Nucleophilic substitution at four-co-ordinate sulphur. Mobility of the leaving group

摘要

468 J.C.S. Perkin I1Nucleophilic Substitution at Four-co-ordinate Sulphur. Mobility of theLeaving GroupBy Ennio Ciuffarin," Lucio Senatore,* and Mauro Isola, lstituto di Chimica Generale, Via Risorgimento 35,56100 Pisa, ItalyThe leaving group effect has been measured for the reactions of benzenesulphonyl halides with aniline, n-butyl-amine, and hydroxide ion. The specific rate constants of displacement a t 25" with I, Br, CI, and F, respectively, asleaving groups, are the following : with aniline, 3.55 x 10-2. 31 -2 x 1 0-2, 4.27 x 1 0-2, and 2-6 x 1 O-7 ; withn-butylamine, 21.9, 103,42-6, and 1-01 x I O-2; and with hydroxide ion, 43.6, 28.9.15.5. and 3.40. The almostidentical leaving group mobility of I, Br, and CI for each nucleophile and the enormous change in relative groupmobility of fluorine on changing the pK, of the nucleophile, point to a mechanism involving an intermediate complexwith bond forming or bond breaking as the rate-limiting step according to the substrate.The activation parametersare also reported and agree with this interpretation.THE neutral and alkaline hydrolysis of sulphonyl deriva-tives has been discussed in terms of S N ~ ,l S N ~ , ~ or SANmechanisms.4 The S N 1 mechanism was later dis-carded2 on various grounds. The SAN mechanism wasalso abandoned because the sulphonyl oxygen atoms donot exchange with those of water during the solvolysisand this was taken as a proof against the formation of anaddition intennediate.5 The mechanism was thusmostly considered to involve a direct displacement onsulphur, as was recently emphasized by Rogne,G sinceall the kinetic features of the reaction are in accord withan SN2 mechanism. Rogne extended the investigationto a number of nucleophiles other than water, in aqueousmedia,718 and the mechanism was identified as a nucleo-phile-catalysed hydrolysis.For reasons given in the Discussion section, we did notconsider the lack of l 8 0 exchange a sufficient proof againstan SAN mechanism in the hydrolysis of sulphonyl deriva-tives.One way to try to distinguish between a directdisplacement and one which occurs via an intermediatecomplex is to determine the relative amount of bondH. Bohme and W. Schurhoff, Chem. Ber., 1951, 84, 28;A. H. Fainberg and S. Winstein, J .Amer. Chem. SOC., 1956, 78,2770; R. V. Vizgert, Zhur. obshchei Khim., 1962, 32, 628.( a ) R. Foon and A. N. Hambly, Austral. J . Chem., 1962, 15,668; (b) R. E. Robertson, B. Rossall, S. E. Sugamori, and L.Treindl, Canad. J . CJaem., 1969, 47, 4199.H. K. Hall, jun., J . Amer. Chem. SOC., 1956, 78, 1450;F. E. Jenkins and A. N. Hambly, Austral. J . Chem., 1961, 14,190, 205.E. 34. Kosower, ' Introduction to Physical Organic Chemis-try,' Wiiey, New York, 1968, p. 65.C. A. Bunton, personal communication cited in ref. 2 ( b ) ;D. R. Christman and S. Oae, Chem. and Ind., 1959, 1251.formation and breaking in the transition state. For thispurpose the Brmnsted coefficients for the nucleophileand for the leaving group, which have been related t o theamount of bond formation and breaking, respectively,in the transition state, might be u ~ e d .~ J ~ However, thevalues reported for nucleophilic substitution at a sul-phony1 sulphur atom (0.45 for the nucleophile' and-0.6 for the leaving groupll) are not sufficient todistinguish between the two mechanisms,129 l3 and Rognelimits himself to the general remark 8 that sulphur occu-pies an intermediate position between a saturated and acarbonyl carbon atom because of the intermediate valueof the nucleophilic Brrztnsted coefficient, and the state-ment ' that such a value should indicate that sulphur,being highly polarizable, should be able to form bondsat a larger distance than carbonyl carbon which usuallygives p values close to 0.8.We report a study of the element effect 14 in nucleo-philic substitutions at four-co-ordinate sulphur. Thiseffect is related to the amount of bond breaking in thetransition state. Similar studies have already been6 0.Rogne, J . Chem. SOC. (B), 1968, 1294.0. Rogne, J . Chem. Soc. (B), 1970, 727. * 0. Rogne, J . Chem. SOC. ( B ) , 1970, 1056.R. F. Hudson, Chimia (Switz.), 1'963, 16, 173.T. C. Bruice and S. Benkovic,Benjamin, New York, 1966, chs. I and IV.Bio-organic Mechanisms,'l1 R. V. Vizgert, Zhur. obshchei Khitn., 1058, 28, 1873.l2 E. Ciuffarin, L. Senatore, and M. Isola, J . Chem. SOC. (B),13 A. R. Fersht and W. P. Jencks, J . Amev. Chem. SOG., 1970,14 J. F. Bunnett, E. W. Garbish, jun., and K. M. Pruitt, J .1971, 2187.92, 5442.Amer.Chem. SOC., 1957, 79, 3861972reported for two-co-ordinate ~ulphur.~~J6 The resultswere discussed in terms of an SAN mechnism with bondforming as the rate-limiting step. Due to the differentreactivity of suphur in the two oxidation states we ex-pected to find a different pattern of leaving groupmobilities .EXPERIMENTALMaterials.-Commercial benzenesulphonyl chloride wasfractionated under vacuum, b.p. 113-115' at 10 mmHg.Benzenesulphonyl fluoride yield 80 ; b.p. 55-58" at5 mmHg ; nDZo 1.490 (lit.,17 1.4922), benzenesulphonylbromide (yield 90; b.p. 79-80" at 1 mmHg),I8 andbenzenesulphonyl iodide yield 90 ; m.p. 4-5" (lit.,la42-45') were prepared according to reported procedures.Aniline and n-butylamine were commercial products whichwere twice distilled from potassium hydroxide pellets.Sodium hydroxide was a reagent grade commercial productand was used without further purification.Deionizedwater was distilled from potassium permanganate. Aceto-nitrile was carefully fractionated.Product Isolation.-The products could not be isolatedfrom the kinetic experiments because of the small concen-trations of substrate used. However, in the same solventmixture, reaction of all substrates in preparative amountswith each nucleophile always gave the expected product inalmost quantitative yield. Products were identified byanalyses and i.r. spectroscopy. n-Butylbenzenesulphon-amide was purified by molecular distillation. An oil wasobtained whose b.p. could not be determined.aO Benzene-sulphonanilide was purified by crystallization (ethanol),m.p.110-111' (lit.,21 108-110).Kinetics.-The reactions with n-butylamine and hydrox-ide ion were followed by U.V. spectroscopy a t the most con-venient wavelength (240-265 nm) either with a Durrumstopped-flow apparatus (for the faster reactions) or with aBeckman U.V. spectrophotometer. In each case the experi-mental infinity spectrum was identical with that of theproducts. The temperature control was better than f 0.loexcept perhaps for the reactions followed with the stopped-flow spectrophotometer above room temperature. In thelatter cases we assumed an accuracy of f0-5'. Thereactions with aniline were followed with a Metrohm E 101conductometer in a j acketted reaction vessel equipped witha platinum conductivity cell.Care was taken to avoidcontact with carbon dioxide of the air. A linear relationwas found between the conductivity and the concentrationof the aniline hydrohalides produced in the reaction.Solutions of all substrates in 90 water-acetonitrile weresufficiently stable toward hydrolysis. They were usedwithin a day after preparation. Solutions of benzenesul-phony1 chloride, bromide, and iodide in 50 acetonitrile-water react slowly but measurably with water. Thesesolutions were prepared immediately before use.All reactions were followed directly in the reaction vesselexcept those between aniline and benzenesulphonyl fluoride,which were followed by sealing portions of the reactingmixture in glass vials, which were withdrawn from thel5 E. Ciuffarin and G.Guaraldi, J . Org. Chem., 1970, 35, 2006.l8 L. Senatore, E. Ciuffarin, and L. Sagramora, J . Chem. SOG.1 7 W. Davies and J. H. Dick, J . Chem. SOG., 1931, 2104.1.3 A. C. Poshkus, J. E. Herweh, and F. A. Magnotta, J . Org.( B ) , 1971, 2191.Chein., 1963, 28, 2766.thermostat at time intervals and chilled.was then measured at 25 "C.The conductivityRESULTSThe rates of reaction of benzenesulphonyl fluoride,chloride, bromide, and iodide with aniline, n-butylamine,and hydroxide ion have been measured in aqueous aceto-nitrile. In all cases the products, in quantitative yield,were those expected for a nucleophilic substitution at sul-phur. The reaction is first order in all substrates andnucleophiles.The reactions were measured under pseudo-first-order conditions and were linear up to 90 completionexcept for the reaction of benzenesulphonyl fluoride withaniline which was followed up to 25 completion. Thedata are in Table 1.TABLE 1Velocity constants for the nucleophilic substitution ofbenzenesulphonyl halides with aniline, n-butylamine,and hydroxide ion aLeavinggroup tpcAniline 1FFc1c1c1BrBrBrIII1001080.012.6525.20.014.725.20.014.525.0n-Butylamine CF 25.0F 49.9c1 22.0c1 25.0c1 32.5Br 25.0Br 48.2I 25.0I 47.8Hydroxide ion CF 21.7F 25.0F 30.8F 48.0c1 22.2c1 25.0c1 30.8c1 48.2Br 25.0Br 30.7Br 48.0I 25.0I 47.8102Nucle ophileM50.033-05.8-1 0.511.6-16'26.7-12.35.3-10.83.2-6.23.2-5.38.5-16.616.6-22'217.A22.111.1-27.53-46-4.700.9 5-2.9 51 * 16-2-500.95-2.951 - 1 6-2.501 * 43-2.831.16-2.501.73-2.352.66-4.271.7 1-4.272.5 6-4.2 71.7 1-2-562.5 6-4.2 71.0-2.02.56-4.271.7 1-2.561.714-271-71-4.271.7 1-4-271.7 1-4-271 . 7 1 4 . 2 7k1 mol-l s-11.35 x 10-51.88 x 10-51.42 x2.54 x4.25 x10.1 x 10-220.4 x30.9 x1-05 x2.22 x 10-23.48 x1.00 x 10-24.64 x 10+38.043.052.510223721.842.02.53.55.012.613.015.321.083.527.544-643.0130220No. ofruns1133333333333553556533333332522640 Substrate concentration, 10-4-10-3~r.6 In acetonitrile-water (9 : 1). C In acetonitrile-water (1 : 1).Because of the solubility properties of the reagents, asingle solvent could not be used. 50 (w/w) Acetonitrile-water was chosen for n-butylamine and hydroxide ion (al9 F. C. Whitmore and N. Thurman, J . Amer. Chem. SOC.,2o W. Ssolonina, J . Russ. Phys. Chem. Ges., 31, 640 (Chem.*l -4. Ginzberg, B e y . , 1903, 36, 2706.1923, 45, 1068.Zentr., 1899, 11, 867)470 J.C.S. Perkin I1smaller proportion of water would not dissolve 0 . 1 ~ -sodium hydroxide). The proportion of water was decreasedto 10 for aniline in order to minimize the rate of the neutralhydrolysis for the slower reactions. The neutral hydrolysiswas in all cases much slower than the measured reactionand therefore did not influence the kinetics.The alkalinehydrolysis was also negligible at the pH values of thesolutions when aniline or n-butylamine were used as nucleo-philes.The average specific rate constants (Table 1) do notpossess the same degree of accuracy. We estimate a rangefrom a minimum of 2-3y0 for reactions with good tempera-ture control, followed up to 90 completion, and a largenumber of runs, to a maximum of 10 in other cases. Sucherrors in the rates are too small to affect the discussion inany way.Enthalpy and entropy of activation are reported inTable 2 along with the maximum error 22 (the small numberTABLE 2Activation parameters and summary of the leaving groupmobility at 25 "C for nucleophilic substitutions ofbenzenesulphonyl halides with aniline, n-butylamine,and hydroxide ionLeaving AHSa -AS$' k bgroup kcal mol-l cal K-l inol-l 1 mol-1 s-1-4niline CI; 11.1 (56)c1 6.5 (0-7)Br 6.6 (0.7)I 7.2 (0-7)n-ButylaniineF 11.4 (0.8)c1 5.1 (1.8)Br 6.3 (1.6)I 4.9 (1.6),51*4 (15.2)43.2 (2.6)38.7 (2.6)41.1 (2.6)29.5 (2.7)34.0 (6.2)28.1 (5.7)36.0 (5.7)0.26 x los64.27 x 10-231.2 x3.55 x 10-21.01 x 10-242.621.9103Hydroxide ionF 10.6 (1.4) 20.5 (4-8) 3.40c1 13.0 (1.5) 9.4 (5.1) 15-5Br 12.1 (1.6) 11.4 (5.5) 28.9I 13.0 (1-7) 7.4 (5.8) 43.6a Maximum error in parentheses. * Calculated from theactivation parameters.c In acetonitrile-water (9 : 1 ) . d Inacetonitrile-water (1 : 1).of experimental temperatures does not permit the calcula-tion of a meaningful standard deviation).However, whileto report the maximum error is sometimes of mechanisticsignificance,22 in the present case the method of jointconfidence regions 23 is more suited. Such a methodclearly shows that, while the entropy and enthalpy ofactivation of C1, Br, and I as leaving groups are the samewithin experimental error for each nucleophile, the activa-tion parameters of benzenesulphonyl fluoride are quitedifferent from those of other sulphonyl halides in all cases.DISCUSSIONThe order of the reaction and the nature of theproducts show that a nucleophilic displacement onsulphur occurs. In the alkaline hydrolysis, which was22 K. B. Wiberg, ' Physical Organic Chemistry,' Wiley, New23 J.Mandel and F. J. Linnig, Analyt. Chem., 1957, 29, 743.1-ork, 1964, pp. 377-379.already identified as a nucleophilic displacement onthe displacement yields the products directlyequation (l). With amines as nucleophiles the re-OH- + ArSO,*X ArSO,*OH + X- (1)R2NH + ArSO,*X --t ArSO,*hHR, + X- (2) A k ArS0,H + R,NH ArSO,*N R,- H - b -Hfaction involves the formation of an unstable sulphonyl-ammonium intermediate equation (2) .i Such anintermediate can either lose a proton to give the corre-sponding ~ulphonamide,~~ or be attacked by water (inaqueous media) to give hydrolysis products.i Whentertiary amines are used, owing to the absence of hydro-gen on the nucleophilic nitrogen atom, only hydrolysisproducts are observed.i When primary or secondaryamines 25 are used, both reactions are possible.How-ever, the fact that the sulphonamides are obtained inquantitative yield indicates that loss of a proton is amuch faster process than attack by water (already con-sidered a fast step '). Thus, the sulphonylammoniumintermediate does not revert to the starting materialsand in each case the reaction is a simple displacement onsulphur. The leaving group mobilities are thereforedirectly related to the first step of the nucleophilic sub-stitution.The leaving group mobility of benzenesulphonylchloride and fluoride in neutral and alkaline hydrolysishas already been measured by Swain and Theydiscussed the kcl : kF ratio of neutral hydrolysis for aseries of substrates in terms of a larger tendency forbond breaking giving a larger kcl : kF ratio.The ratiofor the benzenesulphonyl derivatives is quite large(443 x lo3); therefore the reaction was considered toproceed with a large degree of bond breaking.24 Thekcl: kF ratio for alkaline hydrolysis 24 was found to bemuch smaller (ca. S), an indication that bond breakingis much less pronounced in this case. In the case ofalkaline hydrolysis of benzoyl halides, where the fluoridereacts faster than the chloride, Swain and Scott suggestedthat fluorine facilitates the formation of the bond withthe nucleophile by making the electrophilic centre moreelectron-deficient and positive .24 These ideas are similarto those expressed by Bunnett on the ' element effect '.14However, the element effect can give more detailedinformation because the number of leaving groupsis not limited to two.Swain and Scott's data are inagreement with ours but are insufficient for a com-prehensive discussion of the element effect on sul-phonyl sulphur. For example, the Kcl : k= ratio meas-ured by Swain and Scott for neutral hydrolysis of sul-phonyl halides suggests large bond breaking in the transi-tion state for both leaving groups. Consideration ofmore than two leaving groups shows that, while bond24 G. Swain and C. G. Scott, J . Anzev. Chem. Soc., 1953, 75,25 J. F. Bunnett and J. Y . Bassett, jun., J . Amev. Chem. SOG.,246.1959, 81, 2104; J . Org. Chem., 1962, 27, 23461972 471breaking in the transition state is extensive for thefluoride, it is much less pronounced for the chloride.Table 2 shows that the leaving group mobilities of I,Br, and C1 for each nucleophile are almost identical.The maximum relative mobility changes from 2.8 forhydroxide ion, to 4.7 for n-butylamine, to 8.8 for aniline.On the other hand the relative mobility of fluorine changeson changing the nucleophile.The ka : kF ratios forhydroxide ion, n-butylamine, and aniline, are 4-6, 4.2 xlo3, and 1-65 x lo5. The slight variation relative to theKc, : kF ratio previously determined for the alkalinehydrolysis in aqueous acetone 24 can be attributed to thedifferent solvent. Clearly, on changing the nature ofthe nucleophile, the strength of the S-F bond influencesthe reaction rate to a very different degree.Breakingof the S-F bond is important for aniline and becomes lessimportant for n-butylamine until for hydroxide ion it isalmost of no importance. For the three other leavinggroups on the other hand the strength of the S-halogenbond is always unimportant.The leaving group mobility in the various cases isparalleled by the activation parameters (Table 2).Entlialpies and entropies of activation are the samewithin experimental error for chlorine, bromine, andiodine as leaving groups (considered separately for eachnucleophile). Fluorine has in each case quite differentactivation parameters.The relative mobility and the activation parametersfor the different leaving groups can be accounted for onlyby an SAIN mechanism, which, depending on the leavinggroup, allows either bond forming or bond breaking asthe rate-limiting step.14 Bond forming is the ratelimiting step with all nucleophiles for benzenesulphonylchloride, bromide, and iodide as shown by the almostidentical rate and activation parameters. When thismechanism is operative, the rate of reaction is little in-fluenced by the S-X bond strength.Benzenesulphonylfluoride on the other hand, reacts via an SAN mechanismwith bond breaking as the rate-limiting step as shown bythe different mobilities of fluorine with various nucleo-pliiles. The variation in activation parameters onchanging the pK, value of the nucleophile cannot beeasily interpreted. However, an interesting feature isthe difference in reaction rate between benzenesulphonylfluoride and the other sulphonyl halides, which becomessmaller, the higher the pK, values of the nucleophile.The decomposition of the intermediate towards reagentsor products depends on the relative basicity of nucleo-pliile and leaving group. A nucleophile of higher pK,value is a relatively worse leaving group.Thus, thedecomposition of the intermediate changes from largelyexclusive return to reagents for the reaction of benzene-sulphonyl fluoride with aniline, to a more balanced situa-tion on increasing the pK, value of the nucleophile.A comment about the difference in mobility of I, Br,and C1 found for two- 1 5 9 1 6 and four-co-ordinate sulphur26 I;. -4. Cotton and G. Wilkinson, ' Advanced InorganicChcinistry,' Interscience, Ncw York, 1967, p.402; E. Ciuffarinand .I. Fava, Pvogv. Phys. Ovg. Chew., 1968, 6, 81.seems in order. For sulphenyl sulphur, larger ratios ofleaving group mobilities were found (400 in one case)and the order of leaving group mobility follows theelectronegativity. However, we have assumed for bothtypes of sulphur that iodine, bromine, and chlorine arealmost fully bound to the sulphur atom in the transitionstate. Clearly, while the electronegativity is quite im-portant for two-co-ordinate sulphur, the incomingnucleophile is not affected by it in reactions at sulphonylsulphur where the order of leaving group mobility isdetermined both by the (small) influence of the electro-negativity and by stretching of the S-halogen bond inthe transition state, as shown by the slight inversion inleaving group mobility on changing the nucleophile.As far as two-co-ordinate sulphur is concerned, the in-fluence of a small S-X stretching in the transition stateis overwhelmed by the electronegativity of the leavinggroup, which influences the formation of the inter-mediate much more easily through the much morepolarizable sulphenyl sulphur atomThe Bronsted coefficient for the leaving group meas-ured by Vizgert l1 for alkaline hydrolysis of para-substituted phenyl benzenesulphonates (@ -0.6) shouldbe an indication of fairly large bond breaking in thetransition state.However, we have suggested l6 thatfor reactions which occur via an intermediate complexwith bond making as the rate limiting step, f3 valuesmight be linked to the ability of the leaving group tofacilitate the formation of the intermediate complex.One of the reasons put forward to exclude an inter-mediate complex in the solvolysis of benzenesulphonylderivatives was the absence of l80 exchange betweensubstrate and solvent during the rea~tion.~ However, amechanism which assumes the fast decomposition of theintermediate to products cannot give rise to exchangeof solvent with nucleophile since the intermediate neverreturns to starting materials.Such an exchange mightbe observed in the hydrolysis of benzenesulphonylfluoride, since in this case the intermediate decomposesmore easily to reagents than to products. However,in the trigonal bipyramidal intermediate the enteringand leaving groups lie in apical positions while thesulphonyl oxygen atoms and the remainder of themolecule lie in radial positions.26 Since the two positionsare not equivalent, in order to observe exchange withoutcontradicting the principle of microscopic reversibility,another condition must be fulfilled. The intermediatemust pseudorotate before reverting to reagents. Up tonow, pseudorotation, while very common for phos-p h o r ~ ~ , ~ ~ has not been observed with sulphur sub-strates. Thus, the l80 exchange experiments, whileindicative of intermediate formation when they arepositive, are inconclusive when they are negative.This investigation was supported by C.N.R., Rome.1/1304 Received, 27th JuZy, 1971)2 i E. A. Dennis and F. H. Westheimer, J . Amev. Chem. Soc.,28 R. Tang and K. Alislow, J . Amev. Chew?. Soc., 1969, 91,1966, 88, 3431 ; 3432.5644