General Acid Catalysis in Moderately Concentrated AqueousSulphuric AcidBY A. J. KRESGE, L. E. HAKKA, S. MYLONAKIS AND Y. SATODept. of Chemistry, Illinois Institute of Technology, Chicago, Illinois 6061 6Received 12th January, 1965Rates of aromatic hydrogen exchange in 1,3-dimethoxybenzene are found to be 23-3 timesgreater in moderately concentrated sulphuric acid than in moderately concentrated perchloric acidwhen comparison is made on the basis of ho values. Somewhat smaller differences in the oppositedirection are found for the equilibrium protonation of azulene in these same solutions. Com-bination of these kinetic and equilibrium data provides a measure of the rate of aromatic hydrogenexchange in sulphuric acid through acidic species other than H3O+, and a comparison of this excessrate with concentrations of various solution species indicates that the additional reaction occursby proton transfer from HSOZ and H2S04.This establishes general acid catalysis for aromatichydrogen exchange in moderately concentrated sulphuric acid and shows that there is no fundamentaldifference between proton transfer from strong acids and proton transfer from weak acids.A considerable body of evidence indicates that acid-catalyzed aromatic hydrogenexchange occurs through simple protonation and deprotonation of the aromaticsubstrate.1-* The reaction has a single intermediate, the cationic species HArH+,in which the exchanging hydrogens occupy equivalent positions :H’Ar + HA%H’ArH+ + A-+HAr + H’AA reaction of this type should show general acid catalys-h, and general acid catalysisis observed for aromatic hydrogen exchange in dilute aqueous solutions of weakacids.19 2 Under these conditions, the rate of exchange is the sum of contributionsfrom all acidic species present in the reaction mixture, and these contributions arecorrelated well by the Bronsted relation.However, in concentrated solutions of strong mineral acids, rates of aromatichydrogen exchange seem to bear no simple relationship to the concentrations of theacidic species, and general acid catalysis seems to be absent3 10 The rate is governedinstead by an acidity function which is a measure of the thermodynamic acidity of thesolution. This difference between the behaviour in dilute solution and the behaviourin concentrated solution is unexpected,los 11 and it has led to the suggestion that afundamental difference exists between proton transfer from weak acids and protontransfer from strong acids.12 In order to determine whether such a distinction isnecessary, we have undertaken the examination of aromatic hydrogen exchange inmoderately concentrated solutions of strong acids.It is difficult to say just how general acid catalysis in concentrated solutionsshould be recognized.Proton transfer is an ionic reaction and is certain to shownon-ideal behaviour in solutions containing large amounts of dissociated electrolytes.In concentrated solutions of strong acids, therefore, quite large deviations from pro-portionality between rate and concentration can be expected.Any estimate of thesedeviations will necessarily be uncertain, for added to the normal problems encounteredin treating forces between stable ions at high concentrations is the considerable diffi-culty of handling transition states. It would seem, therefore, that an analysis along75(176 CATALYSIS I N SULPHURIC ACIDthe lines usually carried out for dilute solutions will not be diagnostic. The conclusionwhich has been reached from one analysis of this type, i.e., that general acid catalysisis absent in aromatic hydrogen exchange,g should perhaps be taken with somereservation.A method of detecting general acid catalysis in concentrated solution which seemsmore reliable because it corrects for non-ideal behaviour in an empirical way is thecomparison of reaction rates in solutions of different kinds of acids.In a moderatelyconcentrated aqueous solution of a strong monobasic acid such as perchloric or nitric,the chief acidic species is the hydronium ion, H 3 0 f . In a moderately concentratedsolution of a polybasic acid such as sulphuric or phosphoric, there are appreciableamounts of other acidic species as well as H30+. If these other acids contributeto the rate of reaction, i.e., if the process is catalyzed by general acids, then reactionwill be faster in the polybasic acids than in the monobasic acids. This kind ofcomparison of rates has already been made for a number of reactions which showgeneral acid catalysis in dilute solution, and faster rates in polybasic than in monobasicacids have usually been found.Thus, sulphuric acid is a more effective catalyst thanperchloric acid in the iodination of acetophenone.13 More recently, this test hasbeen applied to several reactions which involve rate-determining proton transferfrom catalyst to substrate and which, therefore, are more closely analagous to aro-matic hydrogen exchange. In the hydrolysis of aryl boronic acids, the effectivenessof strong acid catalysts increases in the order HC104, H2SO4, H3P04;14 similarily,H3P04 and H2S04 are better catalysts than HCl, HN03, and HC104 for the hydrationof mesityl oxide and crotona1dehyde;ls and H2S04 is more effective than HC104in the dehydration of Ph.CH(OH).CH2.COMe and p-N02.C6H4.CH(OH).CH2.COMe,l6 in the formation of 1 ,Zcyclohexadione from its en0l,l7 and in the hydrolysisof vinyl mercuric iodide.18This test has also been applied to aromatic hydrogen exchange, but the resultsare equivocal. With p-chlorophenol, exchange was more rapid in H3PQ4 than inH2SO4, but with p-cresol, exchange was slower in H2SQ4 than in HCl.19 Thisinconsistency, however, may be more apparent than real, for these comparisons weremade on the basis of old values of the acidity function ho * before it was generallyrecognized that acidity function values are strongly dependent on the structure of theindicator bases used to measure them.Recently, ho has been redetermined insulphuric 20 and perchloric 21 acids by spectrophotometric methods with the sameset of primary amine indicators.These new values of ho are significantly differentfrom older values 22 over considerable regions of acidity. They make available,for the first time, a self-consistent and precise acidity function for two acids, and,in so doing, provide a most suitable basis on which to make fine comparisons of acid-base phenomena in two different concentrated acids. As part of our attempt todetermine whether aromatic hydrogen exchange is subject to general acid catalysisin concentrated aqueous acids, we therefore have measured rates of exchange of arepresentative aromatic substrate in perchloric and sulphuric acids.There is still another way in which general acid catalysis in concentrated solutionsmight be detected for aromatic hydrogen exchange.This method is based on thefact that the intermediate in exchange, HArH+, can be detected easily and its con-centration measured accurately when it is present in sufficiently large amounts in* It would seem that an acidity function is the most suitable basis on which to make a com-parison of rates which is designed to detect any reaction through an acidic species other than H3O+.An acidity function contains within it the effect of non-ideal behaviour on equilibrium proton transferfrom H3O+ ; it comes closer, therefore, to making the proper correction for kinetic proton transferfrom H3O+ than other measures of acidity such as CH~O+ or wt. 74 acid which are purely stoichio-metric quantitiesA . J . KRESGE, L. E . HAKKA, S . MYLONAKIS AND Y .SAT0 77acidic solutions.39 49 23 In any acidic solution, the concentration of HArH+ willbe governed by the thermodynamic acidity of that solution toward an aromatic base.If the solution is aqueous, the thermodynamic acidity is determined by the advitiesof the H30+-H20 pair, and the concentration of HArHf is controlled by its rate offormation from HAr and H30+ and its rate of destruction by reaction with H20.The presence of other acids and their conjugate bases may raise the rate of inter-conversion of HAr and HArHf, but it should not alter the concentration of HArHffrom the value determined by H3O+ and H20. This phenomenon can be used todetect aromatic hydrogen exchange by any acid other than H30f, and it can alsoprovide a measure of the additional rate.To do this, rates of aromatic hydrogenexchange and concentrations of MArHf must be measured in two different acids,one monobasic and one polybasic. If the rate of exchange in the polybasic acid isgreater than the rate of exchange in the monobasic acid at the same concentrationof HArH+, then reaction must be occurring through some species other than H@+in the polybasic acid. The extent to which the rate in the polybasic acid exceeds thatin the monobasic acid is a measure of the rate of reaction by other acids.This four-fold comparison of rates and equilibria of aromatic protonation shouldmost properly be performed on a single aromatic substrate. But in the region ofacidity where measureable amounts of HArH+ are formed from a given substrate,rates of aromatic hydrogen exchange for that substrate are very fast. We thereforeused two substrates of somewhat different reactivity and basicity to make this com-parison.For the kinetic experiments, we chose 1,3-dimethoxybenzene labelled withtritium at the 4-position and measured aromatic hydrogen exchange by the rate ofloss of tritium from this substance. This substrate is the least reactive aromaticfor which general acid catalysis has been detected in dilute solutions of weak acidsand for which the parameter a in the Bronsted relation is known ;24 its rate of exchangecan be measured by conventional methods in sulphuric or perchloric acid up to30-40 wt. % acid. None of the simple benzene derivatives is sufficiently basic to giveaccurately measurable amounts of HArHf at these acidities, but azulene is a stronglybasic aromatic whose kinetic and equilibrium protonations are closely similar to thoseof the methoxybenzenes.294 We therefore used azulene for the equilibriumexperiments.EXPERIMENTALMATERTALS.-~ ,3-Dimethoxybenzene-4-t was prepared by treating a solution of 1,3-dimethoxyphenyl magnesium bromide in tetrahydrofuran with tritiated water.Theproduct was isolated in the usual way and was purified by fractional distillation at atmos-pheric pressure. All other materials used were the best available commercial grades;azulene was purified by vacuum sublimation and recrystallization.KINETICSRate measurements were made at 24-69f0.02"C by a method essentially the same asthat already described.1 The wholly aqueous reaction mixtures were prepared by mixingone part of a solution approximately 10-3 M in aromatic with 75 or 100 parts of an acidsolution.Usually, ten 5-ml kinetic samples were taken at time intervals which providedcounting rates ranging from 105 c.p.m. down to a few times background (30 c.g.m.) ; thisfurnished kinetic data over at least ten half-lives. Acidimetric determinations were carriedout directly on samples of reaction mixture.The kinetic data are presented in table 1. They obey the following relationships : inHC104, loglo k = -(3.273 &-0-017)- (1.139 Ilt0.009)Ho ; in H2S04, loglo k = - (2-858 f0.013)- (1.175 f0.023)Ho. (Errors are standard deviations.78 CATALYSIS IN SULPHURlC ACIDTABLE 1.-bTE OF LOSS OF TRITIUM FROM 1,3-DIMETHOXYBENZENE AT 25°Cacid wt.% - H$ l@kl (min-1)HClO4 9-94 0.35 0.134Y Y 18.82 0.91 0-582Y Y 26-82 1-39 2.0 1Y Y 33-95 1-86 7.08¶ Y 37-10 2.1 1 13.540.21 2.37 26-7Hzi04 5-21 - 0.08 0.0983Y Y 10.42 0.34 0.33611 15.62 0.70 0.960¶ Y 23-15 1 *23 3.99Y Y 28.35 1.61 10.5¶ S 20.83 1 -07 2.49values for HC104 from ref. (21) and for &SO4 froin ref. (20).TAELE 2.-EQUILIBRIUM PROTONATION OF AZULENE AT 25°Cacid~ ~ 1 0 ~Y Y¶ YY Y9Ywt. % "8-7511-4213.7816-1418.762 1 *2425-9927.5629.8830.827.008.7911.3113-6715.8 118-3320-2622.8825-1927.5429-273 1.50-If$0.270.440.590-740.901 -051.341-441-581 -640-088.220.400.580-720.901 -831.221.381.541-661 -82a solutions contain 0.5 7; methanol by volume.0 values for HCI04 from ref.(21) and for H2S04 from ref. (20).EQUILIBRIAloglo CHA~H+ KHA~ - 1-29- 1.01- 0.75- 0.49- 0.20 + 0.080.690 821.091.21- 1.58 - 1.41- 1.12- 0.89- 0.67- 0.40 - 0.20 + 0.080.360.620.831.13Absorbance measurements were made at 250+ 0-5"C and 2790 A using a Beck.mannDK-2 spectrophotometer. Replicate determinations and independent measurement of thetemperature coefficient of the extinction coefficient at this wavelength showed that absorb-ances were accurate to at least 0.01 unit. Substrate concentrations (2 x 10-5 M) and opticalpaths (1 cm) were chosen so as to make full use of the 0-1 absorbance scale. All solutionscontained the same amount of substrate and were in an aqueous solvent which contained0.5 % methanol by volume.Acidimetric determinations were performed directly on thesA . J . KRESGE, L . E . HAKKA, S . MYLONAKIS AND Y . S A T 0 79solutions. It was found that azulene is slightly unstable under the conditions of theseexperiments. The greatest amount of decomposition occurred in the region of half-protonation, but even here absorbance decreased at a rate less than 1 % per hr., and thisinstability had no effect on the accuracy of these measurements.The equilibrium data are presented in table 2. In both acids, the relationship betweenloglo (CHArH+/CHAr) and HO is not quite linear. In HC104, the slope of a 1 0 g ~ O ( c ~ H f / c ~ )against HO plot increases from 1.69 to 1.87; in H2SO4, from 1.25 to 1.53.Agreementbetween these data for HC104 and published values 4 is good when account is taken of theslightly different NO scales used.DISCUSSIONAromatic hydrogen exchange, as measured by the rate of loss of tritium from1,3-dimethoxybenzene-4-t, is significantly more rapid in sulphuric acid than in per-chloric acid. Fig. 1 shows that the ratio of the rates in the two acids increasessomewhat with acidity; at HO = -0.3, ~ H , s o , / ~ H C ~ O , = 2.5 and at Ho = - 1.7,kH,sO,/kHc1o4 = 3.0. These ratios are similar to those observed in comparisonsmade on other general acid catalyzed reactions, and the interpretation made in theother cases would seem to be allowed here : in sulphuric acid, there is another reactionin addition to proton transfer from H30f.FIG.1 .-Rates of loss of tritium from 1,3-diniethoxybenzene-4-t at 25°C.Behaviour in perchloric acid different from that in sulphuric acid is also foundfor the equilibrium protonation of azulene. Here, however, the difference is in adirection opposite to that found for the rates of aromatic hydrogen exchange of1,3-dimethoxybenzene, and it is somewhat smaller in magnitude than the ratedifference. Fig. 2 shows that the equilibrium difference also increases with acidity ;at HO = -0.2, JH,SO,/~HC~O, ( I = CHA~H+/CHA~) is almost unity and at Ho = - 1-780 CATALYSIS IN SULPHURIC ACIDIH~soJIHc~o, = 0.34. Because the rate and equilibrium differences are in oppositedirections, they will reinforce one another when rates are compared at the same-L I I I0 I 2- HoFIG.2.-Equilibrium protonation of azulene at 25°C.FIG. 3.-Rates of loss of tritium from 1,3-dimethoxybenzene-4-t compared toequilibrium protonation of azulene.concentration of protonated aromatic, HArH+, in the two acids. Fig. 3 shows thatrates of interconversion of HAr and HArH+ are as much as 5.5 times greater in sulphuriA. J . KRESGE, L. E. HAKKA, S . MYLONAKIS AND Y . S A T 0 81acid than in perchloric acid. The two lines of fig. 3, moreover, are still divergingat the highest acidities employed in the investigation, and the rate difference is likelyto be still greater at higher acidities.Rate constants for the additional reaction in sulphuric acid can be obtained bysubtracting from the observed rates in this acid interpolated values of observed ratesin perchloric acid at the same value of C H ~ H + / C H ~ .These “ excess” rates, presentedin table 3, are seen to increase rapidly with the stoichiometric concentration of sul-phuric acid. The composition of sulphuric acid solutions in this concentrationrange is known from Raman measurements,2~ and values of the concentrations ofH@+ and HSO, and the concentration equilibrium constant, (Kc)~soa, interpolatedfrom the Raman values are also given in table 3. Both the concentration of HSO,and ( K c ) ~ s 0 , also increase rapidly with stoichiometric sulphuric acid concentrationin the region of the kinetic measurements, and both quantities have not yet reachedtheir maximum values at the highest acidity employed for the kinetics.In this sense,the excess rate and the two quantities which would be expected to govern the magnitudeof the exchange reaction through HSO,, CHSO~ and ( K c ) ~ s o i , have parallel behaviour.TABLE 3.-ANALYSIS OF RATE DATA102kHC104b 102kexcess ‘HSO+ ‘HSO; (Kc)HSOa cHSO-[(K~)HSOi] %0.54 0-0521 OW6 1 0.70 0-37 0.28 0.191.13 0.118 0.21 8 1.5 0.77 0.59 0.591.76 0.260 0.700 2.3 1.2 1.00 1.2243 0.582 1.91 3.2 1.7 1.5 2.02.74 0.840 3.15 3.6 1.9 1.7 2.53.48 2.05 8.4 4.6 2.4 2 3 3.6= stoichiometric molar concentration of sulphuric acid.b interpolated values of observed rates in HC104 at equivalent values of CA~H+/CW.d molar concentrations and concentration equilibrium constants interpolated from values pro-(kH~S04) -(kHC104)-vided in ref. (25).If, however, all of the excess rate were the result of exchange through HSO,,then, since the value of Bronsted’s a for aromatic hydrogen exchange in 1,3-dimethoxybenzene is 0.5,24 there should be a close proportionality between the excess rate andthe product cHSO~[(&)HSO,-]~.A comparison of columns 3 and 7 of table 3 shows thatthis is not the case : the excess rate increases by a factor of nearly 200 over the rangeof acidity covered by the measurements while the increase in cHSO&&)HSO$ isten times smaller. This difference is most probably not the result of a rate-acceleratingsalt effect on the bisulphate-catalyzed reaction, for any salt effect on the rate shouldalso affect (K&SO, and will be included in the product CHSO~[(K~)HSO~ 3.It seemsmore reasonable to ascribe this difference to incursion of another reaction, exchangethrough molecular sulphuric acid. In the Raman investigations, none of the speciesH2SO4 could be detected below a stoichiometric sulphuric acid concentration of 14 M.The Raman line assigned to H2S04, however, is closely flanked by HSO, and SO,:!lines ; these are especially strong at low stoichiometric sulphuric acid concentrationsand can completely overshadow a weak H2S0.4 line. The Raman work, therefore,cannot be said to exclude the possibility that small concentrations of H2S04 arepresent in the sulphuric acid solutions of these rate measurements. Sinceis a stronger acid than either H3O+ or HSO,, its catalytic coefficient will be greaterthan that of the other acids present in these solutions, and quite small amounts ofH2S04 will make significant contributions to the excess rate82 CATALYSIS IN SULPHURIC ACIDSince the excess rate can be the sum of contributions from reactions throughHSO, and H2S04, it is not possible to isolate the HSO- reaction and obtain a numericalvalue for the catalytic coefficient of bisulphate ion.It seems likely, however, thatthe &SO4 contribution to the rate will be small at the lowest acidity employed(C = 0.54 M) and that the excess rate here is largely the rate constant for reactionby HSO,. On this assumption, the catalytic coefficients of HSO, and H30+ are ofsimilar magnitude, for, at C = 0.54 M, kHC10,xkexcess and CH~O+% CHSO,-.Thismight seem to be inconsistent with the acid strengths of these two species : at C = 0.54,(KC)~soT = 0.28 whereas the acidity constant of H30+ is usually taken to be 55.5.But the latter is not a true acidity constant ; the ionization constant of H30+ cannotbe measured in aqueous solution, and 55.5 is an assigned value based purely on aformalism. The fact that the catalytic effect of the hydronium ion is usually foundto be lower than the value calculated on the basis of kH30+ = 55.5 by at least an orderof magnitude 26 indicates that this is very probably not the proper value on whichto base rate comparisons.In moderately concentrated aqueous sulphuric acid, then, proton transfer fromHSO, and H2S04 occurs at a rate which is comparable to the rate of proton transferfrom H3O+.It seems, moreover, that as the stoichiometric concentration of sulphuricacid increases, the rates of the reactions through HSO, and H2SO4 increase morerapidly than the rate of reaction through H3O+. But this trend cannot continue,for as the system becomes depleted in water, proton transfer from H3Of must beginto catch up with the other reactions. Complete proton transfer from H30+ liberatesone covalently bound water molecule and several solvating water molecules as well.Proton transfer from MSO, and H2S04, on the other hand, consumes water: theconjugate bases of these acids are ions of greater charge than the acids themselvesand so will require more water for solvation.The same differences will be present,though to a smaller extent, in incomplete proton transfer, i.e., in kinetic protontransfer, from these acids, and a shortage of water, therefore, will favour reactionthrough H3O+ over reaction through HSO, and H2SO4.The former apparent absence of general acid catalysis for aromatic hydrogenexchange in concentrated solutions of strong acids led to the suggestion that there is afundamental difference between proton transfer from weak acids and proton transferfrom strong acids.12 This proposal was based on the observation that the twokinds of acids, the weak acids with which general acid catalysis was observed and thestrong acids for which it was presumably absent, differ markedly in the transferabilityof their protons.The rate of proton transfer between weak acids and water is generallyseveral orders of magnitude smaller than the rate of proton transfer between strongacids and water27 In this investigation, however, we have shown that aromatichydrogen exchange is subject to general acid catalysis by the strong acids HSO, andH2S04. Since the rate of proton transfer between HSO, and water is very rapid,28this observation vitiates the premise on which the above proposal was based andshows that there is no fundamental difference between proton transfer from weakacids and proton transfer from strong acids.This work was supported by grants from the United States Atomic EnergyCommission and the Petroleum Research Fund of the American Chemical Society.1 Kresge and Chiang, J.Amer. Chem. Soc., 1959, 81, 5509 ; 1961,83,2877.2 Colapietro and Long, Chem. and Ind., 1960, 1056. Challis and Long, J. Amer. Chem. SOC..1963, 85, 2524. Schulze and Long, J . -4mer. Chem. Soc., 1964, 86, 331. Thomas and Long,J . Amer. Chem. SOC., 1964, 86, 4770A . J . KRESGE, L. E . HAKKA, S. MYLONAKIS AND Y. SAT0 833 Kresge and Chiang, Proc. Chem. Soc., 1961, 81. Kresge, Barry, Charles and Chiang, J.4 Long and Schulze, J. Amer. Clzem. SOC., 1961, 83, 3340 ; 1964, 86, 322, 327.5Kresge and Chiang, J. Amer. Chem. SOC., 1962, 84, 3976. Kresge, Pure Appl. Chem., 1964,8,243.6 Gold, Lambert and Satchell, Chem. and Ind., 1959, 1312; J. Chem. Soc., 1960, 2461. Battsand Gold, J. Clzem' SOC., 1964, 4284.7 Eaborn and Taylor, J. Chem. SOC., 1960, 3301.8 Melander, Arkiu. Kemi, 1961,17,291 ; 1961, 18, 195.9 Gold and Satchell, J . Chenz. Soc., 1955, 3609.Amer. Cheni. SOC., 1962, 84, 4343.10 Gold in Olah, ed., Friedel Crafts and Related Reactions (Interscience Publishers, New York,11 Melander and Myhre, Arkiu. Kemi, 1959, 13, 507.12 Gold, Proc. Chem. SOC., 1961, 453.13 Zucker and Hammett, J. Amer. Chem. SOC., 1939, 61, 2791.14 Kuivila and Nahabedian, J. Amer. Chem. SOC., 1961, 83, 2159.15 Bell, Preston and Whitney, J. Chem. SOC., 1962, 1167.16 Noyce and Reed, J. Amer. Chem. SOC., 1958, 80, 5539.17 Long and Bakule, J. Amer. Chem. SOC., 1963, 85, 2313.18 Kreevoy and Kretchner, J, Amer. Chem. SOC., 1964, 86, 2435.19 Gold and Satchell, J. Chem. SOC., 1955, 2622.20 Jorgenson and Hartter, J. Amer. Chem. SOC., 1963, 85, 878.21 Yates and Wai, J . Ainer. Chem. SOC., 1964, 86, 5408.22 Long and Paul, Chem. Rev., 1957, 57, 1.23 Schubert and Quacchia, J. Amer. Ckrem. SOC., 1962,84, 3778 ; 1963,85, 1278.24 Kresge and Sato, unpublished work.25 Young and Maranville in Hamer, ed., The Structrire of Electrolytic Solutions (John Wiley26 Bell, Acid-Base Catalysis (Oxford Univ. Press, London, 1941), p. 93.27 Caldin, Fast Reactions in Solution (Blackwell Sci. Publ., Oxford, 1964), p. 263.28 Eigen, Kurtze and Tamm, 2. Elektrochem., 1953, 57, 103.1964), vol. 11, chap. XXIX.and Sons, Inc., New York, 1959), chap. 4