Nucleophilic cleavage of the siliconndash;oxygen bond: acid-catalysed hydrolysis of tributylphenoxysilanes in aqueous organic solvents

摘要

1324 J.C.S. Perkin I1Nucleophilic Cleavage of the Silicon-Oxygen Bond : Acid-catalysedHydrolysis of Tributylphenoxysilanes in Aqueous Organic SolventsBy John R. Chipperfield * and Geoffrey E. Gould, Department of Chemistry, The University, Hull HU6 7RXThe kinetics of acid-catalysed hydrolysis of tributylphenoxysilanes have been studied in aqueous dioxan, aqueouspropan-2-01. and aqueous acetonitrile. In aqueous dioxan, when allowance has been made for changes of acidityfunctions with solvent composition, the hydrolysis is shown to be a t least second-order in water. Substituenteffects are consistent with a mechanism involving fast protonation of the phenoxysilane followed by rate-limitinghydrolysis of the protonated species. The similarity of rates of hydrolysis in each solvent indicates that solvationdifferences between initial and transition states are either small or constant.THERE have been a number of kinetic studies on thesolvolysis of compounds of the type R,MX (R = alkyl oraryl; M = Si or Ge; X includes halogeno, carboxylato,2-pyridyl, hydrido, amino, phenylthio, and phenoxy),and the principal results have been reviewed recent1y.lThe kinetics have generally been interpreted in terms ofa bimolecular solvolysis reaction, similar to the SN2mechanism for nucleophilic substitution at carbon.There is no evidence that these reactions proceed via acation (e.g.R,Si+) as in the S N 1 mechanism at carbon.R. H. Prince, M.T.P. International Reviews of Science:Inorganic Chemistry, Series One, ed.amp;I. L. Tobe, Butterworths,London, 1972, vol. 9, p. 353.The reaction scheme for hydrolysis can be written as inequation (l), where B is a base which can accept H+.When hydrolysis reactions are carried out in a protolyticsolvent such as propan-2-01 the solvent itself can act asthe base B. The reaction is then first-order with respectto R,MX and almost first-order with respect to H,0.293Hydrolysis reactions in mixtures of water and organicsolvents such as acetone, dioxan, and ether, which do notreadily act as B, are first-order with respect to R,MXand the order of reaction (n) with respect to H,O variesR. H. Prince, J . Chem. SOC., 1959, 1783.R. H. Prince and R. E. Timms, Igaovg. Chim. Actu, 1968, 2,2601974 1325between 1 and 6.These n values have been interpretedas evidence that water polymers are formed in thesepartly aqueous solvents, and that these polymers actboth as nucleophile and as base B.3-6The hydrolysis reactions which have been studied tofind n values have all been of the charge type shown inequation (2). The rates of such reactions are verysensitive to the polarity of the solvent as there is a largeused k,, was very small in comparison with k,H*, andthe rate equation could be written as equation (4).With HC10, or HC1 as catalyst, hydrolysis rates weremeasured in aqueous dioxan , aqueous propan-2-01, and - R3MOH + BH' + X- ( 1 ) I+ aqueous acetonitrile. Kinetics were studied under first-order conditions with acid concentrations at least tentimes greater than Bu,SiOPh. Values of ka werecalculated by dividing the observed first-order ratetransition state$neutral molecule + neutral molecule -1 with charge 1- products (2)difference in solvation between the initial state and thetransition state. Values have not been measured forreactions of other charge types (e.g. positive ion +neutral molecule). * There has been comparativelylittle work done on the mechanism of making andbreaking 5-0 bonds. The main work has been on thehydrolysis of organosilicon acetates, R , S ~ O A C , ~ ~ ~ ~ ~ ~ * forwhich kinetics can be easily studied by titration of theacetic acid formed.has shown that hydrolysis of phenoxy-silanes, R,SiOC6H,X (R = alkyl, X = substituent) inaqueous ethanol is catalysed by H+ and OH- and obeysthe rate equation (3), where k, and kb are the acid- andAkermanRate = (k,H+ + k,, + kbOH-)R,SiOPh (3)base-catalytic coefficients and k, is the rate coefficientfor uncatalysed hydrolysis.For the acid-catalysedhydrolysis Akerman proposed that R,SiOPh wasprotonated to give R,SiO(Ph)H+; this was then hydro-lysed to give R,SiOH,+, which lost a proton to giveR,SiOH as product. The rate-limiting step wasprobably the hydrolysis of R,SiO(Ph)H+, although aslow initial protonation was not ruled out.We have studied the kinetics of the acid-catalysedhydrolysis of tri-n-but ylphenoxysilanes in organic sol-vents in order to compare the results with those for un-catalysed hydrolysis of 5-0 bonds and for acid-catalysedhydrolysis of organic esters.RESULTS AND DISCUSSIONTributylphenoxysilane was chosen as a typicalphenoxysilane which was easy to prepare and gaveconvenient hydrolysis rates.Under the conditians* In a recent study of the hydrolysis of trialkyl(phenylthi0)-an n value of 1-51 was obtained in acidic (HCI) aqucous* J. R. Chipperfield and R. H. Prince, J . Chem. Soc., 1963,R. H. Prince and R. E. Timms, Inorg. Chim. Ada, 1967, 1,R. H. Prince and R. E. Timrns, 6norg. Chim. Ada, 1968, 2,silanesdioxan over the very limited range 1.32--2.7 H,O.3567.129.n r m1 separation coefficients, klobs equation (5) by the concentration ofacid present, and are given in Tables 1 and 2. OneRate = k,H+ Bu,SiOPh (4)Rate = klobSBu,SiOPhj (5)product, phenol, was characterised by its U.V.spectrum.The U.V. spectrum in the range 250-300 nm of a solutionof Bu,SiOPh after hydrolysis was compared with theTABLE 1Acid-catalysed hydrolysis of Bu,SiOPh in aqueous dioxan :variation of K , with temperature and with H,Q con-centration (acid = 0 . 0 1 ~ )H,O/ k,/l mol-1 H,Oj/ k,/l mol-1t/"C yo v/v min-la t/"C "/o v/v min-la25.0 10 1.46 40.0 10 3.5125.0 15 1.93 40.0 15 4.8725.0 20 2.62 40.0 25 7.9025.0 40 8.65 40-0 40 17.725.0 50 13.1 45.0 40 23.230.0 1 2-87 50.0 10 6.2930.0 2 2.22 50.0 20 9.930-0 5 1.80 55.0 10 7.630.0 10 1.97 60.0 PO 10.130-0 20 3-47 30.0 10 0.68 b30.0 30 6.20 30.0 20 2.00 h30.0 35 8.35 30.0 30 4.75 b30.0 40 11.5 30.0 35 7.1630.0 45 15.3 30.0 40 10-835.0 20 4.55 30.0 46 15.335.0 40 13.7b HC1 used as catalyst.4 HC10, used as catalyst except where otherwise indicatedspectrum of an equimolar solution of phenol in the samesolvent.The two spectra were identical in both shapeand absorbance, which shows that phenol is one productand that the hydrolysis goes to completion. G.1.c.analysis showed that phenol and Bu,SIOH were theonly products.There is considerable evidence that the Si-0 bond7 R. Danieli and A. Ricci, J.C.S. Perkin 11, 1972, 1471. * G. Schott, H. Kelling, and R. Schild, Chem. Ber., 1966, 99,@ E. Akerrnan. Acta Chern. Scand.. 1956, 10. 298.291; GI S c h t t and K. Deibel, ibid., p, 301.L d l . 10 E. Akerman; Acta Chenz. Scand.; 1957, 11; 3731326 J.C.S. Perkin I1will be broken in the hydrolysis reaction and not thephenolic C-0 bond.The acidic hydrolysis of dimethyl-bk-(D-1-methylpropoxy)silane gives butan-2-01 with thesame optical configuration as the original l-methyl- Blz3SiOPh + HA ~ B ~ , S ~ O ( P ~ ~ ) H + + A- (11)propoxy-group.ll The C-0 bond in aryl ethers is stableto hydrolysis and the breakdown pattern of phenoxy-silanes in a mass spectrometer shows that under theseIf protonation of Bu,SiOPh takes place by reactionwith an acid HA, rather than with Hf equation ( l l ) ,KWHAthe rate of hydrolysis will be given by equation (12).Substitution of H+/KaHA for HA/,4-- gives equationTABLE 2Acicl-catalysed hydroylsis of Bu,SiOPh : variation of K , with H,O concentration in acetonitrile aid propan-8-01 a t 30.0"(acid = 0 .0 1 ~ )1 2 5 10 20 30 40 50 ;H,01/0/6 VIV K,/1 mol-l min-l in MeCN a 44-3 15.8 10.7 8-27 7-8 8.7 9.6K,/1 mol-' min-l in PriOH ZJ 2-65 3.60 4.10 3-70 3.30 3-23 3.30HCl catalyst. 6 HC10, catalyst.conditions the 5-0 bond is broken much more readily (13), and KcqHA/KaHA is equal to Kes, so equation (13)(12)than the c-0 bond.l2 Prince and Timms used l 8 0labelling to show that the Si-0 bond and not the acyl Rate = k2K,,HA,H,01,~,Bu,SiOPhj~HA~~A~C-0 bond is broken in the hydrolysis of Ph,SiOAc. = k,K,,HAH,0nB~3SiOPh H+ /KaHA (13)Uncatalysed hydrolysis reactions of chloro- andacetoxy-silanes have been explained in terms of reactionwith f i water molecules, possibly present as a polymer( H20),,l and a possible scheme for acid-catalysedhydrolysis of phenoxysilanes is shown in equations (6)becomes identical with equation (9) and our resultscannot distinguish between protonation by Hf or HA.By analogy with studies on uncat alysed hydrolysis wewould expect a graph of log ka against log H,O to be astraight line of slope n.Figure 1 shows that this is notK wBu,SiOPh + H+ + Bu,SiO(Ph)H+ (6)and (7). The protonation reaction (6) will be fast ink ,Bu,SiO(Ph)H+ -+ ?iH,O +cornparisoil with the hydrolysis (7), as no substantialelectron reorganisation or change in hybridisation isrequired, and Bu,SiOPh contains no internal hydrogenbonds. Compounds of the type R,SiOPh are weakbases l3 and K,, will be small. This scheme leads to therate equations (8) and (9), and the observed catalyticcoefficient, k,, is given by equation (10).Rate of hydrolysis = k,H,Oj",Bu,SiO(Ph)H'-IBu,SiOH -b PhOH + H,O' + ($2 - 2)H,O (7)(8)= k,KeqHL H,O"Bu,SiOPh (9)0 02 04 0.6 0.8 1.0 1.2 1.4 1.6k , = k2K,,H,072 (10)As expected froin Wkerman's studies we found that the log "/oF,O ( = log H201-0.256)hydrolysis reaction was first-order in Bu,SiOPh and inacid (up to 0.01~).Table 1 shows that HCl and HClO,are equally effective as catalysts in 450/, aqueous dioxan,but that HCl is less effective at lower water concen-trations. HCl0, is completely ionised in aqueousdioxan. Electrochemical measurements show that HC1is completely ionised in 45 aqueous dioxan but isincreasingly present as un-ionised molecules as waterconcentration decreases.14l1 R.H. Krieble and C. A. Burkhard, J . Atner. Clzem. SOC.,l2 G. E. Gould, Ph.D. Thesis, University of Hull, 1971.l3 N. A. Matwiyoff and R. S. Drago, J . Organometallic Claenz.,1947, 69, 2689.1965, 3, 393.FIGURE 1 Acid-catalysed hydrolysis of Bu,SiOPh in aqueousorganic solvents a t 30.0"; plots of log k , against log (H20;(a) djoxan, (b) propan-2-01, (c) acetonitrjlcso for any of the three solvents used. In aqueous dioxanover the range 30-40 water the slope is about 2,suggesting that 2 molecules of water are involved in thehydrolysis.In equation (9) the equilibrium constant, Kes, for theprotonation of Bu,SiOPh equation (6) has beenexpressed in terms of the concentrations of speciespresent. In aqueous organic solvents activity co-l4 H.S. Harned and euro;3. I3. Owen, ' The Physical Chemistry ofElectrolytic Solutions,' 3rd edn., Reinhold, New York, 1958, ch.111974 1327efficients are often far from unity and Keq should beexpressed as in equation (14), where S stands forKeq = cSH-kf2JJL/ug-: . Cg fs (14)Bu,SiOPh. Acidity functions h, have been measured inaqueous dioxan using substituted anilines as indicator.'lo H 20I I I I1 2 5 10 20 40i II FI l i l l ! l l l0.2 0 4 0.6 0.8 1.0 1.2 1.L 1.6log Y o H20 (= log H,O - 0.256)FIGURE 2 Plots of log k , + H, (at contant acid) againstlog H,O: (a) catalyst HClO,, solvent aqueous dioxan; (b)catalyst HC1, solvent aqueous dioxan; (c) catalyst HClO,,solvent aqueous propan-2-01If the ratio fsE+/'s is the same as fia+/hn for theindicator, then Keq can be expressed as in equation (15).K e g = CSHB+/CS .120 (16)Substitution into equation (8) gives (16). By com-parison with equation (4) we get (17), where c is theRate of hydrolysis = K,Keqcamp;oH20n (16)k2Ke,hoH,0n = k,c (17)stoicheiometric concentration of added acid.logarithms in (17) gives (18).TakingHence for a constantlog k, + log Keq - Ho + '12 log H,O =log k, + log c (18)concentration of catalyst, c, a plot of log k, + Ho againstlog H,O should be a straight line of slope n. Hammettacidity functions for HC10, in aqueous dioxan have beenmeasured mainly for c = O-~M, but the data show thath, is a linear function of c up to O - l ~ .~ ~ 9 ~ ~ Hence Hovalues for c = 0 . 0 1 ~ and c = 0 . 1 ~ differ only by aconstant. Thus a plot of log k, + H,, where k, wasdetermined with c = 0 . 0 1 ~ (HClO,) and Ho refers toO.l~-Hc10, against log H,O should have slope n.A fairly linear plot is found from 1 to 45 dioxan withn = 1.8 (Figure 2). A similar plot using k, with HC115 L. L. Leveson and C. W. Thomas, J . Chem. Soc. ( B ) , 1969,1051; J. Koskikallio and A. Ervasti, Acta Chem Scand., 1962,16,701.16 C. W. Thomas and L. L. Leveson, J.C.S. Perkin 11, 1973, 20.17 E. A. Braude and E. S. Stern, J . Chew Soc., 1948, 1976.18 H. Sadek and F. Y . Khalil, 2. phys. C h e w (F~ankfurt), 1968,57, 306.as catalyst and acidity functions for HC117f18 is linearfrom 10 to 450/, aqueous dioxan with n = 2.2.These12 values of around 2 fit in well with results which showthat acid-catalysed hydrolysis of carboxylic esters issecond-order in water.lg One water molecule acts asnucleophile and the other helps to remove a protonfrom the transition state.In hydrolysis reactions in aqueous organic solventswater can have a number of roles. I t is a reactant, itcontributes to the polarity of the medium, and it maybe involved in specific solvent effects in the ground ortransition states. Changes in hydrolysis rates of organo-silicon compounds with water concentration have beenaccounted for in terms of water as a reactant. Ashydrolysis rate is proportional to H,O., n molecules ofwater, possibly present as water polymers, are presumedto be reactant.This treatment does not considerpolarity changes in the solvent. A reaction betweenneutral molecules is strongly accelerated by increase insolvent polarity, and equations similar to (19) (9 and qlog = $(E - 1)/(2E + 1) + q (19)are constants; E is the dielectric constant of the medium)can account for variations of rate with solvent changes.20Over the range 5-40 aqueous dioxan a graph oflog H,O against (E - 1 ) / ( 2 ~ + 1) is a straight line.This leads to the situation where the data of Prince andTimms on the hydrolysis of Pri3SiOAc over the range5-25 aqueous dioxan can be expressed either as (20)or (21). In equation (21) the role of water as a reactantlog k = 4.4 log HZO - 8.3 (20)log k = 24.6(~ - 1 ) / ( 2 ~ + 1) - 13.3 (21)is completely ignored.It would seem reasonable toaccount for the data with a composite equation contain-ing terms for polarity and H20.In acid-catalysed hydrolysis , where the charge typeof the reaction is ' positive ion + neutral molecule,' therate should decrease slightly with increase in solventpolarity. Ingold reports a ten-fold retardation in thereaction of Me.$+ with NMe, when changing fromethanol (E = 24) to water (E = SO). If we assume thatchange in solvent polarity will lower k2 by a factor ofabout 10 as water content increases from 5 to 40 inaqueous dioxan, the observed k,40/ka5 value of 6.5would increase to about 65 if the effect of solventpolarity is eliminated. A plot of log k, + Ho (at con-stant acid concentration) against log H,O has a slopeof about 3 when k, values are used in which the effect ofsolvent polarity has been removed.This indicates that3 molecules of water are involved in reaction (7). Separ-ation of the roles of water as reactant and as a contributor19 A. J. Kirby, in ' Comprehensive Chemical Kinetics,' eds.C. H. Bamford and C. F. H. Tipper, Elsevier, Amsterdam, 1972,vol. 10, p. 106.20 I. A. Koppel and V. A. Palm, 'Advances in Linear FreeEnergy Relationships,' eds. N. B. Chapman and J. Shorter,Plenum Press, Lonton, 1972, ch. 5.21 C. K. Ingold, Structure and Mechanism in Organic Chem-istry,' 2nd edn., Bell, London, 1969, pp. 457-4631328 J.C.S. Perkin I1been attributed to D20 being a weaker nucleophile thanH20, thus reducing kaD/kaH.25 In addition, if theremoval of a proton from the transition state by H,Ois involved in the rate-limiting step (as indicated fromthe order of reaction) , the poorer proton-acceptingability of D,O as compared with H20 will also tend todecrease kaD/kaH.For the three solvents studied the values of K, at 30"in 40 aqueous solvent are: dioxan, 11.5 1 mol-l min-l;acetonitrile, 8.7 1 mol-l min-l; propan-2-01, 3.3 1 mol-lmin-l.This order changes slightly with H,O but theremarkable feature is the small range of k, values found.These are consistent with the small solvent effectsexpected for a reaction between a positively charged ionand a neutral molecule, where there is little change incharge separation between the initial state and thetransition state, and consequently only a small change inthe number of molecules used in solvation.The effect of a substituent X upon the rate of hydro-lysis of Bu,SiOC,H,X is shown in Table 4, which containsTABLE 4Hydrolysis of substituted tributylphenoxysilanes,Bu,SiOC,H,X, in 40 v/v aqueous dioxan at 30.0"(HCIO, = O-OlM)102kMpl kalX A/nm min-1 1 mol-1 min-1 Izar"l$-Me 280 15.62 1.364.18 0-365 :::I 0*01 10.97 0.954o-Me280p-CH215CHH293$:Lo 293 1.00 4.73 0.411m-CHO 325 0.02 7.60 0.660p-Ph 295 0*01 9.14 0.7959-CN 280 1.48 5.18 0.450$-NO2 280 17.6 5.61 0.488o-NO~ 290 7-20 6.24 0.543P-COZEt 285 0.09 5.20 0.462@OMe 305 23.58 2.051.001.0211.5011.70to the polarity of the solvent is difficult as n can bemade greater than 3 if increasing the solvent polarity isconsidered to retard reaction sufficiently.Reaction (7)involves at least two molecules of water, but additionalwater molecules may participate.Values of k, are almost constant between 1 and 40aqueous propan-2-01. Acidity functions for HC10, inaqueous propan-2-01 have not been measured, but as Hovalues for HCl and HC10, behave in a similar way withvariation of water concentration in aqueous dioxan, Hovalues for O-O~M-HC~ in aqueous propan-2-01 may beused as an approximation.22 A plot of log k, + Ho (atconstant acid concentration) against log H20 is astraight line over the range 5 - 4 0 ~ o aqueous propan-2-01with a slope of -0.6. If one molecule of water is areactant (the minimum requirement assuming that H+is removed from the transition state by solvent ratherthan by another water molecule) the plot of log k, + Hoagainst log H2Q should have a slope of 1.For this tobe so the value of ka40/ka5 should be 23, and not 0.8 asmeasured. A 29-fold retardation of rate as the solventchanges from 5 (E = 19) to 40 (E = 37) is reasonable.Although Ho values for 0.01~-HCl in aqueous aceto-nitrile are not available, it is likely that Ho varies withsolvent composition in a similar way to that seen inaqueous dioxan, and changes of k, with water concen-tration can be accounted for in the same way.Bunnett 23 derived the relationship (22) for acid-catalysed hydrolysis in strong acids where a, is theH , + log k = constant + w log a,activity of water in the solution, and w, the slope of agraph of log k + Ho against a,, can be related to themechanism of the hydrolysis reaction.The equationwas derived for aqueous solutions but it has been appliedto hydrolyses in aqueous dioxan.l6 However, using ourdata, together with water activities in aqueous dioxan,=we find that a plot of log k, + Ho against a, is not linearand no conclusions can be drawn about mechanism.Table 3 shows that there is a very small isotope effect(22)TABLE 3Acid-catalysed hydrolysis of Bu,SiOPh in D,O-dioxan :variation of K , with temperature, and with D,Oconcentration (HClOA = 0 . 0 1 ~ )D@I / Aamp;=/t/"C v/v Imol-1 min-l k,H/kaD30.0 40 11.1 1.0435.0 40 13.4 1.0240-0 15 4.18 1.1740-0 20 6.2240.0 25 7.8 1.0240.0 40 17.2 1.0345.0 40 20.5 1.13if D20 is used in place of H20.D+ in D20 is a strongeracid than Hf in H,O by a factor of 3 and hence Kegshould be increased to give kaD/kaE a value of about 3.A value of kaD/kaH of 1-5 has been measured for the acid-catalysed hydrolysis of some organic esters and this has22 A. R Tourky, A. A. Abdel-Hamid, and I. 2. Slim, 2. phys.Chem. (Leifizig), 1972, 250, 61.values of hare1 (k, values with respect to k, for X = H)measured in 40 aqueous dioxan. Values of n areindependent of X in this solvent within experimentalerror (n = 2.4 and 2.6 for X = H and $-NO,, re-spectively) and karel values can be used to show theeffect of X on k, and Keg.Electron release by X shouldenhance Keq and decrease k,. As hare1 decreases whenX is electron-attracting, the effect of X upon Keq out-weighs its effect on k,. The small range of kard valuesimplies either that the effects of X upon K,, and k,almost cancel out for all substituents, or perhaps thatthe reactions are as insensitive to substituent changes asthey are to solvent changes.ortho- and $ara-Methyl substituents should havesimilar electronic effects upon the basicity of the phenolicoxygen atom and so change Kes equally. The value ofk, will probably be lowered by the ortho-methyl groupas a result of steric hindrance in addition to its electronicproperties, which should be similar to those in the para-methyl compound. There should be no steric effects23 J.F. Bunnett, J . Amer. Chem. SOG., 1961, 83, 4956, ef seq.24 A. L. Vierk, 2. anorg. Chrn., 1960,261, 283.25 M. L. Bender, Chem. Rev., 1960, 60, 531974 1329upon Keq in view of the small size of a proton. The valueof hare1 for X = o-Me (0.36) is lower than that forX = $-Me (1-36), which confirms that reaction (7) is therate-limiting step. The value of karel for l-naphthyloxy-silane is low (0-26) and probably arises from a loweringof k, by the large 1-naphthyloxy-group. Plots of log karelagainst Hammett o or o- constants are curved, as oftenfound for reactions with more than one step.26Values of k,, increase rapidly as X becomes moreelectron-withdra~ing,~~ and for fi-NO,C,H,OSiBu, therates of acid-catalysed (H+ = 10-2~) and uncatalysedhydrolysis are similar.Table 5 shows the values ofTABLE 5Values of 12, and K , for the hydrolysis of fi-N0,C,H40SiBu,in aqueous dioxan and in aqueous propan-2-01 at 30.0'Aqueous dioxan Aqueous propan-2-01(HCIO, = 0.0111)7- 7 - 7H,OI/ vlv 30 35 40 20 30 401 02k,,/min-1 4.99 9.44 17.6 8.73 11.8 15.5k,/l mol-l min-l 2.65 3-96 5.61 2.5 2.4 2.5k, and k , found. The slopes of graphs of log k, andlog k , against log H,O over the range 30-40~0aqueous dioxan give n values of 2-6 and 4.4, respectively.The value of 4.4 agrees well with values found for otheruncatalysed hydrolysis reactions of silanes, and 7 = 2-6confirms that protonated phenoxides need fewer watermolecules for hydrolysis. In propan-2-01 the value of neffects of R are similar in acid-catalysed and uncatalysedreactions of R,SiX.The activation parameters obtained (Table 6) arefairly similar to those obtained by Prince and Timms intheir study of the hydrolysis of organosilicon acetates inaqueous d i ~ x a n .~ They attributed the high negativeA S values to the high ordering of water molecules in thetransition state. The mechanism of hydrolysis oforganosilicon compounds in aqueous organic solvents issimilar to that of some organic compounds. Thus nvalues for the hydrolysis of triphenylsilyl acetate aresimilar to those for the hydrolysis of benzyl chloride, butA S for benzyl chloride is of a magnitude normally foundfor bimolecular reactions. Table 6 shows some activ-ation data for several hydrolysis reactions in aqueousorganic solvents.There must be a specific factorassociated with silicon which gives rise to the highnegative AS$ values. Silicon can raise its co-ordinationnumber above 4 (expanding its octet of electrons) andalso forms bonds to fluorine and oxygen which haveenhanced strength because of $,-d, bonding. Siliconwill expand its octet most readily when it is sur-rounded by electronegative atoms. Thus SiF4 willform adducts with amines whereas Me4Si will not. Inthe ground state the silicon atom in a phenoxysilane willbe bonded to three carbon atoms and one oxygen atom.In the transition state of the hydrolysis of a phenoxy-silane there will be two oxygen atoms near to the silicon,TABLE GActivation data for hydrolysis reactionsAH$/ A S /Compound Catalyst Solvent kcal mol-l cal mol-1 K-110.3 - 31.320 v/v H,O-diosan 9.6 - 32.68.4 - 34.37.4 - 37.37-6 - 39.59.9 a -44.5 aHC1 15.7 -25.54040 v/v D,O-dioxanBu,SiOPh HClO,Bu ,SiOPh HC10,Bu,SiOPh HC10,Bu,SiOPh HC10,Bu,SiO (l-naphthyl) HC10, i:t} v/v H,O-dioxan Pri ,SiOAc NoneEtOAcPhCH,Cll01None } 40 v/vH,O-acetone 20.3 G -23-5PhCH,Cl None 40 v/v H,O-dioxan 20.1 - 23.3 Ca From ref.5. b E. Tommila and A. Hella, Ann. Acad. Sci. Fennicae, A I I , 1954, 53. E. Tommila, E. Paakkala, U. K. Virtanen,A. Erva, and S. Varila, Ann. Acad. Sci. Fennicae, A I I , 1959, 91.(0.8) for uncatalysed hydrolysis is in good agreement withthat (1.2) found for hydrolysis of Pri,SiC1 in propan-2-ol.2Only 1 molecule of water is required for hydrolysis aspropan-2-01 itself can accept the proton equation (l).For acid-catalysed hydrolysis in propan-2-01, n = 0, asfound for the unsubstituted phenoxysilane, and thisvalue arises in a similar way.The value of k, for the HC10,-catalysed hydrolysis ofPh,SiOPh in 40 aqueous dioxan at 30" is 3.13 1 mol-lrnin.-l.The relative rate in comparison with thetributyl compound k,(Ph,SiOPh)/k,(Bu,SiOPh) is 0.27.This is fairly similar to the value of 0.17 obtained for theuncatalysed hydrolysis of acetates K( Ph,SiOAc)k(Bu,SiOAc) and shows that steric and electronic26 H. H. JaffC, Chern. Rev., 1953, 55, 191.27 A. A. Humffray and J. J. Ryan, J . Chem. SOG. (B), 1969,1138.those of the departing phenol and of the incoming water.In this situation octet expansion will be favoured (withrespect to the ground state) and also more p,-dn bondingis possible.There is kinetic evidence that nucleophilicsubstitution reactions at silicon proceed through atransition state in which bond making is slightly moreadvanced than bond breaking.,* The transition statefor silicon is thus more complicated (more than fouro-bonds; p,-amp; bonding) and it is possibly this whichgives rise to the large negative A S values for reactions ofsilicon compounds.Low AH$ values for uncatalysed hydrolysis of organo-silicon chlorides and acetates have been explained byproposing that AH$ is composed of two terms, one for theenthalpy of formation of a water polymer, and a second28 C .Eaborn, ' Organosillcon Compounds,' Butterworths,London, 1960, p. 1111330 J.C.S. Perkin 11for the enthalpy of activation for hydrolysis of thesilicon compound by the water polymer. The two termstend to cancel each other and lead to the low observedvalues. Table 7 shows that low AH: values are foundfor both catalysed and uncatalysed hydrolysis of organo-silicon compounds, where reactions probably involvedifferent numbers of water molecules. The low AH$values seem specific to reactions of silicon compoundsand may arise from the energy of the transition statebeing lowered relative to the ground state by formationof more than four o-bonds in the former, where bondformation is more advanced than bond breaking.Debye-Huckel theory predicts that there should beno salt effect when one of the reactants is uncharged.However when salts are added to such reactions takingplace in solvents less polar than water there is a lineardependence of the rate coefficient k according to equation(23),29 where KO is the rate coefficient in the absence ofk = k,(l + bsalt) (23)added salt and b is a constant.For Bu,SiOPh in 40aqueous dioxan at 30" there is a linear relationshipbetween k, and NaClO, up to O-~M-N~C~O,, with b =2.6 1. mol-l. This salt effect probably arises from theincreased activity of HClO, which has been demonstratedon addition of salts to aqueous di~xan.~OOverall the mechanism of acid-catalysed hydrolysis ofphenoxysilanes is similar to that of acid-catalysedhydrolysis of organic esters, with additonal featurescaused by silicon's ability to use 3d orbitals.The base-catalysed hydrolysis of phenoxysilanes equa-tions (24) and (25) has already been ~tudied.~,10,27~3~,3~R,SiOPh + OH- -+ R,SiOH + OPh-OPh- + H,O + PhOH + OH-In alkaline solution kbQH- k,, and rate equation (3)can be simplified to equation (26).No previous workerskb(24)(25)Rate = kbR,SiOPh OH- (26)have studied the effect of variation of water concen-tration upon kb, so we have made a few preliminarymeasurements which are given in Table 7. A graph ofTABLE 7Base-catalysed hydrolysis of Bu,SiOPh in aqueous dioxanat 30.0": variation of K , with H,O concentrationCHaOI I VIV 15 20 26kb/l mol-1 min-1 3.76 4.75 5.25log kb against log H,Q over the range 15-25 aqueousdioxan has a slope of about 0.65.There are formally nowater molecules required for hydrolysis, as under experi-mental conditions OH- R,SiOPh and equilibrium(25) lies well over to the left. Values of n other than28 C. L. Perrin and J. Pressing, J . Amer. Chew. Soc., 1971, 93,3O J. P. H. Boyer, R. J. P. Corriu, and R. J. M. Perz, Tetra-31 R. L. Schowen and K. S. Latham, J . Amer. Ckem. SOC., 1966,5705.hedron, 1971, 27, 5255.88, 3795.0 may arise either from a variation of activity co-efficients of OH- with water concentration, or fromparticipation of water in the reaction, perhaps in assistingthe removal of OPh- from the activated complex.Values of n between 0 and 1 have also been found for thebase-catalysed hydrolysis of ethyl acetate in aqueousdi~xan.,~ We have measured the rates of base-catalysedhydrolysis of Bu,SiOC,H,X (X = H or p-OMe) inaqueous dioxan.Table 8 shows the relative rates ofTABLE 8Base-catalysed solvolysis of R,SiOC,H,X : relative ratesof solvolysis for X = H and X = $-OMe with changeof solvent and RR Solvent t/"C Kbp-*Me/KbH Ref.Ph MeOH 27.4 0.49 32BLI 40 v/v H,O-dioxan 30.0 0.53E t 407; v/bsol;- HZO-EtOH 25.0 0.59 10base-catalysed solvolysis for three systems. The almostconstant value of kbP-OMe/kbH is good evidence that thereis no dramatic change of mechanism among these threesystems, and consequently that there are no specificeffects associated with the butyl groups.EXPERIMENTALTributylphenoxysilanes were prepared by the method ofLar~son,,~ in which ethanol is distilled from a mixture oftributylethoxysilane and the phenol.For phenols con-taining electron-donating substituents a small piece ofsodium (0.2 g) was added to the ethoxysilane-phenolmixture. This promoted the ionisation of the phenol, andaccelerated the reaction. The products were purified byfractional distillation under reduced pressure followed bypreparative g.1.c. The identity and purity of the productswere confirmed by i.r., n.m.r., and mass spectroscopy andg.1.c. Purity was found to be about 97 before preparativeg.1.c. and over 99 after. The main impurity was phenol,and no significant differences in rate coefficients were foundbetween reactions of samples which had received g.1.c.purification, and those of samples which had not.Triphenylphenoxysilane was prepared by stirring to-gether at 120' for 4 h sodium (0.70 g, 0.03 g atom), phenol(30 g, 0-32 mol), and chlorotriphenylsilane (9-3 g, 0.03 mol).The cooled mixture was shaken with water, and the residuefiltered off and extracted with light petroleum (b.p.60-SOo). Filtration (hot) and cooling yielded crystallinetriphenylphenoxysilane. Hydrolysis gave Ph,SiOH andPhOH, confirmed by t.1.c.Dioxan was purified by boiling under reflux with sodiumfor 48 h. It was distilled from the sodium and thendistilled from sodium a second time. Acetonitrile andAnalaR propan-2-01 were used without further purification,as g.1.c. analysis showed them both to be well over 99pure.Solutions of x v/v aqueous solvent were made up bymalting x ml of water (containing HC10, or HCl at anappropriate concentration) up to 100 ml with solvent.The solvent (25 ml) was preheated to the temperature32 R. L. Schowen and K. S. Latham, J . Amer. Chenz. SOC., 1967,89, 4677.33 Calculated from data in E. Tommila, A. Koivisto, J. P.Lyyra, K. Antell, and S. Heimo, Ann. Acad. Sci. Fennicae, A I I ,1952, 47.34 E. Larsson, Chem. Ber., 1953, 86, 13821974required for the kinetic run. The liquid phenoxysilanc(1 or 2 PI) was then added from a Drunimond Microcapspipette, the mixture was shaken, and a 10 mm U.V. spectro-photometer cell was quickly filled. Optical densitymeasurements were taken at suitable times with a UnicamSP 500 spectrophotometer fitted with a thermostatted cellholder. The wavelengths a t which measurements weremade are shown in Table 4. Apparent first-order ratecoefficients, klobs, were calculated in the usual way, andwere reproducible to amp; 3. For phenosysilanes contain-1331ing strongly electron-attracting substituents h,, equation(l) contributes substantially to the observed rate ofhydrolysis ; K , values were measured separately byhydrolysis in the appropriate acicl-free solvent and suitableallowance was made in the calculation of K , values (Table 4).We thank Dr. J. Shorter for comments, and the S.R.C.for a research studentship (to G. E. G.).3/333 Received, 13th Februavy, 1973