摘要:

Ordinary Members PROFESSOR R. J. DONOVAN 1983 PROFESSOR M. C. R. SYMONS 1983 DR G. J. HILLS 1984 PROFESSOR J. M. THOMAS 1983 PROFESSOR A. J. LEADBETTER 1984 DR J. ULSTRUP 1985 DR I . W. M. SMITH 1985 PROFESSOR G. WILLIAMS 1985 PROFESSOR F. L. SWINTON 1983 DR D. A. YOUNG 1984 Honorarj, Secretarj-: DR G. J. HILLS Honorarj- Treasurer : PROFESSOR P. GRAY The President thanked the retiring members of Council, Vice-presidents Professor Sheppard and Professor Wagner, and Ordinary Members Professor King and Professor Purnell, for their services. 5. Reriew of Futurr Acfirifies A programme of future activities of the Division had been tabled and the President drew attention to the forthcoming General Discussions and Symposia. xiOrdinary Members PROFESSOR R. J. DONOVAN 1983 PROFESSOR M. C. R. SYMONS 1983 DR G. J. HILLS 1984 PROFESSOR J. M. THOMAS 1983 PROFESSOR A. J. LEADBETTER 1984 DR J. ULSTRUP 1985 DR I . W. M. SMITH 1985 PROFESSOR G. WILLIAMS 1985 PROFESSOR F. L. SWINTON 1983 DR D. A. YOUNG 1984 Honorarj, Secretarj-: DR G. J. HILLS Honorarj- Treasurer : PROFESSOR P. GRAY The President thanked the retiring members of Council, Vice-presidents Professor Sheppard and Professor Wagner, and Ordinary Members Professor King and Professor Purnell, for their services. 5. Reriew of Futurr Acfirifies A programme of future activities of the Division had been tabled and the President drew attention to the forthcoming General Discussions and Symposia. xi

ISSN:0300-9599    

DOI:10.1039/F198278FX001

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

3 708 REVIEW OF BOOKS is the absence of any reference to possible new and potentially significant applications for polymer latices. Novel applications may well be found in at least two directions, namely, those which exploit the large polymer-aqueous-phase specific surface area of latices, and those which exploit the electrical dissymmetry which is present at the interface between polymer and aqueous phase in the case of electrostatically stabilised latices. No reference is made in this book to the efforts which have so far been made to exploit for medical purposes the adsorptive and binding potentialities of the large area of polymer-aqueous-phase interface in latices. Nor is there any mention of possible catalytic applications of this large interfacial area. So far, catalytic applictions have been confined to those which rely essentially upon enhancement of the counter-ion concentration in regions of the electrical double layer which are near to the polymer surface.However, it is at least possible that the adsorptive capacity of the interface may also be useful in catalytic applications. Some discussion of possibilities such as these would have been welcome. D. C. BLACKLEY Received 14th April, 1982 Shock Waves in Chemistry. Ed. by ASSA LIFSHITZ. (Marcel Dekker, New York, 1981). Pp. ix + 390. Price SFr 182. After a somewhat hesitant start, the use of shock waves to study chemical and physical processes at high temperatures has become an accepted technique and reliable kinetic data can be obtained in this way. Several books have been written, notably by Bradley and by Gaydon and Hurle, which describe not only the underlying principles and the experimental procedures but also give some account of the early results obtained using shock waves to provide high temperatures for short, well defined times in the reactant gases.Inevitably, these books have become rather dated. This new book, edited by Lifshitz, is rather different. It is a collection of self-contained review articles on various aspects of shock waves. The first (by Khandelwal and Skinner) is concerned with hydrocarbon oxidation, and the next (by Tsang) describes the results obtained using the comparative rate technique which he has pioneered. Both these articles include extensive lists of references and represent useful summaries of the present situation.Boyd and Burns have contributed a chapter on dissociation-recombination reactions, while Kiefer describes the laser-schlieren method which he has done so much to develop. There is another chapter by an acknowledged expert, Just, on atomic resonance absorption spectrometry. Under shock-tube conditions it is very seldom that the concentrations of radicals and other species reach a steady state, and so the classical Bodenstein steady-state approximation cannot be used. Instead, it is necessary to integrate the differential equations describing the time-variation of species concentration, and Gardiner, Walker and Wakefield have provided a useful guide to the computational procedures available in this and other aspects of shock-tube work.In addition to these contributions there is another by Bar-Nun on Chemical Aspects of Shock Waves in Planetary Atmospheres which, although interesting in itself, fits rather uneasily with its companions. As is inevitable in a book of this type the standard and style of the chapters varies and there is some overlapping material; none of this, however. represents a serious drawback. What is more difficult to understand is the audience for whom the book is intended. Each chapter is a useful and interesting review which will appeal to a fairly restricted readership, but, in the opinion of this reviewer, the whole volume lacks coherence. The time-honoured phrase ‘should be on the shelves of every library’ probably applies, though the price, over &50 at the current exchange rate, must cause all university librarians to flinch in these days of U.G.C.cuts. There is still room for the definitive up-to-date book to be written on shock waves in chemistry. J. A. BARNARD Received 5th April, 19823 708 REVIEW OF BOOKS is the absence of any reference to possible new and potentially significant applications for polymer latices. Novel applications may well be found in at least two directions, namely, those which exploit the large polymer-aqueous-phase specific surface area of latices, and those which exploit the electrical dissymmetry which is present at the interface between polymer and aqueous phase in the case of electrostatically stabilised latices. No reference is made in this book to the efforts which have so far been made to exploit for medical purposes the adsorptive and binding potentialities of the large area of polymer-aqueous-phase interface in latices.Nor is there any mention of possible catalytic applications of this large interfacial area. So far, catalytic applictions have been confined to those which rely essentially upon enhancement of the counter-ion concentration in regions of the electrical double layer which are near to the polymer surface. However, it is at least possible that the adsorptive capacity of the interface may also be useful in catalytic applications. Some discussion of possibilities such as these would have been welcome. D. C. BLACKLEY Received 14th April, 1982 Shock Waves in Chemistry. Ed. by ASSA LIFSHITZ. (Marcel Dekker, New York, 1981). Pp. ix + 390.Price SFr 182. After a somewhat hesitant start, the use of shock waves to study chemical and physical processes at high temperatures has become an accepted technique and reliable kinetic data can be obtained in this way. Several books have been written, notably by Bradley and by Gaydon and Hurle, which describe not only the underlying principles and the experimental procedures but also give some account of the early results obtained using shock waves to provide high temperatures for short, well defined times in the reactant gases. Inevitably, these books have become rather dated. This new book, edited by Lifshitz, is rather different. It is a collection of self-contained review articles on various aspects of shock waves. The first (by Khandelwal and Skinner) is concerned with hydrocarbon oxidation, and the next (by Tsang) describes the results obtained using the comparative rate technique which he has pioneered.Both these articles include extensive lists of references and represent useful summaries of the present situation. Boyd and Burns have contributed a chapter on dissociation-recombination reactions, while Kiefer describes the laser-schlieren method which he has done so much to develop. There is another chapter by an acknowledged expert, Just, on atomic resonance absorption spectrometry. Under shock-tube conditions it is very seldom that the concentrations of radicals and other species reach a steady state, and so the classical Bodenstein steady-state approximation cannot be used. Instead, it is necessary to integrate the differential equations describing the time-variation of species concentration, and Gardiner, Walker and Wakefield have provided a useful guide to the computational procedures available in this and other aspects of shock-tube work.In addition to these contributions there is another by Bar-Nun on Chemical Aspects of Shock Waves in Planetary Atmospheres which, although interesting in itself, fits rather uneasily with its companions. As is inevitable in a book of this type the standard and style of the chapters varies and there is some overlapping material; none of this, however. represents a serious drawback. What is more difficult to understand is the audience for whom the book is intended. Each chapter is a useful and interesting review which will appeal to a fairly restricted readership, but, in the opinion of this reviewer, the whole volume lacks coherence. The time-honoured phrase ‘should be on the shelves of every library’ probably applies, though the price, over &50 at the current exchange rate, must cause all university librarians to flinch in these days of U.G.C. cuts. There is still room for the definitive up-to-date book to be written on shock waves in chemistry. J. A. BARNARD Received 5th April, 1982

ISSN:0300-9599    

DOI:10.1039/F198278BX003

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1982,78, 7-15 Complex Formation in Molten Salts Association Constants of Lead Halide Complexes in Molten KN0,-Ba(NO,), Eutectic as Solvent BY RAGHUVESH K. GUPTA AND HARISH C. GAUR* Centre of Advanced Study, Department of Chemistry, University of Delhi, Delhi-110 007, India Received 1st July, 1980 Thermodynamic association constants and specific Helmholtz free energies for the formation of the species PbX+, PbX, (X = C1, Br or I) in dilute solutions of Pb(NO,), and KX in molten KN0,-Ba(NO,), (87.6: 12.4 mol%) in the temperature range 568.2-628.2 K are reported. An e.m.f. method involving measurements of activity coefficients of (K,Ba)X and Pb(NO,), using Ag,AgX(s) and Pd-PdO-PbO indicator electrodes, respectively, was employed. Data did not suggest the formation of the dinuclear species.Applicability of the quasi-lattice model equations in predicting temperature coefficients of association constants in the above temperature range has been examined. The near-ideal behaviour of Ag,Ag+ as an indicator electrode in molten alkali and alkaline earth nitrate mixtures as solvents has enabled the potentiometric investigation of silver halide c~rnplexes~-~ in these solvents. Relatively fewer similar studies of complexes of other metal ions, particularly divalent, using indicator electrodes of second kind have been rep0rted.l Although the use of Pd-PdO-CdO as an electrode of a third kind, reversible to Cd2+, was suggested by Inman,8 the applicability of this electrode to other cases does not appear to have been fully exploited.The object of this study was to explore further the possibility of extending the use of the Pd-PdO-MO (M = Pb, Co, Ni, Zn) electrode to potentiometric studies of the halides and other complexes of these metals. In this paper we present a potentiometric study of association equilibria of lead-chloro complexes using the Pd-PdO-PbO indicator electrode, and lead-bromo and iodo complexes using the Ag,AgX(s) (X = Br or I) indicator electrode. EXPERIMENTAL The cell, furnace, method of temperature control and measurement, solvent preparation etc. have been described ear lie^.^ Electrodes were made from silver and palladium wires (1 mm diameter, purity 99.99%, Arora-Matthey) coiled at one end and dipping into a molten-salt solution; the other end of the wire served as the potentiometric lead.Before use, the electrodes were pre-treated in a uniform manner. The palladium wire was polished with an extra fine emery paper moistened with AnalaR acetone and wiped dry. The surface of the wire showed no scratches at x 10 magnification. After a series of measurements the wire had a superficial blue-black surface coating in presence of which the electrode behaviour was sluggish and erratic; it was removed before re-use. The silver wire was pre-treated in the same manner but finally heated in an alcohol flame. The reference half-cell used in the investigation was set up by dipping a coiled silver wire into a solution containing ca. lop3 moles of AgNO, per mole of solvent and an excess of KX (to precipitate AgX), and was isolated in a small portion of the melt contained in a glass tube with a fritted-glass (porosity G-4) bottom dipping in the melt; the rest of the melt in the cell served as the indicator half-cell.7Ag,AgX(s)- KN0,-Ba(NO,), KX(R kx) (4 where n, is the number of moles of species i KN0,-Ba(NO,), PbO, PdO, Pd KX(RKX) (X = C1) Pb(N03)2 (RPb(NOI)2) Agx(s),Ag (X = Br,I) (+I and lZKN03 = 7.06. Y = - nBa(N03)~ At the low solution concentrations (Rd -c lo-,) employed in this study ion molar ratios are not significantly different from the ion fractions. The Ag,AgX(s) (X = Br, I) indicator electrode was set up by the addition of weighed amounts of AgNO, (R,,,,, < lo-, mole fraction) and an equivalent amount of KX. The Pd-PdO-PbO indicator electrode was set up by adding ca.250 mg each of PdO and PbO to the melt. The system was occasionally stirred over 8-10 h to achieve equilibration after which a known amount of KX-Pb(NO,), was added. Complexation was followed by change in e.m.f. of the cell on successive addition of Pb(NO,),-KX. The cell e.m.f. was measured with a L & N (type K-4) potentiometer provided with a L & N (type 9834) direct-current null detector. EVALUATION OF ASSOCIATION CONSTANTS The graphical extrapolation method of Braunstein et aZ.l0 has been employed for analysis of the e.m.f. data; essential details, as needed, are given below. Consider a dilute solution containing both Pb(NO,), and KX in KN0,-Ba(N0,)2 as solvent, and assume that Pb2+ and X- associate to form species of the type PbX+, PbX,, Pb,X3+.. . , the corresponding association constants being Kll, Klz, K,, . . . . In a charge-asymmetric solvent mixture such as this there are different numbers of cations and anions and hence different numbers of sites available to cationic and anionic species; there may thus be ambiguity in assigning Pb-containing species to either the cationic or anionic sites. Instead of using the mixing statistics of charge-asymmetric mixtures, the convention of assigning all Pb-containing species to cationic lattices has been used.ll The total (stoichiometric) concentrations of the solute components* are given by the mass-balance equationsl0? l1 Rpb = Rpb2+ + R p b ~ + RpbXz -I- 2Rpb2x3+ - . . (3) Rx = Rx- +PRp,x+ + 2pRpbx2 +PRpb2~3+ - . . (4) * Subscripts Pb, X and SX refer to the components Pb(NO,),, KX and (K, Ba) X, respectively.R.K. GUPTA AND H. C. GAUR no. of solvent cations = no. of solvent anions y+ 1 where - y + 2' -- We define the activity coefficients of the solutes in the usual form 9 ( 5 ) and RPb l/YPb = - RPb2+ R X VYSX = - RX- and association constants as Kll = RPbX+/RPbZ+ RX- K12 = RPbX2/RPbX+ RX- &i = RPb2X3+/ RPbX+ RPbZ+. (Although the solute is added as KX, its molar ratio, which is < lop3, does not significantly alter the ratio of the two solvent cations and the symbol ysx refers" to the activity coefficients of the pseudo-component Ky12+y Ba112+yX.) Eqn (3) and (4) then take the form '/Yl'b = 1+Kl,RXYSX+KllK12R&Y~X+2KllK21RXRPbYsXYPb+... (6) '/YSX = +PK11RPbYPb+2pK11K12RPbRXYPbYSX+PK11K21 R k b Y h + - .* * (7) Using an iterative procedure on eqn (6) and (7) we get l/?Pb = + Kll R X + K11(2K21 -pKii) RX RPb + Kii K12 R & + - (8) '/?SX = 1 + ~ K l l R P b + ~ K 1 1 ( 2 K 1 2 - K 1 1 ) R X R P b + ~ K ~ ~ K ~ ~ R k b + * . * . (9) * Taking the logarithm of eqn (8), neglecting third- and higher-order terms in the expansion of the individual terms and rearranging, one obtains In ]/?Pb = K1l RX +K11(2K21 -PKii) RX R p t j +K11(K12-~K11) R$ -k . - (10) from which the equations for K,, and K,, using Ag,AgX(s) as indicator electrode are (1 1) s~(o,o) = lim [S,(O>] = lim (T ' ( l ~ ~ ) ) R x = pKll Rx+O Rx-0 and where S,(O) is defined by To evaluate KZl, eqn (9) was re-writtenlO as a function of Rpb in the form (l/ySx- 1) = A R p b + B R k b + . . ..(14) Least-squares fitting was done at several different initial concentrations of Rx. The intercept of the plot of B against Rx on extrapolation to Rx = 0, Bo, enabled the evaluation of K,, lim B = Bo = pKll K21. (15) Rx -*O10 COMPLEX FORMATION IN MOLTEN SALTS Similar equationsll for the Pd-PdO-PbO indicator electrode, based on eqn (lo), are and lim B = B, = K11(~12-4~11) HPb-'O where Spb(0) is defined by The extrapolated limits lead to thermodynamic association constants. RESULTS AND DISCUSSION The liquid-junction potential has been minimised by using the same solvent in both half-cells. The reference half-cell both with Ag,AgX(s), (X = Br, I) and Pd-PdO-PbO, (X = C1) indicator electrodes is the Ag,AgX(s) (X = C1, Br, I) electrode; solubility of silver halide does not interfere with its use as the reference half-cell since the chemical potential of AgNO, in it remains constant owing to its isolation from the indicator half-cell.However, use of Ag,AgCl(s) in the indicator half-cell is excluded owing to the higher solubility12 of AgCl in molten KN0,-Ba(NO,),; the Ag,AgX(s), (X = Br, I) indicator electrode, owing to the lower solubility of AgX in this solvent, has been used successfully. In the absence of Pb(NO,),, the observed Nernst behaviour for the cell with Ag,AgX(s), (X = Br, I) indicator electrode was within k0.2 mV in the concentration range used, indicating that the activity coefficient remains constant and may be defined as unity for an infinitely dilute solution. With the Pd-PdO-PbO electrode (in the absence of KC1) experimental Nernst slopes, Sexpt, were 68.0, 72.0 and 75.0 mV compared with the theoretical values 58.3, 60.3 and 62.3 mV at 588.2, 608.2 and 628.2 K, respectively; the discrepancy in the latter has been ascribede to the possibility of a mixed potential involving oxygen at the palladium electrode. Activity coefficients have been evaluated using eqn (21) where AE,,,, is the change in e.m.f.of the cell on the addition of KX-Pb(NO,),. In the evaluation of the activity coefficient (log ypb) using the Pd-PdO-PbO electrode, Sexpt was employedl37 l4 instead of 2.303 RT/2F. Extrapolation of the Ecell against Rpb2+ plot [before the addition of Pb(NO,),] to Rpb2+ = 0 showed that the concentration due to the dissolution of PbO was no greater than molar ratio; uncertainty of this order in the initial value of R p b would have a negligible effect on the result.Typical data for the variation of YSX/pb as a function of RX/pb were obtained; someR. K. GUPTA AND H. C . GAUR 11 of these data are plotted in fig. l ( a ) and 2(a).* The extrapolated plots are given in fig. l(b) and 2(b). Since the limiting slopes [asB,(o)/aRB,] at all the temperatures employed take near-zero values, it follows from eqn (12) that for lead-bromo association K,, = K11/2. Also the near-zero or negative values of Bo for X = Br or I and the negative slope of (1 +K,,R,,) spb(0) against Rpb plot (not shown) imply the absence of formation of dinuclear species Pb,X3+ (X = C1, Br, I) under the experimental conditions used in this study. 0.2 0.1 :::; 0 1 2 3 4 5 -1.5‘ - 1 :; - 2 0 2 4 6 8 1 104 R~ ab 7 6 1 I 5 t 1 1 I 1 I 0 2 4 6 8 1 1 I 0 4 R~~ FIG.1.-PbCI system. (Temperatures: 0, 588.2; A, 608.2; e, 628.2 K.) (a) Variation of yPb as a function of R,, at different initial values of Rpb ( x lo4): 2.213 (I), 5.524 (II), 5.678 (111). (b) Graphical extrapolation of limiting slopes to evaluate K,, using eqn (16) (0, A, a) and (24) (A). The intercepts give the following values of Spb(O,O): I, 69 2; 11, 73 k 2; 111, 79 f 2. (c) Extrapolation of the coefficient B to evaluate K , , [eqn ( 1911. The plots of (l/ySBr) against R,, are straight lines [fig. 2(a)] at low R,, over the concentration range of lead nitrate employed; the data thus correspond, within experimental error, to eqn (22) [cf. eqn (9)] lim l/ySX = +PK1lRPb).R x -0 Following Braunstein et al. l4 and from the thermodynamic relation * Detailed data (at nine temperatures and 100 initial values of R,,,, with 1112 data points) may be obtained from H. C. G. on request.12 COMPLEX FORMATION I N MOLTEN SALTS 104 R~ 2 .4 2.2 2 .o - ' 1.8 i2 + 1.6 1.4 1.2 1 .o 103 R~,, 104 R~ FIG. 2.-PbX system [X = Br (I, 11, III), X = I (IV, V, VI)]. Temperatures: 0,568.2; 0,588.2; A, 608.2 K. (a) Variation of ysx as a function of Rpb at different initial values of Rx ( x lo4): 12.239 (I), 10.817 (II), 6.121 (111), 5.777 (IV), 9.484 (V), 11.588 (VI). (b) Graphical extrapolation of limiting slopes to evaluate K , , and K,, [eqn (11) and (12)]. The intercepts give the following values of S(0,O): I, Sf;O) = 150::; 11, S$;O) = 135f 1; 111, Sf;') = 1225:; IV, Sio,O) = 2120+:!,; V, S:Ovo) = 2580f80; VI, Sio,O) = 3280+,6:.TABLE 1 .-ASSOCIATION CONSTANTS K,, AND K,, AND SPECFIC HELMHOLTZ FREE ENERGIES - AAll AND - AA,, FOR THE ASSOCIATION OF LEAD AND HALIDE IONS IN MOLTEN KN0,-Ba(NO,), EUTECTIC MIXTURES ~~ temp./K K1lU K1zU - AAllb - A A , , ~ 588.2 608.2 628.2 568.2 558.2 608.2 568.2 588.2 608.2 79+2 73+2 69+2 169:; 152+ 1 1372; 3687';; 2900 +_ 90 2383f!!, 33+ 1 26:; 212; 84f 1 76f 1 69+ 1 14701g 98 1 ?:: 81 1::: PbCl 13.80 & 0.12 13.89 +O. 13 13.38':::; 14.08 k 0.14 12.86'::;: 13.89 k 0.14 PbBr 16.772:::; 17.752:::: 16.86 k 0.03 17.89 f 0.06 Values of Ki, are in (moles per mole of solvent)-'. Values of AAij have been given for 2 = 5 only and are in kJ mol-l.R. K . GUPTA AND H. C.GAUR 13 Thus PKll spb(o) = 1 +pKll R,, by virtue of eqn (20), (22) and (23). This equation, valid at low solute concentrations and in tile absence of inuclear species, can enable evaluation of Kll even from a single set of data. Values of K,, (PbCI+) evaluated using eqn (24) [fig. 1 (b)], are close to the extrapolated values. Values of the association constants are given in table 1. Temperature variation in the values TABLE 2.-sPECIFIC HELMHOLTZ FREE ENERGIES FOR THE ASSOCIATION OF Pb2+ WITH x- (X = C1, Br, I) IN DIFFERENT NITRATE SOLVENTS nitrate solvents (mole %) - AAIla ( Z = 5 ) Li Na K Ba temp./K /kJ rnol-' ref. 50 - 50 - 87.6 - 25 75 50 50 53 47 75 25 50 87.6 100 - - - - - 50 - - - - - 87.6 - x = c1 - 433-473 12.4 568-628 X = Br - 553-573 - 5 13-573 - 528-592 553-573 - 433-473 12.4 568-608 - 51 3-573 X = I 12.4 568-608 - 14.40 13.92 17.07 16.86 16.30 16.63 18.90 16.86 17.21b 31.17 22 this work 24 24 25 24 22 this work this work a Average values in the indicated temperature range have been taken for comparison.* By extrapolation of values in ref. (24). of the Kii was used to evaluate the specific bond free energies the energy required for the formation of associated species PbiXYi-i)+, (X = C1, Br, I) using the following equations for the gener alised quasi-lat tice model' 5-1 Kll = Z(P11- 1) (25) K12 = T [ / ? 1 2 - 2- 1 1 P11- 1 where Pii = exp ( - Acii/RT) and Z is the quasi-lattice coordination number. Since the value of 2 is not known a priori, calculations were made for 2 = 4, 5 and 6 (a reasonable range of values of this parameter).18 Constancy of Acij (table 1) for different values of 2 indicated that the specific entropy of association was negligible and thus Aeij could also be c o n ~ i d e r e d l ~ ~ ~ ~ - ~ ~ as the specific Helmholtz free energies AAij so that the temperature coefficient of Kij was predictable within the temperature range investigated. AAll in this study has also been compared (table 2) with values in other single and binary molten nitrates as solvents.The effect of the presence of Ba2+ ions in the solvent mixture KN0,-Ba(NO,), on the association energy (available from data14 COMPLEX FORMATION I N MOLTEN SALTS for lead-bromo association) has been examined in terms of the reciprocal coulombic effe~f.~l-~, Considering AA,,, (in mixed solvent) to vary according to the equation (27) (where Y is the mole fraction of KNO,) calculations predicted a ‘destabilisation’ to the extent of 3.38 kJ mol-1 in the KN0,-Ba(NO,), eutectic mixture when compared AAmix = YAA,,,3+(l - Y)AABa(No3), 3.01 Ba _c - 1 \ L i with KNO,; the experimental results give 0.35 kJ mol-l.The discrepancy is probably due in part to the neglect of long-range coulombic forces,24 the dispersion and polarisation energiesl89 23 and also the difficulty in assigning21 an ‘effective ionic radius ’ to the nitrate ion. For a given solvent composition the magnitude of destabilisation is in the order Ba2+ > Na+, while stabilisation due to the presence of Li+ is observed (fig. 3). The authors are indebted to the University Grants Commission, New Delhi (India) for a teacher-fellowship at the C.A.S., Department of Chemistry, University of Delhi to R.K.G.* Y. Marcus, Molten Salt Mixtures in Introduction to Liquid State Chemistry (Wiley Interscience, New York, 1977), p. 255. * H. C. Gaur and R. S. Sethi, Trans. Faraday Soc., 1968, 64,445. H. C. Gaur and N. P. Bansal, Indian J. Chem., 1971,9, 1273. H. C. Gaur and N. P. Bansal, J. Chem. Soc., Faraday Trans. I, 1972, 68, 1368. A. K. Adya, K. W. D. Verma, R. S. Sethi, S. K. Jain and H. C. Gaur, Electrochim. Acta, 1979, 24, 267. A. K. Adya, K. W. D. Verma and H. C. Gaur, Indian J. Chem., 1979, 17A, 232. K. W. D. Verma, R. K. Gupta and H. C. Gaur, Trans. Soc. Adv. Electrochem. Sci. Technol., 1980, 15, 1 . H. C. Gaur and W. K. Behl, Electrochim. Acta, 1963, 8, 107. * D. Inman, Nature (London), 1962, 194, 279, lo J. Braunstein, M. Blander and R. M. Lindgren, J . Am. Chem. Soc., 1962, 84, 1529.R. K. GUPTA AND H. C. GAUR 15 l1 H. Braunstein, J. Braunstein, A. S. Minano and R. E. Hagman, Znorg. Chem., 1973, 12, 1407. l2 H. C. Gaur and R. S. Sethi, Electrochim. Actu, 1968, 13, 1737. l3 D. Inman, Electrochim. Actu, 1965, 10, 11. l4 H. Braunstein, J. Braunstein and D. Inman, J. Phys. Chem., 1966,70, 2726. l5 M. Blander, J. Chem. Phys., 1961, 34, 432; J. Phys. Chem., 1959, 63, 1262. l6 M. Blander and J. Braunstein, Ann. N.Y. Acad. Sci., 1960, 79, 838. l7 D. G. Hill, J. Braunstein and M. Blander, J. Phys. Chem., 1960, 64, 1038. M. Blander, Molten Salt Chemistry (Wiley-Interscience, New York, 1964), p. 127. l9 S. H. White, D. Inman and B. Jones, Trans. Furaduy Soc., 1968, 64, 2841. 2o A. Alvarez Funes, J. Braunstein and M. Blander, J. Am. Chem. Soc., 1962,84, 1538. J. Braunstein and J. D. Brill, J. Phys. Chem., 1966, 70, 1261. 22 J. Braunstein and A. S. Minano, Znorg. Chem., 1964, 3, 218. 23 D. L. Manning, R. C. Bansal, J. Braunstein and M. Blander, J. Am. Chem. Soc., 1962, 84, 2028. 24 D. L. Manning, M. Blander and J. Braunstein, Inorg. Chem., 1963, 2, 345. 25 M. Bonneymay and R. Pineaux, C.R. Acad. Sci., 1955, 240, 1774. (PAPER O/ 1027)

ISSN:0300-9599    

DOI:10.1039/F19827800007

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1982, 78, 17-28 Infrared Investigation of Ionic Hydration in Ion-exchange Membranes Part 2.-Alkaline Earth Salts of Grafted Polystyrene Sulphonic Acid Membranes BY LEON Y. J,EVY, MARC Muzzr AND HENRI D. HURWITZ* Laboratoire de Thermodynamique Electrochimique, Faculte des Sciences, C.P. 160, Universite Libre de Bruxelles, 50, av. F. D. Roosevelt, 1050 Brussels, Belgium Receiued 5th September, 1980 Infrared spectroscopic studies have been performed on thin ion-exchange membranes which consist of alkaline earth salts of polystyrene sulphonic acid grafted on a Teflon FEP matrix. The membranes were placed at isopiestic equilibrium with water vapour in the sampling cell. The water absorption isotherm at 25 O C has been determined by measuring the integrated absorbance of the sorbed-water bending vibration.In the case of Mg2+, Ca2+ and Sr2+ salts the isotherms exhibit an absorption step with a hysteresis. The frequency of the symmetrical stretching band of the SO; group passes through a minimum, and the OH stretching band of water in the membrane passes through a maximum at the relative water vapour pressure corresponding to the step displayed by the absorption isotherm. The dependence of the bending vibration band of water is also analysed in terms of solvent structure in the membrane. A model of hydration of alkaline earth salts of an FEP-PSSA membrane which emerges from the spectral analysis is suggested. The ionic interaction and hydration occurring with alkaline salts of polystyrene sulphonic acid grafted on a fluoroethylene propylene matrix (FEP-PSSA) were investigated in a previous pub1ication.l An infrared spectroscopic method was devised for examining these membranes while they were held under isopiestic conditions for the control of water absorption.In conjunction with these measurements, the water content of the film was determined using Karl-Fischer titrimetry.2 The validity of the i.r. spectroscopy applied to thin polyelectrolytic films was emphasized in the pioneering work of Z ~ n d e l . ~ This work and a previous contribution1 suggest that a comprehensive study of water absorption based on spectroscopic arguments should give access to a molecular interpretation of some of the phenomenological properties of ion-exchange membranes, such as swelling and selectivity. The results obtained with Teflon films provide further indication of the significant role played by the hydrophobic matrix in the formation of specific hydration structures.This aspect of the research is of particular interest due to similar effects which might occur in hydrophobic regions of biological membranes. In the case of alkaline earth ions, the shifts of the absorption band maxima of the symmetric stretching vibration vsoL of the sulphonate group and of the stretch- ing vibration vOH of water suggest that with increasing water uptake by the membrane, the cations are peripherally hydrated.l* Some unexpected features were found from the shift of BOH, the maximum of the scissor vibration band of water. This shift reveals the complex interdependence of ionic interaction and water configuration in the membrane pores.In the case of K+ and Cs+, at small water activity, the bending force constant of water is larger than in ice; conversely, with Li+ it is smaller than in liquid 1718 IONIC HYDRATION I N ION-EXCHANGE MEMBRANES - or lb io io i o 50 Qo ;o 80 sb l!O relative humidity (%) FIG. 1 .-Absorption isotherms at 25 OC for the alkaline earth salts of the FEP-PSSA membranes. (>, Mg2+; ., Ca2+; A, Sr2+; A, Ba2+. TABLE I.-NUMBER OF WATER MOLECULES ABSORBED IN THE THOROUGHLY DRIED FEP-PSSA MEMBRANE ion Mg2+ 3.1 Ca2+ 2.1 Sr2+ 1.8 Ba2+ 0.4 TABLE 2.-vIBRATION FREQUENCIES (Cm-') FOR POLYSTYRENE SULPHONIC ACID MEMBRANE AND PSSA GRAFTED ONTO FEP MEMBRANE AS A FUNCTION OF THE COUNTER-ION FOR A RELATIVE HUMIDITY OF 7% '0 H 60, vso; nHzO counter-ion PSSa FEP-PSSA PSSa FEP-PSSA PSSa FEP-PSSA FEP-PSSA Mg2+ 3394 3436 1640 1646 1050 1049 4.33 Ca2+ 3406 3433 1628 1638 1044 1048 2.36 Sr2+ 3412 3458 1625 1636 1039 1040 2.48 Ba2+ 3440 3455 1621 1633 1035 1038 0.86 a After Zundel, ref.(3).L. Y. LEVY, M. MUZZI AND H. D. HURWITZ loo- 19 " 3800 3400 3000 wavenumber/cm -' FIG. 2.-Infrared spectra of the OH stretching mode of the alkaline earth salts of the FEP-PSSA membranes at 7% relative humidity. ( * * * .) Mg2+; (---) Ca*+; (-) Sr2+; (-.--) Ba2+. water.l The specificity of the role of cations involved in water uptake by the membrane is thus clearly emphasized. We endeavour, therefore, in the present publication to extend the investigation to the alkaline earth salts of FEP-PSSA membranes. EXPERIMENTAL Membranes of 62 f 1 pm thickness of polystyrene sulphonic acid grafted on Teflon FEP were used.* A value of 1.15 meq g-l with a scattering of 2.5% was found for the exchange capacity of the various selected membrane samples. An assignment of the i.r.absorption bands found between 3700 and 770 cm-l is given in ref. (1). The accuracy in wavenumber determination reaches f 1 cm-l for the narrow i.r. absorption bands and extends to + 5 cm-l in the most unfavourable cases. The experimental set-up and methods are described in ref. (1). Note that the water content of membranes taken at equilibrium with the laboratory atmosphere was measured by means of a Karl-Fischer titration technique.2 In the case of membranes placed at equilibrium with water vapour in the i.r. spectrophotometric sampling cell, the determination of the amount of sorbed water was achieved by linear interpolation and extrapolation of the relationship which has been established between nHzO, the number of water molecules per equivalent of sites, and AdOH, the integrated absorbance of the water bending absorption band at ca.1640 cm-l. * Progil, France; ref. C50-7-70.20 IONIC HYDRATION IN ION-EXCHANGE MEMBRANES l o o 1 0 1700 1650 1600 1550 wavenumber/cm -' FIG. 3.-Infrared spectra of the OH bending mode of the alkaline earth salts of the FEP-PSSA membranes at 7% relative humidity. Key as in fig. 2. RESULTS The variation of AVOH, the integrated absorbance of the stretching vibration mode of the sorbed water molecules, has been plotted as a function of nHZ0 in fig.4 of ref. (1). This diagram stresses that alkali metal and alkaline earth salts of the FEP-PSSA membranes obey a different relationship. On the other hand, it is shown in fig. 5 of ref. (1) that both types of salts behave identically as regards the change of ABOH, the integrated absorbance of the bending vibration mode, with respect to nHtO. The linear function passes through the origin. The absorption isotherms which have been derived for alkaline earth salts of FEP-PSSA membranes from these results are shown in An estimate of n&,o, the lowest amount of water per equivalent retained in so-called ' thoroughly dried ' membranes prepared under drastic conditions of drying [see ref. (l)], is recorded in table 1.Positions of the maxima of the intense absorption bands corresponding to the symmetric vibration vSOh of the -SO; group, and the stretching vibration vOH and bending vibration do, of the hydration water molecules in FEP-PSSA membranes at 7% relative humidity and 25 O C are reported in table 2. Some group variations fig. 1.L. Y. LEVY, M. MUZZI AND H. D. HURWITZ 21 FIG. 4.-Dependence of the position of the symmetric stretching vibration of the sulphonate group on the degree of hydration (dotted lines represent values at constant relative humidity). Key as in fig. 1. I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 5 10 15 n H 2 0 FIG. 5.-Dependence of the position of the OH stretching vibration of the water of hydration on the degree of hydration. Key as in fig. 1.22 FIG.6. on the degree FIG. 7.-Dependence nHzO of the half-width of the OH stretching vibration band of the water of hydration the degree of hydration. Key as in fig. 1 . onL. Y. LEVY, M. MUZZI AND H. D. HURWITZ 23 obtained by Zunde13 on salts of ungrafted PSS membranes are included in this table for comparison. Selected spectra are shown in fig. 2 and 3, corresponding to the vOH and do, vibration bands of water, respectively. They were obtained under equilibrium conditions with an atmosphere of relative humidity p/po = 7% at 25 OC. The frequency shifts of the maxima of the i.r. absorption bands for the various alkaline earth salts as a function of nHIO are shown in fig. 4, 5 and 6. Futhermore, fig. 7 shows the evolution of the half-width of the vOH band in terms of water uptake.DISCUSSION It has been seen' that the water absorption isotherms of alkali metal salts of FEP-PSSA membranes and PSSA resins cross-linked with 8 % DVB4 present roughly the same pattern. In contrast to this behaviour, there is a considerable difference in shape between isotherms of alkaline earth salts of FEP-PSSA membranes and PSSA resins. The membrane isotherms are much more dependent on the nature of the counter-cation. For most alkaline earth cations the water uptake at low relative water pressure p/po is also much larger whilst, as illustrated in fig. 1, from values o f p / p o x 0.25 onwards a jump in water content is observed. The wavy shape of the isotherm is accentuated in the presence of Mg2+ and is absent in the presence of Ba2+.In the latter case, the curve approximately conforms to the behaviour of PSS resins in the rangeplp, < 0.60. As sketched for Mg2+ in fig. 8, the rehydration of the membrane, starting from the vapour pressure of pure water, po, follows a desorption branch which is higher than the absorption branch in the region of the absorption step. Experimental points of both branches have been recorded in an identical way under conditions of isopiestic equilibrium of the membrane inside the i.r. sampling cell. The hysteresis observed with Mg2+, Ca2+ and Sr2+ salts might suggest either some phase transition of sorbed water molecules, undergoing a considerable mutual attraction, or some capillary condensation in the film.5 According to the experimental results for ' thoroughly dry membranes' it might be '0 1 10 20 30 40 50 60 70 80 90 100 relative humidity (%) FIG.8.-Absorption isotherms at 25 O C for the Mg2+ salt of the FEP-PSSA membrane as a function of hydration (a) and dehydration (H) in terms of relative humidity.24 IONIC HYDRATION I N ION-EXCHANGE MEMBRANES 10 c 0 10 20 30 40 50 60 70 80 90 100 relative humidity (%) FIG. 9.-Absorption isotherm corrected from the initial water uptake as a function of relative humidity. Key as in fig. 6. taken for granted that the initial water content corresponds to a physical absorption process which proceeds differently from any further water uptake and also involves a maximum of molecules, as recorded in table 1. A weaker absorption process sets in after the steep initial rise. Consequently we have subtracted ngz0, the number of water molecules absorbed in the thoroughly dried membrane, from the values of the ordinate of fig.1. The resulting curves drawn in fig. 9 indicate that the isotherms of all alkaline earth systems have a nearly identical slope at a vanishing value of AnHzO = nHZO -ng20. Thus, we might infer that they possess a similar AGE, the initial free energy of absorption for the water molecules involved in this absorption step. An approximate value of AGE can be computed from the slope of AnHz0 against the water activityplp, at vanishingp/p,, as shown in fig. 9. We therefore use the Henry’s law approximation of the Langmuir adsorption isotherm, where a is the fraction of sites accessible in this absorption step. Therefore AnH2,/a is the degree of coverage of the accessible sites.If one assumes that one-half (a = $) or one third (a = i) of the sites are accessible, AGE becomes, respectively, -8.8 and -9.8 kJ mol-l. When compared with the value of ca. - 10 kJ mol-1 obtained for AGE by Glueckauf6 with Li+ salts of PSSA resins, it becomes plausible that an average of of the overall exchange capacity of the grafted PSSA membrane remain available for this absorption process. Furthermore, it is apparent from fig. 9 that the isotherms of Mg2+, Ca2+, Sr2+ and Li+ reproduce the same curve up to AnFIz0 = 3.5. At larger values of water uptake, the absorption step sets in for the alkaline earth salts. We now turn to a spectral analysis of the absorption bands vso;, vOH and vOH. This will provide insight into the physical mechanism responsible for the pattern of the isotherms.L.Y. LEVY, M. MUZZI A N D H. D. HURWITZ 25 SYMMETRIC VIBRATION B A N D : Vso; As already pointed in the most stable ion-pair conformation the centre of the cation is placed in the direction of the SO bond axis of the sulphonate ion. The stronger the ion pairing, the higher the rise in frequencies of vsoT beyond 1040cm-l. As shown in fig. 4, the increase in strength of the interaction between the divalent cation and the fixed anion follows, at any water content, the sequence Ba2+ < Sr2+ < Ca2+ < Mg2+. We also infer from fig. 4 that the interionic bond strength decreases to a minimum with increasing hydration of the membrane for Sr2+, Ca2+ and Mg2+. Note that for the values ofp/p, plotted in fig.4 the minimum appears in the region 0.25 < (p/po)* < 0.33, corresponding to the rise of the absorption step in the isotherm, where (p/po)* is the value ofp/p, at the minimum. STRETCHING VIBRATION B A N D : VOH Notwithstanding the complexity of this band extending from 3200 to 3600 cm-l, some important conclusions can be drawn concerning the OH groups involved in hydrogen bonds. The fact that the membrane contains a negligible number of free water molecules is supported by the absence of any absorption peak at frequencies above 3600 cm-l. The faint shoulder detected at ca. 3600 cm-l is indicative of the small amount of free OH groups pertaining to water molecules hydrogen bonded by their second OH group to the sulphonate ion or to another water molecule absorbed within the membrane.These considerations are valid even at the lowest degree of hydration. On the other hand, the contribution of OH groups involved in hydrogen bonds is very important, as shown by the broadness and intensity of the vOH band. The width of the band is caused by variations in strength and length of the hydrogen Accordingly, it is suggested that broad absorption bands predict a strong induced dipole-ion interaction between the hydrogen bond and the cation. It is furthermore well known that the stronger the hydrogen bond, the larger the shift of the band maximum towards lower frequencies. One also observes in fig. 7 that, at a similar degree of hydration, the ionic interaction with the hydrogen bond decreases as the ionic radius increases.Moreover, the band displays a minimum half-width at (p/po)*, the relative humidity corresponding to the rise of the absorption step in the isotherms. The shifts of vOH plotted in fig. 5 stress the fact that at any given value of nHZq the hydrogen-bond strength also decreases as the ionic radius increases. It is striking that the hydrogen-bond strength passes through a minimum at the same critical value of relative humidity (PIP,,)* as found for the half-width of vOH, vsoT and the isotherms. BENDING VIBRATION B A N D : OH The interaction between a cation and its water of hydration affects predominantly the do, band.13 In order to describe this influence for an electrolytic solution it is necessary to identify two main ionic hydration effects: the building-up effect of a peripheral hydration sheet around the cation and the solvent-restructuring effect in the bulk electrolytic solution beyond the effective radius of the hydrated entity.The magnitude of the decrease in do, is therefore not simply indicative of water molecules bound more firmly to the cation than to other water molecule^.^^ Even though Ba2+ is less peripherally hydrated than Mg2+, it exhibits a more pronounced decrease in do, in electrolytic solutions. This apparent anomalous behaviour parallels some viscosity results13 proving that Ba2+ serves as a centre of disturbance for the structural order of the bulk solvent. Such concepts are also relevant in the case of ionic hydration in the membrane. As shown in fig. 6, the decrease of 2 FAR 126 IONIC HYDRATION IN ION-EXCHANGE MEMBRANES the maximum frequency do, follows the sequence Mg2+ < Ca2+ < Sr2+ Ba2+.It keeps the scissoring vibration frequencies at a value which is higher than that of free water molecules (ca. 1595 cm-l) but for Ca2+, Sr2+ and Ba2+ beneath that of liquid water (ca. 1645 cm-l). Recalling the tendency of these ions to weaken the hydrogen bonding, according to the sequence mentioned above, it becomes apparent that these ions act as entities with increasing ability to disorder water structure, thus producing a collapse around the ion of the ice-like structure of water. Unlike its divalent congeners, Mg2+ produces at low water content a bending vibration which is approximately the same as in liquid water. With increasing hydration, do, has values much above that observed in ice (at 1650 cm-l).This demonstrates a strong ordering effect presumably yielding, in the presence of a sufficient amount of water, a transition towards a highly rigid structure of the surrounding solvent network. We now note the remarkable observation that Li+ and Mg2+ both follow a similar variation in dOH up to frequencies below that of ice within the range 0 < AnH20 < 4-5. MODEL OF HYDRATION OF ALKALINE EARTH SALTS OF FEP-PSSA MEMBRANES Given the fact that water molecules remain absorbed in the dry membrane, we are inclined to adopt the hypothesis that the membrane contains, in limited regions of high graft density, a number of sulphonate sites crosslinked by divalent cations in such a manner as to trap a given amount of water into poorly accessible channels or pores. This view is supported by the strong inhomogeneity of grafting as revealed by electron micrographs,l and the almost irreversible hydration character of the film.Thus, as suggested from the evaluation of AGO,, if ca. 4 of the sulphonate sites remain accessible to reversible water absorption, there are $ of the sites serving to encage the water in the pores. Consequently, the number of water molecules pertaining to the pores per ionic crosslink amounts to an average of 9.3 in the case of Mg2+ and 1.2 in the case of Ba2+. The very weak contribution of free OH group vibration which has been detected suggests that nearly all these water molecules are organized in aggregates of associated molecules. The existence of such aggregates is also confirmed by the values of dOH recorded in the dry membrane and by the value of vOH which predicts that some water molecules undergo hydrogen bonding with the -SO; sites belonging to the neigh- bouring PSSA chains.The tightness of the occluded pores is correlated with the strength of ion association. Simultaneously, the hydrogen-bond donor ability of OH groups is enhanced by their polarization occurring in the strong electric field generated by small divalent ions. Both arguments support an increase of water uptake from Ba2+ to Mg2+. The model suggests further that the equilibrium absorption of the exchangeable water sets in on the remaining accessible sulphonate sites which initially bear no solvent molecules, this being equivalent in the process of absorption irrespective of the amount of water already fixed in the pores.This explains the similarity in slope and pattern of the initial branch of the water absorption isotherms for Mg2+, Ca2+, Sr2+, Li+ and to a lesser extent for Ba2+. As regards the inability of Ba2+ and K+ to reproduce exactly the same behaviour, we assume that a steric effect might impair the water absorption in the case of Ba2+ and note that ionic clustering has been advocated in order to describe the ion-solvent interaction of Na+, K+ and cs+. With increased hydration of the membrane, it is supposed that water molecules form bridges connecting several anionic sites, as depicted in fig. lO(a). In this configuration, the bending vibration mode of water is stiffened and tends towards that of liquid water.Simultaneously, the appropriate orientation of water molecules for hydrogen bondingL. Y. LEVY, M. MUZZI AND H. D. HURWITZ 27 'I' (cF~\.-:-IcF,). F FIG. 10.-Hydration model of the Mg2+ salt of the FEP-PSSA membrane: (a) at a low degree (b) at a degree of hydration corresponding to the absorption step. of hydration; with the sulphonate sites is hindered by the interaction with the cation linking those sites together. Hence a partial cationic hydration shell is built up at the cost of the hydrogen-bond donor properties of the OH groups of water. Similar conclusions have been drawn from the study of alkali metal salts of our grafted membrane9 and of salts of PSSA ungrafted membrane^.^ Zunde13 has developed the argument that the polarizing effect of the cation on the OH band is reduced if spread over several water 2-228 IONIC HYDRATION I N ION-EXCHANGE MEMBRANES molecules.Conversely, the formation of the cationic hydration shell leads to a progressive loosening of the ionic SOT-(cation)-SO; crosslink. It is thus expected that the ion-triplet configuration becomes unstable upon further water uptake. At a given water activity or critical value (p/po)*, the swelling of the membrane coincides with a transition to a more stable conformation. In the building up of the new hydration structure, the decrease in the overall free energy of hydration and the increase in the absorption of water are associated. The absorption isotherm steps rather abruptly to a larger value.For the physical interpretation of this process, one might assume that water molecules insert themselves between the divalent cation and the anion, and contribute to the ionic crosslinking as depicted in fig. 10(b). Such an interpretation of the ionic interaction relies, for instance, on a type of Robinson-Harnedl* localized hydrolysis. This effect leads to the formation of a kind of ion-triplet but differing in that polarized water molecules act as intermediary. The greater the strength of the ionic bond, the more polarized the solvent molecules will be, so the effect would decrease from Mg2+ to Ba2+. This conclusion is consistent with the experimental minima values for the shift of vso, in fig. 4. As such an interaction would enhance the hydrogen-bond donor property of the water molecules, it would also explain why in presence of a proton acceptor like SO; the shift of vOH decieases (fig.5) and the half-width of the band increases (fig. 7) for p / p o > (p/po)* with increasing water uptake. In this respect, it is worth comparing the behaviour of Li+ and Mg2+ salts. At low membrane water contents, it might be suggested that two Li+ and two SO; ions form an ion-pair doublet acting as an ionic crosslink. However, the hydration sheet which is formed around the monovalent cation disrupts the ionic pairs without fostering, at larger water contents, any new type of ion association or hydrogen bonding. The case of Li+ compared with Mg2+ in this work and with K+ in ref. (1) is a good example of the complex behaviour which can be met in ionomeric membranes or resins. We thank Prof. G. Zundel for helpful discussions. We also thank the ‘Fonds National de la Recherche Fondamentale Collective’ and the ‘Fonds National de la Recherche Scientifique ’ of Belgium for having supported this work. L. Y. Levy, A. Jenard and H. D. Hurwitz, J . Chem. SOC., Faraday Trans. I , 1980, 76, 2558. L. Y. Levy, A. Jenard and H. D. Hurwitz, Anal. Chim. Acta, 1977,88, 377. G. Zundel, Hydration and Intermolecular Interaction (Academic Press, New York, 1969). H. P. Gregor, B. R. Sundheim, K. M. Held and M. H. Waxman, J. Colloid Interface Sci., 1952, 7 , 51 1 . J. H. De Boer, The Dynamical Character of Absorption (Oxford University Press, 1953). E. Glueckauf and G. P. Kitt, Proc. R. Soc. London, Ser. A , 1955, 228, 322. G. S. Landsberg and F. S. Baryshanskaya, Izv. Akad. Nauk SSSR, Ser. Fiz., 1946, 10, 509. * Yu. Ya Efimov and Yu I. Naberukhin, Mol. Phys., 1975, 30, 1621. Yu. Ya Efimov and Yu I. Naberukhin, Mol. Phys., 1975, 30, 1627. lo Yu. Ya Efimov and Yu I. Naberukhin, Mol. Phys., 1975, 30, 1635. l1 B. I. Stepanov, Nature (London), 1946, 157, 808. l3 R. E. Nightingale, in Chemical Physics of Ionic Solutions, ed. B. E. Conway and R. G. Barradas l4 R. A. Robinson and H. S. Harned, Chem. Rev., 1942, 28,419. C. A. Coulson and G. N. Robertson, Proc. R. Soc. London, Ser. A, 1974, 337, 167. (J. Wiley, New York, 1966), p. 87. (PAPER O / 1377)

ISSN:0300-9599    

DOI:10.1039/F19827800017

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I, 1982, 78, 29-36 Hydration and Ion-exchange Process in Carboxylic Membranes Part 1 .-Infrared Spectroscopic Investigation of the Acid Membranes BY LEON Y. LEVY, ANDRE JENARD AND HENRI D. HURWITZ* Laboratoire de Thermodynamique Electrochimique, Facult6 des Sciences, C.P. 160, Universitk Libre de Bruxelles, 50, av. F. D. Roosevelt, 1050 Brussels, Belgium Received 5th September, 1980 An infrared investigation of Teflon FEP and PTFE membranes grafted with poly(acry1ic acid) is performed. The changes in the principal absorption bands are examined as a function of the degree of humidity, the density of grafting and the nature of the polymeric matrix of the acid membrane. The prominent role played by the dimerisation of carboxylic groups which gives rise to a network of intermolecular hydrogen bonds is assessed.The specific properties of polymeric membranes containing ionogenic sites are to a large extent determined by ionic and polar interactions. The formation and function of the ionic hydration structures in the membrane also play a prominent role. The influence exerted on the conformational characteristics of these structures by the hydrophobic macromolecular membrane matrix has been assessed in previous investi- gations dealing with the infrared spectroscopy of sulphonic ion-exchange mem- branes.' Analysis of the solvent sorption and desorption pattern in sulphonic systems has led to the detection of an appreciable rearrangement of the solvent network after a certain degree of swelling has been attained.2 With carboxy-containing ion exchangers, the degree of s w e m is much smaller than with sulphonic ion exchangers, and it is well known that this swelling is fundamentally not controlled by ionic hydrati~n.~? This property of carboxylic resins and membranes is now exploited for the production of new types of membranes used in electrochemical cells.In addition, it is inferred from the nature of the fixed carboxy groups and from the influence of the pH on the properties of the membrane that the study of this material is appropriate for the physico-chemical understanding and modelling of biological membranes. The present series of investigation is aimed at elucidating the hydration and the ionic interactions in carboxy-containing ion exchange membranes by means of infrared spectroscopy performed on thin films of poly(acry1ic acid) grafted on different types of perfluorated matrices.This publication deals more specifically with the properties of these membranes in their acid forms as a function of their exchange capacity and water content. EXPERIMENTAL PREPARATION OF SAMPLES The characteristics of the membranes (supplied by Progil, France) are given in table 1. All i.r. spectroscopic measurements were carried out on samples of 32 mm radius with a Beckman 2930 I.R. INVESTIGATION OF ACID MEMBRANES TABLE 1 .-MEMBRANE CHARACTERISTIC thickness capacity membrane matrix h m /meq g-' nH* 0 /es-' 11A FEP 17 12A FEP 17 13A FEP 17 14 PTFE 9 ~ ~ ~ _ _ _ _ _ _ _ _ _ _ _ _ _ 0.97 2.23 1.25 2.92 0.64 1.93 0.69 2.62 I.R. 9 double-beam spectrophotometer. The successive experimental steps are as follows: (1) The membranes are placed successively in three fresh solutions of 1 mol dm-3 HC1, for four hours in each solution.After being removed from the solution, the membranes are washed with distilled water and carefully blotted with filter paper. (2) The membranes are fixed on a membrane holder and left to reach equilibrium with the laboratory atmosphere for one hour. (3) A first spectrum of the membranes is recorded. (4) The membrane remains exposed to the i.r. light in the sampling cell through which a flow of dry air (less than 4 ppm H,O at 25 "C) is passed. A spectrum of the membranes is recorded at regular periods during this drying process. DETERMINATION OF WATER CONTENT I N THE MEMBRANE Membranes at hydration equilibrium with the laboratory atmosphere at 25 OC [as used in step (3) of the experimental procedure] were titrated by the Karl-Fischer method.An automatic KF4 Beckman apparatus was modified slightly in order to satisfy the anhydrous conditions required for the titration agent and the titration cell during the experiment. Unfortunately, the very strong absorption of the C=O stretchingmode vibration at 1700 cm-l prohibits the use of the integrated absorbance of the water bending mode of vibration for determining the water content in the membrane, as can be done with the sulphonic films.5 DETERMINATION OF EXCHANGE CAPACITIES The exchange capacity of the samples prepared for the i.r. spectroscopic investigation was obtained through a coulometric multi-step microtitration carried out following the method described by Levy.* A Pt electrode of 1 cm2 area was used as the hydroxyl ion source.The cell assembly was composed of a combined Ingold microelectrode (type HA405 MJNS) and a Pt counter-electrode in a separate compartment connected to the main compartment by means of a saturated K,SO, agar-agar bridge. A potentiostat (Tacussel PRT 3001) was used as a galvanostatic source providing current intensities of 250 pA with a precision better than 1 pA. An automatic set-up was designed to fix the pH in the well-stirred main compartment at a value of approximately 7. The current injection is stopped whenever pH 8 is reached and starts again whenever the pH falls beneath 6. The coulometry yields the amount of protons released by the membrane sample.RESULTS ASSIGNMENT OF THE INFRARED ABSORPTION BANDS The i.r. spectra of the membrane (1 1 A) in its acid form and salt form are shown in fig. 1. It is well known that five characteristic frequencies 2500-2700, 1700, 1400, 1200-1 300 and 900 cm-l can be ascribed to the functional group of carboxylic acids.' The same absorption bands are observed with our membranes in their acid forms. The most intense absorption band at ca. 1720 cm-l corresponds to the stretching vibrations of the carbonyl group vc=o in the cyclic dimeric or polymeric form ofL. Y. LEVY, A. JENARD AND H. D. HURWITZ 31 4 000 3000 200 0 1600 1200 800 wavenumber/cm-' I I 4000 3000 2000 1600 1200 800 wavenumber/cm-' FIG. 1.-(A) Infrared spectra of an FEP-PAA membrane (1 1A) after various times of drying.( 1 ) Initial hydrated state; (2) 2 h; (3) 24 h. (a) Shoulder at 3400 cm-I; (b) shoulder at 3600 cm-I. (B) Infrared spectra of Na+ salt of a FEP-PAA membrane (11A) after various times of drying. ( 1 ) Initial hydrated state; (2) 3 h; (3) 20 h; (4) 56 h. undissociated acid molecules. The formation of hydrogen bonds with the carbonyl group shifts the vc=o vibration by ca. 15-45 cm-l towards lower wave number^.^^ lo The broadness of this band may be attributed to the superposition of individual peaks due to the heterogeneity of the local molecular configuration around the carbonyl group. This is supported by the fact that the variation in the mutual distance between carbonyl groups affects the vc=o stretching vibration.For instance, oxalic acid absorbs very strongly in the region extending from 1720 to 1690 cm-l, while the spectrum of malonic acid shows two bands at 1710 and 1740 cm-l. With salts of the carboxylic acid, the difference between stretching vibrations with single-bond character at 1320-1210 cm-l (masked in our system by the Teflon absorption band) and with double-bond character at 1720 cm-1 disappears owing to the mesomeric effect and two new absorption bands are found which are then attributed to the carboxylate group.' The absorption band at 1410 cm-l is assigned to the symmetrical stretching vibration of the ionised group COO- whereas the absorption band at 1575 cm-l is assigned to the asymmetrical stretching vibration of COO-. However, the band at 1410 cm-l is very complex and is due also to the bending motion of the >CH, group in the a position of the carboxy group.Concerning the assignment of the broad absorption bands of low intensity near 2650 and 1970 cm-l, the available investigations32 I.R. INVESTIGATION OF ACID MEMBRANES systems (X representing various elements including carbon) indicate that on -X these bands are connected.lo-l3 They correspond to the A and B bands, respectively, in the nomenclature proposed by Sheppard14 and are observed in the case of dimer formation through hydrogen bonding. Their interpretation raises various hypotheses. Band A might belong to the combination vOH with overtone 2doH and to do, + yOH. Band B occurs through Fermi resonance of vOH with 2yOH. Hadzi and Kobilarov12 suggest that the participation of vOH is more important in band A than in band B, which causes a shift of band A towards smaller wavenumbers and a decrease of intensity with increasing strength of the hydrogen bond. Another band at 1280 cm-l can also be ascribed to the dimer formation by carboxylic groups but is masked in our case by the large absorption band at 1200 cm-l due to the CF, groups of the Teflon matrix.As for the very faint peak at 940 cm-l, it is attributed to the out-of-plane OH group vibration in the acid dimer.g A band at 850 cm-l is found only in the salt form of the membrane and corresponds to the bending deformation of the COO- group or the rocking deformation of the CH, group.15 Finally, the broad absorption band between 3000 and 3600 cm-l is due to the stretching vibration of OH groups belonging either to the water of hydration or to the carboxylic acid.Several bands have yet to be assigned. Two bands at 2940 and 2850 cm-l are ascribed to the asymmetric and symmetric stretching vibrations, respectively, of the CH, groups.' A band at 1454 cm-l arises from the combination of the scissoring vibration of CH, and of the coupling of the C-0 stretching vibration with the OH in-plane bending vibrati0n.l' The band at 1330 cm-1 corresponds to the wagging motion (out-of-plane bending) of the CH, group.' Bands at 2380, 1280, 1160, 980, 775, 750, 720, 628, 555 and 519 cm-l are associated with different motions of the CF, or CF, vibration groups as described in ref. (1 6) and (17). 4 0 'OH DISCUSSION Comparison of the i.r.spectra displayed by carboxylic membranes with those displayed by sulphonic membranes shows a great difference in the pattern of the OH band. Unlike the carboxylic acid system, in the sulphonic acid system the bands are not resolved in the high-frequency region, which thus shows a continuum of absorption. This effect was interpreted by Zundel and Weidemann.18 For these authors, the continuous absorption, characteristic of the strong acid spectrum, appears whenever a hydrogen ion solvated structure like H,Oi or HgOi is formed. The formation of a H,Oi grouping implies necessarily that the acid is completely dissociated. This happens already, as shown by Zundel, for two water molecules per site in sulphonic acid membrane^.^^ From the previous consideration, our results lead us to conclude that in the carboxylic acid membranes the acidic groups of the vast majority of carboxylic acids are undissociated.As illustrated in fig. 1 (a), the gradual water uptake of the membrane causes only a slight enhancement of the absorbance in the range 3300-2800 cm-l. This weak increment probably results mainly from the swelling of the film. It is remarkable, however, that the hydration process gives rise, from the outset, to a shoulder at ca. 3600 cm-l and, at a larger water content (corresponding to the gain of approximately two water molecules per site), to a second shoulder at ca. 3400 cm-l. According to our previous investigations, the shoulders at 3600 and 3400 cm-l can be attributed to the stretching vibration mode of, respectively, the free OH groups and OH groups involved in hydrogen bonds linking together water molecules.l9 ' As regards theL.Y. LEVY, A. JENARD AND H. D. HURWITZ TABLE 2.-PRINCIPAL SPECTRAL PROPERTIES OF MEMBRANES 33 shoulder at capacity time of 3400 3600 band A band B vc=o membrane /meq g-l drying/h v,,/cm-l cm-l cm-l /ern-' /cm-l /cm-l 12 A 1.28 1 0 3102 yes yes 2672 weak 1720 2 2 3100 yes yes 2665 weak 1711 3 24 3104 weak yes 2662 1978 1706 4 48 3106 no no 2660 1980 1702 11 A 0.98 1 0 3142 no yes 2690 weak 2 3 3145 no yes 2684 weak 3 24 3146 no no 2681 1972 4 52 3143 no no 2680 1972 13 A 0.6 1 0 3173 no yes 2700 weak 2 2 3169 no no 2696 weak 3 24 3171 no no 2696 weak 716 709 707 706 715 710 708 4 48 3169 no no 2694 weak 1708 14 0.7 1 0 3127 yes yes 2680 weak 1716 2 4 3129 yes no 2679 weak 1709 3 24 3129 weak no 2675 weak 1708 4 48 3132 no no 2675 weak 1708 01 1800 1700 1( wavenumberlcm -' FIG.2,-Infrared spectra of the vc=o band of a FEP-PAA membrane (1 1 A) after various times of drying. (-) Initial hydrated state; (---) 2 h; (. * . 3) 24 h.34 I.R. INVESTIGATION OF ACID MEMBRANES position of the absorption band at ca. 3150 cm-l, it can thus be assigned almost exclusively to the hydroxy groups vibration vOH pertaining to the carboxylic acid. The lowering of the stretching vibration frequencies vOH and vc=o are an indication of the strength of the hydrogen-bond interaction. The data recorded in table 2 show that the humidity of the membrane does not affect the wavenumber of the absorption maximum of vOH.This makes it clear that all OH groups of the carboxylic acid are involved in hydrogen bonding, whatever the degree of humidity. As observed in table 2, in contrast to the behaviour of the vOH absorption band, the vcz0 band shifts abruptly towards smaller wavenumbers from the beginning of the dehydration process (fig. 2). Such a change reveals a strengthening of the hydrogen bonds acting on the >C=O oscillator. With regard to this shift of vco, the following is already known from extensive experimental and theoretical investigations :lo, 20-22 the v c=o vibration mode of a carboxylic group is more strongly affected by hydrogen bonding with an OH group pertaining to another carboxylic group than to water. For instance, in acrylic acid the absorption band of free C=O is found at 1725 cm-l, that of C=O involved in direct dimerisation of carboxylic groups at 1660 cm-l and that of C=O forming a dimeric structure mediated through hydrogen bonding with one water molecule per carboxylic group at ca.1690 cm-l (see, for example, the structure appearing on the right of case I in fig. 2).209 21 In the Raman spectrum of a perfectly dry poly(acry1ic acid), Simon et aI.l0 observe a broadening of the band at 1660 cm-l and a shift of the band at 1725 cm-l towards 1740 cm-l. In an aqueous solution of 80 % polyacrylic acid, the bands at 1660 and 1690 cm-l transform into a broad band extending between 1670 and 1740 cm-l.l0 In the i.r. spectrum of an aqueous solution of poly(acry1ic acid), this band is found at ca. 1723 cm-l.l0 On the grounds of these results, we believe that the low values of vcz0 frequencies ranging from 1702 to 1708 cm-l, as displayed by the membranes at their lowest degree of humidity, suggest the formation of cyclic acid dimers without the involvement of water molecules.The low values of frequencies might be explained on the grounds that the hydrophobic matrix restrains the movement of the carboxy groups and thereby gives rise to stronger contacts at the intermolecular level. The sharpening displayed by bands A and B and the band at 940 cm-l as a function of drying provides further evidence for the existence of a carboxylic acid dimer configuration. The consideration developed above allows us to propose fig. 3 as a model of the distribution of intermolecular bonds.Structure (111) illustrates the presence of an extensive network of interchain contacts via hydrogen bonds. The passage from structure (111) to structure (11) is consistent with the observation of a sudden increase of the absorption shoulder at 3600 cm-l and the fact that the position of the vOH band remains nearly unaffected by the degree of humidity. The transition from structure (11) to structure (I) demonstrates the tendency of the carboxylic dimers to open and accept hydrogen-bonded water molecules. This model agrees with the appearance of the shoulder at 3400 cm-l, with the weakening of bands A and B and the band at 940 cm-l and with the increase of the vce0 band frequencies. The latter increase is also found in the above-mentioned Raman spectroscopic observation on acrylic and polymeric acids.It is also inferred from the evolution of bands A and B and the vco and vOH absorption bands recorded in table 2 that the hydrogen bonds linking together the carboxylic groups in the dimer configuration are getting stronger as the exchange capacity of the membrane increases. The reason for this tendency is found in the fact that the dipolar interaction between polarizable bonds is more pronounced in the case of a large graft density, which causes a reinforcement of the hydrogen bonds. The aggregation of carboxylic groups in networks of very strong interchain contact and cyclic dimerisation is thus promoted. The membranes 13A and 14 both possess anFIG. 3.-Model L. Y. LEVY, A. JENARD AND H. D. HURWITZ 0 H, I of distribution in intermolecular f--", H ,O ....H- O..-.- H-0, 0-H ..... 0 - H .... 0 ,R -R-C, *C - R .c\o H' ;I bonds as a function of water content of 35 membrane. equivalent capacity but differ in their Teflon matrix. The results of table 2 suggest that the existence of a lateral -CF, group in the FEP matrix exerts a steric hindrance which significantly weakens the hydrogen bonds linking two carboxylic groups. With respect to this observation, it shows that, whatever the nature of the matrix or the graft density, it is essentially vOH, and thus the hydrogen-bond donor property of the carboxylic group, which is affected. CONCLUSIONS The information resulting from the i.r. spectroscopic investigation has lead to an understanding of some of the properties of the carboxy-containing ion exchangers.The building-up of a network of hydrogen bonds crosslinking the polymeric chains affords some flexibility of the PAA polymers between their graft sites on the Teflon matrix. Moreover, the carboxylic groups must possess some rotational freedom in order to orient themselves favourably for hydrogen bonding. The dimer association of the acid groups c o n k s to the system a rigidity prohibiting, for example, the out-of-plane wagging vibration mode of the CH, groups which absorb at36 I.R. INVESTIGATION OF ACID MEMBRANES ca. 1330 cm-l. The presence of the extensive lattice of hydrogen bonds hinders the diffusion of water in the membrane and the hydration of the membrane occurs in such a way as to affect the intermolecular association of the polymeric chains to only a very slight degree.This restraint explains the small water uptake of the membrane. Since the carboxylic resins show an equivalent small degree of swelling, one might suppose that hydrogen bonding of the functional groups is an essential characteristic of solid carboxy-containing It is thus understandable that any ionic exchange can happen only in as much as it destroys this network of hydrogen bonds. Consequently, the selectivity of the ion exchangers towards the various counter-ions is usually weak compared with their affinity for the hydrogen ion. This point will be considered in a forthcoming publication. We thank the ‘Fonds National de la Recherche Fondamentale Collective’ and the ‘Fonds National de la Recherche Scientifique’ of Belgium for having supported this work.L. Y. Levy, A. Jenard and H. D. Hurwitz, J. Chem. Soc., Faraday Trans. I , 1980, 76, 2558. L. Y. Levy, M. Muzzi and H. D. Hurwitz, J. Chem. Soc., Faraday Trans. 1, 1982, 78, 17. V. I. Soldatov and L. V. Novitskaya, Russ. J. Phys. Chem., 1965, 39, 1453. H. P. Gregor, L. B. Luttinger and E. M. Loebl, J. Phys. Chem., 1965, 59, 34. L. Y. Levy, A. Jenard and H. D. Hurwitz, Anal. Chim. Acta, 1977, 88, 377. L. Y. Levy, Ph.D. Thesis (Universite Libre Brussels, 1979). R. Kiss-Eross, Analytical Infrared Spectroscopy (Elsevier, Amsterdam, 1976), vol. 6. E. P. Otocka and T. K. Kwei, Macromolecules, 1968, 1, 244. * W. J. MacKnight, L. W. MacKenna, B. E. Read and R. S. Stein, J. Phys. Chem., 1968, 72, 1122. lo A. Simon, M. Mucklich, D. Kunath and C. Heinz, J. Polym. Sci., 1958, 30, 201. l 1 J. De Villepin and A. Novak, Spectrochem. Acta, Part A, 1971, 27, 1259. l 2 D. Hadzi and N. Kobilarov, J. Chem. Soc. A, 1966, 439. l 3 D. Hadzi and S. Bratos, in The Hydrogen Bond, ed. P. Schuster, G. Zundel and C. Sandorfy (North Holland, Amsterdam, 1976), vol. 11, p. 568. l 4 N. Sheppard, in Hydrogen Bonding, ed. D. Hadzi (Pergamon Press, 1959). l5 A. I. Grigorev, Zh. Neorg. Khim., 1963, 8, 802. l6 J. F. Rabolt and B. Fanconi, Macromolecules, 1978, 11, 740. M. J. Hannon, F. J. Boerio and J. K. Koenig, J. Chem. Phys., 1969, 50, 2829. G. Zundel and E. G. Weidemann, Trans. Faraday SOC., 1970, 66, 1941. G. Zundel, Hydration and Intermolecular Interaction (Academic Press, New York, 1969). 2o S. Feneant and M. J. Cabannes, C.R. Acad. Sci., 1952, 235, 240. *l S. Feneant and M. J. Cabannes, C.R. Acad. Sci., 1952, 235, 1292. 22 L. Bardet, G. Cassanas-Fabre and M. Alain, J. Mol. Struct., 1975, 24, 153. 23 M. G. Marina, Yu. B. Monakov and S. R. Rafikov, Russ. Chem. Rev., 1979, 48, 389. (PAPER O / 1376)

ISSN:0300-9599    

DOI:10.1039/F19827800029

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I, 1982, 78, 37-42 Solute-Solvent Interactions in Water + t-Butyl Alcohol Mixtures Part 1 1 .-Enthalpies of Transfer of Ammonium and Tetra-alkylammonium Salts BY JEAN JUILLARD Laboratoire d’Etude des Interactions Solutis-Solvants, Universite de Clermont 11, BP 45, 63170 Aubiere, France Received 29th September, 1980 Enthalpies of solution of tetra-alkylammonium halides (from ammonium to tetrapentylammonium) in Enthalpies of transfer thus obtained are discussed in terms of solvent structure and solute-solvent mixtures of water and t-butyl alcohol (from 0 to 40% by weight) are reported. interactions. There have been many studies of solute properties in water + t-butyl alcohol (TBA) mixtures. As the largest alcohol which is entirely miscible with water, TBA added to water induces some striking features.Most of them have been attributed to a strong enhancement of the water network when the first amounts of TBA are added. It is accepted by many authors that this increase reaches a maximum and that the so-called ‘ structuration maximum ’ corresponds to extrema in the curves of the thermodynamic properties of the solutes. This whole question has been amply discussed in previous papers of this series.’ As far as enthalpies of transfer are concerned many types of solute have been studied. All of these enthalpies, except those corresponding to tetrabutylammonium bromide, studied by Alhuwalia,2 go through a maximum when successive fractions of TBA are added to water. This unusual behaviour prompted us to study the whole series of the tetra-alkylammonium bromides.EXPERIMENTAL t-Butyl alcohol and water were purified as previously stated.l Ammonium chloride and tetra-alkylammonium bromides were pure commercial products (Fluka puriss) used as received. Tetrapentylammonium bromide was kindly provided by G. Perron and J. E. Desnoyers. Heats of solution were obtained using an LKB solution calorimeter following a procedure already described. RESULTS The standard molal enthalpies of solution for ammonium chloride, tetramethyl-, tetraethyl-, tetrapropyl-, tetrabutyl- and tetrapentyl-ammonium bromides are reported in table 1. From the previously reported4 values of pK, of the ammonium ion it can be ascertained that there is no significant hydrolysis of the ammonium chloride under the experimental conditions of the calorimetric measurements.Data supplied by Alhuwalia et aZ.2+ were available for tetrabutyl- and tetrapentyl-ammonium bromides; for these two salts our investigations were restricted first to four mixtures in order to compare both our methods and results with theirs. There was excellent agreement between our results and those of Alhuwalia et al. for NBu,Br. The same was not true 3738 INTERACTIONS IN WATER -I- t-BUOH MIXTURES TABLE 1 .-ENTHALPIES OF SOLUTION OF TETRA-ALKYLAMMONIUM BROMIDES AND AMMONIUM CHLORIDE weight ma AHsb AHFC AH? % TBA mol kg-' /kJ mol-l /kJ mo1-I /kJ mo1-I 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 0 10 20 30 40 0 5 10 15 20 25 30 40 tetramethylammonium bromide 1.1 24.53 f 0.05 24.4 1.3 25.96 f 0.04 25.8 1.3 27.53 k0.04 27.3 1.3 28.82 f 0.02 28.5 1.2 29.61 f 0.03 29.3 1.2 29.87 k0.03 29.5 1.4 29.86 f 0.04 29.3 1.3 29.73 & 0.05 29.1 1.3 29.45 f 0.04 28.7 tetraethylammonium bromide 1 .o 6.30 f 0.01 6.1 1 .o 8.53 k0.03 8.3 1.2 1 1.32 f 0.01 11.1 1.1 14.04 f 0.03 13.8 1 .o 15.73 fO.01 15.4 1 .o 16.47 f 0.02 16.1 1.2 16.97 & 0.02 16.5 1.2 17.25 & 0.01 16.7 1.1 17.49 & 0.02 16.8 tetrapropylammonium bromide 1.1 - 4.29 f 0.01 -4.5 1.1 0.00 & 0.02 - 0.2 1.1 5.61 fO.01 5.4 1.1 12.21 fO.01 11.9 1.2 16.95 f 0.04 16.6 1.1 19.08 f0.02 18.7 1.2 20.30 f 0.03 19.8 1.2 21.14 & 0.03 20.6 1.2 22.20 f 0.02 21.5 tetrabutylammonium bromide 0.75 8.32 f 0.02 8.1 0.70 3 1.36 f 0.03 31.1 0.74 35.35 f 0.03 34.9 0.8 1 35.95 f 0.02 35.3 tetrapentylammonium bromide 0.55 3.83 +O.Ol 3.8 0.53 14.57 k0.03 14.4 0.53 3 1.02 f 0.02 30.8 0.50 52.26 f 0.06 52.0 0.50 68.78 f 0.16 68.5 0.47 68.30 f 0.05 68.0 0.53 64.63 & 0.09 64.3 0.53 59.20 0.10 58.7 0.85 -8.31 fO.O1 - 8.5 1.4 2.9 4.1 4.9 5.1 4.9 4.7 4.3 2.2 5.0 7.7 9.3 10.0 10.4 10.6 10.7 4.3 9.9 16.4 21.1 23.2 24.3 25.1 26.0 16.6 39.6 43.4 43.8 10.6 27.0 48.2 64.7 64.2 60.5 54.9J.JUILLARD 39 TABLE 1. (cont.) ~~ ~~ weight ma AHsb AH?' AH? % TBA mol kg-l /kJ mol-l /kJ mol-l /kJ mol-1 0 5 10 15 20 25 30 35 40 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .O 1 .o ammonium chloride 15.01 15.89 16.85 17.93 19.06 19.24 19.01 18.57 17.94 14.85 15.69 0.84 16.61 1.76 17.65 2.8 1 18.73 3.88 18.85 4.05 18.54 3.70 18.02 3.17 17.28 2.44 a Approximate mean concentration of the various measurements; * mean value of 2 or 3 maximum uncertainty 0.1 (except for NH,Cl, measurements with maximum uncertainty ; 0.02).I 8o t I 1 I I 0.05 0.10 0.15 X FIG. 1 .-Standard molal enthalpies of solution of tetrabutylammonium bromide (lower curve) and tetra-n-pentylammonium bromide (upper curve) plotted against mole fraction of TBA in the water +TBA mixtures ( x , our data; 0, data of Alhuwalia et U I . ~ . ~ ) ) .40 INTER ACTIONS I N WATER + t-BUOH MIXTURES " I --- 0 OlO 5 0:10 X FIG. 2.-Standard molal enthalpies of transfer of 0, ammonium; 0, tetramethylammonium; ., tetraethylammonium ; A, tetrapropylammonium; V, tetrabutylammonium; and 0, tetrapentylammonium bromides from water to water + TBA mixtures plotted against TBA mole fraction.for NPen,Br as is shown in fig. 1, and other data have therefore been recorded in order to confirm our results. The standard molal enthalpies AH? are obtained from the molal enthalpies AHs at finite concentration (also given in table 1) corrected for the heat of dilution. This last term is calculated using the Debye-Hiickel extended law as applied to enthalpies in the form we have already proposed.6 From these values of standard molal enthalpies of solution, enthalpies of transfer of the various salts from water to the mixtures have been obtained. In order to facilitate a comparison of the ammonium ion with the tetra-alkylammonium ion the values for NH,Br were calculated from the corresponding enthalpies of transfer of NH,Cl, KC16 and KBr.la In fig.2 the enthalpies of transfer for all the bromides are plotted against the mole fraction of t-butyl alcohol in the mixtures. Tetra-alkylammonium bromide-t-butyl alcohol interactions in water have been previously studied by Perron et at.' Such studies require either measurements taken in solution of NR,Br in t-butyl alcohol+water mixtures rich in water (up to mTBA = 1) or measurements taken in solutions of t-butyl alcohol in NR,Br+water mixtures rich in water. Both measurements are supposed to give, as the first term in a series, the same electrolyte-non-electrolyte pair-interaction parameter in water. Here the enthalpy of transfer is related to the pair- and triplet-interaction parameters (hNE and h"E, respectively) by7 AHP/rnN = 4hNE +6h"~ mN where m is the molality and where the subscript E stands for the electrolyte and N for the non-electrolyte.Although it was not the purpose of the present study to determine these parameters,J. JUILLARD 41 as they were already known, the coefficient h,, can be obtained from the slope of the tangent at the origin of the curve AH,e =f(m,). The values thus obtained are in good agreement with the values reported by Perron et al.:7 for tetramethyl- ammonium bromide, 500 as compared with 480 ; for tetraethylammonium bromide, 800 as compared with 808 ; and for tetrapropylammonium bromide, 1500 as compared with 1420 (all coefficients in J mo1-2 kg). Such a determination is not possible for tetrabutylammonium and tetra- pentylammonium bromides owing to the rapid deviation from the limiting slope which is observed for these two electrolytes; however, the results of Perron et ai., h,, = 2340 and 3280 J mo1-2 kg, respectively, are compatible with our data.In addition, the value 290 J moF2 kg is obtained here for the parameter h,, corresponding to the value for ammonium chloride with an acceptable accuracy (k 5). This value is slightly less than the value reported for LiC17 (302-3 12). DISCUSSION Fig. 2 shows that the enthalpy of transfer is positive for all the electrolytes studied here and increases with the size of the cation. In media rich in water enthalpic pair-interaction parameters h, are representative of the interactions between both ions (bromide and tetra-alkylammonium) and t-butyl alcohol in water. Their meaning and interpretation have been amply discussed by Perron et Most of the thermodynamic interaction parameters of the species studied here can be predicted using the scaled-particle theory (s.p.t.) in so far as the term corresponding to the formation of the cavity, calculated according to this theory, is the leading one.In the case of both g N E and h,, the interaction terms are frequently important. In any case, for such large ions as the tetra-alkylammonium cation it may be expected that most of the heat of transfer should be due to the change in the heat of cavity formation. Such calculations have been attempted by Desrosiers and Desnoyers for Bu,NBr.s They give a correct estimation of the trend of the process, at least in water-rich media. The increase observed here in the hNE parameters with the size of the cation is also in agreement with such a picture. Calculations have been made for all these cations.They show that results are very sensitive to the choice of hard-sphere diameter for t-butyl alcohol since unfortunately this diameter is not known for certain8 There is at least a qualitative agreement between the variations in calculated and experimental h,, parameters and it can be ascertained that much of the enthalpic effect observed in the transfer of tetra-alkyl ions from water to t-butyl alcohol mixtures is related to the cavity effect and thus increases with size. One question remains, namely whether the description of water + TBA media as mixtures of hard spheres is realistic, even if the parameters used in the calculations (density and expansibility) are those of the bulk.These mixtures are frequently depicted as highly structured and in the previous papers of this series, we have favoured an interpretation of our former data using the 'fluctuating-cage In this model it is supposed that the water + t-butyl alcohol mixtures retain something of the structure of the solid clathrate. Recent investigations by Iwasaki and Fujiyama'O using light-scattering techniques support the picture of the mixtures as given by this model. As far as our data on enthalpies of transfer of solutes3* 6 + l 1 l 3 are concerned, a maximum is always observed in the curves of the enthalpies plotted against the concentration of t-butyl alcohol; the location as well as the amplitude of the maximum could generally be explained within the scope of the 'fluctuating-cage model ' .1 1 9 l3 Here, except for the two extremes in size, ammonium and tetrapentylammonium42 INTERACTIONS I N WATER+t-BUOH MIXTURES bromide, there is no maximum in the AH, = f ( x ) curves. However, it may be that tetra-alkylammonium ions are as able as t-butyl alcohol to form clathrate-like structures with water and that there is some sort of competition between the destruction of the water clathrate of TBA, which is supposed to be endothermic, and the build-up of the water clathrate of NR:, which can be supposed exothermic. If such assumptions are retained there is no way of predicting apriuri the shape of the curves which represent the variation of the enthalpy of transfer with the TBA concentration.At most it can be noted that for NPen,Br the first effect would be the dominant one. Thus, arguments in favour of or against the quasi-clathrate model do not arise from these results but such a model can be reconciled with our present findings. A more thorough discussion would involve distinguishing between the effects of the two ions, tetra-alkylammonium and bromide, in the enthalpy of transfer of these electrolytes; this will be presented in the following paper. (a) Part 9, Y. Pointud and J. Juillard, J. Chem. SOC., Faraday Trans. I , 1977, 73, 1907. (b) Part 10, N. Dollet, J. Juillard and R. Zana, J. Solution Chem., 1980, 9, 827. R. K. Mohanty, T. S. Sarma, S. Subramanian and J. C. Ahluwalia, Trans. Furaday Soc., 1971, 67, 305. L. Avedikian, J. Juillard, J-P. Morel and M. Ducros, Thermochim. Acta, 1973, 6, 283. Y. Pointud, H. Gillet and J. Juillard, Tulunta, 1973, 23, 741. R. K. Mohanty, S. Sunder and J. C. Ahluwalia, J. Phys. Chem., 1973, 76, 577. Y. Pointud, J. Juillard, L. Avedikian, J-P. Morel and M. Ducros, Thermochim. Acta, 1974, 8, 423. ' G. Perron, D. Joly, J-E. Desnoyers, L. Avedikian and J-P. Morel, Can. J. Chem., 1978, 56, 552. N. Desrosiers and J. E. Desnoyers, Can. J. Chem., 1976, 54, 3800. * E. K. Baumgartner and G. Atkinson, J. Phys. Chem., 1971, 75, 2236. lo K. Iwasaki and T. Fujiyama, J. Phys. Chem., 1977, 81, 1908. l2 Y. Pointud, J-P. Morel and J. Juillard, J. Phys. Chem., 1976, 80, 381. l3 Y. Pointud and J. Juillard, J. Chem. Soc., Furaday Trans. I , 1977, 73, 1907. N. Dollet and J. Juillard, J. Solution Chem., 1976, 5, 77. (PAPER O / 1482)

ISSN:0300-9599    

DOI:10.1039/F19827800037

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1982, 78, 43-52 Solute-Solvent Interactions in Water + t-Butyl Alcohol Mixtures Part 12.-Single-ion Enthalpies of Transfer using the Tetraphenylarsonium Te trap hen yl bora te Assump ti on BY JEAN JUILLARD Laboratoire d’Etude des Interactions Solutes-Solvants, Universite de Clermont 11, B.P. 45, 63170 Aubiere, France Received 29th September, 1980 Heats of solution of tetraphenylarsonium chloride (4,AsCl) and sodium tetraphenylboride (NaB4,) in water and water+t-butyl alcohol mixtures (from 0 to 40% by weight) have been measured. From these data and the previous values for NaCl, enthalpies of transfer of tetraphenylarsonium tetraphenylboride are estimated. Using the classical assumption of equivalent change of solvation with solvent media of the two tetra-aryl ions, ionic enthalpies of transfer are estimated for various cations and anions.This reveals a completely distinct behaviour for the two types of ions. These results are discussed in terms of the structure of the water + t-butyl alcohol mixtures. In previous papers we have reported data for some thermodynamic properties of transfer of various solutes from water to an exemplary series of water + organic media, the water + t-butyl alcohol mixtures. The interpretation of the experimental results is made difficult because each species involved in the process of solvation cannot be considered separately. Therefore it was decided to carry out a splitting into ionic contributions using various extra thermodynamic assumptions. A review of the various processes used in such a splitting into individual ionic transfer and solvation properties has been recently provided by C0nway.l Various methods can be used: (1) Extrapolation of salt values, with respect to ionic radius or ionic volume (or mass), to zero ion size or zero reciprocal radius.Such a method used data for homologous series of salts having either a common anion or a common cation. (2) Use of the assumption that an ion and a neutral molecule similar in size, shape and surface have the same properties of transfer. (3) Splitting the value for a salt composed of ions of similar radius and shape into its ionic component values. Such a method is considered to be reliable for large and chemically similar cations and anions in a salt, e.g. d4P+B4; or b4As+B&.As far as the first type of splitting method is considered, many types of extrapolation were applied to the present problem using both alkali metal halides and tetra- alkylammonium halides. None of them corresponds to linear plots or gives consistent results. In the second type of method some hope was placed in the idea that the enthalpies of transfer of both the cryptates of alkali ions and the corresponding free ligands could be considered as identical. However, it has been shown recently by Abraham,2 who studied enthalpies of transfer from water to methanol in such systems, that this assumption is not relevant. Finally a method of the third type, which is now considered to give more reliable results, was used : data concerning enthalpies of transfer of tetraphenylarsonium 4344 INTERACTIONS I N WATER + t-BUOH MIXTURES tetraphenylboride (B#,#,As+) and their use for obtaining ionic enthalpies of transfer for numerous cations and anions in water+TBA mixtures are reported here.Tetraphenylboride tetraphenylphosphonium could have been considered as well but it has been proved in some cases that the two assumptions are eq~ivalent.~ Such separations using tetra-aryl ions have been criticised by Treiner, on the basis of ' scaled-particle theory' calculations of the Gibbs free energy of transfer. However, as emphasised by Abraham,5 Treiner's criticisms result from a misunderstanding of the nature of this assumption. Some data were already available on enthalpies of transfer of tetraphenylarsonium tetraphenylboride (#4A~B$4) in water + TBA media.These enthalpies were obtained from measurements of the heat of precipitation of this electrolyte in various mixtures by Bright and Jezorek.6 They can also be calculated using data reported by Arnett'? for sodium tetraphenylborate (NaB4,) and tetraphenylarsonium chloride (4,AsCl). Preliminary calculations made from these two sets of data give such surprising results for the enthalpies of transfer of anions and cations that it appears necessary to confirm these results before discussing further what can be deduced from them. Therefore, in order to check and extend these results to mixtures with higher alcohol concentrations and to specify the location of the maximum, we measured the heat of solution of NaB4, and #,AsCl in water+TBA mixtures from 0 to 40% by weight. From the additivity of ionic contributions to the enthalpies of transfer, AHt of #,AsB#, can be obtained and the ionic enthalpy of transfer can thus be estimated for various cations and anions from the AH, values of electrolytes previously reported.9-16 EXPERIMENTAL PART Sodium tetraphenylborate and tetraphenylarsonium chloride were pure commercial products (Fluka puriss and pro analysis) used as received. Purification of solvents and calorimetric measurement procedures have already been described.l0 RESULTS Results obtained for the heats of solution of NaB#, and #,AsCl are given in table 1. The standard enthalpies of solution were obtained using the Debye-Huckel calculation of the heat of dilution as previously de~cribed.~ Fig.1 shows clearly that TABLE 1 .-ENTHALPIES OF SOLUTION OF SODIUM TETRAPHENY LBORIDE AND TETRAPHENYLARSONIUM CHLORIDE FROM WATER TO WATER+t-BUTYL ALCOHOL MIXTURES AT 25 "c XU Xb NaBq4," q3,AsCI" 0 5 10 15 20 25 30 35 40 0.0000 0.01 26 0.0263 0.041 1 0.0573 0.0749 0.0943 0.1157 0.1394 - 18.59 8.82 -8.10 21.64 +6.16 36.97 f41.51 50.40 + 49.34 47.96 + 17.03 37.18 + 9.24 - + 4.40 34.15 + 30.53 - a Weight % of alcohol in the mixture; mole fraction of alcohol in the mixture; AH? in kJ mo1-I (mean uncertainty k0.05).J . JUILLARD 45 A 90 80 - 70 - h 60- . 5 0 - - h 24 3 40 - 30 - 20 - FIG. 1.-Enthalpies of transfer from water to waterfTBA mixtures (mole fraction x of TBA) of: (l), tetraphenylarsonium chloride (right-hand scale): A, our values; A, corrected values of Arnett’ (see text) ; (2) sodium tetraphenylborate (left-hand scale); a, our values; 0, from Arnett and McKelvey;* (3) tetraphenylarsonium tetraphenylborate (left-hand scale) : 0 , calculated from our measurements [eqn (I)] ; 0, from heat of precipitation (Bright and Jezorek); 0, calculated from Arnett’s data [eqn (2)].data obtained by previous authors61 are roughly in agreement with ours. Values of enthalpies of transfer of tetraphenylarsonium tetraphenylboride from water to water+TBA mixtures have thus been confirmed and specified. For NaB4, there is a good agreement between our values and those of Arnett, as shown in fig. 1 , in which both sets of enthalpies of transfer are compared. For #,AsCl, Arnett’s values estimated from fig. 1 of ref. (7) and corrected for the transfer of H,017 are slightly lower than ours. Owing to the uncertainty of such a procedure, the discrepancies observed can be considered as acceptable.Enthalpies of transfer of #,AsB4, calculated from our data and eqn (1) are compared in fig. 1 with the value obtained by Bright and Jezorek6 from enthalpies of precipitation and also with the values calculated using eqn (2) and the data of Arnett and McKelvey7~* as proposed by Holterman and Eng1e~ts.l~ Enthalpies of transfer of NaCl and H,O used in the calculations are taken, respectively, from ref. (9) and (16): AHt(#,Asq5,B) = AH,(Na#,B) + AHt(q5,AsCl) - AH,(NaCl) (1)46 INTERACTIONS I N WATER + t-BUOH MIXTURES AHt(4,As4,B) = AH,(Nad,B) + AHt(4,AsC1, 2H,O) - 2AHt(H,0) - AH,(NaCl). (2) A mean curve was drawn using both Jezorek’s data and ours, and the enthalpies of transfer of the two ions were calculated from eqn (3): Another procedure could have been used (for example, considering only our data), but in any case a separation of ionic contributions should not give exact values but only show any trend in variations of the ionic enthalpies of transfer.From the C1- and Na+ contributions thus obtained, data of ref. (9) give the enthalpies of transfer of K+, Rb+ and Cs+, data of ref. (12) give H+ and Li+, and data of ref. (14) give Br-, I- and C10;. The values for NO; are obtained from Pointud’s unpublished results for AH, of KNO,. The enthalpies of transfer of OH- are obtained from the values previously calculated for H+ and from the dissociation enthalpy and partial molal enthalpy of water in water + TBA media.16 AHt for substituted benzoates are deduced from data reported in ref.(13) and AHt for acetate, propionate and isobutyrate are obtained from AH, of acetic, propionic and isobutyric acidslo combined with the changes in the enthalpies of dissociation of these acids.ll Finally, ammonium and tetra-alkylammonium ionic enthalpies of transfer are obtained from AH, of their halides reported in ref. (15). All these results are presented in fig. 2-5. There are no other determinations of individual ionic enthalpies of transfer from water to water+ TBA mixtures in the literature. However, some arguments confirming the trend in the variations observed here can be based on recent determinations by Gilletl8 -15 - 10 - 5 - 0 0.05 0.10 FIG.2.-Enthalpies of transfer of some inorganic monovalent cations from water to water + TBA mixtures (mole fraction x of TBA): dashed line, Na+; full lines, (1) H+, (2) NH: (left-hand scale), (3) Li+, (4) Cs+, (5) Rb+, (6) K+ (right-hand scale).J. JUILLARD 47 5 t 0.05 0.10 X FIG. 3.-Enthalpies of transfer of some inorganic monovalent anions: dashed lined, Br-; full lines, ( I ) I-, (2) ClO;, (3) NO;, (4) C1-, (5) OH-. 0.05 0.10 FIG. 4.-Enthalpies of transfer of some carboxylate ions: (1) acetate, (2) propionate, (3) isobutyrate, (4) benzoate, (5) p-iodobenzoate.48 INTER A C T I 0 N S I N WATER + t-BUOH MIXTURES 40 30 20 t o 1 I a 0.05 0.10 FIG. 5.-Enthalpies of transfer of tetra-alkyl and tetra-aryl ions: dashed line, TPB- and TPAs+; full lines, (1) NH:, (2) NMe:, (3) NEt:, (4) NPra, (5) NB:, (6) NPen:.of the enthalpies of acid dissociation of anilinium and pyridinium ions in these mixtures. It is frequently accepted that variations in the solvation of anilines and corresponding anilinium cations with the nature of the solvent media are analogous. Such an assumption is at the root of the use of ' Hammet indicators' in defining acidity functions. This means that changes in enthalpy of dissociation of anilinium would be roughly equal to the enthalpy of transfer of H+. A comparison of enthalpies estimated in this way and enthalpies of transfer obtained here is comforting. Even if the values estimated from treatment of anilinium dissociation data are lower than values obtained here, the same trend can be observed (a maximum in AHt for x z 0.05 but of ca.7 as compared with 15 kJ mol-l). DISCUSSION The results obtained are, on the whole, quite surprising. The most striking fact is the very distinct behaviour on the one hand of inorganic cations and tetra- alkylammonium ions and on the other hand of inorganic anions and even of small organic anions. In all the curves of transfer enthalpies as a function of composition, three parts can be distinguished which probably correspond to zones of distinct structure types for the water + t-butyl alcohol mixtures. Zone 1 : For mole fractions from 0 to 0.03 the enthalpy changes are frequently weak and generally positive, both for cations and anions. This zone is very rich in water; the water structure is frequently described as being enhanced by the presence of the alcohol.Zone 2: For mole fractions from 0.03 to 0.08 notable variations in the enthalpies of transfer can be observed, underlined by the existence of large extrema (maximumJ. JUILLARD 49 or minimum) which are located at an alcohol mole fraction of ca. 0.05; according to the charge, the enthalpy changes with content are very different. There is an endothermic maximum for the cations and an exothermic minimum for the anions. This range of concentration is frequently described as a region where there is a maximum structuring of the mixtures, which corresponds to the formation of an original water + alcohol structure. Zone 3 : For mole fractions from 0.08 to 0.15 there can again be observed a distinct type of behaviour according to the charge of the inorganic ion.When the alcohol concentration increases the enthalpy of transfer becomes more and more exothermic for the cations and more and more endothermic for the anions, except for those which have a hydrophobic radical. In this region there is an increase in free molecules or aggregates of TBA. STRUCTURE OF THE WATER+TBA MIXTURES The structure of water + alcohol and specially of water + t-butyl alcohol mixtures has been discussed by many a u t h o r ~ , l ~ - ~ ~ on the basis of both thermodynamic and spectroscopic data. Alcohols are solutes which bear both hydrophilic and hydrophobic groups. These latter determine their ‘ structure-promoting ’ characteristic which is of the same type as the one observed with apolar solutes.The ‘structure-promoting’ ability of the alcohols increases with the size of the alkyl group. It is generally accepted that when adding alcohol to water there is a reinforcement in the stability of the ‘clusters’. The real type of the structure itself is not yet well-known. However, numerous arguments have been presented for the formation, as in the case of apolar solutes, of a structure analogous to the solid clathrate, but of a more labile type. This was first proposed by Arnettlg to explain the properties of the t-butyl alcohol mixtures and formalised by A t k i n ~ o n ~ ~ in terms of the fluctuating-cage model. The use of this model has been supported by work carried out recently by Iwasaki and F ~ j i y a m a ~ ~ on concentration fluctuations from light-scattering spectra and by Tamura et a1.26 on ultrasonic absorption of the mixtures.The picture of the structure of the mixtures which results is not in clear contradiction to that previously proposed by Symons and Blandamer.21 If one accepts the idea of a two-state mixture model for water it can be said that the introduction of an alcohol molecule will determine a stabilisation of the ‘clusters’ which corresponds to a gradual organisation into fluctuating cages ofwater surrounding an alcohol molecule. These cages have a well-defined stoichiometry. The water structure, at first little affected (zone l), becomes more and more organised when the alcohol concentration is raised, up to a concentration corresponding to the stoichiometry of the cages (zone 2). This stoichiometry, 1 : 17 for a solid clathrate of type 11, is doubtless slightly different for the solution, taking into account the degree of freedom of the cages one with another; the work of Iwasaki et al.25 suggests a value of 1:21.A solid t-butyl alcohol +water clathrate is not known, but a compound where such a clathrate is stabilised by two molecules of H2S for one of TBA and 17 of water has been described.27 Clathrate hydrates of type 1128 present two types of holes, large and small, able to encage molecules, and their structure is stabilised when the small holes are also occupied. Above a certain composition of the mixtures which corresponds to a maximum formation of the cages and thus to a structuration maximum, two phenomena occur: the apparition of free TBA molecules and the condensation of the cages25 thus forming ‘merged clathrates’21 (zone 3).50 INTERACTIONS IN WATER + t-BUOH MIXTURES VARIATION OF THE ENTHALPY OF TRANSFER WITH ALCOHOL CONTENT FOR THE VARIOUS TYPES OF IONS For all the ions studied here, an extremum in AH, (minimum or more frequently maximum) is obtained for TBA contents between mole fractions x = 0.045 (1 : 22) and x = 0.055 (1 : 18).In fact the location of the maximum is not very accurate, taking into account the various calculations involved in estimating ionic enthalpies of transfer; nevertheless such results permit us to discuss our results in terms of the model of TBA + water mixtures presented above (the model of fluctuating cages) even if the stoichiometry of the cages cannot be determined precisely from our data.The results which concern the cations are plotted in fig. 2 (inorganic monovalent cations) and 5 (tetra-alkylammonium ions). For the first results (for x < 0.03) only small changes in the enthalpy of transfer are observed initially and these can be attributed to the persistence of the types of water-cation interactions which exist in water, the cages appearing in the medium having only one effect, that of restricting the quantity of water able to participate in these interactions. The same effect is observed for the first tetra-alkylammonium ions; however, as the size increases, adding t-butyl alcohol to water causes an increasingly rapid variation in the enthalpy of transfer. We underlined in the previous paper (after Desrosiers and Desn~yers~~) that the enthalpy corresponding to the formation of the cavity, according to the size of the ion, is sufficient to explain the magnitude of the effect observed.Whatever the type of cation, a large increase in the enthalpy of transfer is observed above x = 0.03. This enthalpy goes through an endothermic maximum for a composition of the mixture which corresponds roughly to the stoichiometry of the ' pseudo-clathrate'. For the tetra-alkylammonium cations the amplitude of the maximum is a function of the size of the cation. It can be suggested that the introduction of such cations can only be carried out by the (at least partial) destruction of the local structure, this destruction being more and more pronounced as the size of the cation increases.This effect would be strongest for an alcohol content which corresponds to the highest structuring, i.e. to the greatest cage density. For a different reason H+, NH; and the alkali metal cations would also, as a consequence of their ability to attract water molecules, disrupt the structure of the mixtures where most of the water molecules are engaged in the formation of the cages. The sequence of the observed effects for the alkali metal cations, Na+ > K+ > Rb+ > Cs+ > NHZ, is in agreement with the electrostrictive power of these ions (as reflected by the difference between aqueous molar volumes and crystallographic ones30). Compared with this series both H+ and Li+ show a slightly anomalous behaviour. It can be supposed that these small ions can occupy an interstitial location in the structure.Beyond this zone of maximal structuration the enthalpies of transfer decrease more rapidly for the inorganic ions than for the tetra-alkylammonium ions when the alcohol is added. This could be due to interactions between cations and the free TBA molecules which present some basicity as indicated by data from Kolthoff and Chantooni for H+ solvation in alcohols.31 A smaller decrease, due allowance being made in the transfer enthalpy for the tetra-alkylammonium ions, could result from the interference of various processes : a decrease with the size of ions in the interactions related to the basicity of TBA, an increase in the hydrophobic interactions between TBA and the tetra-alkylammonium ions, and competition between TBA and tetra-alkylammonium ions in the building of water clathrates.It is not possible to know the respective participation of these various effects in the enthalpy changes observed.J. JUILLARD 51 If we now consider the inorganic anions (fig. 3) and the small organic anions (fig. 4) the variations in the enthalpies of transfer are analogous to what is observed for cations in the same zone, rich as it is in water (zone 1): there are only small endothermic variations whose weakness is probably related to the lack of large changes in the water-anion interactions in this range of composition. These variations in enthalpy nevertheless appear to increase with the size of the organic ion. This can again be attributed to the increase in the enthalpy of cavity formation with the size of the anion.What happens for anions in zone 2 is more difficult to explain. There is here, for all but the larger anions, an enthalpy minimum which also appears at the ‘magic fraction’. Inversion occurs only for anions as large as bromobenzoate. If we again refer to the fluctuating-cage model, it can be said that anions are more easily accommodated in the clathrate-like mixture than in water but increasingly less so as their size increases. It may be suggested that both inorganic and small organic anions can occupy interstitial locations between the cages where they, like H2S in the solid clathrate, strengthen the structure. Both shape and size would probably determine their ability to fill the vacant holes in the fluctuating structure.Influence of size is clear in the case of small organic ions (fig. 5 ) for which enthalpy of transfer from water to the mixture (x = 0.05) is negative, but increasingly smaller when the size of the ions increase; for inorganic ions there is no clear link between the size and the amplitude of the minimum. Except for the larger ions, the variations in AH, observed in zone 3 (x = 0.08-0.15) are again opposed for anions and cations. The increase in AH, when adding alcohol would be related to the weakness of the interactions between anions and free TBA molecules. The fact that such an effect is smaller for nitrate, perchlorate and benzoate ions is probably related to the development of interactions through dispersion forces between TBA molecules and these anions, in which electron delocalisation is high and which are consequently highly solvated in organic media.CONCLUSIONS To sum up, as far as enthalpies of transfer are concerned, the behaviour of cations and anions, at least of the small ones, is very different. This fact, not always easy to explain, might cause some doubts as to the relevance of the extrathermodynamic assumption used here. But it must be mentioned that the ionic molar volume separation realised by Dollet et al. in the same media,32 this time using a rigorous method developed by Zana and ye age^-,^^ leads to the same sort of observation: there is a decrease in the partial molal volumes of the halide anions and an increase in the partial molal volumes of the alkali cations when going from water to the clathrate-like mixture.In addition separation by Wells3* of ionic contributions to Gibbs free energies of transfer, using other assumptions, leads also to a very different effect for cations and anions: AG, values are positive for anions and negative for cations. Our most recent investigations confirm this It must therefore be accepted that anions and cations are not accommodated in the same way in water+t-butyl alcohol mixtures. Considering the formation of a quasi-clathrate fluctuating structure followed by its destruction allows us, more or less, to explain the effect observed in the change of the solvation of ions, as far as enthalpic effects are concerned. This cannot be considered as proof of the validity of such a structural description of the mixtures, but only as an attempt to fit thermodynamic data on the solvation of ions in these media to what seems at the moment to best reflect the structure and interactions in water + t-butyl alcohol mixtures.J. JUILLARD 51 If we now consider the inorganic anions (fig.3) and the small organic anions (fig. 4) the variations in the enthalpies of transfer are analogous to what is observed for cations in the same zone, rich as it is in water (zone 1): there are only small endothermic variations whose weakness is probably related to the lack of large changes in the water-anion interactions in this range of composition. These variations in enthalpy nevertheless appear to increase with the size of the organic ion. This can again be attributed to the increase in the enthalpy of cavity formation with the size of the anion. What happens for anions in zone 2 is more difficult to explain.There is here, for all but the larger anions, an enthalpy minimum which also appears at the ‘magic fraction’. Inversion occurs only for anions as large as bromobenzoate. If we again refer to the fluctuating-cage model, it can be said that anions are more easily accommodated in the clathrate-like mixture than in water but increasingly less so as their size increases. It may be suggested that both inorganic and small organic anions can occupy interstitial locations between the cages where they, like H2S in the solid clathrate, strengthen the structure. Both shape and size would probably determine their ability to fill the vacant holes in the fluctuating structure.Influence of size is clear in the case of small organic ions (fig. 5 ) for which enthalpy of transfer from water to the mixture (x = 0.05) is negative, but increasingly smaller when the size of the ions increase; for inorganic ions there is no clear link between the size and the amplitude of the minimum. Except for the larger ions, the variations in AH, observed in zone 3 (x = 0.08-0.15) are again opposed for anions and cations. The increase in AH, when adding alcohol would be related to the weakness of the interactions between anions and free TBA molecules. The fact that such an effect is smaller for nitrate, perchlorate and benzoate ions is probably related to the development of interactions through dispersion forces between TBA molecules and these anions, in which electron delocalisation is high and which are consequently highly solvated in organic media. CONCLUSIONS To sum up, as far as enthalpies of transfer are concerned, the behaviour of cations and anions, at least of the small ones, is very different. This fact, not always easy to explain, might cause some doubts as to the relevance of the extrathermodynamic assumption used here. But it must be mentioned that the ionic molar volume separation realised by Dollet et al. in the same media,32 this time using a rigorous method developed by Zana and ye age^-,^^ leads to the same sort of observation: there is a decrease in the partial molal volumes of the halide anions and an increase in the partial molal volumes of the alkali cations when going from water to the clathrate-like mixture. In addition separation by Wells3* of ionic contributions to Gibbs free energies of transfer, using other assumptions, leads also to a very different effect for cations and anions: AG, values are positive for anions and negative for cations. Our most recent investigations confirm this It must therefore be accepted that anions and cations are not accommodated in the same way in water+t-butyl alcohol mixtures. Considering the formation of a quasi-clathrate fluctuating structure followed by its destruction allows us, more or less, to explain the effect observed in the change of the solvation of ions, as far as enthalpic effects are concerned. This cannot be considered as proof of the validity of such a structural description of the mixtures, but only as an attempt to fit thermodynamic data on the solvation of ions in these media to what seems at the moment to best reflect the structure and interactions in water + t-butyl alcohol mixtures.

ISSN:0300-9599    

DOI:10.1039/F19827800043

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1982,18, 53-60 Microcatalytic Study of the Depolymerization of 2,4,6-Trimethyl- 1,3,5-trioxan (Paraldehyde) over Mordenite Surfaces BY PAUL JOE CHONG AND GEOFFREY CURTHOYS* Department of Chemistry, University of Newcastle, New South Wales, Australia 2308 Received 8th October, 1980 The catalytic depolymerization of paraldehyde (2,4,6-trimethyl- 1,3,5-trioxan) to acetaldehyde over Na+-, Ca2+- and La3+-exchanged mordenites has been investigated by means of conventional pulse-flow microcatalytic chromatography. The injection port of a gas chromatograph was used as a microcatalytic reactor chamber. The apparent activation energy was found to be 37.0 kJ mol-' for NaM and 16.0 kJ mol-l for CaM and LaM. The mordenite-catalysed depolymerization of paraldehyde fits the Langmuir- Hinshelwood kinetic model, which can be described by the following pathway (see Discussion for symbols) : k+i k+, k+3 (PA),+E + [PA...El* + [AA.. -El* + (AA)g+X k-1 slow fast fast Microcatalytic chromatography was first developed by Kokes et al.' and has been modified since by several investigator^.^.^ In the usual operation of this method, a pulse of reactant is injected into a stream of a carrier gas, passed through a microreactor containing catalyst and then through an analytical column. Bassett and Habgood5 proposed gas-solid microcatalytic chromatography for the kinetic study of first-order surface-catalysed reactions. They used the catalyst for dual functioning, both for the catalytic reaction and for the chromatographic separation. The pulsed version of this method has been theoretically analysed by Roginskii and RozentaP and several Ettre and Brenner8 proposed the attachment of a precolumn microreactor for the pulsed-flow method.Recently Schmiegel et aL9 have reported the use of a gas-chromatography injection port for the microcatalytic determination of the activation energy and Fowler et al.l0 for the quantitative recovery of adsorbed species such as pollutant molecules. In the present work the catalytic properties of the mono(Na+)-, di(Ca2+)- and tri( La3+)-valent cation exchanged mordeni tes have been investigated and the apparent activation energies (E,) were l 1 Since many surface-catalysed reactions are accompanied by side reactions, which complicate the kinetic analysis, depoly- merization of paraldehyde was chosen as a probe reaction.Paraldehyde undergoes clean conversion into a single product species, acetaldehyde. Langer et a1.12 stated that depolymerization of paraldehyde is suitable for testing acidic properties of solid surfaces. EXPERIMENTAL AND RESULTS AnalaR grade paraldehyde or 2,4,6-trimethyl- 1,3,5-trioxan, (CH,CHO),, ex. May and Baker, was used as reactant, the purity being confirmed chromatographically. 5354 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE Synthetic NaM (M = Mordenite), ex. Norton, was used as parent adsorbent. CaM and LaM were prepared from NaM by ion-exchange, using conventional exchange procedures.13 From the elemental analysis of the zeolites1* the unit-cell compositions were determined, and are shown in table 1.For use in the catalytic experiments, the cation-exchanged mordenite powders were compressed into pellets (35-45 mesh). An aliquot (10-12 mg) of the resulting catalyst was loaded into the microcatalytic reactor, and activated initially at 350 O C for 16 h in vacuo and then finally at 500 OC for 0.5 h in an oxygen atmosphere. The amount of catalyst chosen was small enough to prevent prolonged surface holding of reactant-product species, which otherwise causes the elution peaks to be diffuse and broadened? For evaluation of catalytic activity at different temperatures, conventional pulsed-flow microcatalytic chromatography was used with an injection port as a microcatalytic chamber as shown in fig. 1 . 5 t 9* l1 A Packard model 427 gas chromatograph having a Gow-Mac katharometer was used TABLE 1 .-UNIT-CELL COMPOSITIONS OF MORDENITE CATALYSTS13* l4 cation type structural composition ~ ~~~ % cation exchange (from) NaM (Na20),(A1203)4(Si02)4~.3023.2H20 100 (1 mol dm-3 NaC1) CaM (Ca0)3.4dNa2O)O. 54(A1203)4(Si02)41. 3023-2H20 86.6 (0.02 mol dm-3 CaC1,) LaM (La203)1.05(Na20)0.85(A1203~4~si02~41,3023~2H20 78.4 (0.1 mol dmP3 LaCl,) with an analytical column (length 196 cm; i.d. 0.23 cm; 0.d. 0.63 cm) of Carbowax 20 M. At a column temperature of 110 OC and flow rate of Fo = 12.0 cm3 min-l there was no measurable decomposition of paraldehyde during elution. The catalytic effect of the reactor wall was found to be negligible by comparison with blank runs, which were carried out using glass beads. The apparent activation energy, E,, was determined using the following equation :5-7 where Fo is the corrected flow rate of He carrier gas at 273 K (cm min-l), m is the mass of catalyst (g), Xis the fraction of reactant converted, Q is the fraction of reactant remaining unchanged, A' is a pseudo-frequency factor (constant), which includes the Arrhenius pre-exponential factor ( A ) , catalyst void volume ( v ) and integral constant of the van't Hoff equation ( I ) , and the other symbols have their usual meaning. The value of E, can be determined by the measurement of X / Q as a function of the reaction temperature.Fig. 2 shows typical chromatograms of the decomposition of paraldehyde on NaM over the temperature range 413-533 K. Fig. 3 shows the plots of The left-hand side term of eqn (1) is a measure of catalyticP. J .CHONG AND G . CURTHOYS 55 FIG. 1 .-Schematic diagram of pre-column microcatalytic reactor: A, A1 injection port; B, silicone septum (high-temperature durable); C, s/s ring (grooved); D, thermocouple; E, glass-wool plug (preheating zone); F, cartridge heater; G, microcatalytic reactor (etched exterior wall); H, catalyst bed; I, glass-wool plug; J, asbestos insulation; K, reactor-level adjuster; L, ferrule; M, Swagelok; N, g.c. analytical column; P, flow regulator (micro); Q, carrier gas (He); R, wall of g.c. oven cabinet; S, g.c. thermostat bath. In [(F*/m) In (1 + X / Q ) ] against 1 / 7’ for NaM, CaM and LaM. From the slopes and the Y-intercepts the values of Ea and A’ were obtained, as shown in table 2. DISCUSSION First-order surface-catalysed reactions can be investigated quantitatively by pulsed- flow microcatalytic chromatography, provided that the fractional conversion is independent of the input pulse or partial pressure of The chromatograms obtained in the manner as described are quite symmetrical, their integrated peak areas varying with the depolymerization temperature, which permits a quantitative evaluation of the catalytic effect by analysis of the resulting elution peaks.4* l 1 9 l5 As shown in fig.3, the Arrhenius plots were found to be linear over the range of temperature investigated. The Ea value derived for NaM was much higher than those obtained for CaM and LaM, the latter two cases being identical. The values of A’ varied between the mordenite samples but no simple correlation was found.In general,56 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE I II 90 60 30 0 90 60 30 0 retention time/s FIG. 2.-Typical chromatograms of catalytic decomposition of paraldehyde over sodium mordenite surface (10.0 mg) at (A) 240 and (B) 150 O C : (a) paraldehyde, (b) acetaldehyde and (c) air. FO(He) = 12.0 cm3 min-', sample size = 0.5 mm3, chart speed = 10 cm min-', TCD = 350 OC, column temp. = 110 OC. the true activation energy for surface reactions following first-order kinetics can be found from the measurement of retention v01umes.~~ l6 Since a relatively small packing of the catalysts was required for developing satisfactory chromatograms, the present method is not amenable to the determination of the true activation energy.NaM is known to be catalytically inactive for acid-catalysed reactions, while CaM and LaM are active due to the greater polarizing power of the respective exchanged cations.17* l8 The evidence indicates that the depolymerization of paraldehyde is an acid-catalysed reaction, the magnitude of E, being related to the strength of the surface acidity. As the role of a catalyst is to lower E,, it is clear that the polyvalent cation exchanged mordenites are catalytically more effective than the monovalent cation exchanged mordenites.16t In studying the catalytic depolymerization of paraldehyde in static media, Walvekar and Halgerilg stated that only silica/alumina surfaces having acid strength of pK, < -3.0 are catalytically effective and found that the catalytic activity of silica-based oxides was much stronger than that of alumina-based oxides.In this respect the highly siliceous nature of mordenites would be a contributing factor in the decomposition of paraldehyde. For a kinetic analysis in terms of classical models,20-22 the surface-catalysed depolymerization of paraldehyde is assumed to occur in the sequence adsorption- surface-reaction-desorption, tacitly ignoring the influence of diffusion to and from the surface. According to the absolute rate theory the behaviour of adsorbate molecules may be defined either by the Rideal mechanism, which refers to the surface interaction between the adsorbed and the unadsorbed species, or by the Langmuir- Hinshelwood model, which refers to the catalytic interaction proceeding only between adsorbed species.Both hypotheses may be operative for the depolymerization of paraldehyde in static media, but their validity in zeolite cavities is subjected to certain geometric restrictions.17* 239.0 - h M % + 8 . 0 1 - v E: - h E 1 5 Y E: - 7.0 6.C P. J. CHONG AND G. CURTHOYS 1 I 1 I I I a 57 2.00 2.20 2.40 2.60 2.80 3.00 lo3 KIT FIG. 3.-Plots of In [(Fo/rn) In (1 +X/Q)] against reciprocal of absolute temperature: 0, LaM; 0, CaM and (>, NaM. TABLE 2.-KINETIC PARAMETERS DETERMINED FOR THE DEPOLYMERIZATION OF PARALDEHYDE OVER DIFFERENT MORDENITE CATALYSTS cation apparent activation energy, pseudo-frequency type EJkJ mol-' factor, A' NaM CaM LaM 37.0 16.0 16.0 3 . 5 ~ 107 3 . 2 ~ 105 4.5 x 105 Mordenite has main channels parallel to the c-axis and side-pockets to the a-axis (fig.4).17 The main channels consist of elliptically distorted 12-membered rings, having a free aperture of 0.581 nm x 0.695 nm. They may admit paraldehyde (kinetic diameter 0.8 nm)24 or even larger molecules, owing to framework distortion and/or molecular def0rmabi1ity.l~ However, the main channel cannot pass more than one paraldehyde molecule at a time,23 although repeated interactions with the surface sites along the channel walls would be quite possible. Topologically the free rotation of a paraldehyde molecule is forbidden in the mordenite cavities. The side-pocket (0.387 nm x 0.472 nm) cannot accommodate paraldehyde 23 In view of the above facts, the Langmuir-Hinshelwood kinetic model is regarded as the preferred reaction 2 5 7 26 The depolymerization of paraldehyde over mordenite surfaces proceeds with the formation of carbocations as surface intermediate^.^^^ 27 3 FAR 158 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE [3 &m coplanar faces:.,carbon: 0 , h y d r o g e n : O oxygen.FIG. 4.-Framework unit cell structure of mordenite (cross-sectional planar view from c-axis; schematic encasing of paraldehyde molecule in main pore shown, not to scale).l79 24 Units in A. Undoubtedly the surface-catalysed depolymerization is initiated through the oxygen atom of paraldehyde molecule, possibly via H-bonding with surface acid sites. As a result the C-0 bond in the heterocyclic linkage is ruptured, creating a carbocation. Through subsequent rearrangement of the vicinal-bond electrons a series of chain scissions occurs, each with formation of monomeric fragments. A hydride ion is abstracted from the terminal monomeric unit and transferred to the surface, whereupon the acid site is restored.This will continue, as long as the surface sites remain catalytically active (fig. 5). All the evidence supports the fact that the ring-opening process is surface-catalysed and is rate-determining.21y 22$ 27 From the surface reaction occurring in the mordenite pores, as schematically depicted in fig. 6, the general stoichiometric relation may be expressed in the following way: k+l k*2 k+3 (PA),+X $ [PA. - *XI* [AA. - *XI* + (AA),+C k-1 k-2 k--3 where PA refers to paraldehyde, I= the vacant surface sites, [ ]* the activated surface intermediate, ( ), the molecules in a gas phase and k + , - to k , , the rate constants for the respective steps. As detailed by Hutchinson et aZ.,O several rate equations can be derived from the above relations, depending upon the state of equilibrium and the rate-determining steps and their derivation may be quite complicated.For the depolymerization under chromatographic conditions, however, simplification is possible through the following rationalization:7* 11, 22 (i) the reversal of the surface reaction step is negligible, viz. k-, LLI 0, (ii) the product species is removed from the surface immediately upon formation, uiz. k-, N 0, and (iii) the surface reaction is regarded as rate-determining,P. J. CHONG A N D G. CURTHOYS 0 1; C H 3 A 0 ' L C H , 0 0 0 0 'si/ \iL/ \ / \ ./ \JO / \ / \ ,si\ 7'\ Y\ c l I M (OH)',"_, H 0 0 0 0 0 Al Si Si A1 \si/ \-/ \ / \ / \-/ / \ / \ / \ / \ I\ 59 0 0 0 0 0 \ ./ \-/ \ ./ \ ./ \-/ SI Al SI SI AL / ' / \ / \ / \ / \ FIG.5.-Scheme for mechanism of depolymerization of paraldehyde over zeolite surface: (M = exchange cation with valence of n). a 11 a -+ free molecules (vapour phase) mass transfer 11 (diffusion) surface- bound ,,B pore-mou t h area mordenite cavity @g@--- s u r f a c e active centres methyl groups - - - _ _ FIG. 6.-Pictorial representation of depolymerization of paraldehyde in mordenite cage (hypothetical). 3-260 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE viz. d[PA+AA]*/dt 2: 0. This reduces the rate expression to an equation of the Langmuir-type, uiz. d(PA)g - k w w g - -- dt 1 +K(PA), where k is the surface reaction rate constant and K is the Langmuir adsorption constant.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 25 24 25 26 27 R. J. Kokes, H. Tobin and P. H. Emmett, J . Am. Chem. Soc., 1955, 77, 5860. W. K. Hall and P. H. Emmett. J . Am. Chem. Soc., 1957, 78, 2091. D. M. Nace, Ind. Eng. Chem., Prod. Res. Deu., 1969, 8, 1. W. A. Blaton Jr, C. H. Byers and R. P. Merrill, Ind. Eng. Chem. Fundam., 1968, 7, 61 1 ; 154th Natl Meeting of the Am. Chem. SOC. Chicago, 1967. D. W. Bassett and H. W. Habgood, J . Phys. Chem., 1960, 64, 769. S. Z. Roginskii and A. L. Rozental, Dokl. Akad. Nauk SSSR, 1962, 146, 152; Kinet. Katal., 1964, 5, 104. ( a ) G. A. Gaziev, V. Yu. Filinovskii and M. I. Yanovskii, Kinet. Katal., 1963, 4, 688; ( 6 ) T.Hattori and Y. J. Murakami, J . Catal., 1968, 10, 114. L. S. Ettre and N. Brenner, J . Chromatogr., 1960, 3, 524. W. W. Schmiegel, F. A. Litt and D. 0. Cowan, J. Org. Chem., 1968, 33, 3334. W. K. Fowler, C. H. Duffey and M. C. Miller, Anal. Chem., 1979, 15, 14, 2333. (a) J. R. Conder and C. L. Young, Physicochemical Measurement by Gas Chromatography (J. Wiley, New York, 1979), p. 550; (b) p. 531. S. H. Langer, J. Y. Yurchak and C. M. Shaughnessy, Anal. Chem., 1968, 40, 1747. T. A. Egerton and F. S . Stone, J . Chem. Soc., Faraday Trans. I , 1973, 69, 22. W. W. Marsh Jr, R. M. Heistand and J. 0. Rice, Reprints-General Papers, A-Petroleum Chemistry (Am. Chem. SOC., Petrol. Chem. Inc., Chicago, 1970), vol. 15(3), A-123. J. C . Bartlet and D. M. Smith, Can. J . Chem., 1960, 38, 2057. D. Atkinson and G. Curthoys, J. Chem. Ed., 1978, 55, 564; 1979, 56, 802. D. W. Breck, Zeolite Molecular Sieves (J. Wiley, New York, 1974), p. 29. D. Atkinson and G. Curthoys, J . Phys. Chem., 1980, 84, 1358. S. P. Walvekar and A. B. Halgeri, J . Res. Inst. Catal., Hokkaido Uniu., 1972, 20, 219. H. L. Hutchinson, P. L. Barrick and L. F. Brown, Chem. Eng. Prog. Symp. Ser., 1967, 63, 18. A. Clark, The Theory of Adsorption and Catalysis (Academic Press, New York, 1970), p. 239. J. W. Hightower and W. K. Hall, J . Phys. Chem., 1968, 72, 4555. A. H. Keough and L. B. Sand, J . Am. Chem. Soc., 1961, 83, 3536. D. C. Carpenter and L. 0. Brockway, J . Am. Chem. Soc., 1936, 58, 1270. P. B. Venuto and P. S. Landis, Adu. Catal., 1968, 18, 259. K. A. Becker, H. G. Karge and W. D. Streubel, J . Catal., 1973, 28, 403. M. L. Poutsma, Zeolite Chemistry and Catalysis, ed. J. Rabo. (Am. Chem. SOC., Washington, 1976), p. 529. (PAPER O/ 1540)

ISSN:0300-9599    

DOI:10.1039/F19827800053

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1982, 78, 61-73 Controlled Wettability of Quartz Surfaces BY ROBERT N. LAMB AND D. NEIL FURLONG*? Colloid and Surface Chemistry Group, Department of Physical Chemistry, University of Melbourne, Parkville, Victoria 3052, Australia Received 4th November, 1980 The wettability of flat quartz-crystal surfaces has been assessed by measurement at 25 "C of contact angle at the water-vapour/water-drop/quartz-plate three-phase line. Plates pretreated by heating in vacuum gave angles, measured through the drop 1 min after removal from vacuum, of 0-44 " as pretreatment temperature was increased from 200 to 1000 "C. Fully hydroxylated, therefore hydrophilic, quartz surfaces are progressively rendered hydrophobic by mutual condensation of surface hydroxyls to form siloxane bridges.Hysteresis was at a maximum after heating at 700-800°C, indicating that maximum surface chemical heterogeneity was produced by heating in this temperature range. Plates methylated subsequent to heat treatment gave angles that were constant at ca. 80" up to treatment at 600 "C and that decreased from 80 to ca. 47 O as the pretreatment temperature was further increased to 1000 OC. This variation with temperature is consistent with a mechanism for methylation in which only non-hydrogen-bonded surface hydroxyl groups on quartz are reactive towards the methylating reagent. Contact angles on both heat-treated and methylated plates were observed to decrease following extended exposure to water vapour. The wettability of solid surfaces, in particular those of finely divided minerals, is central to collection processes such as froth flotation, selective agglomeration and transfer between bulk fluid phases.The complexity of such processes has led to much emphasis being put on the use of model solids1 in wetting studies in much the same way as model colloids (polymer latices, for example) have been used in studies of colloid Various forms of silica and silicate glasses have for some time retained their popularity as model substrates in wetting ~tudies.~ The hydrophilic/ hydrophobic nature of silica surfaces has often been investigated6 and it is well-known that silica, like many other inorganic oxides, owes its inherent hydrophilicity to surface hydroxyl groups, groups which can be removed by treatment at elevated temperature' or by chemical reaction.*Y9 Such treatments have often been assessed on high- surface-area silica powders or granular materials whose amorphous and porous nature sometimes leads to uncertainty in interpretation.The aim of the present study was to determine the wettability of well-defined macroscopic quartz surfaces via measurement of contact angle following outgassing at up to 1000 OC and/or chemical modification by reaction with trimethylchlorosilane. In particular, it was intended to examine how these two types of treatment could be systematically varied and combined so as to produce surfaces with a range of wettability. The study was also undertaken to increase understanding of the inter- pretation of the wettability of heterogeneous silica surfaces and to assess their value as model surfaces in wettability studies.t Currently at C.S.I.R.O. Division of Applied Organic Chemistry, G.P.O. Box 4331, Melbourne, Victoria 3001, Australia. 6162 CONTROLLED WETTABILITY OF SURFACES EXPERIMENTAL Laboratory-reagent trimethylchlorosilane (99 %) was used. All other organic liquids and all inorganic cleaning reagents used were analytical grade (A.R.) chemicals. Triply distilled water produced from an all-glass still in which the second stage contained alkaline permanganate was used. This water had a conductivity < 1.3 x R-'cm-l and a pH of 5.4f0.2 when equilibrated with air. High-purity nitrogen gas (99.99%) was used as supplied. FIG. 1 .-Heat-treatment/methylation apparatus : 1, thermometer; 2, distillation bulb; 3, glass-to-metal seal; 4, kel-F tap; 5, copper tubing; 6, vacuum line; 7, water-circulation cooling coils; 8, O-ring joint; 9, quartz reaction tube; 10, quartz plate-holder; 11, quartz plate; 12, cylindrical furnace.Natural transparent quartz crystals (Arkansas rock crystal, courtesy of Prof. D. W. Fuerstenau) were z-cut into plates of thickness 1 .O-1.5 mm using a diamond-cut saw. Each plate was mounted in wax on a microscope slide for ease of handling and the exposed face wet-polished on a rotating disc. The sequence of polishing media was carborundum powder, fine diamond paste and a series of cerium oxide powders (ca. 4, 3.2 and 2.8 pm average particle size). TheR. N. LAMB AND D. N. FURLONG 63 plates so polished showed no sign of scratches under an optical microscope at 150 times magnification.Following polishing, the plates were washed in warm hexane to remove the wax and then rinsed in warm ethanol and allowed to dry. Plates were then cleaned with ammoniacal hydrogen peroxide solution for ca. 4 min prior to rinsing with copious amounts of triply distilled water and drying under a jet of nitrogen. Plate cleanliness was assessed by the so-called 'steam test',10 in which water vapour was condensed onto the cleaned silica plate and the nature of the condensed film observed. Vig'l has shown that interference fringes occur during evaporation when the contact angle of water on silica is < 4 O. All cleaned plates used in this study gave excellent fringes during the steam test. The cleaning procedure described above, whilst being very efficient for the removal of contaminants, does not produce macroscopic etching of quartz surfaces.12 Glassware exposed to trimethylchlorosilane (TMCS) was cleaned by soaking in warm 30% KOH solutions for 30s.For all other surfaces the washing sequence chloroform, water, concentrated HNO,, water, concentrated NH, and water was used. All surfaces were blown dry with nitrogen. Plates were heated under vacuum ( < 0.13 N m-2) and/or treated with TMCS using the apparatus shown in fig. 1. The temperature of the quartz tube was raised at CQ. 1 O C mine', a rate sufficiently slow to prevent fracture of the quartz plate as it passed through the a-p transition temperature at ca. 570 OC. The plate was kept at the desired temperature for ca.24 h before being cooled to room temperature at ca. 1 "C min-l. The entire system was then purged with nitrogen. For experiments involving heat treatment only, the quartz plate was then transferred immediately to the contact-angle apparatus. For ' methylation' the plate was completely immersed for 15 min in a 10% by volume solution of TMCS in hexane. TMCS in hexane was distilled directly into the quartz reaction tube via a special kel-F tap. The direct distillation methylation procedure was used to minimise the possibility of polymeric material being deposited onto the quartz plates. Methylated plates were washed in hexane to remove excess TMCS, blown dry and transferred to the contact-angle apparatus. Contact angles (8) were measured at 25 f 2 OC by the sessile-drop technique with a system isolated from vibration.The cell for measurement was sealed from the laboratory atmosphere and contained a reservoir of water to aid in saturation of its volume and to restrict evaporation from the sessile drop. Advancing (8,) and receding (8,) contact angles were determined by adding and removing water, respectively, from the sessile drop using a microsyringe. Angles > 50 O were determined using the arc extension method of Neumann and Good1, to locate accurately the tangent to the drop at the line of solid/liquid/gas contact. Smaller angles were determined by fitting the photographed drop profile to the known shape of sessile drops. A modified version of a computer program developed by Maze and BurnetlO was used to perform this curve fitting and to calculate the contact angle from the drop shape.The methods of analysis used enabled contact angles > 5' to be determined from photographs to within lo. RESULTS AND DISCUSSION All contact angles were measured at the quartz-plate/water-drop/water-vapour three-phase contact line. Contact angles on a clean quartz plate for increasing residence time in the contact angle cell are given in fig. 2. The sessile drop used to assess contact angle was removed from the plate following each measurement. The plate, initially water-wet, remained so for a period > 20 h. At longer times both OA and 8, increased, most probably due to atmospheric contamination. It was therefore considered that 20 h represented the upper limit for time-dependent studies using the present experimental set-up.Contact angles are shown with an uncertainty of *2*. This represented the combined effects of drop asymmetry (eleft - Oright), surfacc non-reproducibility, varying drop position on a plate and parallax errors in drop photography. The latter error was estimated from the geometry of the experimental apparatus while the other errors were determined from repeated measurements on numerous quartz plates.64 CONTROLLED WETTABILITY OF SURFACES time/h FIG. 2.4ontamination test on quartz plates: 5, advancing contact angle; Q , receding contact angle. HEAT TREATMENT Contact angles (el) on quartz plates outgassed at temperatures from 200 to 1000 OC are given in fig. 3. These angles were measured ca. 1 min after the plates were removed from the outgassing/methylation cell, hence the superscript. One minute was the minimum total time required for transfer of a plate to the contact-angle cell, positioning of the sessile drop and for photographing advancing and receding angles.Some results of White15 for fused silica plates heated in oxygen are also shown in fig. 3. Although the contact angles reported by White appear to be in good agreement with the advancing angles determined in the present study, comparison is difficult because White gave no details of his methods or accuracy of measurement or if the angles he reported were advancing or receding or if, in fact, he observed hysteresis at all. Both the advancing and receding contact angles, O i and Ok in fig. 3, increased with increasing outgassing temperature, the largest increases occurring between 650 and 700 OC.Contact-angle hysteresis exhibited a maximum in the temperature range 700 to 800 OC (fig. 4). The increase in hydrophobicity of the quartz plates with outgassing temperature is believed due to the progressive removal of surface silanol groups by mutual condensation to form surface siloxane bridges. There is a good deal of reported evidence, for example from studies of interactions with water vapour16- l7 and heats of immersion,18 showing siloxane bridges to be inherently hydrophobic at a molecular scale relative to silanols. Detailed discussion of the nature of this molecular hydro- phobicity will therefore not be given here. Fig. 3 represents the only study in which the ' macroscopic' hydrophobicity of quartz surfaces (i.e.the hydrophobicity detected via contact-angle measurement) has been systematically studied as a function of treatment temperature. The changes in macroscopic hydrophobicity shown in fig. 3 have resulted from changes in the molecular nature of the quartz plates, in particular the removal of surface silanol groups. However, because of the relatively low number of surface groups present on a macroscopic quartz-crystal surface, it was not possible in the present study to perform analyses of surface populations of silanol and siloxane groups after heat treatment at various temperatures. In order to relate the contact-angleR. N. LAMB AND D. N. FURLONG I FIG. 3 . 4 o n t a c t 200 LOO 600 800 1000 outgassing temperature/OC 65 , receding (Bk); (---) data of 200 Loo 600 800 loo0 outgassing ternperature/"C FIG. 4.-Contact-angle hysteresis on heat-treated quartz plates.66 CONTROLLED WETTABILITY OF SURFACES data to the population of surface groups on silica it is necessary therefore to use population data reported for silica powders.Twenty different studies on silica powders have recently been summarized by Knozinger' in the form of a plot of surface number-density of silanol groups as a function of outgassing temperature over the range 0-1000 O C . The data plotted in this manner show considerable scatter due, no 1.0- 0.9 - 0.8 - $ 0.7- h U m L-( - 5 0.6- .* k .d 2 0.5- & c1 u 0.4 - (d 0.3 - 0.2 - outgassing temperature/'C FIG. 5.-Calculated silanol populations on silica surfaces: 1111, summary of literature data on silica powders; (a) from equilibrium contact angles on heat-treated quartz plates; (b) from equilibrium contact angles on heat-treated/methylated quartz plates.doubt, to variations in the origin, pretreatment, porosity and cleanliness of the materials studied, as well as the relative precisions of the various spectroscopic and thermogravimetric techniques used. As rigorous discussion of the wettability of heterogeneous surfaces is based upon surface-area heter~geneity,'~ the silanol number- density data summarized by Knozinger have been converted to 'area fraction of silanols' and, as such, are shown as a band in fig. 5 . For the conversion, it was assumed that one siloxane bridge occupied the area of two silanols, an assumption consistent with the formation of siloxane bridges via the mutual condensation of silanols. The scatter in the area fraction plot (fig.5) is small and indicates that the variations in the reported silanol number density/temperature data stem in the main from variations in the total number of hydroxyls on a fully hydrated surface.2o The data band in fig. 5 predicts that silica surfaces outgassed at 1000 *C should beR. N. LAMB AND D. N. FURLONG 67 chemically homogeneous on a molecular scale, in that they consist almost entirely of siloxane bridges. As it is customary to consider the major causes of contact-angle hysteresis to be surface roughness and chemical heterogeneity,21 the degree of hysteresis observed (fig. 4) following outgassing at 1000 OC, ca.5 O, can therefore be attributed to surface roughness alone. This level of hysteresis is quite sma1119 and is a good indication that the polishing techniques used were successful in producing essentially smooth surfaces. The data band in fig. 5 also predicts that the molecular surface chemical heterogeneity passes through a maximum as the outgassing temperature is raised from 200 to 1000 OC ; i.e. with increasing temperature the surface changes from one consisting entirely of silanols to one with a mixture of silanols and siloxanes to one dominated by siloxane bridges. The general form of the contact-angle hysteresis (fig. 4) is consistent with this prediction. However, maximum contact-angle hysteresis occurred at 700-800 O C , compared with the predicted maximum in surface chemical heterogeneity ‘ seen ’ in contact-angle studies varied with outgassing temperature in a similar manner to powdered silicas previously studied, the nature of the surface at any given temperature was not as previously determined.It has been proposed by Cassie22 that the equilibrium contact angle (8,) on a heterogeneous surface consisting of microscopic patches of two homogeneous com- ponents can be related to the composition of that surface as [a(sioH) - a(si-o--si)] at 400 O C . It therefore appears that, whilst the quartz surface cos 6E = a1 cos 8 E l + a2 cos 8E2 (1) where a, and a, are the fractions of components 1 and 2, respectively, and 8E1 and 8,2 the equilibrium contact angles on homogeneous surfaces of components 1 and 2.Although application of eqn (1) is sometimes limited,lW it has been used to describe the nature of surfaces such as those of interest in the present study, i.e. to derive a surface population profile of a mixed surface (al,a2) from observed contact angles (8E,8El,8E2). The quartz plates of fig. 3 were bifunctional from a wettability point of view, consisting of silanol groups of various typess for which &(SiOH) = &@OH) = 8l,(SiOH) = 0’ and siloxane bridges. In the present study of treated quartz surfaces, advancing and receding, rather than ‘equilibrium’, contact angles were measured. Thus in order to use eqn (1) it is necessary to evaluate an equilibrium angle from the measured advancing and receding angles. Wolfram and F a u ~ t ~ ~ have recently provided experi- mental and simple theoretical justification for the use of the equation for systems wherecontact-angle hysteresis is observed. r is the Wenzel surface-roughness which can be set equal to 1 1 9 7 25 when hysteresis due to surface roughness alone is < loo, as concluded above for the quartz plates used in the experiments of fig.3. Eqn (2) can simply be regarded as a means of averaging 8 A and 8,; the key to its use in combination with eqn (1) is the identification of OY, the angle defined by Young’s equation, with the equilibrium contact angle. Johnson and Dettre19 have proposed a model to describe contact-angle hysteresis on heterogeneous surfaces and show how 8, is related to BAand 8, for some surface types. The essential point of their theory is that many metastable angles exist for heterogeneous surfaces, the energy barriers between states becoming smaller the smaller the size of surface heterogeneities and68 CONTROLLED WETTABILITY OF SURFACES the more random their distribution.Dettre and Johnson have advancing and receding angles on variously coated titanium dioxide surfaces for which OEl = 54' and OEz = 0"; calculation of 6, using eqn (2) shows it to be within 10' over a wide range of surface coverage, with 6, calculated using eqn (1). Assuming the quartz plates of the present study to be completely dehydroxylated during outgassing at 1000 O C gives &(Si-O-Si) = 44' and 6k(Si-0-Si) = 39O.These values are similar to the OE2 value of Johnson and Dettre. The system of variously dehydroxylated quartz surfaces has then similar OEl and 8,, values to those of the experimental system of Dettre and and we have therefore used eqn (2) in conjunction with the data of fig.3 to calculate Ol, at the various outgassing temperatures used. Eqn (1) can be rewritten for these experiments as (3) cos el, - 0.743 0.257 a ( S i O d = TABLE 1 .-CONTACT-ANGLE-TIME STUDIES ON HEAT-TREATED QUARTZ outgassing 9,io wo temperature 1°C 1 min 2.5 h 10 h 1 min 2.5 h 10 h ~~ 625 20 20 20 0 3 5 865 35 26 26 21 18 17 1000 44 44 42 39 40 38 and used to calculate the variation of surface silanol area coverage with outgassing temperature from the contact-angle data of fig. 3. The resultant curve (a) in fig. 5 shows that the quartz plates contained a significantly higher density of surface silanols at all outgassing temperatures when compared with the silica powders studied previously.Curve 5 (a) also indicates that maximum chemical heterogeneity occurred at ca. 800 O C , in good agreement with the occurrence of maximum contact-angle hysteresis at 700-800 O C (fig. 4). It therefore seems that the nature of the quartz plates as seen by contact-angle measurements is well-represented by curve (a) in fig. 5, despite the approximations used in its derivation. Note that the contact-angle data of White15 would also correspond to a surface more hydroxylated at any outgassing temperature than would be predicted from studies on high-surface-area silicas. It is possible that flat silica surfaces, the quartz plates of the present study or the fused discs used by White, do in fact respond differently to heat treatment than do silica powders. It is more likely, however, that in some instances rehydroxylation has occurred in the time interval between heating and measurement of the contact angle. The contact angles on quartz plates in fig.3 were measured at between 1 min, for advancing angles, and 1.5 min, for receding angles, after removal from vacuum and after ca. 30-60 s, respectively, of water-drop residence. Some examples of the subsequent stability of the contact angle on exposure to near-saturated water vapour are given in table 1. Following treatment at 625 O C (and for all temperatures below) no changes were detected in 6, during 10 h of ageing, whilst 6, increased slightly. By contrast, both 6, and 6, decreased significantly during the first 2.5 h of ageing of the 865 OC treated plate.The initial and final values for OA on this plate were within experimental error of those given recently by Pashley and KitcheneP for a plate heated at 875 O C and subsequently aged. No significant changes in 6, or 6, were detected during 10 h ageing of a plate heated in vacuum at 1000 O C .R. N. LAMB A N D D. N. FURLONG 69 It therefore appears that for the quartz plates heated at up to 625 O C considerable rehydroxylation occurred before and/or during the measurement of contact angles. At higher temperatures less rehydroxylation occurred. Such a temperature dependence for the rate of rehydroxylation is consistent with many previous studies on high- surface-area silicas.l79 18* 27 Thus the maximum in chemical heterogeneity observed at 700-800 O C in contact-angle studies represents a surface that at the time ofmeasurement had rehydroxylated to a 1 : 1 silanol: siloxane area ratio from a surface immediately after heat treatment that had a lower area ratio.outgassing temperaturelo C FIG. 6.-Summary of the controlled wettability of quartz: $ 0 , heat-treated quartz; I) @, heat-treated/ methylated quartz. Round symbols, advancing contact angle; square symbols, receding contact angle. HEAT TREATMENT/METHYLATION Contact angles on quartz plates methylated subsequent to outgassing at temperatures from 250 to 1000 O C are given in the upper half of fig. 6. Both advancing and receding angles were constant up to 400 OC and decreased above 700 OC.Both the advancing and receding angles on quartz methylated subsequent to heat treatment at 1000 "C were only 6" greater than on surfaces subjected to the heat treatment alone. This provides confirmation of the assumption used above in application of the Cassie equation to contact angles on heat treated quartz, viz. surfaces treated at 1000 O C were very sparsely populated with surface silanol groups [qSiOH) = 0.0251.70 CONTROLLED WETTABILITY OF SURFACES In an attempt to relate this behaviour to that previously reported, reference can be made to numerous previous studies26*28-33 in which estimates of the air/water contact angle on silica methylated using TMCS have been made (table 2). Often, TABLE 2.-sUMMARY OF CONTACT-ANGLE DATA ON METHYLATED SILICA material method contact angle/O ref.vitreous silica glass microscope vitreous silica Cabosil MY platea slideb platea Aerosil 200a ground quartza quartz crystala quartz crystal captive bubble sessile drop sessile drop ( ?) sessile drop on pressed disc calculation from coverage rise up a packed bed captive bubble sessile drop 70-75 28 8, x 85 eR x o 8, 95- 100 8, 80-90 - 130 29 30 31 70 32 75 33 (x 10% coverage) 8, 80 hysteresis x 5 26 8,88 current work 8,72 a Liquid-phase methylation of Laskowski and Kitchener ; refluxing with liquid chlorosilane. vapour-phase methylation; however, the measurement of contact angle was ancillary to the main aim of each study, which might have been investigation of film thickness,26* 30 water or immersion and partition beha~iour,~~ and as a result not well-described.Published contact angles for various types of methylated silica (without heat treatment) fall between 70 and 1 30°, the one exception being a zero receding angle given by Herzberg et 4Lz9 These workers determined OR using an evaporating sessile drop and their data must therefore be suspect. Three previous 2 9 9 30 in which an advancing angle was measured gave an average value of 88 f go, in excellent agreement with the present value for non-heat-treated quartz. Only Blake and Kitchener30 seem to have also measured a receding angle; their value of 80-90° is higher than that reported here. One study31 in which methylation was achieved by 'refluxing in liquid chlorosilane' reported a contact angle some 40° greater than other studies.It can only be supposed that such treatment resulted in the deposition of polymeric material, particularly as abnormally high surface densities of trimethyl silyl groups were also found. Laskowski and KitchenerZ8 have reported a drop in contact angle from ca. 50' to ca. 40' when a fused-silica disc was methylated following treatment at room temperature and 450 O C , respectively. However, drying conditions were not specified in both experiments and hence their change cannot be truly compared with the results of the present study. It has been p r ~ p o s e d ~ ? ~ ~ that reaction of TMCS with fully hydroxylated silica surfaces proceeds via non-hydrogen-bonded hydroxyl groups. Therefore only removal by outgassing of these isolated and geminal hydroxyls, believed to occur above ca.500 OC, should influence the subsequent population of surface methyl groups. ThisR. N. LAMB AND D. N. FURLONG 71 is confirmed by the contact-angle data in fig. 6. If it is proposed, therefore, that the quartz surfaces heated at above 500 OC and then methylated consist only of siloxane bridges (formed by removal of vicinal hydroxyls below 500 "C) and trimethyl silyl groups, eqn (3) can, for these surfaces, be written as where 8, (Si-0-Si) is taken as 42' (fig. 3), the equilibrium contact angle on a surface consisting only of methyl groups as 1 and where a(SiOH) corresponds to the area density of silanols prior to methylation. The resultant curve of a(SiOH), calculated from the methylation contact-angle data of fig. 6 in conjunction with eqn (2), is included in fig.5 [curve (b)]. Curve 5(b) shows that the maximum area coverage of the quartz surface with trimethyl silyl groups was 54 4 1 %, a coverage in excellent agreement with that calculated by Laskowski and Kitchener2* using Kiselev's area36 of 0.42 nm2 per -Si(CH,), group and assuming reaction with only isolated and geminal silanol groups. Such good agreement indicates that the Cassie treatment leading to eqn (4) does indeed provide a realistic picture of the composition of mixed siloxane-trimethyl silyl surfaces. Curve 5 (b) indicates that heat-treated/methylated quartz surfaces were ca. 1 : 1 in siloxane: trimethyl silyl groups at ca. 500 'C, whilst at higher temperatures they were more heavily populated by siloxane bridges. This prediction of decreasing chemical heterogeneity as the outgassing temperature was increased above 500 OC is confirmed qualitatively by the observed hysteresis in contact angles, although the experimental trend in hysteresis was not very clear. Quartz plates heated at below 500 OC and then methylated will contain siloxane bridges and trimethyl silyl groups as well as residual hydroxyls.There are, however, insufficient contact-angle data in this temperature region (fig. 6) to justify any attempt to describe the degree of surface heterogeneity. The variation with time of contact angles on methylated plates, shown in table 3, TABLE 3 .<ONTACT-ANGLE-TIME STUDIES ON HEAT-TREATEDIMETHYLATED QUARTZ outgassing W0 wo temperature 1°C 1 min 2.5 h 10 h 1 min 2.5 h 10 h ~~ ~~ 82 1 70 63 58 48 39 35 922 55 44 42 48 41 38 indicates that the surfaces became more hydrophilic following exposure to water vapour.Whether such surfaces eventually become fully hydrophilic, as has been previously ~uggested,~~ has not been evaluated in the present study. CONCLUSIONS (1) Quartz plates heated in vacuum at temperatures between 200 and 1000 'C exhibited contact angles that increased progressively from 0 to 44O with increasing temperature. Methylation subsequent to heat treatment gave angles that decreased from ca. 80 to ca. 47'. Therefore, by combination of heat treatment and methylation, quartz can be made to exhibit a wide range of wettability (fig. 6). (2) The wettability of treated quartz surfaces was explained in terms of the relative hydrophobicities of surface silanol, siloxane and trimethyl silyl groups.Analysis of72 CONTROLLED WETTABILITY OF SURFACES contact angles enabled the construction of surface chemical heterogeneity profiles following various treatments, profiles which were in qualitative agreement with observed contact -angle hysteresis, (3) Some contact angles decreased following exposure to water vapour, indicating that some quartz plates had undergone rehydroxylation subsequent to treatment but prior to contact-angle measurement. For heat-treated plates this was also indicated following a comparison of the calculated silanol population with those of previous studies of silica. Rehydroxylation reduces the accessible range of wettability available for such treated surfaces and may limit their usefulness as model surfaces in wettability studies.The authors acknowledge support from the Australian Research Grants Committee (ARGC) and the National Energy Research, Development and Demonstration Council (NERDDC) of the Department of National Development, Commonwealth of Australia. We thank Prof. T. W. Healy for many valuable discussions and for constructive criticism of this work. J. M. Haynes, Wetting, Spreading and Adhesion, ed. J. F. Padday (Academic Press, London, 1978), p. 469. A. Kotera, K. Furusawa and K. Kuoto, Kolloid 2. 2. Polym., 1970, 240, 837, T. W. Healy, A. Homola, R. 0. James and R. J. Hunter, Polymer Colloids 11, ed. R. M. Fitch (Plenum Press, New York, 1980), p. 527. H. Sasaki, E. Matijevic and E. Barouch, J. Colloid Interface Sci., 1980, 76, 319.J. F. Padday, Wetting, Spreading and Adhesion, ed. J. F. Padday (Academic Press, London, 1978), p. 464. I3 R. K. Iler, The Chemistry of Silica (Wiley-Interscience, New York, 1979), chap. 6, pp. 622-729. ' H. Knozinger, The Hydrogen Bond, ed. P. Schuster, G. Zundel and C. Sanderfy (North Holland, Amsterdam, 1976), vol. 3, chap. 27, p. 1270. M. L. Hair and W. Hertl, J. Phys. Chem., 1969, 73, 2372. C. G. Armistead and J. A. Hockey, Trans., Faraday Soc., 1967, 63, 2549. lo M. L. White, Proc. Annu. Freq. Control. Symp., 1973, 27, 79. l1 J. R. Vig, J. W. Lebus and F. L. Filler, Proc. Annu. Freq. Control Symp., 1975, 29, 220. l2 R. M. Pashley, PhD Thesis (Imperial College, London, 1978), chap. 5, pp. 94-104. l3 A. W. Neumann and R. J. Good, Surface and Colloid Science, ed.R. J. Good and R. R. Stromberg l4 C. Maze and G. Burnet, Surf. Sci., 1969, 13, 451. lF, M. L. White, Clean Surfaces, ed. G. Goldfinger (Marcel Dekker, New York, 1970), chap. 18, pp. l6 G. J. Young, J. Colloid Sci., 1958, 13, 67. (Plenum Press, New York, 1979), vol. 1 1 , chap. 2, p. 36. 36 1-373. K. Klier and A. C. Zettlemoyer, J. Colloid Interface Sci., 1977, 58, 216. G. J. Young and T. P. Bursh, J. Colloid Sci., 1960, 15, 361. New York, 1969), chap. 2, pp. 85-153. lB R. E. Johnson and R. H. Dettre, Surface and Colloid Science, ed. E. Matijevic and F. Eirich (Wiley, 2o D. E. Yates, F. Grieser, R. Cooper and T. W. Healy, Aust. J. Chem., 1977, 30, 1655. 21 R. J. Good, Surface and Colloid Science, ed. R. J. Good and R. R. Stromberg (Plenum Press, New 22 A. B. D. Cassie, Discuss. Faraday SOC., 1948, 3, 11. 23 E. Wolfram and R. Faust, Wetting, Spreading and Adhesion, ed. J. F. Padday (Academic Press, 24 R. N. Wenzel, J. Phys. Chem., 1949, 53, 1466. 25 N. K. Adam and G. Jessop, J. Chem. Soc., 1925, 1863. 26 R. M. Pashley and J. A. Kitchener, J. Colloid Interface Sci., 1979, 71, 491. 27 R. H. Dettre and R. E. Johnson, J. Phys. Chem., 1965, 69, 1507. 28 J. Laskowski and J. A. Kitchener, J. Colloid Interface Sci., 1969, 29, 670. 2B W. J. Herzberg, J. E. Marian and T. Vermeulen, J. Colloid Interface Sci., 1970, 33, 164. T. D. Blake and J. A. Kitchener, Trans. Faraday Soc., 1972, 68, 1435. 31 W. T. Yen, R. S. Chahal and T. Salman, Can. Metall. Q., 1973, 12, 231. York, 1979), vol. 1 1 , chap. 1, pp. 10-13. London, 1978), chap. 10, p. 213.R. N. LAMB AND D. N. FURLONG 73 32 L. Alzamora, S. Contreras and J. Cortes, J . Colloid Interface Sci., 1975, 50, 503. 33 S. Garhsva, S. Contreras and J. Goldfarb, Colloid Polym. Sci., 1978, 256, 241. 34 V. Ya. Davydov, A. V. Kiselev and L. T. Zhuravlev, Trans. Faraday SOC., 1964, 60, 2254. 36 N. K. Adam, A h . Chem. Ser., 1964,43, 52. 38 A. V. Kiselev, N. V. Kovaleva, A. Ya. Korolev and K. A. Shcherbakova, Dokl. Akad. Nauk SSSR, 1959, 124, 617. (PAPER O / 1682)

ISSN:0300-9599    

DOI:10.1039/F19827800061

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Tians. I , 1982,78, 75-88 Kinetics and Mechanism of the Quaternization of Poly(4-vinyl pyridine) with Alkyl and Arylalkyl Bromides in Sulpholane BY ERNEST A. BOUCHER* AND CHRISTOPHER C. MOLLETT School of Molecular Sciences, University of Sussex, Brighton BNl 9QJ Received 6th November, 1980 Quaternization reactions of poly(4-vinyl pyridine) with several organic bromides show retardation of reaction in excess of that predicted by second-order kinetics. Kinetic expressions from a neighbouring-group model, in which a reaction'is characterized by the rate constants ki for reaction of a pyridyl group having i = 0, 1 or 2 already reacted neighbours, have been applied to the experimental results. Quantities obtained include k,, K = k,/k,, L = k,/k,, activation energies E,( = E, = E2) and pre-exponential factors A,.The ratios K and L for the bromides range from 0.95 and 0.54, respectively, for ethyl bromide to 0.55 and 0.28 for 1-bromo-3-phenylpropane. On average K = 2L. The dependence of k , values on the nature of the organic group of the bromides is discussed in terms of steric hindrance in the transition state; similar reasoning can be applied to molecules with a single pyridyl group. Linear empirical relationships have been found between K and L and the extended lengths of the alkyl and arylalkyl groups of the bromides. The deviation from second-order kinetics ( K < 1, L < 1) is explained in terms of steric hindrance, involving the transition state of a reacting pyridyl group and the alkyl group on a neighbouring already reacted pyridyl group.Electrostatic effects are not thought to influence the reaction kinetics. In 1955, Coleman and FUOSS~ reported that the quaternization of poly(4-vinyl pyridine), P4VP, with n-butyl bromide in sulpholane departed from overall second- order kinetic behaviour. They found retardation of the reaction which was attributed to a neighbouring-group effect with the reactivity of a pyridyl group depending on whether or not its neighbours have reacted. The quaternization of P4VP by a variety of alkyl and arylalkyl halides is a good example of a polymer-transformation reaction which yields a polyelectrolyte. In a wider context, polymer-transformation reactions have been reviewed2 with emphasis on kinetic and statistical treatments.Several groups of workers3-* have previously studied some of the many possible quaternization reactions of P4VP, but for the reasons given in the following summary few of these studies seemed to be reliable for elucidating the cause of the retardation as reaction proceeds. The general consensus of opinion found in previous reports was that retardation was due to electrostatic effects, although one study attributed retardation to steric hindrance (on unreliable grounds). Fuoss et aL3 assumed that the reactivity of a group with one reacted nearest neighbour was the same as that for a group with neither neighbour reacted. They ascribed all of the retardation to the lower reactivity, due to electrostatic effects, of a group having both neighbours already quaternized.Morcellet-Sauvage and Loucheux4~ also analysed kinetic measurements for n-butyl bromide in sulpholane by the approximate methods of Fuoss and coworkers, and thought that electrostatic effects were involved. Arends9 analysed some of the data of Coleman using more appropriate mathematical expressions, but again attributed the effect to groups having both neighbours already reacted. He formed the opinion that steric hindrance was involved because the activation energy (based on experiments at three temperatures) 7576 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) for reaction of groups with both neighbours reacted was reckoned to be the same as that for a group with two unreacted neighbours. However, the prior assumptions that, in the notation used herein, k, = k, and that k,/k, dQes not depend on temperature ensures that the activation energies are equal: E, = El = E,. Boucher and MollettlO and Boucher et d.ll have analysed experimental data for the quaternization of P4VP with several alkyl bromides in sulpholane using rigorously derived kinetic expressions based on the neighbouring-group model involving rate constants k,, k, and k,.These studies showed that the existence of one reacted neighbour affected the reactivity of a group, i.e. k , # k,. Furthermore, for reasons which are elaborated on below, it was concluded that the retardation is due mainly to steric hindrance and that the magnitude of the effect increases with increase in the size of the organic bromide. In a complementary study,12 where viscosity changes of the reacting systems were measured as a function of the extent of reaction, it was found that the addition of N-ethyl pyridinium bromide suppressed expansion of the polymer molecules, but did not usually affect the kinetics.This provided a major piece of evidence against the importance of electrostatic effects in causing retardation. We have extended the kinetic study to a wider range of alkyl and arylalkyl bromides and are able to quantify the neighbouring-group effect. The use of different organic bromides also throws more light on the fundamental nature of these examples of the Menschutkin reaction. Many of the older studies of the quaternization reactions of small molecules were not particularly thorough, but the essentially S,2 character was no doubt correctly deduced.A mystery regarding many quaternization reactions has been the nature of the colours sometimes found as quaternization proceeds (reported, for example, by Hantzsch13). A separate account will show that the products of the quaternization reactions involving polymer and small pyridyl molecules are probably charge-transfer complexes, the electronic absorption properties of which depend on the species of pyridyl molecules, the halide species and the reaction conditions (e.g. temperature and solvent species). NEIGHBOURING-GROUP MODEL The development of neighbouring-group models has recently been reviewed in detail and the most important kinetic expressions summarized.2 However, it is appropriate to give a brief account leading to the method of data analysis used in this study.The pyridyl groups along a polymer chain are represented by circles, where 0 denotes an unreacted group and 0 denotes a reacted group. A portion of the chain, . . . 0 0 0 0 0 0 0 0 0 0 0 0 . - t t t k2 kl k0 signifies that the probability of reaction of a group in the time interval t to t+dt is k, dt, k, dt or k , dt, depending on whether none, one or two neighbours of the group have already reacted at time t. The fractional extent of reaction of the pyridyl groups for large molecules (strictly for the number of monomer units rn -+ a) is given by2 5 = 1 - 2KaL J: (1 - exp [ - 2( 1 - K ) (1 - u)] du + 2aAJ: (1 - u) ~ 2 ~ - ~ exp [ - 2( 1 - K ) (1 - u)] du -(2-a) a2Kexp[-2(1 -K)(l--a)].E. A. BOUCHER AND C . C . MOLLETT 77 Integration can either be carried out by numerical computation or analytically, as here, after the exponentials have been expanded.The time variable a in eqn (1) is given r rt 1 by and K = k,/k,, L = k,/k,. (3) In eqn (2), c, and c: are the instantaneous and initial values of the concentration of the bromide reagent. When c, = c: is constant, a = exp ( - k , t), and eqn (1) would apply to a unimolecular reaction or to a pseudo-first-order bimolecular reaction. In the present reactions, eqn (2) allows for the depletion of reagent as reaction proceeds. An overall second-order bimolecular reaction corresponds to K = L = 1. The pro- cedure used to obtain a consisted of plotting c,/c: against actual time t on a large scale and using this for integration over suitable time intervals, with k, having been obtained separately in the limit t -+ 0 from second-order plots.Plots according to the usual form f = (c: - c;)-l In (c: c,/c:c,) = k, t (4) are expected to curve towards the time axis when there is retardation over and above that due to reducing concentrations cp of polymer groups and c, of reagent. We estimate the limits of precision on individual k, values to be ca. & 2%. However, values quoted for a given halide species under standard conditions are obtained by interpolation of values at several temperatures and are more precise (table 1). When eqn (1) is expanded and integrated, K(l -aa+l)+[K(b-2)+ 1](1 -aa+2) a + 2 < = 1 -2aLe-b -(2-a)a2Kexp[-b(l-a)]; a = 2K-L-1, b = 2(1-K). (5) This series converges sufficiently rapidly for only the first eight terms in a to be required.Having obtained a suitable range of values of a for a particular reaction, the next step in comparing theory with experiment is to evaluate r from eqn ( 5 ) for several pairs of K and L values. To do this, large-scale plots were made of robs against tabs, and the theoretical plots of 5 against t superimposed. The ' goodness of fit' was judged by eye and tested by evaluating the deviation of the experimental points from the theoretical plots. We have paid particular attention to testing the reliability of K and L values by systematically varying k,, K and L from best-fit values. It is not possible to illustrate the changes on small-scale reproductions, but it clearly emerges that if k , has an unreliable value, then changes in K and L will not compensate for this and restore the goodness of fit.Furthermore, K dominates the shape of the curves in the approximate range r = 0.3-0.7, and L is dominant over the range 0.7-1 ; Kand L values are not strongly correlated. To test the above procedure, deviations between the fitted curve and experimental values of < have been expressed as the standard deviation CT for a few temperatures. Values of o confirmed that best-fit curves were obtained; typical values were for 2-bromoethylbenzene at 344.1 K. = 7 x for i-butyl bromide at 338.9 K and 6.8 xro C TABLE 1 .-SUMMARY OF KINETIC RESULTS FOR QUATERNIZATION OF POLY(4-VINYL PYRIDINE) IN SULPHOLANE 4 lo5 k, range /dm3 mo1-I s-' K L 107 A , EO temp. no. of 4 range r-t range key / K expts /mol dm-3 reagent letter /mol dmP3 at 340 K (f0.03) (k0.03) /dm3 rno1-I s-' /kJ mol-l 306.2-343.2 307.9-348.2 308.6-353.2 324.2-349.1 306.2-353.2 304.3-323.7 308.5-349.0 323.7-348.0 306.2-328.2 308.4-349.4 3 13.0-348.9 3 30.0-348.9 9 9 9 6 4 8 8 7 4 9 6 6 0.049-0.227 0.107-0.159 0.105-0.21 4 0.087-0.107 0.17 1-0.195 0.034-0.036 0.105-0.107 0.105-0.106 0.042-0.044 0.105-0.108 0.104-0.106 0.104-0.106 ethyl bromidea n-propyl bromide n-butyl bromideb n-pentyl bromide n-hexyl bromidea ally1 bromide i-butyl bromide bromometh ylcyclohexane benzyl bromidea 2-bromoethylbenzene 1 -bromo-3-phenylpropane 1 -bromo-2-phenylpropane A B D F G C E H I J M N 0.200-1.869 0.383-0.635 0.215-0.428 0.380-0.394 0.195-0.378 0.I09 0.406 0.383 0.392 0.382 0.382 0.042-0.066 91.02 k 0.90 0.95 0.54 31.57f0.31 0.80 0.38 27.47 _+ 0.26 0.70 0.3 1 31.28k0.42 0.65 0.30 33.13 k 0.32 0.55 0.27 3736k5.2 0.80 0.40 1.721 fO.O1 0.60 0.37 1.22k0.016 0.65 0.33 5730 k 50 0.75 0.34 16.01 f0.15 0.65 0.32 29.16f0.23 0.55 0.28 0.602 & 0.01 0.60 0.32 3.25 k 2.52 0.45 & 0.30 0.34f 0.24 0.42 k 0.3 1 0.44k0.32 2.84 1.30 0.028 k 0.01 0.01 7 k 0.010 0.67 k 0.52 0.16 f 0.08 0.29 0.19 0.0086 & 0.004 68.8 & 3.3 66.2 4.3 65.8 k 4.0 66.1 & 1.1 66.0 &- 3.5 57.9 4 1.2 66.6 4 0.9 66.1 k3.1 52.6 k 4.9 65.3 k 2.3 65.2 & 2.3 66.2 4 2.7 a Analysis of Groves' data; analysis of Fletcher's data.=3 0 z 0, v 0 rE. A. BOUCHER AND C. C. MOLLETT 79 EXPERIMENTAL The basic techniques of polymer preparation, and of making kinetic measurements, based on bromide ion determinations by Volhard’s method, have been described previously.There are, however, new reagents involved, and some aspects of technique are emphasized. The concentration ranges of the polymer and reagent (organic bromide) species, the reaction temperature ranges and the number of kinetic runs are given in table 1. The purification techniques are mainly based on the monograph by Perrin et all4 Ally1 bromide (Aldrich) was purified by washing with aqueous sodium bicarbonate (10% w/v) and distilled water, dried over CaC1, and distilled (298 K at 130 Pa). n-Pentyl bromide (Aldrich) was washed with concentrated H,SO,, water, aqueous sodium carbonate (10% w/v) and water again, then it was dried (CaCl,) and distilled (305 K at 130 Pa). Neopentyl bromide (Fluka) was used without further purification.Bromomethylcyclohexane (Aldrich), was dried (CaCl,) and distilled (303 K at 13 Pa), as were 2-bromoethylbenzene (Aldrich, 307 K, 13 Pa), 1 -bromo-3-phenylpropane (Aldrich, 308 K, 13 Pa) and 1 -bromo-2-phenylpropane (Aldrich, 310 K, 13 Pa). All reagents were stored in the dark. The basic method of making kinetic measurementslo’ll was modified for the very rapid quaternization reactions of poly(4-vinyl pyridine) with allyl bromide. A known weight of P4VP was dissolved under dry N, in a known volume of sulpholane. The reaction vessel was equilibrated in a thermostatted water bath for 1.8 ks (30 min), and reagent added from a calibrated syringe. At intervals 2 cm3 aliquots of reaction mixture were removed with a syringe (calibrated at the reaction temperature) and the bromide ion concentration determined by Volhard’s method.The samples of bulk free-radical polymer were not dissolved in t-butyl alcohol during purification, as has been customary in previous studies.lo9 l1 Khosravi-Babadi16 has shown that polymer recovered from t-butyl alcohol contains entrapped solvent. The polymer used in the present study was extruded into toluene (2 dm3) and then thoroughly washed with fresh toluene. Elemental analysis of the polymer samples has been found to be unreliable, attributed to incomplete combustion of the samples. However, reaction involving the samples of P4VP have consistently reached fractional extents ( of reaction of 0.99 kO.01, which is taken, together with u.v., i.r. and n.m.r. spectra, as a good indication of polymer purity.The viscosity-average molecular weight of the polymer was determined in absolute ethanol at 298 K using Mark-Houwink constants,16 K = 2.5 x cm3 g-l, a = 0.68. The molecular weight was (576 & 14) x lo3 for the polymer recovered by extrusion into toluene, compared with (301 f 6) x lo3 for samples involving recovery from t-butyl alcohol. RESULTS Table 1 shows the kinetic results for the quaternization of P4VP with several alkyl and arylalkyl bromides. The time dependence of the fractional extent of reaction for four typical systems differing in reaction rate constants and activation energies are shown in fig. 1-4. The curves were obtained from the theoretical analysis summarized above. Fig. 5 shows a comparison of the time dependence of the extent of reaction of P4VP with n-pentyl bromide in sulpholane at 339.6 K ( K = 0.65, L = 0.30) with that for a true second-order reaction ( K = 1, L = 1).A typical example of a second-order plot is given in fig. 6 for the quaternization of P4VP with allyl bromide in sulpholane at 305.6 K. The initial rate constant k, is given by the t -+ 0 limiting tangent to the curve. The filled-in circles and the curve were obtained from the theoretical time dependence of 5. SUMMARY OF KINETIC RESULTS Table 1 shows the ratios K and L defined by eqn (5) for all the systems studied. The fractional error on L is approximately twice that on K. Fig. 7 shows the80 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) FIG. 1 .-Time dependence of the fractional extent 5: of reaction for P4VP + ally1 bromide in sulpholane at: (a) 323.7, (b) 315.7 and (c) 305.6 K.All curves in fig. 1-5 are from the theory. f I I 1 I I 0 5 10 15 20 25 tlks FIG, 2.-Dependence of < on t for PllVP+i-butyl bromide in sulpholane at: (a) 349.2, (b) 339.2 and (c) 325.2 K.E. A. BOUCHER AND C. C. MOLLETT 81 FIG. 3.-Dependence of 5 on t for P4VP+2-bromoethylbenzene in sulpholane at: (a) 323.7, (b) 317.8 and (c) 308.4 K. tlks FIG. 4.-Dependence of < on t for P4W+ I-bromo-3-phenylpropane in sulpholane at: (a) 348.9, (b) 340.6 and (c) 333.2 K.82 QUA TE RNI Z A T I ON OF P 0 L Y(4-V IN Y L P Y R ID I N E) FIG. 5-Dependence of 5 on t for P4VP+n-pentyl bromide in sulpholane at 339.6 K for K = L = 1. and theoretical curve FIG. 6.-Second-order plot for P4VP+allyl bromide in sulpholane at 305.6 K. The open circles are experimental and the filled circles are from the theory. The straight-line limiting, t/ks -+ 0, tangent gives ko.dependence of K and L on n, the number of carbon atoms in the organic group attached to the bromine atom of the reagent. There is a surprising degree of systematic variation for such a widespread set of molecules, with the alkyl species clearly distinct from the others. The values of k, given in table 1 are all for 340.2 K and were obtained by interpolation of Arrhenius plots based on the temperature ranges shown, withE. A. BOUCHER AND C. C. MOLLETT 0 / (0) 83 O.* t O * * l 0 2 I 6 8 10 n FIG. 7.-Dependence of the ratios K , sets (a), and L, sets (b), on the number n of carbon atoms in the organic bromides.comparable but not exactly identical initial concentrations of reagent and polymer. Some of the kinetic results have previously been givenll for 319.2 K. Assuming that there are three activation energies E,, El and E2 corresponding to k,, k, and k, in the neighbouring-group model, we can write ki = Aiexp(-E,/RT), i = 0, 1, 2. ( 6 ) The values of A , and E, and the associated experimental errors were given by least-squares determinations of straight lines fitting Ink, as a function of 103K/T. Similar plots of In k, and In k , showed that, within experimental error, El and E2 are identical to E,, i.e. K and L are not themselves temperature dependent. Table 2 gives results for quaternization with n-propyl bromide and benzyl bromide in sulpholane, propylene carbonate and 2,4-dimethyl sulpholane.The values of K and L are not dependent on the solvent species, although k, can be expected to depend on solvent for the reasons already discussed. l2 TABLE 2.--KINETIC QUANTITIES FOR QUATERNIZATION OF P4vP 104k, reagent solvent T/K /dm3 mol-l s-l K L n-propyl bromide sulpholane 333.1 17.39 0.80 0.38 n-propyl bromide propylene 333.5 11.35 0.80 0.36 carbonate sulpholane n-propyl bromide 2,4-dimethyl 333.2 6.74 - - benzyl bromide sulpholane 306.2 691.0 0.75 0.34 benzyl bromide 2,4-dimethyl 306.2 2 5 4 . 0 0.75 0.36 sulpholane84 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) The initial rate constants k , in table 1 show that values for the C, to c6 n-alkyl bromides are reasonably constant, but that k, for ethyl bromide is ca.3 times that for the others. The value for 1 -bromo-3-phenylpropane is comparable with those for the C, to c6 alkyl molecules, and that for 2-bromoethylbenzene is about one half the value. Rate constants for i-butyl bromide, bromomethylcyclohexane and 1 -bromo- 2-phenylpropane, on the other hand, are very approximately 1/30 of those for the C,-C, molecules, and those for allyl bromide and benzyl bromide are ca. 100 times greater. The following discussion deals with the magnitude of the rate constant for a particular organic bromide, as well as with the values of the ratios K and L. The activation energies E, are the same within experimental error, except for values for allyl bromide and benzyl bromide. It is more difficult to explain the pre-exponential A , values because the errors on them are relatively large: a frequently occurring difficulty with this quantity.DISCUSSION THIS STUDY The quaternization reactions have been found to go to completion, with most experimental limits being = 0.99+0.01. There is no evidence that these reactions are reversible. If they were they would be expected to reach an equilibrium limit 5 < 1 . The mechanism of quaternization, involving nucleophilic substitution (S,2) of the bromide on the a-carbon of the reagent by the pyridyl base, is thought to be exactly analogous to quaternization reactions of small pyridyl species.17 The rate-determining step is the formation of the transition state. This is represented schematically in fig. 8, where C, and C, are the a-carbon and a-carbon of the organic bromide, A, B and C are substituents on C,, Y is the leaving bromine and X is the entering pyridyl species.FIG. 8.--Schematic representation without implied scale of the transition state for a quaternization reaction. See text for explanation. We first explain why ethyl bromide reacts more rapidly than n-propyl bromide (and the C,-C6 bromides: table 1). The activation energies are constant indicating that the transition state for n-propyl bromide is not distorted for the C,Br compared with the C,Br,18 e.g. were C,-X and C,-Y distances larger for C,Br to reduce steric interactions with C, substituents, E, would increase. Magat and coworker^^^^ 2o pointed out that if the methyl group on C, cannot be rotated so that A, B and C canE.A. BOUCHER AND C. C. MOLLETT 85 equally well occupy all three of the positions shown in fig. 8, then the entropy of activation, and k,, should be smaller compared with a transition state where all three positions are equally probable. The enthalpy AH and entropy A S of activation2’ can be related to Ei and 2, by AH= Eo-RT A , = (kT/h)exp(AS/R) (7) where k is Boltzmann’s constant and k is Planck’s constant. A H is associated with binding forces, and A S with the freedom of movement of the atoms. Clearly, therefore, variations in A , and A S will be important in all the reactions for which E, is constant. We suppose therefore that in the transition state, the substituents on Cp of the n-propyl group are restricted to a unique conformation (position), whereas with the ethyl group, three equally probable positions exist, and the ratio of the rate constants is expected to be 1 : 3, compared with 3 1.6 : 9 1 .O found experimentally (table 1).Similar arguments apply to the rate constants for the n-C,-C, bromides. The ratios of the rate constants for i-butyl and n-butyl bromides to ethyl bromide, 0.02 and 0.40, respectively, with Eo virtually identical, suggest that there is severe steric hindrance involving the two methyl groups on C, and the X and Y groups. When there are three methyl groups on C, (neopentyl bromide), the steric hindrance completely inhibits reaction (no bromide ions could be detected after 42 days). With ally1 bromide and benzyl bromide the large rate constants are in part associated with lower E, values than for the n-alkyl bromides. The transition state is more readily formed because of n-bond overlap between X and Y and the organic part of the halide molecule (cf.similarly large rate constants compared with that for ethyl bromide found by Clarke and Rothwel122 with the quaternization of 4-ethyl pyridine). By comparison, the rate constants for 2-bromoethylbenzene and 1 -bromo- 3-phenyl propane relative to that for ethyl bromide are 0.18 and 0.36, respectively, but E, values are comparable for all three bromides. It is concluded that when one or more carbon atoms exist between a benzene ring and C,, there is no enhanced rate constant due to a favourable transition state, i.e. k, values are similar to those for n-C, to C , bromides. With bromomethylcyclohexane the low rate of reaction is due to increased steric hindrance in the transition state compared with the n-alkyl bromides, whilst there is no decrease in the activation energy as observed with benzyl bromide.The low reactivity of 1 -bromo-2-phenylpropane is not unexpected and is analogous with that of i-butyl bromide in the aliphatic series of bromides. The systematic decrease of K and L ratios with increasing size of organic group attached to the bromide is in itself evidence that steric hindrance plays a part, and could be the sole factor in the neighbouring-group effect. An electrostatic effect would be expected to be virtually independent of the nature of the organic group. We realize that errors on pre-exponential factors are relatively large, but if it is accepted that Eo = El = E,, then retardation of reaction due to the neighbouring-group effect must arise for the condition A , > A , > A,.Therefore, for a particular organic bromide, as the reaction proceeds there is increasing steric hindrance restricting the number of configurations which can be adopted by the organic group in the transition state due to the presence of already quaternized nearest neighbours. It seems reasonable that the ‘reach’, or the longest distance that the organic group can influence, is a measure of its ability to be involved sterically. Fig, 9 shows K and L plotted against the extended lengths of the molecules relative to ethyl bromide as obtained from space-filling scale models. The distinction between the n-alkyl bromides and the other molecules is no longer evident, and both sets of results (for K and L)86 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) 0 4 1 I Q- I Q- 0.4- 1 (b) 8 1 *w.@\ 0.2 - 0 I I 1 1.5 2 .o 2.5 I I 1 length ratio FIG. 9.-Dependence of K (a) and L (b) on the length of the organic groups of the bromide reagents in table 1. fall on reasonably straight lines. It is also noted that the benzene ring of benzyl bromide ( K = 0.75) provides less steric hindrance than the ring in bromomethyl- cyclohexane ( K = 0.65) which can exist in chair or boat forms. The values of K (0.75) are the same for n-propyl and allyl bromides, i.e. there are unlikely to be inductive effects since any reduction in the rate of reaction of a pyridyl group having a reacted neighbour would be expected to be less for reaction with allyl bromide compared with n-propyl bromide.Comparison of the K and L values for these quaternization reactions with values obtained by other workers is of limited value. Previous analyses of kinetic data have in the main used less rigorous model equations and have assumed that K = 1, which disagrees with our own findings. Necessarily, the values of L obtained on this basis disagree with the values given in table 1. The mean value of K / L for the twelve reagents used is within the limits of precision 2.0. This is taken as support for a mechanism of retardation involving steric hindrance of the alkyl group attached to the a-carbon in the transition state. If the organic group attached to an adjacent reacted pyridyl group is capable of preventing the organic group of a reagent molecule from occupying certain positions in the transition state, then it is to be expected that the presence of a second reacted neighbour will exert an approximately equal and additive effect.RELATED WORK Previous studies on polymer quaternization have been summarized in the introduc- tion and in a recent review.2 Generally direct comparison of rate constants is not possible due to differing reaction conditions or methods of data analysis. Comparison has already been made1' between rate constants for polymer quaternization and those for the quaternization of 4-ethyl pyridine.E. A. BOUCHER AND C. C. MOLLETT 87 TABLE 3.-cOMPARISON OF RATE CONSTANTS RELATIVE TO CORRESPONDING ETHYL HALIDE VALUE KI + P4VP + organic organic chlorides reagent bromides [ref.(23) group (this study) and (24)] n-propy 1 0.38 0.41 ally1 41 .O 31.6 n-butyl 0.32 0.40 n-pentyl 0.34 0.56 n-hexyl 0.36 0.52 benzyl 63 80 ethyl benzene 0.18 0.48 4-Ethyl pyridine + n-propyl and n-butyl bromides give 0.35 and 0.30, respectively. Comparison can also be made with a few analogous reactions,23! 24 especially regarding ratios of reaction rates for organic halides, as shown in table 3 for the halide exchange between KI and organic chlorides. There is a broad measure of agreement among values relative to those for reaction involving ethyl bromide and ethyl chloride, but ratios for organic chlorides exceed those for the corresponding bromides. Hammett and coworker^^^^^^ found that the ratio of the rate constants for the reaction of n-propyl halides with sodium thiosulphate in ethanol S,O;-+ C3H7Br -+ C3H7S,0; + Br- Bunte salt relative to the rate constant for the corresponding ethyl halide was ca.0.3, and that the activation energies were approximately the same for a particular halide. They explained these findings in terms of steric hindrance using a model similar to that used above for the polymer reactions. However, they found that an increase in the activation energy, associated with an increase in the halide to a-carbon bond distance, was responsible for low reactivity of i-butyl halides; this has not been found for quaternization with i-butyl bromide. Dostrovsky and Hughesz7 found that ethyl, i-butyl and neopentyl bromides reacted with sodium ethoxide (giving the corresponding ether + sodium bromide) in the ratio of rate constants of 1 : 0.04: which is comparable with our ratio for quaternization of the polymer, namely, 1 : 0.02: 0.CONCLUSIONS The wide range of systems studied has greatly increased the amount of quantitative information on quaternization reactions. There is evidence that differences in reactivity of an isolated pyridyl group towards the organic bromides depends significantly on the degree of steric hindrance in the transition state. The retardation of the overall quaternization reaction as the reaction proceeds is explained in terms of steric hindrance between the organic group on the reacting group and that on an already reacted neighbour. There is no support in this study for an electrostatic effect being responsible for the retardation.88 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) C .C . M. acknowledges with gratitude a studentship from the States of Guernsey Education Department. B. D. Coleman and R. M. Fuoss, J. Am. Chem. Soc., 1955,77, 5472. E. A. Boucher, Prog. Polym. Sci., 1978, 6, 63. R. M. Fuoss, M. Watanabe and B. D. Coleman, J. Polym. Sci., 1960, 48, 5 . J. Morcellet-Sauvage and C. Loucheux, C.R. Acad. Sci., Ser. C, 1972, 275, 873, J. Morcellett-Sauvage and C. Loucheux, Makromol. Chem., 1975, 176, 31 5. E. Tsuchida and I. Irie, J. Polym. Sci., Polym. Chem., 1973, 11, 789. C. L. Arcus and W. A. Hall, J. Chem. SOC., Suppl., 1964, 2, 5995. N. A. PlatC, Pure Appl. Chem., 1976, 46, 49. C. B. Arends, J. Chem. Phys., 1963, 39, 1903. lo E. A. Boucher and C. C. Mollett, J. Polym. Sci., Polym. Phys. Ed., 1977, 15, 283. l1 E. A. Boucher, J. A. Groves, C. C. Mollett and P. W. Fletcher, J. Chem. Soc., Faraday Trans. I , 1977, l2 E. A. Boucher, E. Khosravi-Babadi and C. C. Mollett, J. Chem. SOC., Faraday Trans. 1, 1979, 75, l 3 A. Hantzsch, Berichte, 1909, 42, 81. l4 D. D. Perrin, W. L. F. Armarego and D. R. Perrin, Purification ofLaboratory Chemicals (Pergamon, l5 E. Khosravi-Babadi, 1979, unpublished results. l6 J. B. Berkowitz, M. Yamin and R. M. Fuoss, J. Polym. Sci., 1958, 28, 69. l7 E. A. Boucher, E. Khosravi-Babadi and C. C. Mollett, J. Chem. Soc., Faraday Trans. 1,1978,74,427. lo N. Ivanoff and M. Magat, J. Chim. Phys., 1950, 47, 914. 2o E. Bauer and M. Magat, J. Chim. Phys., 1950, 47, 922. 21 A. Streitwieser, Soluolytic Displacement Reactions (McGraw-Hill, New York, 1962). 22 K. Clarke and K. Rothwell, J. Chem. Soc., 1960, 1885. 23 J. B. Conant and W. R. Kirner, J. Am. Chem. Soc., 1924, 46, 232. 24 J. B. Conant and R. E. Hussey, J. Am. Chem. Soc., 1925,47, 476. 25 T. I. Crowell and L. P. Hammett, J. Am. Chem. SOC., 1948, 70, 3444. 26 P. M. Dunbar and L. P. Hammett, J. Am. Chem. Soc., 1950, 72, 109. *' I. Dostrovsky and E. D. Hughes, J. Chem. Soc., 1946, 157. 73, 1629. 1728. Oxford, 1966). J. P. Hine, Physical Organic Chemistry (McGraw-Hill, Tokyo, 2nd edn., 1962). (PAPER O / 1698)

ISSN:0300-9599    

DOI:10.1039/F19827800075

出版商:RSC    

年代:1982

数据来源: RSC