September 1 979 The Analyst Vol. 104 No. 1242 Ionic Polymerisation as a Means of End-point Indication in Non-aqueous Thermometric Titrimetry Part X.” Acrylonitrile Indicator Reactions in the . Determination of Acids E. J. Greenhow and L. A. Dajer de Torrijos Department of Chemistry, Chelsea College, University of London, Manresa Road, London, SW3 6LX The mechanism of the reaction marking the end-point when acrylonitrile is used as the indicator reagent in the thermometric titration of acids with solutions of potassium hydroxide and alkosides in alkanols has been investi- gated. Two indicator reactions occur : cyanoethylation of alcohols and anionic polymerisation of the acrylonitrile. The relative contribution of the two reactions is shown to depend on the nature and concentration of alcohols in the titrand solution.Cyanoethylation is mofe important with respect to end-point sharpness, when primary and secondary alcohols are present, and polymerisation predominates in the presence of tertiary alcohols. Anomalous titration values obtained in the determination of thiols and sulphanilamide can be explained by the effect of the structure of the sample compound on the reversibility of its reaction with acrylonitrile. The experimental findings are used to devise combinations of titrand solvents and titrants leading to “ideal” titration curves. Keywords : Non-aqueous thermometric titrimetry ; end-point indication ; ionic polymerisation ; acrylonitrile indicator reactions ; weak acid determination In the initial investigation of the use of acrylonitrile as an indicator reagent for the thermo- metric titration of weak acids,1$2 it was suggested that both anionic polymerisation of the acrylonitrile and cyanoethylation reactions could be responsible for the rise in temperature marking the end-point.However, in subsequent studies of the applications of this indicator reagent, it has been assumed that the polymerisation reaction predominates. In a recent preliminary communication3 we reported that significant amounts of cyano- ethylation products were present in the final titration solutions when acrylonitrile was used as the indicator reagent, and primary and secondary alcohols were introduced into the sample solution, either as the titrant solvent or as an original constituent of the solution.This investigation is concerned with the nature of the indicator reaction and the means by which it can be modified to obtain sharp end-point inflections in the titration of weak acids. Particular attention is paid to the effects of the titrant, titrant solvent and sample solvent on the shape of the titration curve. An attempt is made to explain these effects in terms of the mechanisms and rates of the reactions involved. Weak acids, such as phenols, thiols and sulphonamides, can undergo cyanoethylation and, in earlier paper^,^^^ it was suggested that the low reaction stoicheiometries observed in the titration of some thiols and sulphanilamide could be explained in terms of the complete or partial cyanoethylation of the weak acid. This analytical problem is examined in this paper.Experimental Reagents Toluene, pyridine, methanol, propan-2-01 and n-butanol were of analytical-reagent grade ; acrylonitrile, morpholine and other solvents were of laboratory-reagent grade. All liquid reagents were dried over molecular sieve 4A before use. * For Part IX, see Analyst, 1976, 101, 777. 801802 GREENHOW, DAJER DE TORRIJOS IONIC POLYMERISATION FOR END-POINT Analyst, VOL. 104 Titrants Potassium hydroxide and alkoxide solutions were prepared in the usual manner by dissolving the solid hydroxide and metal, respectively, in the appropriate alkanol. Tetra- methylammonium hydroxide reagent, 0.025 M in pyridine - propan-2-01 (1 + l), was pre- pared from a 25% aqueous solution by evaporating off most of the water a t 28 “C under reduced pressure, adding pyridine, continuing the evaporation until all of the water had been removed and diluting the pyridine solution with pyridine and propan-2-01.Cyanoethyl Derivatives These derivatives were prepared by the addition of acrylonitrile to the alkanols, phenol, 2-mercaptobenzothiazole and sulphanilamide in the presence of benzyltrimethylammonium hydroxide. The procedures described by Utermohlen,6 Bachman and Levine,’ Hurd and Gershbeins and Misra and Raog were used for the preparation of 3-alkoxypropionitriles, 3-phenoxypropionitrile, S-(2’-cyanoethyl)-2-mercaptobenzothiazole and NN-bis(Z’-cyano- ethyl)-4-aminobenzenesulphonamide, respectively. Apparatus The automatic apparatus described in Part IIIlO was used, but with unsilvered Dewar beakers (capacities 14 and 30 ml) instead of foam-insulated titration flasks.A variable- speed syringe, driven by a stepper motor, was used in investigations of the effects of titration rate on observed reaction stoicheiometries. Procedure Thermometric titrimetry Prepare a solution of benzoic acid in acrylonitrile, pipette an aliquot of the solution into the titration beaker and add additional acrylonitrile, if required, and any other solvents. Introduce the titrant a t a constant rate (0.2 ml min-l) unless otherwise required, and record the temperature and titrant volume on a millivolt chart recorder (500-mV scale) a t a chart speed of 600 mm h-l. Gas - liquid chromatogyaphy After completion of the catalytic thermometric titration, acidify the titration solution with glacial acetic acid (2 ml), filter off any precipitated polymer and inject about 0.5 p1 of the filtrate into the gas chromatograph. Use a stainless-steel column (2 m x 2 mm i.d.) packed with 2.5% OV-17 on Chromosorb G (80-100 mesh), and a flame-ionisation detector.Programme the column temperature from 60 “C to an upper value appropriate for the higher boiling constituents of the sample solution. Results and Discussion Effect of Alcohols in the Titrant and Titrand Solutions Acrylonitrile alone and solutions of benzoic acid in acrylonitrile were titrated with solutions of potassium hydroxide in methanol and propan-2-01, and potassium alkoxides in the corresponding alkanols. The titration curves obtained are shown in Fig. 1. The anions present in the potassium hydroxide solutions are essentially alkoxide and not hydroxide because the equilibrium in the titrant solution, represented by the equation OH- + ROH .+ OR- + H,O strongly favours the right-hand side.ll As observed in earlier studies of solvent effects,lZ the titration curves for the solvents only [Fig.l(a)-(e)] reveal more clearly the influence of different reagents on end-point sharpness than do the corresponding curves for the benzoic acid solutions [Fig. l(a’)-(e’)]. In the blank titrations of acrylonitrile, steeper rises in temperature are obtained with potassium hydroxide in methanol, potassium n-butoxide in n-butanol and potassium tert- butoxide in tert-butanol than with potassium hydroxide in propan-2-01 and potassium sec- butoxide in sec-butanol as the titrants.The induction period before the indicator reactionSeptember, 1979 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART X 803 is initiated is about 1 min when the last two titrants are used, and about 0.3, 0 and 0 min with the potassium n-butoxide in n-butan-ol, potassium hydroxide in methanol and potassium tert-butoxide in tert-butanol, respectively. There is evidence13J4 that the catalytic "/ I I I I I I I I I I I I ;\ a I!, a I d / , Titrant/ml (1 division = 1 rnl) Fig. 1. Effect of the nature of the alcohol titrant solvent on the titration graph. Titrand solvent: acrylonitrile, 10 ml. Sample: benzoic acid, 61.1 mg. Titrants (0.5 M) : a, potassium hydroxide in methanol ; b, potassium hydroxide in propan-2- 01; c, potassium n-butoxide in n-butanol; d, potassium sec- butoxide in sec-butanol; and e, potassium tert-butoxide in tert-butanol.Graphs a-e are the blank titrations and a'-e' are the sample titrations. activity of alkali metal alkoxides for the cyanoethylation reaction increases in the order CH30- < n-C,H,O- rn n-C,H,O- < iso-C,H,O- < tert-C,H,O-. On the other hand, the acidities of the corresponding alcohols increase in the reverse order.14 Thus, although the methoxide ion may be a less effective catalyst than the secondary and tertiary alkoxide ions, the titrant solvent, methanol, is more reactive than the secondary and tertiary alcohols. It is these alcohols that are required to complete the cyanoethylation process : RO- + CH,=CHCN + ROCH2-6HCN . . . . . . (1) ROCH~CHCN + ROH + ROCH,CH,CN + RO- .. ' . (2) and it seems probable that the sharp rise in temperature seen in titration curves Fig. l(a) and (c) is caused by cyanoethylation rather than polymerisation. In contrast, the sharp temperature rise in curve (e) in Fig. 1 must be attributed to anionic polymerisation because tert-butanol does not yield cyanoethylation products at the temperature of the titration process. Titration curves (b) and (d) in Fig. 1 could represent a combination of both cyanoethyla- tion and anionic polymerisation, although there is considerable evidence from kinetic and other s t u d i e ~ ~ ~ - l * that, in general, cyanoethylatiori will precede the anionic polymerisation of acrylonitrile when primary and secondary alcohols are present. Gas - liquid chromatographic analysis of the titration solutions, after they have been acidified with acetic acid to prevent decyanoethylation reactions on the column of the chromatograph, confirms the absence of a cyanoethylation product when the tert-butoxide titrant is used, and the presence of a significant amount of cyanoethylation product when titrants containing primary and secondary alkanols are employed [Table I, experiments l(a)-(e) and l(a')-(e')].The higher yield of polymer when potassium hydroxide (effectively isopropoxide) in propan-2-01 is the titrant suggests that the relatively long induction period corresponds more to polymerisation than to a cyarioethylation reaction. This possibility is supported by the appearance of a yellow colour, associated with the formation of p l y - acrylonitrile,15 which is seen to develop when the end-point is indicated.804 GREENHOW, DAJER DE TORRI JOS IONIC POLYMERISATION FOR END-POINT Analyst, Val.104 TABLE I COMPOSITION OF TITRATION SOLUTIONS IN THE CATALYTIC THERMOMETRIC TITRATION OF SOLVENT MIXTURES AND BENZOIC ACID SOLUTIONS WITH ACRYLONITRILE AS INDICATOR Conditions: the final solutions from titrations described in Figs. 1 and 2 [l(a)-(d’) and 2(d)-(f’)] are used after acidification with acetic acid (2 ml). Composition of the final solution Titrantt M K N S B B B B M N S B B B Initial solvent system v-----7 Acrylo- nitrilelg Co-solvent:/g 8.06 - 8.06 8.06 - 8.06 - 8.06 c 8.06 R1, 2.09 8.06 R2, 1.02 8.06 R3, 0.95 8.06 8.06 - 8.06 - 8.06 RI, 2.09 8.06 R2, 0.91 8.06 R3, 0.80 - - Acrylo- nitrile/g 7.06 5.58 7.84 7.71 7.90 6.10 5.00 2.05 5.60 7.59 7.31 5.60 4.60 4.98 Alkanol/g Trace 0.11 0.03 0.097 0.53 0.00 0.45 0.09 0.21 0.07 0.21 0.00 0.00 0.08 Cyano- ethylation products/g 0.37 0.40 0.48 0.67 0.00 3.11 0.74 0.00 3.16 1.03 1.44 3.11 1.17 0.00 Polymer$/g 0.77 2.29 0.02 0.07 0.16 0.94 2.90 6.01 0.49 0.04 0.15 1.44 3.19 3.08 * 1 and 2 refer to Fig.1 and Fig. 2, respectively. Curves l(a)-2(f) are for the solvent mixtures only; t 0.5 M titrants: M = KOH in methanol; K = KOH in propan-2-01; N = n-BuOK in n-butanol; S = $ R1 = benzyl alcohol; R2 = diphenylmethanol; R3 = triphenylmethanol. 5 By difference. curves l(a’)-2(f’) are for solutions containing 61.1 mg of benzoic acid. sec-BuOK in sec-butanol; R = tert-BuOK in tert-butanol. Although a significant amount of polymer is formed in addition to 3-methoxypropio- nitrile in the titration of acrylonitrile with potassium hydroxide (methoxide) in methanol, there seems little doubt, from the kinetic studies of Feit and Bigon,16 that cyanoethylation must be the initial step.Not only did they establish that cyanoethylation was a faster reaction than anionic polymerisation, but by quenching the reaction at an appropriate time they were able to show the presence of the cyanoethylation product before any polymerisa- tion had occurred. Consideration of the gas - liquid chromatographic analysis of the titration solution and the shape of the titration curve for the titration of acrylonitrile, when the potassium sec- butoxide reagent is used [Fig. l(d)], leads US to believe that the indicator reaction in this instance is mainly a relatively slow cyanoethylation reaction.The titration curves for the solutions of benzoic acid in acrylonitrile [Fig. l(a’)-(e’)] do not differ as much as the blank titration curves, possibly because at the end-point there is present a significant amount (about 10% by volume) of the alkanol, which makes the con- ditions favourable for cyanoethylation. The encl-point inflection is sharpest when potassium n-butoxide is the titrant, and is the least s h a q with potassium sec-butoxide as the reagent. The differences in the shape of titration curt’es (a’) and (c’) in Fig. 1 can be explained if it is assumed that the highly reactive methanol added in the titration represented in Fig. l(a’) yields a cyanoethylation product before the alkoxide appears in excess in the titration solution, i.e., the cyanoethylation is catalysed by the weakly basic benzoate ion.The polymer formed does not precipitate from the :solution and must have a low degree of polymer- isation. The inflections in titration curves (c’) and (d’) in Fig. 1, at approximately the half- neutralisation stage, correspond to the point where potassium benzoate is precipitated from the titration solution. The effect of including primary, secondary and tertiary alkanols and arylalkanols in the original titrand solution has also been investigated. Cyanoethylation reactions involving the titrant have been avoided by using the solution of potassium tert-butoxide in tert-butanol.Se$tember, 1979 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY.PART x 805 Titration curves and gas-chromatographic analyses are shown in Fig. 2 and Table I [2(d)-(f) and 2(d’)-(f’)], respectively. As in Fig. 1, it can be seen that a very rounded end-point is obtained in the blank titration when sec-butanol is present in the titration solution [Fig, 2(b)], whereas the end-point is rounded, but not to the same extent, in the titration of the benzoic acid [Fig. 2(b’)]. This small improvement in end-point sharpness can be attributed to the effect of the added tert-butanol. Sharp end-point inflections are achieved not only when the primary alcohols n-butanol and benzyl alcohol are present, but also in the presence of diphenylmethanol, a secondary alcohol. I t is interesting that virtually all of the added diphenylmethanol is cyanoethylated in the benzoic acid titration [Table I, 2(e’)] and there is a high yield of cyanoethylation product also in the blank titration [2(e)].Benzyl alcohol is completely cyanoethylated in both the blank and benzoic acid titrations [Table I, 2(d) and (d’)]. On the basis of these analyses, there is strong evidence that the indicator reaction is cyanoethylation when these arylalkanols are present. In contrast to the results obtained when the primary and secondary arylalkanols are present, the presence of the tertiary aryl- alkanol, triphenylmethanol, as a co-solvent leads to rounded end-point inflections [Fig. 2(f) and (f’)], and there is no evidence for the formation of a cyanoethylation derivative. Only about 10% of the triphenylmethanol can be accounted for in the gas - liquid chromatographic analysis, and there is a high yield of polymer [Table I, 2(f) and (f’)].I t must be concluded that the triphenylmethoxy anion initiates the polymerisation of acrylonitrile, and this reaction, which is slower than cyanoethylation , is responsible for the rounded end-point inflections. - 0.5 M Potassium tert-butoxide in tert-butanollrnl ( 1 division = 1 ml) Fig. 2. Effect of the nature of the alcohol titrand solvent on the titration graph. Sample solvent: a, 5 ml of acrylonitrile plus 2 ml of n-butanol; b, 5 ml of acrylonitrile plus 2 ml of sec- butanol; c, 5 ml of acrylonitrile plus 2 ml of tert-butanol; d, 10 ml of acrylonitrile plus 2 ml of benzyl alcohol; e, 10 ml of acryloni- trile plus 1.0 g of diphenylmethanol; and f , 10 ml of acrylonitrile plus 0.8 g of triphenylmethanol.Sample: benzoic acid, 61.1 mg. Titrant : 0.5 M potassium tert-butoxide in tert-butanol. Graphs a-f are the blank titrations and a’-f’ are the sample titrations. In the titrations summarised in Fig. 2, equilibria will be established when the tevt-butoxide titrant mixes with the co-solvent in the titrand solution and the co-solvent alkoxide is likely to be an important constituent of the catalyst system. The effect of increasing the concentration of alkanol in the titrand solution is demonstrated in Fig. 3. sec-Butanol was chosen as the co-solvent because it is known to affect the end- point inflection adversely. Fig. 3 shows that an increase in alkanol concentration not only reduces the end-point sharpness but also causes an increasing rise in temperature before the temperature rise associated with the presence of an excess of titrant.In the blank titrations [Fig. 3(a)-(d)] there is at first a gradual increase in temperature when sec-butanol is present, and later the temperature increases more rapidly. In the absence of the alcohol [Fig. 3(a)] there is an induction period before the temperature rise, and it can be assumed that a slow cyanoethylation reaction is initiated immediately in the presence of the alkanol. The effect806 GREENHOW, DAJER DE TORRIJOS: IONIC POLYMERISATION FOR END-POINT Analyst, VOZ. 104 of increasing the content of the sec-butanol is clearly to reduce the rate of the indicator reaction. This is in accord with the findings of Feit and Bigon16 and Feit et aZ.,18 who established that the rate of cyanoethylation o-E methanol was inversely proportional to the a’ I- 0.5 M Potassium tert-butoxide in rert-butanollml ( 1 division = 1 ml) Effect of the ratio of acrylonitrile t o sec-butanol in the titrand solvent.Sample solvent: a, 4 ml of acrylonitrile; b, 3 ml of acrylonitrile plus 1 ml of sec-butanol; c, 2 ml of acrylo- nitrile plus 2 ml of sec-butanol; tl, 1 ml of acrylonitrile plus 3 ml of sec-butanol. Sample : benzoic acid, 61.1 mg. Titrant : 0.5 M potassium tert-butoxide in tert-butanol. Graphs a-d are blank titrations and a’-d’ are the sample titrations. nth power of the methanol concentration, where n depends on the nature and concentration of other solvents present in the cyanoethylation reaction.They consider that solvation of the alkoxide ion by the methanol reduces the catalytic activity of the former. The value of n is related to the desolvating power of other solvents in the reaction mixture; thus dimethylformamide, an effective “desolvating” solvent because it readily solvates alkanols and prevents them from solvating the alkoxide ions, causes a reduction in n when it is included in the cyanoethylation mixture. In addition, dimethylformamide solvates the potassium cation but not the alkoxide anion, thereby enhancing the dissociation of the catalyst ion pair and increasing the catalytic activity of the anion. The effect of including dimethylformamide in the titrand solution is shown in Fig. 4. By Fig. 3. 0.5 M Potassium tert-butoxide in tert-butanollml (1 division = 1 ml) Fig.4. Effect of the acrylonitrile : sec-butanol : dimethyl- formamide ratio in the titrand solvent. Sample solvent (millilitres of acrylonitrile : sec-butanol : dimethylformamide) : a, 4 : 0 : 2; b, 3 : 1 : 2; c, 2 : 2 : 2 ; d, 1 : 3 : 2. Sample: benzoic acid, 61.1 mg. Titrant: 0.5 M potassium tert-butoxide in tert- butanol. Graphs a-d are the blank titrations and a’-d’ are the sample titrations.September, 1979 INDICATION I N NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART x 807 comparing Fig. 4(a)-(d) with Fig. 3(a)-(d) it can be seen that the addition of dimethyl- formamide to mixtures of acrylonitrile and sec-butanol gives rise to considerably sharper end-point inflections. The effect is similar in the titrations of the benzoic acid solutions [compare Fig.4(b’)-(d’) with Fig. 3(b’)-(d’)]. When the sec-butanol is omitted [Figs. 4(a’) and 3(a’)] the dimethylformamide has the effect of initiating the anionic polymerisation before the end-point is reached [Fig. 4(a’)]. The decreasing influence of the dimethyl- formamide with increasing concentration of sec-butanol in both the blank and benzoic acid titrations is an indication of the limit of the desolvating power of this dipolar solvent; apparently dimethylformamide will “desolvate” about an equal volume of sec-butanol [compare the slopes of curves (c) and (d) or (c’) and (d’) in Fig. 41. Greenhow and Nadjafi19 observed that the rise in temperature marking the end-point in the titration of benzoic acid with tetramethylammonium hydroxide reagent when pyridine is used as the co-solvent with acrylonitrile was much less than when alcohols or dimethyl- formamide were the co-solvents.A “normal” rise in temperature at the end-point was achieved by adding a significant amount of propan-2-01 to the initial sample solution (e.g., 1 ml per 2 ml of pyridine). The retarding effect of pyridine on the indicator reaction was believed to be due to the pyridine forming an adduct with the growing polymer chain : This type of adduct formation was proposed by Yagi et aL20 to explain the retarding effect of pyridine on the anionic polymerisation of styrene. We find that a further, significant, rise in temperature occurs immediately on the addition of propan-2-01 to the titration solution after the end-point is indicated [Fig. 5(a,), (a2) and (a3)].Apparently, the propan- 2-01 displaces pyridine from the adduct and undergoes chain transfer to continue the cyano- ethylation, oligomerisation and polymerisation reactions at a faster rate. There is evidence, therefore, that alkanols can be useful in preventing certain solvents from retarding the acrylonitrile reaction. In titrations of benzoic acid with n- and sec-alkoxides in the corresponding alkanols, the end-point sharpness was found to be influenced by the titrant molarity. Thus, it can be seen in Fig. 6 that whereas a sharp end-point inflection is obtained by using 0.1 M potassium hydroxide in propan-2-01 for the titration of 0.1 M solutions of benzoic acid, significant rises a2 -- I I I 0.025 M Tetrarnethylarnrnoniurn hydroxide reagendrnl ( 1 division = 0.5 ml) Fig.5. Effect of propan-2-01 when pyridine is a titrand solvent. Sample solvent : 4 ml of acrylonitrile plus 2 ml of pyridine. Propan- 2-01 added (1 ml) : a (solid line), none; al-a3 (broken lines), a t the points indicated by the arrows; b, to the original sample solution before the titration. Sample: benzoic acid, 1.5 mg. Titrant: 0.025 M tetramethylammonium hydroxide in pyridine - propan-2-01 (1 + 1).808 GREENHOW, DAJER DE TORRI JOS : IONIC POLYMERISATION FOR END-POINT Analyst, VOt?. 104 in temperature occur before the inflection point when 0.5 and 1.0 M titrants are used for the titration of 0.5 and 1.0 M benzoic acid solutions, respectively. These premature rises in temperature are not apparent when 0.5 M potassium tert-butoxide in tert-butanol is used as the titrant [Fig.6(d)]. I t must be concluded that the acrylonitrile competes with the benzoic acid for the alkoxide ion when the titrant is present, momentarily, in high concentra- tion. This hypothesis was confirmed by adding the 0.5 and 1.0 M potassium hydroxide titrants at a slower rate, e.g., 0.02 instead of 0.2 ml min-l, when titration curves similar to to that in Fig. 6(c) are obtained. Titrant/mI ( 1 division = 1 ml) Fig. 6. Effect of titrant molarity. Sample: benzoic acid in 2 ml of acrylonitrile plus 2 ml of pyridine. Titrant, milligrams of benzoic acid: a 0.5 M potassium hydroxide solution in propan-2-01, 61.1 ; 11, 1.0 M potassium hydroxide solution in propan-2-01, 122.1; c, 0.1 M potassium hydroxide solution in propan-2-01, 12.2; d, 0.5 M potassium tert-butoxide in tert-butanol, 61.1.Cyanoethylation Effects in the Determination of Phenols, Thiols and Sulphanil- amide As phenols, thiols and some sulphonamides can be cyanoethylated with acrylonitrile in reactions that replace their acidic hydrogen functions by the non-acidic cyanoethyl group, it might be expected that the use of acrylonitrile as an indicator reagent could lead to erratic or negligible titration values in the determination of these weak acids. In our previous studies we found no evidence of a reduction in the acidity of phenols when they were titrated in the presence of acrylonitrile, and the titration values obtained by the catalytic thermo- metric method were comparable with, and sometimes higher than, those obtained by non- aqueous potentiometric titration.However, alkylthiols and thiophenols could not be titrated when acrylonitrile was present4 and a. titration value corresponding to a stoicheio- metric reaction was obtained for sulphanilamide only with certain titrant - solvent combina- t i o n ~ . ~ Unexpected results were obtained iin the titration of some heterocyclic thiols, including 2-mercaptobenzothiazole.4 These compounds could be titrated without difficulty to give a titration value corresponding to the neutralisation of the thiol function, although they are known to be cyanoethylated readily by acrylonitrile.8 The effect of cyanoethylation on titration values was investigated by preparing and titrating the cyanoethylation derivatives of phenol, sulphanilamide and 2-mercaptobenzo- thiazole. The titration curves, and titration values, obtained are compared with those of the parent compounds in Figs.7 and 8. I t can be observed (Fig. 7) that cyanoethylated and uncyanoethylated phenol and 2-mercaptobenzothiazole give rise to very similar titration curves and identical titration values corresponcling to the neutralisation of one acidic function. It must be concluded that decyanoethylation of the two cyanoethylation derivatives occurs readily under the conditions obtaining in the titration. Titration curves for an arylalkyl thiol, toluene-a-thiol, and an arylthiol, toluene-4-thiol, are included in Fig. 7 for comparison ;September, 1979 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY.PART x 809 d'el . i / I / C \ 1 0.5 M Potassium hydroxide in propan -2-01 ( 1 division = 1 ml) Titration graphs for cyanoethylated and uncyano- ethylated weak acids. Sample (in 4 ml of acrylonitrile plus 2 ml of dimethylformamide), mg : a, 2-mercaptobenzothiazole, 83.6 ; b, S-(2-cyanoethyl)-2-mercaptobenzothiazole, 110.0; c , phenol, 47.1 ; d, 3-phenoxypropionitrile, 73,5 ; e, toluene- a-thiol, 58.2; f, toluene-4-thio1, 523.0. Titrant: 0.5 M potassium hydroxide solution in propan-2-01. Fig. 7. these compounds clearly form very stable cyanoethylation derivatives. Titration curves for the biscyanoethylsulphanilamide in two solvent systems (Fig. 8) show that only partial decyanoethylation occurs during the titration. Although the rounded form of the curves suggests that the decyanoethylation process is occurring slowly, no significant increase in the measured reaction stoicheiometry was achieved by reducing the rate of titrant addition to 0.02 ml min-l.Another difficulty that can arise when acrylonitrile is used as the indicator reagent is the ready cyanoethylation, in the absence of a catalyst, of some solvents that are used success- fully in the non-aqueous potentiometric titration of weak acids, namely primary and secondary alkylamines. The use of such solvents leads to a large rise in temperature during the course of the thermometric titration, before the end-point is indicated.21 However, in a re-examination of this problem we found that, although the use of an amine solvent I 0.1 M Potassium hydroxide in propan-2-01 (1 division = 1 ml) Fig.8. Titration graphs for dicyanoethylated and uncyanoethylated sulphanilamide. Sample : a and b, NN-bis(2'-cyanoethyl)-4-aminobenzenesulphonamide, 27.8 mg; c and d, sulphanilamide, 17.2 mg. Sample solvents : a and d, 3 ml of acrylonitrile plus1 ml of dimethyl- formamide; b and c, 4 ml of acrylonitrile. Titrant: 0.1 M potassium hydroxide solution in propan-2-01.810 GREENHOW, DAJER DE TORRIJOS: IONIC POLYMERISATION FOR END-POINT Analyst, VoZ. 104 (morpholine) causes a premature temperature rise, acceptably sharp end-point inflections can still be achieved with solvent mixtures containing morpholine and acrylonitrile in equal volume, particularly when dimethylformamide is present (Fig. 9). Conclusions This investigation has revealed that primary and secondary alcohols present in the titrant or sample solutions during the catalytic thermometric titration undergo cyanoethylation when acrylonitrile is the indicator reagent.There is strong evidence that these cyanoethylation reactions occur before the anionic polymerisa tion of acrylonitrile. When highly reactive alcohols are present some cyanoethylation may occur before the end-point of the determina- tive reaction, and this will affect the end-point adversely. Thus, although it is desirable that the primary or secondary alcohols should be reactive they should not be too reactive. The results obtained with potassium tert-butoxide in tert-butanol as the titrant show that cyanoethylation is not an essential requirement for an acceptably sharp end-point inflection, provided that the titrant is sufficiently active to initiate the anionic polymerisation after a very brief induction period.0.1 M Potassium hydroxide in propan-2-ol/ml (1 division = 1 ml) Fig. 9. Effect of morpholine in the titrand solvent. Sample: benzoic acid, 12.2 mg. Sample solvent (millilitres of acrylonitrile - morpholine - dimethylformamide) : a, 2 : 0 : 4; b, 2 : 1 : 3; c, 2 : 2 : 2; d, 2 : 4 : 0. Titrant: 0.1 M potassium hydroxide solution in propan-2-01. The solvating properties of the alcohols and other co-solvents have an important effect on the rate of the indicator reaction and, therefore, on the sharpness of the end-point inflection. As solvation of the catalyst anion by alcohols reduces the rate of the cyanoethylation reaction the concentration of alcohols should not be too high.In addition, the presence of dipolar aprotic solvents, such as dimethylformamide, that reduce the solvating effect of thc alcohols and solvate the cation but not the anion of the titrant - catalyst is advantageous When 0.5 M titrants are used, there is evidence that the presence of dimethylformamide can promote premature rises in temperature, before the end-point, when alcohols are absent from the sample solution. This suggests that alcohols have a modifying effect on the dimethylformamide. The premature rises in temperature with more concentrated titrants is related to the competition between the determinative and indicator reactions, and they can be reduced or eliminated by adding the titrants at a slower rate than is used with less concentrated titrants.In devising a combination of titrant and sample solvent in order to achieve the ideal end- point inflection in these titrations, the aim is to minimise the extent of reactions other than the determinative reaction before the end-point, and to maximise the rate of the indicator reaction immediately the titrant is present in excess. Suitable reagent - solvent combina- tions will have, as the titrant, either a higher primary alkoxide in a higher primary alkanol or a tertiary alkoxide in a tertiary alkanol arid, as the sample solvent, a mixture of acrylo-September, 1979 81 1 nitrile, a dipolar aprotic solvent (dimethylformamide, dimethyl sulphoxide or hexamethyl- phosphorotriamide) and, when the tertiary alkoxide reagent is used, a higher primary alkanol to promote a rapid cyanoethylation reaction.Titration curves obtained with these titrant - solvent combinations are shown in Fig. 10. INDICATION I N NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART x 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. a I: Titrant/ml (1 division = 1 ml) Fig. 10. Ideal titration curves. Sample: benzoic acid, 12.2 mg.* Sample solvent? and titrant:: a, I, A; b, 11, B; c, 111, A; d, IV, C. *d = 61.1 mg. TI = 4 ml of acrylonitrile plus 2 ml of dimethylformamide; I1 = 3 ml of acrylonitrile plus 2 ml of dimethylformamide plus 1 ml of n-butanol; I11 = 4 ml of acrylonitrile plus 2 in1 of dimethyl sulph- oxide; IV = 4 ml of acrylonitrile plus 1 ml of dimethylformamide plus 1 ml of n-butanol. :A = 0.1 M potassium n-butoxide in n-butanol; B = 0.1 M potas- sium tert-butoxide in tert-butanol; C = 0.5 M potassium tert-butoxide in tert-butanol. References Greenhow, E. J., Chemy Ind., 1972, 422. Greenhow, E. J.. Chemy Ind., 1972, 466. Greenhow, E. J., Nadjafy, A., and Dajer de Torrijos, L. A., Analyst, 1978, 103, 411. Greenhow, E. J., and Loo, L. H., Analyst, 1974, 99, 360. Greenhow, E. J., and Spencer, L. E., Analyt. Chem., 1975, 47, 1384. Utermohlen, W. P., J . Am. Chem. SOC., 1945, 67, 1505. Bachman, G. B., and Levine, H. A., J . Am. Chem. SOC., 1948, 70, 599. Hurd, C. D., and Gershbein, L. L., J. Am. Chem. SOC., 1947, 69, 2328. Misra, G. S., and Rao, M. V. R., J . Indian Chem. SOC., 1976, 53, 953. Greenhow, E. J., and Spencer, L. E., Analyst, 1973, 98, 98. Caldin, E. F., and Long, G., J . Chem. SOC., 1954, 3737. Greenhow, E. J., and Shafi, A. A . , Analyst, 1976, 101, 421. Feit, €3. A., and Zilkha, A., J . Org. Chem., 1963, 28, 406. Zilkha, A., Feit, B. A,, and Frankel, M., J . Chem. SOC., 1959, 928. Maerker, G., Kenney, I i . E., and Donahue, E. T., J . Am. Oil Chem. SOC., 1968, 45, 72. Feit, B. A,, and Bigon, Z., J . Org. Chem., 1969, 34, 3942. Feit, B. A,, and Zilka, A,, J . Appl. Polym. Sci., 1963, 7, 281. Feit, €3. A., Sinnreich, J., and Zilkha, A., J . Org. Chem., 1967, 32, 2570. Greenhow, E. J . , and Nadjafi, A,, t o be published. Yagi, K., Tsuyama, S., Toda, F., and lwakura, Y., J . Polym. Sci., Polym. Chem. Ed., 1976, 14, Greenhow, E. J., and Spencer, L. E., Analyst, 1973, 98, 90. 1097. NOTE-References 4, 10, 12 and 21 are t o Parts VI, 111, VIII and I1 of this series, respectively. Received January 25th, 1979 Accepted April 6th, 1979