to ex- press thanks to Dr. Alexander MacDonald for his invaluable suggestions concerning the gas chromatography portion of the research. iii TABLE OF CONTENTS Page I. INTRODUCTION . . . . . . . . . . . . . 1 II. HISTORICAL BACKGROUND . . . . 6 III. EXPERIMENTAL . . . . . . . . . . 25 A. Materials . . . . . . . . 25 B. Apparatus . . . . . . . . . . . . . . . 26 1. The Reaction Vessel . . . . . . . . 26 2. The Furnace . . . . . . . . . . . . 26 3. The Vacuum System . . . . . . . . . 27 4. The Toepler Pump . . . . . . . . . . 27 5. The Gas Chromatographic System . . . 34 C. Experimental Procedure . . . . . . . . . 34 1. The Sample Introduction . . . . . . 34 2. The Kinetics of the Decomposition . 35 3. Product Analysis . . . . . . . . . 36 4. Isotopic Content of the Hydrogen . . 37 IV. RESULTS AND DISCUSSION A. Kinetics. . . . . B. Product Analysis . . . . . . . . . . . . 44 o o o o o o o o o o o o 46 C. Studies using DCOOH and HCOOD . . . . . . 59 V. BIBLIOGRAPHY . . . . . iv LIST OF TABLES Table Page 1 Data for decomposition of formic acid . . . . 8 2 Initial first-order rate constants as a function of pressure at the three temperatures 49 3 First and second-order rate constants for two typical runs . . . . . . . . . . . . . . 56 4. Product analysis . . . . . . . . . . . . . . 58 5 Kinetic isotope effect . . . . . . . . . . . 61 6 Isotopic hydrogen analysis . . . . . . . . . 62 LIS T 'OF FIGURES Figure Page 1. Surface Catalyzed Decomposition of Formic Acid 10 2. The Vacuum System . . . . . . . . . . . . . . 28 3. Relay Control Circuit for Toepler Pump . . . 30 4. The Reaction Vessel and Introduction System . 32 5. The Gas Chromatographic System . . . . . . . 39 6. 'Apparatus for Equilibration of H2 and D2 . . 42 7. 'Typical Kinetic Runs . . . . . . . . . . . . 47 8. Initial First-order Rate Constants as a Func- tion of Pressure at 451°C . . . . . . . . . 50 9. Initial First-order Rate Constants as a Func- tion of Pressure at 478°C . . . . . . . . . 52 10. Initial First—order Rate Constants as a Func- tion of Pressure at 505°C . . . . . . . . . 54 vi I. INTRODUCTION Theories relating reaction rates to molecular param- eters have been developed for elementary unimolecular processes in the gas-phase.1 Considerable interest at- taches to these theories because of the assumptions concern- ing intermolecular and intramolecular energy exchange processes, vibrational anharmonicity, the "strong colli- sion" assumption, etc. These assumptions are important in other areas and in particular for the understanding of other rate processes. The nature and validity of the as— sumptions are subject to direct investigation and test in unimolecular reactions by comparison of experimental re- sults with those calculated from the various theories; only a limited number of unimolecular reactions are known for which detailed comparisons have been made, particularly over wide ranges of temperature and pressure. Most mole— cules which undergo unimolecular reaction in the gas-phase are large molecules of considerable complexity for which only the first-order region is accessible for study, and for which the calculations are onerous. Considerable more information can be obtained from the unimolecular reactions <>f small molecules, which can be studied in the second- cxrder and/or the fall-off region of pressure. Formic acid decomposes under a variety of conditions ir1 the gas-phase by paths (1) and (2): ant I > co2 + H2 (1) HCOOH | > co + H20 (2) The decomposition can be carried out both photochemically and thermally. Recent photochemical studies2 suggest that the reaction occurs almost exclusively via the primary process: HCOOH + hV '—-———> HCO' + 'OH followed by secondary chain processes leading to final products. Molecular processes such as represented by (I) and (2) are of negligible importance under these conditions. In the thermal decomposition, it has been found that both paths (1) and (2) are subject to catalysis by a number of substances, and the decomposition has been the subject of many studies on selective catalysis.3'9 However, most substances catalyze the dehydrogenation (Path 1), and most of the catalysis work has been on this path. Hirota and co--workersl°"12 have obtained evidence for the presence of formate ion on the catalyst surface, and they propose a Inechanism involving a rather complex series of radical and surface reactions which explain all of the main features of the reaction, including the deuterium isotope effects anud the distribution of isotopic hydrogen in the products in the decomposition of the various deuterated acids. Recently, Blake and Hinshelwood13 reported that in re- act:ion vessels coated with a specially prepared carbon sur- face, the surface reactions and radical reactions were 3 inhibited, and that homogeneous components of both path (1) and (2) occur. They find that under these conditions, path (1) is first-order and path (2) very nearly second— order (except at very low initial pressures of formic acid, where the observed order increases somewhat for both pro- cesses). They suggest that path (1) consists of a uni- molecular decomposition of monomer HCOOH, whereas path (2) corresponds to a unimolecular decomposition of dimer (HCOOH)2. Their evidence concerning the molecular character of the decomposition consists of the observations that the rate was unaffected by added radical inhibitors (propylene, isobutene, NO) in modest amounts, and by an 8-fold increase in the surface/volume ratio. Also, the dependence of rate on initial formic acid pressure was much different than observed on uncoated vessels. The kinetic parameters found for path (2) assuming a unimolecular decomposition of dimer, log10 k (sec‘l) = 13.58 — 42600/2.303 RT C0 are reasonable and support the assumption. However, those observed for path (1), 10910 k (sec'l) = 4.8 - 30600/2.303 RT co2 are very abnormal for a unimolecular reaction; in particular, the pre-exponential factor, 104-8, is much lower than the "typical" value of 1013, indicating an abnormally small entropy of activation. Blake and Hinshelwood suggest that 4 a possible eXplanation for the low pre—exponential factor is that path (1) actually consists of the decarboxylation of ion pairs (HCOO-)(H+) present in minute concentration, but no supporting evidence is offered. Since carbon and active charcoal were found to catalyze the decomposition, especially that of path (2), and since the evidence concerning the molecular character of the de- composition is indirect only, the interpretation of Blake and Hinshelwood is questionable. If either path (1) or (2) or both were found to be molecular under the conditions of Blake and Hinshelwood, they could serve as important new reasons for study, and it was felt worthwhile to investi— gate in more detail the molecular character of the decomposi- tion. More direct evidence on the molecular character of the decomposition can be obtained by the use of deuterated acids. If path (1) and (2) represent the unimolecular de- composition of monomer and dimer, respectively, then in the decomposition of dfformic acid (DCOOH) and formic acidjg (HCOOD), HD and HDO would be the sole hydrogen—containing products, providing (a) the water gas equilibrium H20+CO > H2 + co2 < was not mobile, and that the exchange processes > 2HD and > 2HDO 5 did not occur, under the conditions of the experiment. It was established that neither (a) nor (b) occurred in the experiments, but that (c) did occur; therefore, the experi- ments directed toward establishing the molecular character of path (1) by determination of the isotopic content of the hydrogen produced in the decomposition of DCOOH and HCOOD. In addition, a complete product analysis was undertaken in order to establish that equivalent amounts of CO and H20 and of C02 and Hz were produced, since these are demanded for molecular processes, but not necessarily for radical processes; discrepancies are noted, for instance, in the photochemical decomposition,2 which has been shown to occur primarily by radical processes. II. HISTORICAL BACKGROUND The vapor—phase decomposition of formic acid has been studied extensively by photochemical and thermal methods. Formic acid vapor decomposes by two simultaneous paths: (1) dehydrogenation and (2) dehydration: Y HCOOH -————> co + H20 Most of thermal studies have been concerned with the selec- tive catalysis by various substances, by studying the ratio of rates for the two paths on a given substance, and by studying the rates of the individual paths (1) and (2) on various substances (most such studies have been made on path (1), the dehydrogenation). Metals, especially noble metals, such as Ag, Pt, Au, Rh, Pd, Cu, as well as some oxides ($5., ZnO;, see Table I), catalyze the dehydrogena- tion, path (1), predominantly, while many other substances show varying catalysis for both reactions. In fact, a Vihcde spectrum of substances ranging from purely dehydrators 'tllrough catalysts of varying selective activity, to purely de'hydrogenators, has been defined by the work of many Chemists. The original suggestion by Sabatier”, that the Selective catalytic activity was an intrinsic chemical E>Jr HCOO- (ads) + H+ (ads) (a) HCOOH (surface) for metals with unstable formates and a rate determining step (b) > C02 + H (ads) + e (b) Hcoo' (ads) for metals with stable formates. This mechanism is not consistent with all of the available data, and has more recently been largely diSproved by Schwab and Watson9 lasing deuterated formic acids. The isotOpe effect (here ciefined as the ratio of rate constants for reactions of :isotopic molecules) gives directly information concerned vvith the rate determining step. It is easy to show that 'the deuterium isotope effect in the decomposition of d- :Eormic acid (DCOOH) should be primary for a reaction (b) 23nd secondary for reaction (a) £42,, k(HCOOHVk(DCOOH) >> 1 :Eor (b) and :1 for (a). Conversely, for formic acid-d (HCOOD), one expects k(HCOOH)/k(HCOOD) >> 1 for (a) and Sfi.for (b). Schwab and Watson found that kH/kD was consider- al>ly greater than unity for all of the deuterated acids IiCLOOD, DCOOH and DCOOD on all catalysts investigated, some CD15 which had stable formates and some of which did not. Other evidence against a Sachtler-type mechanism is aIlfszorded by the work of Hirota and co-workersz° who studied the decomposition of formic acid on several catalysts in Since tildxe presence of D2 and observed no HD in the products. a H2 + co2 > H2 'I' 2C02 'I' 2 e 2Hcoo' (ads.) HCOOH (ads.) > H2 + co2 HCOOH (ads.) + HCOO—(ads.)-—> H2 + co2 + HCOO- (ads.) 2H+ (ads.) + 2 e“ > H2 It seems that it is fairly well established that the (Zaitalytic dehydrogenation of formic acid does not occur hi3? a simple molecular process, but by a rather more compli- C=Eited series of surface reactions involving formate ion and perhaps other absorbed species. Much less is known about the catalytic dehydration, 1>I11: the available evidence21-22 suggests this may also pro- Qeed by a complex mechanism. 15 Even less is known about the detailed mechanism of the decomposition of formic acid on surfaces which catalyze both the dehydrogenation and the dehydration. One such catalyst of particular importance for this study is active charcoal (C-active), Table I. This very porous, mostly divided sub- stance shows a selective catalytic behavior which is very dependent upon the method of preparation. There has also been a considerable number of studies, involving the photolysis of formic acid.2:23726 This re- action tends to be more complex, the sole products being again H2, C0, C02, H20, but the product yields being very dependent on wavelength, temperature, and rate of photolysis. Most of the photolysis work was done at wavelengths corre- sponding to an excitation energy of 50 Kcal/mole, consider- ably above that available in the thermal reactions, with an increased probability of the presence of radicals in the reaction. Since most of the studies were carried out near room temperature, and at pressures at which formic acid vapor is known to associate considerably27, hence the presence of the dimer is also an important factor in photol— ysis experiments. In an early paper Taylor and Gorin24 photolyzed formic acid under optimum conditions for dimeri- zation, and suggested that the dimer decomposed molecularly to form exclusively C02 and H2. They reported no evidence for the presence of radicals when the photolysis was per- formed in the presence of para—rich hydrogen; no ortho-para conversion was detectable by thermal conductivity measurements. 16 Burton25 also obtained no evidence for atoms or radicals when he photolyzed formic acid in the presence of an anti- mony mirror. However, Terenin28 observed an emission Spec— trum of the OH radical in photolysis experiments, and on this basis Burton25 proposed that the primary process was HCOOH + hV > HCOOH* > HCO° + °OH Gordon and Ausloss2 were the first researchers to es- tablish the importance of radicals in the direct and mer- cury photosensitized photolysis of formic acid (HCOOH and DCOOH). They found that the yield of Hz was in all cases smaller than the yield of C02, and that the production of H2 was markedly reduced in the presence of efficient H-atom scavengers (C2H4 and 02). Thus it is clearly indicated that processes other than direct molecular—split mechanisms were involved in the photolysis. The possible primary processes proposed were: HCOOH + hv > co2 + H2 (1) > co + H20 (2) > Hco- + -OH (3) > H + -COOH (4) > HCOO- + H (5). The low yields of H2 relative to C02 indicate that process (1) cannot account for more than 5% of the overall yield. Moreover, the fact that the yield of H2 is greatly reduced in the presence of H-atom scavengers, whereas that of C02 is not affected, indicates that processes other than (1) 17 are more important for the production of H2 and C02. Many other observations suggest that, other than a small contri- bution from the molecular reactions (1) and (2), the import- ant primary process is process (3), the production of formyl and hydroxyl radicals. Secondary processes involving H abstraction from HCOOH by OH to form H20 and either .COOH or HCOO- (both of which are presumed to yield H and C02 in subsequent steps, es- pecially at high temperatures), decomposition of excited formyl radical to yield CO and H,and an abstraction reaction of H with HCOOH to form H20 and HCO- account for most of the results observed by Ausloss and Gordon at ordinary and moderate temperatures. Termination is assumed to occur by hydrogen atom recombination. At higher temperatures, a sharp rise in rates of production of H2 and C02 indicates the im- portance of chain reactions involving H atom abstraction from formic acid by H to form H2 and COOH or HCOO which form C02 and H. More recent work7 is consistent with the proposed radi- cal reactions. It thus seems that the photolytic decomposi- tion of formic acid vapor occurs almost exclusively by radi- cal and radical-chain processes, and that molecular mechan- isms for dehydration and/or dehydrogenation are relatively unimportant at higher energies. Blake and Hinshelwood13 have recently suggested that the heterogeneous surface and/or radical paths for dehydrogen- ation and dehydration of formic acid vapor are suppressed when formic acid is decomposed thermally in reaction 18 vessels which had been especially coated with carbon, and that under these conditions, the decomposition of either process consists of the molecular components. They further propose that the reaction to produce COZand H2 is a uni- molecular decomposition of monomer formic acid, and that the reaction to form HéOand.CO. proceeds by a unimolecular decomposition of dimer molecules. This seems an unlikely possibility, since carbon generally shows varying selec- tive catalytic effects for both reactions. However, the surface obtained by Blake and Hinshelwood by cracking iso- butylene at 600°C followed by a baking at 700°C would more likely form a synthetic carbon surface with greatly reduced area and porosity than active charcoal, then by analogy with the behavior of other substances upon similar treat- ments, the catalytic effect expected would be one of in- creased dehydrogenation; the Opposite effect is observed; at nearly all pressures, the dominant reaction was the forma- tion of CO. It thus seems likely that either carbon is anomolous in its behavior, or the reaction in the Blake— Hinshelwood reactor is indeed of a different character. If the molecular components of either the dehydration or dehydrogenation of formic acid could be isolated for study, such studies would be very important, for formic acid is a relatively simple molecule for which detailed theoretical rate calculations are feasible. Hence an example of a reaction which could be compared with theory in detail may 19 be afforded. This research was undertaken to examine in some detail the molecular character of the decomposition reactions of formic acid vapor in carbon-coated vessels. Before outlining the experiments undertaken, a review of the evidence presented by Blake and Hinshelwood to support the assumption of molecular reactions seems appropriate. The Blake-Hinshelwood reaction vessel consisted of a silica vessel coated with carbon from the decomposition of six charges of isobutylene at 600°C followed by a final baking at 700°C. After careful pump-out and seasoning with a small amount of formic acid for several hours fol- lowed by a final evacuation, formic acid vapor was decomposed at temperatures of 436 to 532°C The rate of the decompo- sition was followed by measuring the total pressure of the system. The pressure increase was extrapolated to obtain initial rates based formally on a first-order rate equation. The homogeneity of the reaction was established by the fact that the rate was only slightly affected (in fact, the rate was slightly decreased) in a reaction vessel packed with small glass chips so that an eight-fold increase in surface volume ratio was effected. In several experiments, modest amounts (maximum ratio inhibitor :HCOOH = 2:1) of radical-chain inhibitors (iso- butene, propene, NO) were added, with no noticeable de- crease in rate. In fact, a small increase in the case of added NO was noted, and attributed to the reaction of NO with the carbon surface to expose the silica surface. The 20 rate was established to be much faster on the uncoated silica vessel. The product composition (CO/C02 ratio) was also substantially unchanged by addition of propylene. These results were interpreted by Blake and Hinshelwood as making it very improbable that radical-chain processes contribute significantly to the reaction under these con- ditions. However, experiments in which much larger amounts of inhibitors were added would be needed to definitely rule out a radical process. The results in the presence of modest amounts of inhibitors probably indicate only that long chains are not involved in the reaction. Short chain radical processes, such as those proposed for the photolytic reaction, may not be seriously affected by small amount of inhibitors. Product analysis was limited to determining the CO/C02 ratio as the reaction proceeded. In separate experiments it was shown that the water gas equilibrium C0 + H20 -::: H2 + C02 was not mobile under the conditions of the ex- periments, so that the CO/C02 ratio may be used to partition the overall observed rate constant k into that describing the dehydration, k and that for dehydrogenation, kC CO’ 02' The initial decomposition rate was found to have a dependence on initial formic acid pressure represented by an average order slightly less than 2 over the pressure range 1-650 mm pressure, although the order is slightly higher than 2 below about 50 mm. The individual rate con- stants RC0 and kC obtained by partitioning the rate 02 21 constant by the observed product ratio.CO/C02, were found to have very different dependences on initial formic acid r : - - - ' ' p essure kCO2 15, Within experimental error first-order at all pressures, whereas kC is slightly less than 2nd 0 order above 50 mm, and very nearly 2nd order below 50 mm. At all pressures above 10 mm, the reaction forming CO (dehydration) is dominant. At lower pressure, dehydrogena- tion becomes dominant, but at high pressure this reaction is negligible. It is thus possible by studying the tempera- ture dependence in the high and low pressure region, to obtain apparent activation energies for the two processes; these were found to be ECO = 28.5 and ECO2 = 30.6 Kcal/mole. The product distribution, CO/COz, is thus nearly independent of temperature. On an uncoated silica vessel, the reaction rate was approximately one-half order in initial formic acid pres- sure, and the product ratio CO/C02 was >> 1 at all pressures, even below 10 mm. The reaction is thus very different in the coated vessel, and Blake and Hinshelwood conclude that both reactions are homogeneous, molecular processes under the conditions of the experiments. The reaction forming CO is nearly second-order in initial formic acid pressure, and if the small deviation is neglected, the results expressed as a bimolecular rate constant, it is found that kCOI -1 kCO = 3 x 107 exp(28500/RT) 1 mole"1 sec , 22 implying a steric factor of ~t10-4, somewhat lower than would be expected. The interpretation proposed is that the sebond-order reaction forming CO is really a unimolecular decomposition of dimer molecules present in the system. Correcting for the concentration of dimer molecules and for the heat of dimerization one obtains an apparent first-order rate constant for unimolecular decomposition of dimer mole- cules: kggmer = 4 x 1013 exp(-42600/RT) sec-1, and the Arrhenius parameters are in the proper range for an unimolecular process. The slight deviation from second-order behavior is presumed due to a small contribution from a first-order decomposition of monomer to form CO and H20. The reaction forming C02 is first-order at all pres- sures and assumed to be molecular, with a rate expression of k = 104°8 exp(—30600/RT) sec—1. co2 The pre-exponential factor differs so significantly from the usual value of 1013 sec—1, that a unimolecular de- composition of monomer seems unlikely. Indeed, it is dif- ficult to reconcile two different unimolecular decomposi- tions of the monomer--one to form CO as assumed to account for the deviation of kCO from second-order, and one forming C02 and Hz with a very different rate. A very speculative explanation is offered, that the reaction forming C02 actually 23 consists of a rate determining decarboxylation of ion pairs (HCOO-, H+) present in minute concentration. The evidence of the molecular character of the de- composition of formic acid vapor in carbon-coated vessels consists mainly of (1) the apparent homogeneity of the de- composition, (2) the lack of evidence for radical-chain processes, and (3) the lack of suitable alternative mechanisms to account for the kinetics and product composition. Also Blake and Hinshelwood report only results on the initial rates, and no mention is made of detailed study of the kinetics over the extent of the reaction; therefore, a more detailed kinetic study is desirable. It has already been noted that in many of the catalytic and photolytic studies, an excess of-C02 over H2 was observed, thus arguing against a molecular-split mechanism for this mode of decomposition of formic acid. Blake and Hinshel- wood apparently did not investigate the HZ/COZ product ratio for the reaction in carbon-coated vessels, although such information could be very important for their con- clusions. The product composition H2:CO:C02 was one of the goals of the present investigation. In the catalytic studies, the deuterium isotope effects and especially the isotopic distribution of hydrogens formed from decomposition of HCOOD and/or DCOOH have been among the most important facts leading to an understanding of the nature of the reaction. Such studies are equally if not Inore important for establishing the molecular character of 24 the decomposition of HCOOH in carbon-coated vessels. It was established during this study that the exchange reac- > HD does not occur under the conditions <— tion H2 + D2 of the experiments, so that if the reaction forming C02 and H2 is molecular (whether the actual process is a unimolecu- lar decomposition of monomer molecules or decarboxylation of ion pairs in the gas phase), the only isotopic hydrogen which should be produced in the decomposition of either HCOOD or DCOOH is HD. In addition, the kinetics are easily studied, so that deuterium kinetic isotope effects between HCOOH and all three deuterated acids are readily measured by comparative rate methods (competitive rate methods would be possible only between HCOOH and DCOOH, since the carbox- ylic hydrogen exchanges readily), and could offer valuable information concerning the rate-determining step of the reactions. II I . EXPERIMENTAL A. Materials _Formic acid (99%) obtained from K & K Chemical Labora- tories was purified by at least four trap-to-trap vacuum distillations shortly before use in a run. After this procedure, the purity was checked by gas chromatography on a 6 ft. x 1/4 in. column of 100-120 mesh Porapak Q (Waters Associates, Framingham, Mass.). No impurities other than water were detected, and the vacuum purification was re- peated until no water was detectable (estimated 1 1 mole % max., impurities). Formic acid-d (HCOOD) and d-formic acid (DCOOH) of 99+% purity were obtained from Merck, Sharp, and Dohme, Ltd., Montreal, Canada. No further purification was required for the labeled acids. Isobutylene (C.P. grade 99.0 minimum % purity) was obtained from Matheson Co., and used without further puri- fication. COpper oxide wire, Fisher Certified Reagent, was sifted through a coarse cloth to remove dust, and used to fill the combustion tubes. A plug of copper gauze (Cenco), washed 'with hot ethanol and acetone to remove the lacquer and oxi— dized in an oxidizing flame, was used as a stop at each end of the combustion tube. The combustion tube was re-oxidized 'with tankuoxygen for at least one hour at 700°C before use. 25 26 No detectable impurities were added to the gas mixture from the combustion furnace. The Activated Alumina (Grade F—l, Mesh 40-80) was purchased from the Aluminum Company of America. The Ferric Chloride (Reagent A.C.S. lumps of 99% purity) was obtained from Matheson, Coleman and Bell. The Ammonium Hydroxide was Reagent Grade 99+% purity from Fisher Scientific Company. B. Apparatus 1. The Reaction Vessel consisted of a quartz cylinder, 23 cm long and 2.8 cm i.d., joined by a Vycor-Pyrex seal to a short length of 2 mm Pyrex capillary tubing, and a 2 mm capillary vacuum stopcock to the introduction and vacuum system. In addition, a side arm from the neck of the reac- tion vessel connected with a mercury manometer for measuring pressures. (Figure 4,p. 32.) 2. The Furnace used for heating the reaction vessel was a single tube furnace heavy insulating type 13559 (Central Scientific Company) closed at one end with a 1-inch plug of asbestos. The space remaining at the other end was closed by wrapping 1-inch asbestos tape about the neck of the reaction vessel so that a good fit was obtained when the reaction vessel was inserted into the furnace. A 7.5 sump manually-controlled variac was used to power the furnace. 27 3. The Vacuum System consisted of three sections: the manifold and trapping section, the sample introduction sys- tem, and the gas-analysis and product handling system. The design was conventional, consisting of Pyrex tubing, high- vacuum stopcocks, U-traps, and a Toepler pump for handling non-condensable gases. Details will be given under the section on experimental procedure (Figure 2). 4. The Toepler Pump was automatic of conventional de- sign, purchased from Delmar Scientific Co.. However, it was found that the relay control supplied by the manufacturer caused arcing across the leads in the pump, which is very undesirable, especially when pumping combustible gases such as hydrogen, so that the pump was operated by a relay con- trol circuit shown in Figure 3. The control circuit Oper- ates as follows: the lower chamber of the Toepler pump is fitted with a stopcock which was connected to a Y connector, one connection going to a vacuum pump (Cenco Hyvac-2), the other connection to one end of Hoke needle valve. The other end of the needle valve is arranged so that it is firmly closed by an arm connected to the power relay PR of the circuit with that relay in the closed position; otherwise it was open to the atmosphere, and the valve served as a vari- able—rate leak valve. When connection is made between the common and the upper lead of the Toepler pump by mercury (Figure 3), the relay PR is activated; the leak is closed, and the 110 outlet to the pump is activated; the lower 28 Figure 2 The Vacuum System 29 H )‘.‘z \J H EEIVL IHHIH awn—“H. Ud<0> W All H0 d? m .INI m M HH gran. r 0&3 U<> Ll a Hmmmm> 0..) we? cam _ . AI Nam cofluommm momcusm . mm. 050 Iv . ~ I \J U(> :Fl . m? H E H mm momsusm [A hm Iv. l 30 Figure 3. Relay Control Circuit and Toepler Pump 31 ‘V I Toepler Pump 32 Figure 4. The Reaction Vessel and Introduction System 33 AH“ mmmmw> OHDUMOM OUMGHSW __ as 34 chamber evacuates, and the mercury level in the pump lowers until the mercury makes contact between the common and and 1933; leads, at which point the relay PR is de-activated, the pump is shut off, and the leak to atmosphere is open, forcing the mercury up in the pump until contact is again made with the 3222; lead, and the cycle is repeated. The circuit employs a 22.5 v potential through 106 ohms across the contacts in the pump, which will not cause arcing. 5. The Gas Chromatographic System consisted of an Aerograph Model 90-P manually programmed chromatograph, modi- fied so that a gas inlet system could be interposed in the He inlet line, and so that an external column (with an empty dummy column substituted in the chromatograph) could be used for the isotopic hydrogen analysis, where immersion of the column in liquid nitrogen was necessary. The temperature of the injector, columneand detector were always the same (200°C for the analysis of the formic acids, or 150°C for the analysis of the isotopic hydrogens), and a flow rate of 80-100 ml/min of He were used in all cases. C. Experimental Procedure 1. The Sample_;ntroduction was effected by expansion of a known amount of formic acid from bulb-Bl, (Figure 4) maintained near 100°C by boiling water in a Dewar flask, and by nichrome heating wire wrapped around the section be- tween stopcocks S-1 and S-2, through stopcock 81 directly 35 into the preheated reaction vessel. Fifteen seconds were allowed for equilibration, and 31 was closed. Formic acid was contained in a storage bulb B2. Prior to a run, equilibration of formic acid vapor from B2 into the expansion bulb B3 (various size bulbs were used as ex- pansion bulbs B3 in different runs) was effected at room temperature. The vapor trapped in B3 was then transferred to bulb Bl by vacuum distillation for later expansion into the reaction vessel. 2. The Kinetics of the Decomposition were followed by measuring the total pressure in the reaction vessel as the reaction proceeded. For this purpose, one arm of a capil- lary manometer was connected to the neck of the reaction vessel as shown in Figure 4. The mercury level in the ma-- nometer was adjustable by means of a leveling bulb and was adjusted prior to a run so that the mercury was very close to the neck of the reaction vessel, so as to expose as small an area of uncoated glass as possible. The other arm of this null manometer NM was connected to a measuring manom- eter, MM, as shown in Figure 4. The 200 cc ballast bulb B4 on top of the high pressure arm of MM could be connected by stOpcock S4 to vacuum or, by means of a very fine capillary, to air. During a run, the expected initial pressure of HCOOH was calculated from the amount transferred to bulb B1 and the temperature of the furnace, and this pressure of air was leaked into B4 and MM by S4. After expanding the 36 HCOOH into the reaction vessel and closing stopcock $1, the timer was started, and stOpcock S3 connecting NM and MM Opened, and the level in NM brought back to its null posi— tion by adjusting the pressure in MM by S4. The pressure in MM and the time were recorded. Similarly, measurements of P were made by this null technique at various times t during the reaction. 3. Product Analysis was made usually only after the reaction had proceeded to completion. The reaction was quenched by Opening stopcocks $1, 82, S5, S6, (See Figure 2) and allowing the reaction mixture to expand into traps T1 (room temp.) and T2 cooled in dry-ice--i—propanol bath (-70°C). The volume of traps T1 and T2 were each about 50 cc. The toepler pump was started immediately, and the entire contents of the reaction vessel were pumped through the traps. Water and any unreacted HCOOH were retained in T2, and the remaining gases noncondensable at -70°C (C02. CO, H2), were pumped into the volume C1 and connecting volumes. After all of the gas had been transferred, stop- cock 58 was closed. The volume C1 was constructed of a 100 ml graduated cylinder, and was attached at the lower end to a leveling bulb of Hg, so that the Hg level in C1 could be adjusted to any level (h) and read. The volume of C1 and the connecting tubing to the capillary closed-end manometer M2 was calibrated as a function of level h in C1 and pressure P measured on M2, so that the absolute quantity of gas contained in the volume could be determined 37 at any level h of C1. Generally, the Hg level in C1 was raised until a pressure P on M2 convenient for measurement with a small error, was obtained. The quantity of gas was then determined. Initially this corresponds to n + n + co co2 nHz, where n represents moles of gas. Subsequently, T1 was cooled to -196°C by liquid nitro- gen, and the gas was cycled from C1 via $11, $12, bypassing the furnace, through 813 and $14, and finally through T1 (where C02 is retained) and T2 back to C1. After a suffici- ent time for cycling, $12 was closed and the remaining gas (CO and H2) transferred back into C1 and remeasured, yield- ing nCO + nH2° Finally, the CO and H2 were cycled through the CuO combustion furnace at 500°C (where they were burned to C02 and H20) through T2 (H20 retained), back to C1, and finally remeasured, yielding n (measured as C02 at this CO point). The quantities of each of the three products C0, C02, Hz were thus obtained by direct measurement, and, if desired, the quantity of H20 could be obtained by difference, CO, H20, C02, Hz were determined to be the sole products. 4. Isotopic Content of the Hydrogen produced in the decompositions of the deuterated acids HCOOD and DCOOH was determined by gas chromatography. The products noncondensable at -196°C (CO and Hz) were removed through the port at $11 in a small storage bulb and removed for chromatographic analysis. 38 The Hz-CO was introduced through the gas-inlet port on the chromatographic system shown in Figure 5. Helium car- rier gas from a cylinder C swept the sample in the specially prepared Fe-Alumina column maintained at -196°C by immersion in liquid nitrogen. The CO was retained and separation of H2, HD, and D2 were effected. The isotopic hydrogens were then burned to the corresponding isotopic water by passage through a CuO combustion tube at 500°C, and finally were analyzed by thermal conductivity in the detector of the chromatograph. The column was prepared from 8 ft. x 1/4 in. Cu tubing packed with the prepared packing material. The packing was prepared according to the method described by G. F. Shipman.29 Alcoa Alumina F-l, mesh 40-80, was ignited in air at 500°C for 16 hours. To 80 ml of this alumina were added in small increments 40 ml of 1.8 M_FeC13, followed by addition of 100 ml H20. The mixture was then titrated with 6M;NH4OH to pH 7 (Beckmann pH meter). The resulting mixture was de- canted, filtered, and washed with H20 until the washings were pH 6 (pHydrion paper). The resulting packing material was dried in air at 120°C for 24 hours and packed in the column. The column was conditioned for 12 hours with.C02 and for 24 hours with He at room temperature prior to use. Re- conditioning with C02 was accomplished periodically, especi- ally if the system were shut off for a period of time. After several samples of HZ-CO were run, the column was allowed 39 Figure 5. The Gas Chromatographic System ammumoumfiouso muomé Eco: .238 40 smumsm uchH mmo Hmpuoowm ammumouwfl « i i momsusm O .O O..@ 050 ohm nu nHHHHu Iii nHHHHU E a i -—L_I an 41 to warm to room temperature in order to liberate the trap- ped CO. The temperature of the column is difficult to maintain exactly at -196°C because of rapid evaporation of N2 when He is flowing through the column. The peak shapes and retention times were markedly sensitive to the temperature of the column. The best overall results were obtained by maintaining the dewar in which the column was immersed three-fourths full of liquid N2. The average temperature of the column was not known under these conditions, but a temperature of -175 to -190°C seems likely. The retention times under these conditions were 24 min (H2), 27.5 min (HD), 34 min (D2). There was a small overlap of the.H2-HD peaks, and no overlap of the HD-Dz peaks. The peaks tailed only slightly under these conditions. The equilibrium Hz/Dz/HD mixture for checking the chromatographic column was prepared from H2 (Prepurified Grade 99.95 min. % purity) and D2 (C.P. Grade 99.5 min. % purity) by the method described by P. P. Hunt and H. H. Smith3°. The H2 and D2 were placed in a bulb with a nichrome wire connected to tungsten leads sealed in separated wells on the bulb as shown in Figure 6; the tungsten leads were then connected to a Variac and the nichrome wire heated to dull red for 12 hours. The equilibrium constant K (HD)1/(H2)(D2) was determined by mass spectrometry as K 3.55, which corresponds to a temperature32 Of about 450°C, which is very reasonable. 42 Figure 6. Apparatus for Equilibration of H2 and D2 IV. RESULTS AND DISCUSSION A. Kinetics The rate of decomposition of formic acid in the speci- ally coated carbon vessel was followed at 451°C, 478°C and 505°C, by measuring the increase of pressure with a manom- eter communicating with the reactor, as described in the last chapter. In order to facilitate comparison of the results with those of Blake and Hinshelwood13, the data was treated formally by a first-order rate expression The first-order rate constant was determined by two methods; in the first method, the integrated first-order rate expression k1 = -1/t ln (1 - f) . where f is the fraction of reaction obtained at time t, was used to calculate first-order rate constants cor— responding to each pressure—time data point. A plot of kIl/z yg, t was then constructed and extrapolated to t = 0 to obtain the initial first-order rate constants. The fraction f of reaction was calculated from the relation31 f=P-Po/PCD-Po where P is the total pressure, Po the initial pressure of formic acid (obtained by extrapolation of a plot of P Xén t to t = 0), Poo is the final pressure Generally,‘ the pressure was followed for only one or two half-lives, and hence Poo was not measured. Since the only products 44 45 are H2, C02, H20 and CO, the final pressure should be Poo = 2P0 under the conditions of the eXperiments. In those runs in which a final pressure reading was established, a value close to 2P0 was obtained. In the second method, P0 was obtained from extrapolation of a plot of P Kg. t to t = 0, and values of the pressure increase AP = P - P0 was calculated at each time. The initial rate R0 is ob— tained as the limiting slope of a plot of AP yg, t near t = 0. The first-order rate constant is then given by k1 = Ro/Po - Rate constants determined by the two methods always agreed very closely, but the second method proved somewhat more reliable. Forcxmparison, second-order rate constants were also determined as follows: The integrated rate expression as— suming second-order dependence on formic acid (represented as A) is: 1/A - 1/Ao = kt. multiplying by A0, substituting 1 — f = A/Ao, and simpli— fying yields f/l - f = Aokt assuming POD = 2P0 , f = P - Po/Po - AP/Po and using pressure units for the initial concentration of formic acid, one obtained the second-order rate constant k = 1 AP P 2 POt 1-9.2 Po Figure 7 shows typical pressure XE: time curves ob- tained at each of the three temperatures. Table 2 is a summary of the initial first-order rate constants k1 as a function of initial pressure at each temperature. Fig- ures 8, 9, and 10 show plots of k1.y§. P0 at 451, 478, and 505°C, reSpectively. Included for comparison on Figures 9 and 10 are data Obtained by Blake and Hinshelwood (BH)13 at 476.8 and 500°C, respectively. The agreement is excellent, and shows conclusively that the experimental conditions of BH were reproduced in this study. The rate constant k1 is extremely sensitive to the condition of the coating, especially at low pressure, as found by BH and verified in this study. Table 3 shows raw data from two runs, one at 451°C and one at 505°C, as well as values of k1 and k2. The over- all reaction order can be seen to be very nearly second order. Detailed examination of all of the data indicate an order slightly less than two, in agreement with the re- sults of BH. B. Product Analysis The relative amounts of H2, C0, C02 produced in several runs at 505°C were determined by the method described in Chapter III. The results are shown in Table 4 There is 47 Figure 7. Typical Kinetic RUns EI 451°C A 478°C 0 505°C F, f? c E. U C a) :1 H :5 m In (D H a. 15.0-- “ 14.0 J J_ 1_ l I J l 41 J 1 OJ L l I I I I I 1 1 r I I . f I I I 0 200 400 600 800 1000 1200 1400 Time (secs.) Table 2. 49 Initial first-order rate constants as a function of pressure at the three temperatures. ' Run # Initial Pressure(mm) 1st-Order Rate Constants(sec-1) 36 37 38 40 39 12 15 14 16 21 22 31 23 44 49 42 45 46 47 48 37.8 104.4 132.6 149.4 183.4 ,26.6 '49k3 56.8 64.4 176.4 93.6 101.4 106.8 118.5 160.7 164.0 205.0 39.5 54.7 58.6 95.3 110.2 137.2 140.2 Temperature ‘ Temperature Temperature = 505°C 451°C 478°C 4.1 5.3 4.8 7.6 8.5 10.9 10.2 xxxxxxxxxxxx XXXXX N >4 X >4 X >< X 10 10’ 10‘ 10' 10' #:5an 10"4 10"4 10" 10" 10'4 10‘ 10" 10' 10‘ 10' I bibnblb 10'4 50 Figure 8. Initial First-order Rate Constants as a function of Initial Pressure at 451°C. 51 AEEV mnsmmmum HOHDHGH 0mm. CNN. omH -b -- P b b _ q 4 d OWH OOH om om ow ON 0 _ u b— L . . _ H _ 4 q o l _or x ‘x ass) I,_ 1E HD4511“! ! L 52 Figure 9. Initial First-Order Rate Constants as a Function of Initial Pressure at 478°C. A Reference 13 at 476.8°C O This research at 478°C 53 AEEV mmfimmmnm HMHDHCH CON CNN owH _ w h . p . _ 4 q 4 O¢H -I- OOH -I— rum Otxtx (I_oes) =I . 54 Figure 10. Initial First-order Rate Constants as a Function of Initial Pressure at 505°C. A Reference 13 at 500°C 0 This research at 505°C 55 AEEV Ondmmmum HOHDHCH . owm . one . ova . _ _ .4 _ _ . oov _I O£m _ _ q _ own owN . - - d O _.N H _..O 00 --O V _..o N 0 ac Ira ITNH (I_DSS) *_01 x Ix .ION Table 3. The first and second-order rate constants as a function of time for two typical runs. I525) P(cm) AP(cm) k1(sef;1) (cm-1k:ec- ) - x 10 x 10'5 Run#38(451°C) 0 13.26 55 13.40 0.14 1.93 1.46 140 13.70 0.44 2.41 1.85 240 14.00 0.74 2.39 1.86 340 14.30 1.04 2.40 1.89 430 14.60 1.34 2.48 1.97 560 15.10 1.84 2.67 2.17 750 15.50 2.24 2.47 2.04 940 15.90 2.64 2.36 1.99 1230 16.50 3.24 2.28 1.98 1560 16.95 3.69 2.09 1.86 1940 17.45 4.19 1.96 1.80 2410 18.00 4.74 1.84 1.74 k1 (Method 1, extrapolation) = 2.8 x 10-4 sec—1 k1 (Method 2, initial rate) - 2.6 x 10"4 sec"1 k1 (Ref. 13, smoothed results) = "" 1 Run # 48 (505°C) 0 14.02 I ' 70 15.10 1.08 11.45 9.71 90 15.50 1.48 12.40 9.35 115 15.90 1.88 12.52 9.60 145 16.40 2.38 12.83 10.16 Table 3 (Continued) 57 I333) Pic“) AP H- + ~COOH (1a) Initiation HCOOH -———-—-> HCO° + 'OH (1b) -—-> HCOO- + H- (1c) followed by unimolecular radical decompositions: ‘__.‘ > C0 + ‘OH (23) 'COOH'—— > cog *+ H- (2b) Hco- > CC + H' (3a) > C02 + H' (43) HCOO: -—- > 0- + HCO' —> c02+ H- (4b) and the several abstraction reactions: > H2 + -c00H (5a) H- + HCOOH-—— > H2 + HCOO' (5b) > H20 + Hco- (SC) > H20 + 'COOH (6a) °OH + HCOOH «—— > H20 + Hcoo- (6b) and possible termination reactions: H + H' > H2 (7a) H- + -OH ‘> H20 (7b) 65 It seems unlikely that reaction (4b) contributes sig- nificantly to the reaction, since BH Observed that even a small addition of oxygen to the system caused a very large acceleration of the rate. The other reactions correSpond to a radical population in the reacting system which con- tains a large fraction of H atoms. The reaction is propagated; then, very largely by the abstraction reactions 5, and the product distribution will reflect this fact. If it is assumed that the abstraction by H from the car- boxyl end of the molecule occurs preferentially by -OH abstraction, i423, reaction 5c more important than 5b, then the mechanism is consistent with the isotopic distri- bution of hydrogens observed for DCOOH and HCOOD. The corresponding abstraction reactions important for the two isotopic acids are then for DCOOH: H' + DCOOH “( D' + DCOOH -—( and for HCOOD: H' + HCOOD —-( D' + HCOOD -( and it is seen that in the case of DCOOH, the deuterium > HD + ‘COOH > H20 + DCO ’ > D2 + 'COOH > HOD + DCOO' > H2 + °COOD > HOD + HCOO‘ > HD + 'COOD > D20 + HCOO ° 66 preferentially ends up in the hydrogen product, so that relatively larger values for the HD/H2 and Dz/HD product ratios should be observed than in the case of HCOOD, where the deuterium ends up preferentially in the water product. Table 6 shows that this is precisely what is Observed. At high pressures, the abstraction reactions will be relatively more important than the decomposition reactions 2 and 3 for determining product composition than at lower pressures. It has already been assumed that 5b is less V I probable than 5c. If, in addition, 5a and 5c are of com- parable importance, then since 3a occurs subsequent to 5c to produce CO, and reaction(s) 2b (and 2a) occur sub- sequent to 5a to produce C02 (andCO), it is possible for the CO/COz product ratio to increase slightly with pressure as observed (Table 4). The Hz/COZ product ratio at low pressures (where reactions 2, 3, 4, and 7 are relatively more important) would be expected to be comparable to that at high.pressures (where reactions 2, 3, 5a, 5c, are rela- tively more important). A value for Hz/COZ close to unity as Observed (Table 4) is not inconsistent with the mechanism. In addition, molecular split mechanisms or bimolecular mechanisms may also be of some importance in the reaction, and may help explain some of the results. In particular, a bimolecular reaction Of formic acid (or a unimolecular decomposition of dimer) to produce CO and H20 products would also lead to an increased CO/COZ product ratio with in— creasedjpressure. 67 The kinetic isotope effects are difficult to interpret on the prOposed mechanism, but are rather small and about the same for both acids. Both of these results are con- sistent with the radical mechanism. The iSotope effect, if associated with abstraction, would be expected to be AIJ2 due to the mass effect on breaking the H- residue bond. One has still to explain why BH observed no noticeable surface/volume effect and no significant effect when radi- cal inhibitors were added to the system, if a radical mech- anism is assumed. However, these results do not, we believe, rule out the possibility of a radical reaction. The change in surface/volume ratio of BH was quite small, and did re- sult in a small decrease in rate, which could occur if the radical terminated at the wall, which is not inconceivable on a carbon surface. The modest amounts of inhibitors used by BH would lead to a significant change in rate only if long chains were involved. For the present mechanism, only a small change in rate for this level of inhibitors would be expected, but the yield of hydrogen would probably be significantly reduced in the presence of a radical scavenger, as observed by Gordon and Ausloos2 in the photolytic de- composition of formic acid. A more detailed and extensive study Of the product dis- tribution, isotope effects, isotOpic content of hydrogen, yield of hydrogen in scavenger experiments, and a more ex- haustive study of the kinetics over larger extents of reaction would be useful for a more complete characterization of the reaction mechanism. 10. 11. 12. 13. 14. 15. 16. 17. N. G. J. G. K. P. . C. Tingey and C. N. Hinshelwood, J. Chem. Soc. 121, . N. Hinshelwood and B. Topley, J. Chem. Soc., 123, . M. H. Sachtler, J. Fahrenfort, and L. L. van Reijen, V. BIBLIOGRAPHY B. Slater, “Theory of Unimolecular Reactions", Ithaca, New York, Cornell Univ. Press., 1959. . Gordon, Jr., and P. Auslosscv J. Phys. Chem., 65, 1033 (1961). . N. Hinshelwood, H. Hartley and B. Topley, Proc. Roy. Soc. (London) 100A, 575 (1922). 1668 (1922). 1014 (1923). M. Schwab and E. Schwab-Agallidis, J. A. C. S., 11, 1806 (1949). . Otaki, J. Chem. Soc. Japan, 80, 255 (1959). Block and H. Kral, Z. Electrochem, 63, 182 (1959). M. Schwab and A. M. Watson, Trans. Faraday Soc., .QQ. 1833 (1964). . Hirota and T. Otaki, J. Chem. Soc. Japan, 22, 1288 (1958). . Hirota, K. Kuwata, and Y. Nakai, Bull. Chem.-Soc. Japan, 31, 861 (1958). Hirota, K. Kuwata, and S. Asai, J. Chem. Soc. Japan, .89” 701 (1959). G. Blake and C. N. Hinshelwood, Proc. Roy. Soc., A 255, 444-455 (1960). . Sabatier, "Catalysis in Organic Chemistry", Leipzig, 1927. . Adkins, J. A. c. S., 43, 2175 (1922). . M. H. Sachtler and J. Fahrenfort, Congr. Catalysis (Paris) I., 37 (1960). Z. Electrochem., 64, 216 (1960). 68 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 69 W. M. H. Sachtler and W. J. M. Rootsaert, Z. Physik. Chem., 26, 16 (1960). W. M. H. Sachtler and N. H. deBoer, J. Phys. Chem., 64, 1579 (1960). K. Hirota, K. Kuwata, T. Otaki and S. Asai, Congr. Catalysis (Paris) I, 36 (1960). "Annual Review of Physical Chemistry", V. 11, 466, Annual Revies Inc., Palo Alto, California (1960). "Annual Review of Physical Chemistry", V. 12, 129, Annual Reviews Inc., Palo Alto, California (1961). H. C. Ramsper er and C. W. Porter, J. A. C. S.,‘4§, 1267 (1926 . E. Gorin and H. S. Taylor, J. A. C. S.,‘_§, 2042 (1934). M. Burton, J. A. C. 8., 58, 1655 (1936). P. E. Yankwich and E. F. Steigelmann, J. Phys. Chem., ‘61, 757 (1963). . s. Coolidge, J. A. c. s.,.§g, 2166 (1928). A A. Terenin, Acta Physicochim. U.R.S.S., 3“ 181 (1935). G. F. Shipman, Anal. Chem., 34, 877 (1962). P . P. Hunt and H. A. Smith, J. Phys. Chem., 65, 87 (1961). A. A. Frost and R. G. Pearson, "Kinetics and Mechanism", New York, John Wiley and Sons, 2nd Edition, 1961, pp. 28-30. S. Akhtar and H. A. Smith, Chem. Rev., pg, 261 (1964).