A Study of the Catalytic Hydrogenation of ... - ACS Publications

BY HILTON A. SMITH AND BILLY L. STUMP. RECEIVED JANUARY 26, 1961. The catalytic hydrogenation of phenol, catechol, resorcinol, hydroquinone, ...
3 downloads 0 Views 599KB Size
fune20, 1961

HYDROGENATION OF HYDROXYBENZENES OVER PLATINUM AND RHODIUM

converted to disintegration per minute (d.p.m.) by measurernent of the exact counting efficiencies with calibrated tritium solutions.

Acknowledgments.-The author is grateful to Dr. V P. Guinn for furnishing the radioactivity measurements, t o E. L. Sanborn and J. R. Weaver

[CONTRIBUTION FROM

THE

2739

for emission spectroscopic analyses of the magnesiulll metals, to p. A. wadsworth, Jr,, for mass spectronietric analyses. Appreciation is also expressed to Drs. R. D. Mullineallx, T. P. Rudy and H. E. De La Mare for criticism of this manuscript before publication.

DEPARTMENT OF CHEXISTRY, THEUSIVERSITVOF TESNESSEE,KNOXVILLE, TESS.]

A Study of the Catalytic Hydrogenation of Hydroxybenzenes over Platinum and Rhodium Catalysts BY HILTON A. SMITH AND BILLYL. STUMP RECEIVED JANUARY 26, 1961 The catalytic hydrogenation of phenol, catechol, resorcinol, hydroquinone, phloroglucinol and pyrogallol has been studied n ith platinum and rhodium catalysts, The rates of reduction are of the same order as those found for corresponding methyland methoxybenzenes. Some cleavage occurs in the reduction process, the amount depending on both temperature and catalyst. While ketones are produced during the reduction of the phenols, the kinetic evidence indicates that the mechanism of formation of cyclohexanols from the phenols does not require ketone intermediates. It is suggested that cyclohexenols are formed as intermediates, and t h a t these mav be reduced to form cyclohexanols, cleaved to form cyclohexanes, or isomerized to form cyclohexanones.

It has been recognized for many years t h a t when phenol undergoes hydrogenation, the predominant product may be cyclohexanol, benzene or cyclohexane, depending on the catalyst employed and the reaction. Under certain conditions and with specific catalysts, cyclohexanone may be isolated during the course of the reaction, and the role of this compound as an “intermediate” in phenol hydrogenation has been investigated. Vavon and Berton’ were the first to isolate cyclohexanone as an “intermediate” in the hydrogenation of phenol. They hydrogenated phenol and the cresols over platinum black and isolated the cyclohexanones formed by reaction of the ketone with semicarbazide to produce the semicarbazone derivative. Coussemant and JungersZ made a thorough study of the kinetics of catalytic hydrogenation of phenol over Raney nickel. On the basis of the results obtained in these investigations, the following reaction mechanism has been proposed

tematic study of the cleavage reaction of the methoxyl group from an aromatic nucleus during hydrogenation, and it was felt t h a t an extension of this study to include hydroxyl cleavage would lead to a further understanding of the reaction. A possible mechanism for the cleavage of the carbon-tooxygen bond during such catalytic hydrogenations is: The methoxybenzene, anisole for example, is adsorbed on the catalyst surface where hydrogenation takes place to give a niethoxycyclohexene as a short-lived intermediate. This intermediate may undergo cleavage and hydrogenation to give cyclohexane and methanol, or i t may undergo simple hydrogenation to give methoxycyclohexane. These two reactions would probably occur simultaneously until the double-bonded molecules were saturated. Finally, it was the purpose of this research to extend the kinetic data for catalytic hydrogenation of hydroxybenzenes. Experimental

Wicker3 hydrogenated 0-,m-,p-cresol, and the corresponding methylcyclohexanones in acetic acid over Adams platinum catalyst a t 20’. Stereochemical evidence obtained indicated t h a t these hydroxy compounds do not hydrogenate via a ketone intermediate. I t was the purpose of this research to extend the study of phenol hydrogenations to include the isomeric dihydroxybenzenes, catechol, resorcinol and hydroquinone, to find out whether ketones or diketones are formed during the course of the reaction and, if so, in what amounts. Of interest also was the cleavage of hydroxyl groups from the aromatic ring during the course of hydrogenation. Smith and Thompson4 made a sys(1) G . Vavon and A . L. Berton, Bull. SOC. chim., 37, 296 (1925). ( 2 ) F. Coussemant and J. C. Jungers, Bull. soc. chim. Belgcs, 59, 295 ( 1950). (3) R . J. Wicker, J. C h i n . SOC.,3289 (1957). (4) H. A. Smith and R. G. Thompson, “Advances i n Catalysis,” 1’01. IX, Academic Prcss, Inc., h-ew York, N. Y . , 1057, p. 727.

Comnicrcial prcpar:itii~nsof pl:ttinum oxide werc obtaiiictl froin the -4nicrican Platinuin TT’orks, Xewirk, N . J., aiitl Goldsmith Brothers, Chicago, Ill. The 5% rhodium-onalumina catalyst was obtained from Baker and Co., IIIC., Newark, 9. J. Glacial acetic acid i i a s purified by fractionation of du Pont C.P. acid through an 8-foot Viereux column. Commercial hydrogen o b t i n e d from the Cational Cylinder Gas Co. was used without further purification. This hydrogen lias been shown pre\.iously t o be satisfactory for kinetic studies. Matheson (Coleman and Bell Division) phenol was fractionated through a 2-foot 1-igreux column and the fraction boiling at 181.6’ (745.8 mm.) was used. Eastman Kodak C o . white label catechol and RIerck and Co., Inc., hydroquinone were purified by recrystallization from benzene. Melting points are 108.5’ and 175.4-176.6’, respectively. Matheson (Coleman and Bell Division) resorcinol was recrystallized from toluene (m.p. 111.2-111.6°). Eastman Kodak Co. white label phloroglucinol and cyclohexanone were used without further purification. Baker Chemical Co. pyrogallol was recrystallized from a n ethanol-benzene solution (m.p. 136’). A much simpler method than has been reported previously was used to prepire dihydroresot-cinol. Ten grams of resorcinol was hydrogetiated a t room tcrnperature in 20 ml. of water containing 4.30 g . of sodium hydroxide over 2.0

2740

HILTON

A.

SMITH AND

BILLS L.

Vol. 83

STUhlP

g. of 59" rhodium-on-alumina catalyst.

The initial pressure Experimental Calculations was 50 p.s.i.g. The reaction ceased after 37y0 of tlie The hydrogenation of hydroxybenzenes was theoretical amouut of hydrogen had been absorbed. The catalyst was removed by filtration and the reaction tnix- found to be first order with respect to hydrogen ture was treated with concentrated hydrochloric acid until pressure, zero order with respect to concentration ilie solution was acidic t o congo red test paper. Chilling the solution in the refrigerator produced 8.90 g. (87.47; of hydrogen acceptor and directly proportionril to the weight of catalyst used. The reaction rate is yield) of dihydroresorcinol, m.p. 104-!05°. Recrystallizntion from benzene was necessary to remove traces of sodium given bya - d r j d t kfJ/'.!i chloride from the product, but the meiting point mas not raised.6 where P is the pressure, k is the specific r'A. t e conSeveral hydrogenations ol catechol were made in methanol over 570 rhodium-on-alumina catalyst, and the 1,ercentage stant, 17 is the volume of the system and t is time. ketone present was studied as a function of percentage reacValues of k were obtained by multiplying by tion. This was accomplished by stopping the hydrogena-2.3031' the slope of the straight line oi a log I-' tions at various degrees of completion, removing the catalyst arid solvent, and treating the residue xhich remained with V S . f piot. All rate constants are referred to 1.0 g. of catalyst (kl$, the units of I Z ~being . ~ liters,/minute. phenylhydrazine. 2-Hydroxycyclohexanone phenylhydrazone, m.p. 117.4-118°,6 was isolated in each case. T h e perAs previously noted in cetalytic hjdrogenatims, ccrit.age ketone present iii the reaction mixture r e x h e d a there was often a s!ow tirift fruiii linearity after maximum of 16 at SOYc completion. n'hen the reactions CO-SO% rextion. This was uiidoubtedly due to a were repeated using acetic acid solvent in place of methanol, poisoning or decay of the active catalyst surface. the percentage ketone present in the reaction mixture reached a maximum of 23%, again at 509; reaction. It apRate constants were reprodtlcible within 5%. pears, then, t h a t 2-hgdroxycyclohexanone is formed in the Activation energies were obtained from the usual hydrogenation of catechol, rather than the diketone 1,2ilrrhenius plots. cyclohexanedione. Catechol was hydrogenated in acetic acid over Adanis Results and Discussicn platinum and in the presence of semicarbazide hydrochloride. In platinum-catalyzed hydrogenations Of b m A white solid was isolated from this rcaction mixture which iiielted a t 198-199" a n d , from an elemental analysis, ap- zene, methylbenzenes and rnethox) benzenes (acetic acid solvent, ordinary pressures and ternperapeared t o be 2-hydroxycyclohexanone semicarbazone dihydrate. tures), it has been shown that the reactions are ,4naZ. Calcd. for C,HI?K~OI: C, 40.56; €1, 8.27: N, first order in hydrogen pressure, zero order in con20.28. Found: C, 40.6; H, 7.9; N, 20.42. centration of acceptor and directly proportional to Several hydrogenations OF reiorcinol were carried to vary- catalyst eight.^!^ The same kinetic picture was ing degrees of completion over both Adams platiiiuin and exhibited by the hydroxybenzenes. The hytlro5 % rhodium-on-alurnin~cat:tiysth, but in no case wvns i t genation rates for methylbenzcnes decrease with possible to isolate a ketone intermediate. the number of substituents, so that the rate for Hydroquinone was hydrogenated to 40%? completion in benzene toluene> o-xylene lieniir~ielli~~n~~ acetic acid over 5'74 rhodium-on-:tluniitia. The residue which remained after removing the catalyst and solvent WHS For compounds with the same number of substidissolved in ethanol and treated with 2,4-diriitrophenyltuents, the one with a symmetrical arrangement hydrazine. Crystals were obtained which melted a t 157" has the highest rate, the vicinal isomer has the a:ter recrystallization from a methanol-water mixture. lomest rate, and the unsymmetrical isomer an The derivative was analyzed as 94-hydroxycyclohexanone 2,4dinitrophenylhydrazone. intermediate rate: p-xylene nz-xylene a-xylene. A n d . Calcd. for C12R11?;IOL: C, 48.9s; €I, 4.80; K, There exists an overlapping of these factors such 19.04. Found: C,48.93; H, 4.71; X, 19.10. that a symmetrical compound with a number of

>

>

>

>

This reaction estab5sled the fact t h a t 4-hydroxycycloliexanone rather than 1,4-cyclohexanerlione is the ketone formed during the liydrogeiiation of hydroquinone. The ketones cyclohexanone and 1,3-cgclohexanedione were hydrogenated o i w both ylatinllni oxide and 57" rhodiumon-alumina. A modified Parr low pressure reduction apparatus was employed for all hydrogenation reactions. The general procedure was the same as t h a t previously discussed.' Initial hydrogen pressure was approxiniately 50 p.s.i.g. with changes in pressure of 5 to 15 p.s.i. during the course of a reaction. As solvent, 25 ml. of glacial acetic acid was employed for all runs. The weights of catalyst and acceptor were 0.10.3 and 1.0 g., respectively. During each run the reaction bottle was enclosed in a metal jacket through which water from a constant temperature bath (*0.1") was circulated. The relationship between pressure drop and moles of hydrogen reacted was determined by the hydrogenation of benzoic acid which is known t o require three moles of hydrogen per mole. Determination of the rate constant for benzoic acid hydrogenation with platinum oxide allowed a comparison of the rate constants with previous results with this catalyst. (5) Since preparation of this article, this procedure has also been reported by E. Esch and H . J. Schaeffsr, J . A m . Pharm. Assoc., Sci. E d . , 49, 786 (1960). (ti) W. Pritzkow, Chcm. Rcr., 87, 1668 (1954). (7) H. A. Smith, D. M. Alderman and F. W. Nadig, J. Am. Chcm. Soc., 67. 272 (1948); H . A. Smith and E. F. II. Pcnnekamp, i b i d . , 6'7, 276, 279 (1945).

substituents has a higher rate than a vicinal isonier with one less substituent: rnesitylene > a-xylene. Siiiiilar behavior was noted for rnetho~ybenzenes.~ As shown in Table I, the rate constants for tllf hydrugeiia,tion of hydroxybenzenes reveal the salile effect of s;-iniiietry, number of substituents, etc.. as for t h e iiiethylbenzciics. Tile relative rates arc' TABLE I COXPARISON Compound

CONSTANTS .4T 30" GENATIOSS OVER PLATINUhC

O F IIATR

ki.85

k!lydioivl kmethvlb

FOR

H\WRO-

f_ihidrn\i I

kmerhaxvlc

0.90 1.02 Phenol 0.1620 1.43 1.58 .1330 Catechol 1.59 ,1611 1.12 Resorcinol 1.10 1.29 .2!)72 Hydroquinone 2.26 2.11 .0951 Pyrogallol a The k l . o is iii 1. g.? inin.-' and is given per unit weiglit of catalyst, * Calculated using data obtained from Smith and Penneka1np.7 Calculated using data obtained from Smith and Thompson.'

not only in the same order, but fair agreement i l l actual values also exists. exception is found (8) II. A. Smith and J. 12. Fuzek, J . A n t . Chem. Soc., 70, 3743 (1918).

JuncL'O, 1961

HYDROGENATION 01'HYDROXYBENZENES OVER PLATINUM AND RHODIUM

with the trisuhstitut.ed compounds, where the hydroxyl coiiipound hydrogenwtes about twice as rapidly as the methyl coinpound. It is interesting to note, too, that reduction of catechol is somewhat faster than that for o-xylene. Any difference in steric requiretnents would be manifested with these two compounds, and this probably accounts for the results observed here. The correspondencc between hydroxy- and methoxybenzene rate constants can also be seen from the data in Table I. The 0- and rrz- dihydroxy compounds hydrogenate about 1.ii times as rapidly as the corresponding methoxyl compounds. This difference in rates may be attributable to the difference in size of the hydroxyl arid methoxyl groups, but the extensive forination of ketones during the course of the hydrogenation of the hydrosybenzenes could be an equally important factor. Especially noteworthy is the fact that the ratio of the rate constants for heminiellitene and 1,2,3-trimethoxybenzene is about one, while the ratio of the rate constant fur pyrogallol to either of these compounds is greater than two. The rates of hydrogenation at 30' for the hydroxybenzenes using 5% rhodium-on-alumina show the same general effect of symmetry as observed for the corresponding compounds with platinum. Table I1 shows a comparison of the r'Jt e constants of hyciroxybenzenes with those of methoxybenzenes over 5% rliocliurri-on-alum in^. The relative rates are in the same order, and fair correspondence of values also exists.

TIONS OF

I'hcnol 0.06dr 0.78 Catechoi .0146 .46 Resorciiiol ,0311 .86 Hydroquinoiic .113::3 1.OU ' 1 The k!.o in 1. g . 3 viiii. -1 and is given per t i i i i t iveiglit or total cat:T.ly,st. CLt!cul:tted ~ I ~ J I Idata I obtained from Smith aiid lhorripsori.'

I n Table TI1 arc shown the moles of liydrogen absorbed for each rriole of compound hydrogenated. Included are values bot.11 for platinum and for 5% rhodium-on-alumina catalysts. Hydrogenation of TABLE I11 CLEAVAGE OP HYDROXYL GROUPS AS A FUNCTION OF TEMPERATURE Over A d a m platinum ---Temperatures----.

Compound Phenol Catechol Resorcinol Hydroquinone Phloroglucinol Pyrogallol

respectively. I t is obvious that the extent of cleavage is dependent upon both the catalyst and the temperature. Hydrogenolysis is quite extensive over Adams platinum catalyst and appears to be a linear function of reaction temperature. The results obtained indicate that the higher reaction temperature and use of platinum oxide catalyst promote cleavage, while low temperatures and the use of rhodium on alumina conibine to retard cleavage during hydrogenation. I t can also be seen from the data in Table 111 that the o-isomer undergoes less cleavage than the in- or p - isomers. This may be due to a slight steric inhibition of cleavage for the o-isomer. These results parallel those reported earlier for hydrogenation of methoxpbenzenes over rhodium and platinum catalyst^.^ Table I V shows the hydrogenation rates a s a function of temperature for the hydroxybenzenes. TABLE IV REACTION RATESAND ACTIVATIONENERGIES FOR GEXATIONS OF HYDROIYRESZENES

Compound

Phenol

Catechol

20" 30' 40" 50" 3 . 1 9 3 . 2 2 3.31 3.38 3.51 3.53 3.66 3.68 3.91 4.00 4 . 0 3 4.07 3.95 3.97 4 05 4 . 0 5 3.28 3.48 3.54 3 . 6 4

3.57 3.70 3 . 7 2 3.83

Over 5 % Rh-on-alumina ---Temperature20' 30' 40° 50" 3 . 0 4 3 . 0 1 2.99 3.04

3.04 3.07 3.29 2.78 3.12

2.94 3.10 3.18 2.85 3.23

3.10 3.16 3.14 2.88 3.24

3.13 3.16 3.21 2.81 3.24

Values are given in moles of hydrogen absorbed per mole of acceptor.

the ring requires three moles of hydrogen per mole of acceptor, while ring-hydrogenation accompanied by complete hydroxyl cleavage requires 4, 5 and A moles for the mono-, di- and trihydroxybenzenes,

T, "C.

20 30 40 50 20 30 30

Ir Resorcinol COSSTANTS A r ::oo FOR HYDROI~YDROXYUENZENES ASD > I E ~ I I O X Y I ! E N Z U ~ K S 'r.2nI.s

C O h l P A R I S o X 1)F R A T E

2741

50 20 30 40

50 Hydroquinouc 20 30 40 50 Phloroglucii~ol 20 30 40 50 Pyrogallol 20 30 40 50

Over Adams -platinumAH,, kt.o, kcal. 1. m h - 1 mole-'

0.13G1 .1620 ,2050 ,243.4 .1015 . 1330 ,1748 .2281 ,1177 .1611 ,2129 ,2856 . 1571 .2072 ,2611 ,3175 ,0211 ,0378 ,9614 ,1001 ,0628 ,0951 ,1390 ,1687

4.49

4.62

5.61

4.40

9.67

G.54

--

Ei's'DRll-

-.

Over 574, Rh-onalumina-.

km, 1. min.-l

0.0445 ,0687 . 10-14 .1249 .0102 . 0 145 ,0211 ,0282 ,0177 ,0311 ,0442 .(lGS3 ,0281 ,0363 ,0445 .0653 ,00471 ,00807 .0160 ,0263 ,00463 ,00757 .0118 .0174

AHa.

kc.,!. mole-1

6.61

6.43

7.8-1

ti, 49

10.82

8.25

Activation energies were calculated from the "least-squares" slopes of activation energy plots. For each compound, the activation energy was greater for hydrogenation with the supported rhodium catalyst than with Adams platinum. The low rate constants and high activation energies for phloroglucinol may be attributed to resonance stabilization of this compound in the keto form. The results reported here indicate that 2hydroxy- and 4-hydroxycyclohexanone are formed during the hydrogenation of catechol and hydroquinone, respectively. No evidence for ketone formation was found in the hydrogenation of resorcinol. Kinetically speaking, cyclohexanone and 1,3-cyclohexanedione could be intermediates in the platinum-catalyzed hydrogenations of phenol and resorcinol, since the rate constants for these

HILTON A. SMITH A N D BILLYL. STUMP

9-12

TAELE Ir 30" F O R I ~ Y ' D R O G E N A T I O S SO F Adams platinum----. ----

C O M P A K I S O S S OF l

"

when desorbed from the catalyst into the reaction mixture. Experiments performed dwing this research have shown that 2-hydrox>-cyclohexanoneis formed during the hydrogenation of catechol. Intermediate A has a double-bond 3-7 to a hydroxyl group and renders this group labile to hydrogenolysis. Intermediate D makes the 1ea.st contribution to the over-all reaction. If steric factors are important a t all in determining the course'of the reaction, intermediate B would be expected to predominate. This would account for the fact t h a t catechol cleaves less rapidly than do resorcinol and hydroquinone. Acknowledgment.-This research was supported by the Petroleum Research Fund of the hrnericaii Chemical Society.

CHEIIISTRY, TOTVA S T A T E UNIVERSITY, .%XES, IO\1.A:

11. Eliminations of Axial and Equatorial Acetates'

BY C. H. DEPUYASD R. W.KING RECEIVED FEBRUARY 16, 1861 cis- and ti.ans-1-methyl-i-t-butylcyclohexgl acetates have been prepared and n study made of the relative amounts of exo and ettdo olefin formed in the pyrolyses of these conformationally pure esters. I t was shown t h a t there are no confornzntio?za~ e f e c t s on the direction of elinzinetion in the pyrolysis of these esters, both isomers giving the same esojendo product ratio as was found in the pyrolysis of 1-methylcyclohexyl acetate. Relative rate studies on these same two acetates and on cis- and finns-4-t-butylcyclohex).l acetate showed t h a t the ease of pyrolysis depends mainly m i the ground state energy of the acetcites, since with the 1-methyl-4-t-butylcyclohexyl acetates the equatorial ester undergres eliniination more readily, wliile the axial ester is the more reactive of the 4-t-butylcyclohesyl acetates.

SVhen 1-methylcyclohexyl acetate (I) is pyrolyzed a t 400°, it is smoothly converted into a mixture of methylenecyclohexane (11) and l-methylcyclohexene (111). Although i t was a t first thought that the former of these olefins was the predominant product of this elimination, more recent studies 0 C H I OCCH,

\/

C'HZ

I,

CH,

I

U I

11, 2W0

111, 747,

(1) P a r t I, C. H. DePuy, C. A. Bishop and C. N . Goeders, J . A m . C h c m . Soc., 83, 2151 (1961). This research was supported by a g r a n t from t h e Petroleum Research F u n d administered by t h e American Chemical Society. Grateful acknonledgment is hereby made t o the donors of this fund.

have shown that the internal olefin I11 makes up 74Yc of the olefinic product.2 The preponderance of 1-methylcyclohexene has been ascribed to the fact that pyrolytic eliminations from esters have transition states with appreciable double bond character? and that the well-known greater stability of iiiternal over external olefins in the cyclohexane series4 accounts for its ease of formation. Nevertheless, in more complicated systems e m olefin has been reported to be the major product of elimination. Thus Maali acetate (IV) decomposes (2) (a) D . H. Froemsdorf, C. H. Collins, G. S. Hammond and C. H . DePuy, ibid., 81, 043 (1959); ( b ) W J . Bailey and W. F. Hale, ibid., 81, 851 (1959); (c) R. A. Benkeser a n d J. J. Hazdra, ibid., 81, 228 (1959). (3) C. H. DePuy and R. W. Ring, Chem. Reo.., 60, 431 (1960) (4) H. C. Brown, J. H . Rrervster and H. Shechter, J . Am. Chem. S o c . , 7 6 , 467 (1954).