230
H.H. STORCH,C. 0. HAWKAND W. E. O'NEILL
either lithium or sodium chloride, more nearly coincides with the effect of initial concentration. According to the predictions of Eyring,12 provided the ionic mechanism persists, the reaction of sodium n-butyl mercaptide with n-butyl bromide should proceed more rapidly in alcohol-benLene mixtures than in pure alcohol. It is evident from Fig. 1 that in methyl alcohol-benzene mixtures the rate constant decreases with increasing mole fraction benzene, but does not appear to approach zero in pure benzene. This indicates that an ion-pair reaction comes into play in alcoholbenzene mixtures in addition t u the ionic reaction. The ion-pair reaction requires a much higher activation energy than the ionic reaction, since there is involved a separation of the ion-pair against a strong electrostatic attraction. It appears, then. that the etherification reactions studied here are capable of proceeding by two distinct mechanisms depending on the nature of the solvent. In solvents of high dielectric constaiit the reactions are ionic, in solvents of very low dielectric constant an ion-pair mechanism predominates, while in solvents of moderate dielectric constant both mechanisms are involved, anti the
[CONTRIBUTION FROM THE CENTR.4L
Vol, 64
measured rate constant depends on the fraction of total reaction occurring by each mechanism.
summary 1. The rates of reaction of n-butyl bromide with the sodium salts of phenol, thiophenol and n-butyl mercaptan have been determined under comparable conditions, and activation energies have been determined with both methyl and ethyl alcohols as solvents. 2 . The order of reactivity with n-butyl bromide was found to be BuSNa > PhSNa > PhONa. 3. Rate constants were found to decrease with increasing initial concentration of reactants. 4. Neutral salts were found to exert an effect similar in sign and magnitude to the initial concentration effect. 5. The order of reactivity of sodium n-butyl mercaptide with n-butyl bromide in different solvents, other factors being comparable, was found to be CzHjOH > CH30H > CeHB. A. -4 mechanism for the reactions is postulated and discussed in the light of the experimental data and certain theoretical considerations. EMORY UNIVERSITY, GA. RECEIVED AUGUST29, 1941
EXPERIMENT STATION, BUREAUOF
MINES,
u. s. DEPARTMENT OF THE INTERIOR]
Kinetics of Hydrogen Consumption, Oxygen Elimination and Liquefaction in Coal Hydrogenation. Nature of the Catalytic Reactions1 BY €I. 13. STORCH, C. 0. HAWK, AND W. E. O'NEILL
In previous papers of this series2J,*on coal hydrogenation in the presence of a stannous sulfide catalyst, the rates of change of readily measurable quantities [namely : hydrogen absorbed, oxygen removal as water and carbon oxides, and solubility in benzene) were determined. 'The temperature coefficient of the rate of hydrogen absorption was relatively low, aiid the rate appeared to be independent of those of oxygen elimination and liquefaction. It was assumed, therefore, that the slow step in the hydrogen consumption was the diffusion of hydrogen through a liquid film on the (1) Published by permission of the Director, Bureau of Mines. U. S. Department of the Interior. Piot copyrighted. (2) H. H. Stotch, C. H. Fisher, A. Eisner and L. Clarke, I n d . Eng. Chenz., 32, 346-53 (1940). (3) C. H. Fisher, G. C. Sprunk, A. Eisner, L. Clarke, M. L. Fein and H.€1. Storch, Fuel, 19, 13--10, 51-55, 67-69 (1940). (4) C. H. Fisher, A. Eisner. L, Clarke 31. L. Fein and H. H. $torch. Ebid , 20, 5--l:; i l 0 4 i
catalyst surface. -4s will be shown below, additional data and a more careful analysis show that the nature of the rate-determining step varies with the temperature. From the earlier work2 on the rates of oxygen removal and of liquefaction a t various temperatures in the presence of a stannous sulfide catalyst, four stages in the coal-hydrogenation process could be discerned, namely: (a) Solution or extraction of the coal. This is practically the only process occurring a t a moderate, measurable rate below 370'. Its temperature coefficient is about 1.2/15', and the slow step is probably a depolymerization of the coal by solvation. (b) -4 primary decomposition of the coal at temperatures above 370' involving the rapid elimination of about 60% of the oxygen of the
Feb., 1942
CATALYTIC REACTION IN COALHYDROGENATION
coal as water and carbon dioxide. This reaction has a temperature coefficient of about 3.0/15' and is probably non-catalytic. (c) A reaction with a considerably lower temperature coefficient than that of (b) is responsible for the slower elimination of the remaining 40% of the oxygen. It was suggested that this reaction might be a catalytic process. (d) Condensation of the dissolved primary coal decomposition products to form molecules more stable than the original coal. This process becomes important at temperatures beyond 400' and results in an apparent decrease in rate with increasing temperature for the later stages of the liquefaction and oxygen-elimination reactions. Above 440' this process results in coke formation, even in the presence of 180 atmospheres of hydrogen. To obtain some information concerning the nature of the catalytic reactions involved in coal hydrogenation, a series of experiments was performed with one bituminous coal a t various temperatures and contact times, in the presence and absence of a catalyst. To avoid possible catalytic effects of the autoclave walls, Pyrex glass liners similar to those described by the British Fuel Research Laboratory6 were used in all but one of the experiments described below. A detailed description of the hydrogenation equipment and method has been published.@ The coal sample (50 g.), 65 g. of tetrahydronaphthalene, and 0.5 g. of stannous sulfide were placed in the Pyrex glass liner of the 1200-cc. stainless steel bomb. In several special tests naphthalene was used as the vehicle, iodoform as the catalyst, and nitrogen or natural gas in place of hydrogen. Air was removed from the bomb by flushing with hydrogen, and hydrogen was then added until the pressure a t 20" was 1000 pounds per square inch, 70.3 kg. per sq. cm. The rotating bomb and its contents were heated to reaction temperature in about one and one-half hours. Experiments of six hours duration or less were completed with only one charging of the bomb with hydrogen. Nine-, twelve- and fifteen-hour experiments were made, respectively, in two reaction periods of four and one-half, six and seven and one-half hours each. At the end of the first period, the bomb was cooled to room temperature and the gases were discharged through dilute sulfuric acid, sodium hydroxide and a wet test meter, after which fresh hydrogen was introduced to 1000 pounds per square inch at 20" for the second period. Small samples of the bomb gases were collected for Orsat analysis. The liquid and solid products were transferred to a 250cc. centrifuge bottle and centrifuged. The centrifuge residue was washed six times with acetone at room temperature. The centrifuged oil and the f i s t three acetone washings were distilled through a well-insulated, indented, 6-inch column to 220'. The distillate collected at 150 (5) British Fuel Research Board Report for the Year Ended March 31,1937, H. M. S. O5ce, London, 1937, pp. 149-151. (6) C. H. Fisher and A. Eisner, I n d . Eng. Ckem., 29, 1371-1376 (1937).
23 1
t o 220°, chiefly tetrahydronaphthalene, was analyzed for tar acids, bases, olefins, and aromatics. The distillation residues, which were black, semisolid pitches, were analyzed for ash, carbon, hydrogen, nitrogen, sulfur, and oxygen (by difference) The amount of oxygen not removed by hydrogenation was calculated from the oxygen content of the coal, pitch, and residue. Reacted or eliminated oxygen was defined as that removed as water, carbon dioxide, carbon monoxide, and phenols and carboxylic acids boiling below 220 '. As'most of the eliminated oxygen of the coal was converted into water and carbon dioxide, similar results would have been obtained by considering the oxygen in phenols and carboxylic acids (boiling below 220") as inert and unremoved by hydrogenation. The amount of oxygen removed by cracking during distillation of the centrifuged oil is believed to be comparatively small.
I* Bb 6
%
0
0
'2 b a 8M 4
3
R
g 2
i
8
0
8 12 16 20 Time, hours. Fig. 1.-Rate of hydrogen absorption in presence of stannous sulfide catalwt. Note that point of zero absorption is shifted to make each curve distinct.
4
8
"07
2 k
5 5
k
8 33
t$
8 1
80
8 12 16 20 Time, hours. Fig. 2.-Rates of hydrogen absorption, infiuence of catalyst. Note that point of zero absorption is shifted on ordinate to make each pair of curves distinct. 4
An outline of the general features of the coal-hydrogenation reactions may be obtained from studying the data of
Vol. (if 'i'A1li.E I BATCHHYDROGENATION OF PITTSBURGH BED Coai. TO DETERMINE EFFECTOF VEHICLE,CATALYST, ETC. Temperature, 4Ch3"; initial pressure. 1000 lb. per square inch w h r i ~ li m n b is cold; contact time, 1 hour in all twts.
< T e s t no.
0.76 ! ) , 00 0.00 2.45 2.50
1.06 2.46 2.88 2.50 1 . RO 2.26 2.85 3 . oii
[Iroyeii or of hcr gas. I t is apparent from the discus~ion@\-en above that tetrahydronaphthalene i b the most essential constituent of this solvent, nnll that its efficacy as a liquefaction agent is rehtctl, at least in part, to its ability to transfer li\-(lrqe:i tl) the coal substance. Further compariioti oi eiperiment H-26 with H-41 shows that naphthalene plus hydrogen gas is a much pciorer liqueiaction agent than tetrahydronapht haletie n-ithout hydrogen. Indeed, in the prescmte o i an c~xccssof tetrahydronaphthalene, the degree of liquefaction and of oxygen removal is :bout the same with nitrogen (experiment H-41) with hydrogen (H-49). As may be expected irom the tiiscusqion given above, both in the and absence of catalysts and in a hydro~ e i iatmosphere a much larger degree of liquel,ioil is obtained when tetrahydronaphthalene is prc3?cii! than 17 hen naphthalene is the vehicle; conip~i-eH-24 nith H-?.iA, arid H-26 with H-6. I 11, iewlts of H-2 i and H - 2 U also show that in til, prtscnct' of wcloiorm as a catalyst, more hy\iroyeii is consumed per 100 g. of coal when naphthaliiic. I > the i-tJhicle than when tetrahydroua;>!ithalene 15 used. When no catalyst is emcliinination. ._ l h e comparison of €I-26 v;ith 11-41 is of special ployell (compare H-26 with H-UA, and H-O), e of hydrogen are coninterest in connection with the data of Pott arid about tltc ~ i i amounts siitrieti A ith either naphthalene or tetrahydroFiroche, which sliow that 1-irtually complete liquefaction of bitmninuus coals 1n2i\. b c , I ;i)taiticil riq~htlialenras vehicles. At 4009, in the absence oi -1 catalyst, the rate of hydrogenatioi of naphhen they are heated at :sFiil t:) 4 I:?' with ;lil equal weight of tetrah~dronapiittialeiie--plieiiol t1idc:ic is x cry ~ 1 0 ; v . ~This fact, co isidere3 naphthalene, 2 : 1: 2 , soh-ent u:idcr the pressure of &tiq irith the experiments of Table I i : ~licates oi the inu.,t important functio.is of the t ? solvent ~ T,'apors, t h a t i s , :.-it11 no atdrleti h y i n coal-hydrogenation reactions is to 17) A . Pott. H Bror
Table I. In all of the experiments of this table thi. contact. time was one hour and the tcmperature -10 J". niay be seen from Fig. 2 (discussion of this is given in a later section), the efiects of catalysts a t -!OO. and one hour reaction time arc well within t h c limits oi reproducibility of the experiments. Hence comparison-. involving only two experiments, one with and one witlioui catalyst, have doubtful significance unless auhstantiatd by nttier evidence. comparison of experiment 11-36, in which naphthalene was used as vehicle, with 13-28 and H-29, where the same conditions prevailed except that nitrogen and natural gas were nubstituted for hydrogen in H-29 and H-28 (Table I), respectively, shows clearly that hvdrogen is essential for primary coal liquefaction and vlimination of osygcn at 4iiO'-. T h e degree of liqilefaciion in H-20 was more than double that in li-28 a.tid 1 - 1 - 3 , and virtually no oxygen was eliminated in the latter whereas about 10% of the total c a h oxygen was rem in S 3 i A comparison of H-26 with FI -1I , in which a?! css of tetrahydronaplithaleue wa; used and tiitrogeu rehydrogen, shows that the dcgrec of Iiquefai-Lioil and elimination in 13-41 was 2 t o :_i tiitici a i gr:aaf :is in It is therefore probable that oiic o i thc SIC) in coal hydrogenation in the absmce of :i catal>-iti i the formation of hydroaromatic conipo~l1:ds which >crvr. as hydrogen carriers. It is nece,sary f i r :hi5 argunient to i z l l attention to the fact h a t thc amount ol hydrogen present in H-26 ( ~ ~ O L 3.7' I L g . ) greatly cscecdcd that nccczsary (about 2 g 1' for complcrc. liq:ief the members of the ccial-analysis section, and oi the experimental coal-h!-(lrogenation plant of the Central Experiment Station of the Bureau of \Iines
Summary An analysis is made of data on the effect of a stannous sulfide catalyst on the rate of change of (10) P Erasmus, " b b e r die Btldung und den chemischen Baii der Rohlen " Ferdinand Fmke Stiitt,:art 1438 ' ) O pi,
The rate of the mutarotation ol glucose has been studied extensively in water solution, both when catalyzed by the solvent, and when catalyzed by strong or weak acids and bases a t various temperatures.2 In solvents other than water, tneasurements have usually been made a t one temperature only, and only a few catalysts liave been used.3 It should be of interest, therefore, to measure the rate in various water-methanol inixtures, and to determine the temperature coefficients both for the solveiit.-catalyxect and i ~ the r acid-catalyzed reaction. (1) This paper is in pnrt taken from a t h e s i s submitted by H . I$. t o t h e Graduate School of Arts and Sciences of Duke UnivtiDurham, Korth Cartilina, in partial fdfillment of t h e requirct s for t h e degree of Doctor of Philosophy. June. 1941. ) Hudson, THIS JOOKNAI., 33, 859 ( l O J i ) ) , Osrika, $ e m , 36, 061 (1900); I x m r y and S m i t h J . C h e ~ 7Sor . ,2 uhn and Jacob, 2 . ph?.sik. Chern., 113, 3x0 (1924); R r Guggenheim, THISJI.X:RNAL; 49, 2524 I.1CjY7iI Kendrew and AIoelwyn-Hughes. Proc. Roy. S O L (London). A176. 322 !1!240); Moelwyn-Hughes and Johnson, ?'runs. Fiirrdny Suc., 37, 289 11941) ; Oyas
.md many others. (3) Ti'orley and Andrewn. J . , ~ l \T'nulkncr :md 1 u v - Y . .I C
Vol. 61
readily measurable quantities in the hydrogenation of a bituminous coal in a small autoclave. The results show that the rate-determining step for the hydrogen absorption varies with temperature. Below 300C it is apparently a chemical reaction between a hydrogen carrier and unsaturated groups in the coal; between 300 and 370" diffusion of hydrogen through a liquid film on the surface of the coal arid catalyst is the slowest step; above *i70Qthe rate-determining step is a chemical reaction beti+ccn a hydroaromatic and the oxygen ant1 unsaturated groups in the products of the priniarv cteconipo~itionof the coal. The chief ruiiction ot the catalyst is to increase the rate of rc.generation 01 a hydrogen carrier, which is a hytiroarc matic cornpound such as hexahydronaphIhalenf: 1' he reactions of the hydrogeti carrier 11 ith c~\;ygeiiand unsaturated groups are largely rioii-catalytic 'the e8ects of the catalyst on the rate 01 oxygen elirriination and on that of liquefact r i m are verl; similar, a d it is probable that both processes are intimately associated in one reaction or group of reactions. I'II r5TjI IIGII,
Rr CEIVED OCTOBER 9, 1941
P\
Farticular interest attaches to this reaction in view of the statement of Nelson and Beegle4that the equilibrium in this reversible isomerization i7 not changed by temperature. This statement has been accepted recently by Kendrew and Moelwyi-Hughes2 in a paper appearing while the work was in process, although their data would spem to us t o point strongly against its truth. If i t were true, the entire free energy change of the reaction would necessarily be an entropy change, and the determination of the free en q i e 5 and entropies of activation should be particularly interesting. Our work, however, s e f n i s to show definitely that the equilibrium constant does change with temperature, albeit in a rather unusual manner. The temperature coefficients of both the rates and the equilibrium constants will be shown to exhibit parallel anomalies wliich will riced to be considered in any mechanisin proposed for this reaction. le
7 HIS
JOURVAL
41, 559 (1910)
