ISSN:0306-0012    

DOI:10.1039/CS98514FX005

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ISSN 0306-0012 CSRVBR 14(2) 121-196 (1985) Chemical Society Reviews Vol 14 No 2 1985 Page CENTENARY LECTURE Phase Equilibrium and Interfacial Structure By B. Widom 121 TILDEN LECTURE The Collision Dynamics of Vibrationally Excited Molecules By Ian W. M. Smith 141 The B12dependent Isomerase Enzymes; How the Protein Controls the Active Site By John M. Pratt 161 S-Nitrosation and the Reactions of S-Nitroso Compounds By D. Lyn H. Williams 171 The Royal Society of Chemistry London

ISSN:0306-0012    

DOI:10.1039/CS98514BX007

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

ISSN 0306-0012 CSRVBR 14(2) 121-196 (1985) Chemical Society Reviews Vol 14 No 2 1985 Page CENTENARY LECTURE Phase Equilibrium and Interfacial Structure By B. Widom 121 TILDEN LECTURE The Collision Dynamics of Vibrationally Excited Molecules By Ian W. M. Smith 141 The B zdependent Isomerase Enzymes; How the Protein Controls the Active Site By John M. Pratt 161 S-Nitrosation and the Reactions of S-Nitroso Compounds By D. Lyn H. Williams 171 The Royal Society of ChemistryLondon Chemical Society Reviews EDITORIAL BOARD Professor K. W. Bagnall, B.Sc., Ph.D., D.Sc., C.Chem., F.R.S.C. Professor B. T. Golding, B.Sc., M.Sc., Ph.D., C.Chem., F.R.S.C. Professor G. Pattenden, Ph.D., C.Chem., F.R.S.C. Professor P.A. H. Wyatt, B.Sc., Ph.D., C.Chem., F.R.S.C. (Chairman) Dr. D. A. Young, Ph.D., D.Sc., C.Phys., M.Inst. P. Editor: K. J. Wilkinson, B.Sc., M.Phil. Chemical Society Reviews (ISSN 0306-0012) appears quarterly and comprises approximately 20 articles (ca. 500 pp) per annum. It is intended that each review article shall be of interest to chemists in general, and not merely to those with a specialist interest in the subject under review. The articles range over the whole of chemistry and its interfaces with other disciplines. Although the majority of articles are intended to be specially commissioned, the Society is always prepared to consider offers of articles for publication. In such cases a short synopsis, rather than the completed article, should be submitted to the Managing Editor, Books and Reviews Section, The Royal Society of Chemistry, Burlington House, Piccadilly, London, W 1 V OBN.Members of the Royal Society of Chemistry may subscribe to Chemical Society Reviews at 515.50 per annum; they should place their orders on the Annual Subscription renewal forms in the usual way. All other orders accompanied with payment should be sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letch- worth, Herts. SG6 1HN England. 1985 annual subscription rate U.K. f45.00, Rest of World €47.50, U.S.A. $87.00. Air freight and mailing in the U.S.A. by Publica- tions Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. U.S.A. Postmaster: Send address changes to Chemical Society Reviews, Publications Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. Second class postage is paid at Jamaica, New York 11431. All other despatches outside the U.K. by Bulk Airmail within Europe, Accelerated Surface Post outside Europe. 0The Royal Society of Chemistry, 1985 AII Rights Reserved No part of this book may be reproduced or transmitted in any form or by any means -graphic, electronic, including photocopying, recording, taping, or information storage andretrievalsystems- without written permission from The Royal Society ofChemistry Published by The Royal Society of Chemistry, Burlington House, London, W1V OBN Printed in England by Richard Clay (The Chaucer Press) Ltd, Bungay, Suffolk

ISSN:0306-0012    

DOI:10.1039/CS98514FP009

出版商:RSC    

年代:1985

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS98514BP011

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

CENTENARY LECTURE * Phase Equilibrium and Interfacial Structure By B. Widom DEPARTMENT OF CHEMISTRY, CORNELL UNIVERSITY, ITHACA, NEW YORK 14853, USA 1 Introduction We shall discuss the equilibrium of fluid phases and the properties of the interfaces between them. We outline the theory that is current and see some of its successes, but the primary aim of this Lecture is to call attention to questions that are still unresolved. Of particular (but not exclusive) concern are the critical points of those phase equilibria. We recall briefly in Part 2 how the properties of fluids and their interfaces vary near critical points; i.e., what the critical indices (exponents) are. In Part 3 we describe the mean-field theory of inhomogeneous fluids, from which one obtains (approximately) the structure and thermodynamic properties of the interfaces between phases.Both these topics have been reviewed thoroughly and often,’-’ but we set out in Parts 2 and 3 the major ideas and formulae we need to refer to later, so that this Lecture may be self-contained. It is then through a sequence of special topics, which we treat in the subsequent parts, that we see what some of the theory’s successes and failures have been, what extensions of it are needed, and what some of the unsolved problems are. 2 Phase Equilibria and their Critical Points In Figure la we see schematically the equilibrium between two phases a and p. This could be the equilibrium of a liquid p with its vapour a, or that between two incompletely miscible liquids.In Figure lb is pictured the equilibrium of three phases, a, p, and y, separated in pairs by two interfaces. Here p and y may be two incompletely miscible liquids and a their common vapour, or all three may be liquids. The critical points of phase equilibria are limiting states in which phases become identical and the interface between them disappears. In Figure la these may be the * Delivered at a Symposium of the Faraday Division of the Royal Society of Chemistry on 10 December, 1984 at Imperial College, London. J. S. Rowlinson and B. Widom, ‘Molecular Theory of Capillarity’ Oxford, 1982.’D. Jasnow, Rep. Prog. Phys., 1984, 47, 1059. K. Binder, in ‘Phase Transitions and Critical Phenomena’, Vol. 8, ed. C. Domb and J.L. Lebowitz, Academic Press, 1983, pp. 1-144,467474. J. S. Rowlinson, Chem. SOC.Rev., 1983, 12, 251. B. Widom, Furuduy Symp. Chem. SOC.,1981, 16, 7. Phase Equilibrium and Interfacial Structure n U 1III1111I11111111 P IIIIIIIIIIIIIIIIIII 7 u (6) Figure 1 familiar critical points of liquid-vapour equilibrium or the consolute points of liquid-liquid equilibrium. In Figure 16, the state in which the two phases p and y become one at a critical point while the phase a remains in equilibrium with them but distinct, is called a critical endpoint; and there may similarly be a critical endpoint in which a and p become identical while in equilibrium with the distinct phase y. The limiting state in which the two critical endpoints occur simultaneously, so that all three of a, p, and y become one while both interfaces disappear, is called the tricritical point of the three-phase equilibrium.When the interface between two phases disappears at a critical point the tension (free energy per unit area) of that interface vanishes, and does so proportionally to a positive power of the distance from the critical point. If we express this distance, as is most often done, as the difference T, -Tbetween the temperature T of the phase equilibrium and the critical temperature T,, then the surface (interfacial) tension <T behaves as as T-T,, where q,and p are positive parameters. We assume in (1) an ‘upper’ critical point, so that distinct phases exist for T < T,. For a ‘lower’ critical point we replace 1 -T/T, in (1) by T/T, -1.For all ordinary critical points of two-phase equilibrium the exponent p has the universal value p = 1.26 (as closely as we can tell from current theory and experiment I), but the coefficient oois not universal. By the scaling theory of critical 0 is related to the coherence length 6 of density or composition fluctuations, and also to the difference Ap in density or composition of the coexisting phases and to the compressibility (mechanical or osmotic) x of the Widom phases. Each of these also behaves as a power of 1 -T/T, as the critical point is approached, (We must be careful not to confuse the exponents p and y with the names of the phases p and y.) Here pcis the density or composition at the critical point; v, p, and y are universal exponents: v = 0.63, f3 = 0.325, y = 1.24; while go, B, and xo are non-universal amplitudes.We shall later see the scaling-theory connections of cs to 5, Ap, and x. These will then imply relations among the universal exponents p, v, p, and y, as well as universal relations 'p7 among the otherwise non-universal amplitudes cso, to,B, and xo,and we shall examine these to see how well they hold in theoretical models and in experiment. A critical endpoint is an ordinary critical point. At an ap critical endpoint in the system in Figure lb, for example, the y phase is merely a spectator to the ap critical point and has no more effect on it than does the container, which is indeed yet another phase.Thus, (1)--(4) and the scaling relations among those quantities all continue to hold for the critical phases (but not for the spectator phase) at a critical endpoint, and they do so with the values of the exponents already quoted. A tricritical point, which, as we noted, is the coincidence of two critical endpoints and is a limit of three-phase coexistence (in contrast to an ordinary critical point or critical endpoint, which is a limit of two-phase coexistence), is different. While the relevant o,g,Ap, and x near a tricritical point are again as in (1)-(4) (with T, now the tricritical-point temperature), with exponents that are again universal, the values of the latter differ from those at an ordinary critical point: for tricritical points' p = 2, v = 1, p = 1/2, y = 2.We shall need to recall all this later, but for now we change the subject and outline the mean-field theory of inhomogeneous fluids. 3 Mean-field Theory The structure and tension of interfaces may in principle be calculated from statistical mechanics once the intermolecular forces are known; but in practice this is done only approximately, most often with the van der Waals theory or one or another of its immediate extensions, which, collectively, are called the mean-field theory (or appr~ximation).'-~ We shall see the theory here in its simplest context, that of liquid-vapour equilibrium across a planar interface in a one-component fluid. We think of the molecules as attracting hard spheres of diameter b.The potential energy of attraction as a function of the distance r between centres is cp(r),with the convention (p(.o) = 0. With z the distance perpendicular to and through the interface, the D. Stauffer, M. Ferer, and M. Wortis, Phys. Rev. Lett., 1972, 29, 345.'D. Stauffer, Phys. Left.A, 1973, 44, 261. Phase Equilibrium and Interfacial Structure density p is z-dependent, p = p(z), as in Figure 2, varying from the bulk gas-phase density p( -co) = peto the bulk liquid-phase density p(co) = p1as z varies from -00 to co. I -1----1-I I I I-.I liquid I I Iu I I I 6 Pl Figure 2 If we ignore correlations in the positions of the molecules other than the hard-sphere exclusions, the mean potential energy of attraction at any point Pat depth z is Jc cp(r)p(z')d~,r>b where dz is an element of volume at a variable point, z' is the depth at that point, r is the distance between that point and the point P, and the integration is over all such points outside a sphere of radius b about P (Figure 3).The mean-field approximation takes this potential energy to be the contribution the attractive forces make to the chemical potential of the fluid; there is imagined to be no entropic component of that contribution. At the same time the hard-sphere repulsions are taken to contribute phs[p(z)J to the chemical potential, where the function ph*(p)is the chemical potential of a fluid of the same hard spheres without attraction [q(r) = 01 at a uniform density p and at the temperature of the model fluid with attractions.If we represent by p [not to be confused with the exponent in (l)], with no superscript and no indicated argument, the uniform chemical potential of the two-phase fluid, then our approximation is The terms on the right-hand side of (5) are separately z-dependent but their sum is not.The contributions of the hard-sphere repulsions and of the attractions are Widom z c dT Figure 3 being taken to be separable (except for the hard-sphere exclusion in the range of integration) and additive. The former contribution is being taken to be local- phs(p) evaluated at the local p(z)-which is an approximation that would be accurate only if the density gradient dp(z)/dz was so small that p(z) was nearly constant over the distance 6.The effect of the attractions is treated in (5) as fully non-local-molecules at all distances r from P for which q(r) is sensibly different from zero contribute to the potential at P-but is otherwise approximated as that of a mean field. Equation (5) is a functional equation for the density profile p(z) in terms of the presumed known phs(p)and a given q(r). If the fluid were (hypothetically) constrained to be uniform with the density p its chemical potential, M(p),would by (5) be the analytic function in which a is van der Waals's a-parameter, In this mean-field theory the densities pgand p, of the bulk gas and liquid phases, at the chemical potential p, satisfy Also, in terms of this same function M(p)defined by (6), the functional equation (5) is P = M"l + l,bcp(r)CP(z') -P(Z)ldT (9) The profile p(z) is the solution of (5) or (9) that satisfies p(-co) = pg and p(co) = pI when pg and p1 are solutions of (8).When p(z) is slowly varying (as it is near the critical point, where the interface is diffuse), (9) may be further approximated by expanding p(z') about z' = z and Phase Equilibrium and Interfacial Structure truncating after second order. Then where m = br2q(r)dt-&l, [The coefficient of the first derivative, dp(z)/dz, vanishes by symmetry.] This is the simplest form of the van der Waals By analogy with the dynamics of a particle on a line, in which rn plays the role of the mass, z that of the time, p that of the co-ordinate, and M(p)-p that of the force, it is readily seen that the solution of (10) that satisfies p( f00) = ps,, when these, in turn, satisfy (S), is as in Figure 2.The p(z) determined by (9) may be seen by the methods of the variational calculus also to be that which minimizes a functional 0, = Jm {FCP(41 + +P(Z> I,bCp(r>CP(Z’)-p(z)ldrIdz (12) -00 where the function F(p) is defined in terms of M(p) -p by This together with (8) amounts to Maxwell’s equal-areas construction, which de- termines the pe,p,, and p of the coexisting phases once the temperature is specified. The integrand in the z-integration in (12) is the mean-field theory’s approximation to the excess free-energy density due to the inhomogeneity, i.e., arising from the z-dependence of p.Then o,the integral over z, is the excess free energy per unit area, which is the interfacial tension; and the equilibrium profile p(z) is then that which minimizes the tension. This identification of the minimal oin (12) with the equilibrium surface tension is confirmed by noting that it satisfies the condition, required by the Gibbs adsorption equation, that d(o/T)/d(l/T) be the excess surface energy per unit area when the Gibbs dividing surface is that of vanishing adsorption. Since the minimal CJ is extremal with respect to variation of p(z), in differentiating it with respect to T we need consider only the explicit temperature dependence of the integrand in (12) at fixed p(z) and not the implicit dependence arising from the T-dependence of the equilibrium p(z) itself.Furthermore, [phs(p) -p]/T may be expressed as the difference between a function of p alone, independent of T, and a function of Talone, independent of p. With these two observations we have from (6)-(8) and (12j( 14), Widom If we define a function pBI(z)to be pBwhen z is on the gas side of a dividing surface and pI when z is on the liquid side, then the location of the dividing surface of vanishing adsorption is determined by There is no reference in (15) to any dividing surface; but if we now specify (16) then (15) becomes = *Imd(a/T)/d(l/T) -CC J&~)CP(Z)P(Z’) -~~d4~1d.r (17)dz which is manifestly the excess surface energy per unit area in mean-field approximation.In the same square-gradient approximation that led to (lo), and after integration by parts converts -p(z)d2p(z)/dz2 to [dp(z)/dzI2, the functional 0in (12) becomes This is the ‘action’ in the dynamical analogy; the equilibrium 0may then be equally well expressed as the integral of the ‘momentum’ over the ‘co-ordinate’, In much of the current work on inhomogeneous fluids (3,(9), or (12) (they are equivalent), or some direct extension of them, is the starting point. They are immediately generalizable to potentials that include external fields, to multi- component systems, and to inhomogeneities that are more than one-dimensional, where the densities depend on two or all three spatial co-ordinates. An important example of a two-dimensional inhomogeneity is the line in which three phases meet, where the several densities (in a multi-component system) depend on the two co- ordinates in any plane perpendicular to that line.’ While the generalizations to include external fields, to multi-component systems, and to multi-dimensional inhomogeneities are straightforward, other generaliza- tions are not, and may call for correction to the basic mean-field idea.Even then this remains a useful theoretical framework. In the remaining parts of this Lecture, as we treat a succession of problems and call attention to some unanswered questions, we refer repeatedly to these ideas and formulae, as well as to those in Part 2 relating to critical points. 4 Surface Tension Near Critical Points We may use the theory of the preceding section to discuss further the critical-point behaviour of surface tension, which we outlined in Part 2.Near the critical point the interface is diffuse and the density or composition gradients are small, so for many purposes the version of the theory in (18) and (19) is adequate. Phase Equilibrium and Interfacial Structure When p(z) is the equilibrium profile the two terms SFdz and +mJ(dp/dz)2dz in (18) contribute equally to 6.Furthermore, on comparing this theory of interfacial structure with the Ornstein-Zernike theory of density or composition fluctuations near critical points,' one sees that the distance through the interface over which the variation of the composition from that of one bulk phase to that of the other mainly occurs-hence, the interfacial thickness, or the range of z over which the integrand in (18) differs sensibly from &is the same as the coherence (correlation) length 5of the fluctuations in the bulk phases, to which we referred in Part 2.Then from (18) we estimate o -Km(Ap)*/s, where K is some dimensionless proportionality constant and where Ap (= p1 -pe for a liquid-vapour interface) is just the density or composition difference referred to in Part 2. From the Ornstein-Zernike theory of fluctuations one knows that near the critical point with x the compressibility, as in Part 2. Thus, Here 6 and x may be evaluated in either bulk phase; near the critical point 5 is the same in the two phases, as is x. If we take the mean-field theory literally we may evaluate Ap/p, and x from (6)-(8) and (14) [and the thermodynamic identity x-l = p2(ap/ap),J, m from (ll), E, from (20), and 6from (6), (13), and (19).We find (lj(4) to hold with the classical values p = 3/2, v = 1/2, p = 1/2, and y = 1 of the exponents; we find explicit values of the coefficients oo,go, B, and xo;and for the dimensionless coefficient Kin (21), which is related to the coefficients in (1)-(4) by we find the explicit and universal value K = 1/6. These values of the exponents p, v, p, and y are not those quoted in Part 2; the mean-field theory does not give a quantitatively correct account of critical-point behaviour. We may reasonably suppose that if in (18) or (19) we used a function F(p)that we knew contained the right critical-point singularities, instead of that obtained from (6) via (13), we would get the right answers. [We would also need to replace the potential q(r)in(l1) by -kTtimes the direct correlation function c(r)to obtain m correctly; k is Boltzmann's constant. It is with this m, more generally, that the interfacial thickness 5,or the correlation length 5 of the Ornstein-Zernike theory, is related to x by (20).) The result is again of the form (21), but now with non-classical values of the exponents p, v, p, and y, and with a generally different but again universal value of Krelated to the non-universal amplitudes oo,xo, B, and co by (22).From (21) we find the relation p= -v+y+2p (23) Widom among these four exponents, which we may verify holds with the values of the exponents quoted earlier.Fluctuation theory gives (sp)l = p2kTx/v for the density fluctuations 6p in a macroscopic sub-volume v of a fluid. By critical-point scaling theory the fluctuations in density in a volume gd (d = dimensionality) in either of two coexisting, near-critical phases are of the order of the difference, Ap, of the densities of those phases. Thus, (Ap/pJ2 -K'kT, x/tdwith some universal proportionality constant K'. Also by the scaling theory the free energy associated with the elementary, coherent density fluctuation of coherence length 5 is kT; while the magnitude of the typical density fluctuation in the volume cd is Ap, so the free energy associated with it is at the same time of order 05"'.Hence, 06"' -K"kTc, with still another universal proportionality constant K". These give us the further exponent relations which are consistent with (23), and the further universal relations among non- universal amplitudes, which are consistent with (22) with K = K"/K'. The relations (24) with d = 3 may also be verified to hold with the quoted values of the exponents, which is an important success of the theory; but there is an as yet unresolved discrepancy between theory and experiment with respect to the universal relations among amplitudes that include 0,;i.e., in the values of K and K". Careful analyses of the experimental results '-' reveal that o0~,/B2C,and oo~~-'/kTcare indeed universal, but that their universal values are about twice as great as the theoretical estimates of these quantities ''-I3 in models whose critical properties have been thought to be identical with those of real fluids.The discrepancy is believed' to reside in 0,.Until it is resolved the present theory of surface tension near critical points cannot be accepted as final. We saw in Part 2 that the critical-point exponent for the vanishing of the interfacial tensions at a tricritical point is supposed theoretically to be p = 2, rather than p = 1.26 as at an ordinary critical point of two-phase equilibrium. Together with the other values of the tricritical-point exponents quoted in Part 2, the value p = 2 satisfies (23), and also (24) with d = 3. There has been no experimental determination of the exponent p for a tricritical point of ordinary three-phase equilibrium, so this theoretical prediction remains untested, except indirectly: the * M.B. Schneider, personal communication (1984). M. R. Moldover, Phys. Rev. A, 1985, 31, 1022. lo H. L. Gielen, 0.B. Verbeke, and J. Thoen, J. Chem. Phys., 1984, 81, 6154. K. Binder, Phys. Rev. A, 1982, 25, 1699. E. Brezin and S. Feng, Phys. Rev. B, 1984, 29, 472. l3 K. K. Mon and D. Jasnow, Phys. Rev. A, 1984, 30, 670. Phase Equilibrium and Interfacial Structure equilibrium between the superfluid and non-superfluid liquid phases of 3He-4He mixtures is believed to be thermodynamically equivalent to three-phase equilibrium in classical (we shall say more about this for pure 4He in Part 5); and measurements of the interfacial tension near the liquid-liquid consolute point of those mixtures l6 are consistent with p = 2.In the phase equilibria of microemulsions one often finds a microemulsion, containing large amounts of both oil and water (or brine) in a homogeneous solution with a surfactant, in equilibrium with a phase that is almost pure oil, or one that is almost pure water (or brine-r with both, in a three-phase equilibrium. The composition of the microemulsion is far from that of the phase or phases with which it is in equilibrium, so in that sense one is far from a critical point; yet the interfacial tensions are extraordinarily low 17-19 (l~-’-lO-~ dyne cm-’), as when near a critical point in systems without surfactant.We may understand that ’O from the theory in Part 3. From (19), where E is a typical value of F in the interface, or even, for order-of-magnitude estimates, the maximum value. Here p is any measure of chemical composition while F(p),from (6) and (13), is as in Figure 4, with p, and pethe values of p in the bulk phases. Near a critical point F*and Ap are both small. In the microemulsion phase equilibria that we are contemplating Ap is not small, but F(p)is nevertheless extremely smal121 over the whole interval pa Ip Ipe; so p* is small, and thence also o. F(p) Figure 4 5 Non-critical Interfaces Near Critical Endpoints Let us imagine in Figure lb that a is the vapour in equilibrium with two liquids p and y that are near their consolute point; or that a is a wall, or a solid adsorbent, while p and y are the liquid and vapour phases of a fluid near its critical point or l4 R.B. Griffiths, Phys. Rev. Left.,1970, 24, 715. R. B. Griffiths, Phys. Rev. B, 1973, 7, 545. l6 P. Leiderer, H. Poisel, and M. Wanner, J. Low Temp. Phys., 1977, 28, 167. l7 A. M. Bellocq, D. Bourbon, and B. Lemanceau, J. Disp. Sci. Tech., 1981, 2, 27. A. Pouchelon, D. Chatenay, J. Meunier, and D. Langevin, J. Coll. InterJ Sci., 1981, 82, 418. l9 A. M. Cazabat, D. Langevin, J. Meunier, and A. Pouchelon, Adv. Coll. Interf Sci., 1982, 16, 175. 2o H. T. Davis and L. E. Scriven, presented at the 55th Annual Fall Technical Conference and Exhibition of the Society of Petroleum Engineers of the AIME, Dallas, Sept. 21--24, 1980.” Y.Talmon and S. Prager, Nature, 1977, 267, 333. Widom again two liquid phases near a consolute point. In each case a is the spectator phase at the By critical point (critical endpoint; Part 2). The By interface is the one we discussed in Part 4, but here we concentrate on the non-critical ap or ay interfaces, and we ask, in particular, how the By critical point manifests itself in the tensions of those interfaces . We may again use the theoretical framework of Part 3, now with F(p) as in Figure 5, or even an explicitly two-density version of the theory for the two- TC TC T+ T+ Figure 6 component liquid rnixt~re.~~*~~ are as in Figure 6. Here G,,~~The results1*22*23 is the tension of the interface between the spectator phase a and the single Pr phase above the Py critical point at the critical density or composition, while oaBand oav are the separate tensions of the af3 and ay interfaces in the three-phase region below the By critical point. The latter tensions are related to each other and to the tension aBVof the critical interface by (Antonow’s rule l), as follows from Figure 5 and equation 19 (with ‘pg) and ‘pl’ taken to be pa, pe, or pV,as appropriate).Figure 6a is found when the fly critical point is incorporated in F(p)with classical l2 M. M. Telo da Gama, R. Evans, and I. Hadjiagapiou, Mol. Phys., 1984,52, 573. 23 M. E. Costas, C. Varea, and A. Robledo, Phys. Rev. Letf., 1983, 51, 2394. Phase Equilibrium and Interfacial Structure (mean-field-theory) exponents and Figure 66 when the exponents are non-classical.oay,and G,,~~,In each of the figures all three curves, o,,.,, are of the form in the immediate neighbourhood of the critical point, where p, as before, is the exponent in the critical surface tension oBy(so p = 3/2 or 1.26 in the classical and non-classical versions of the theory, respectively), and A is a common constant for all three curves, which are thus predicted to share a common tangent (the dashed line in each figure), while C is a different constant for each curve and of opposite sign on the two sides of the critical point. Thus, the deviations of the curves from their common tangent is predicted to be of one sign on one side of the critical point and of opposite sign on the other, as shown in the figures.This last feature, which appears as an inevitable and universal consequence of the theory, is not (or has not yet been) verified by experiment. There is an indication from theory that in Figure 6b the curve oaymay quickly turn upward on departing from the critical point and mostly lie above the tangent except very near the critical point itself. That could greatly complicate the interpretation of experiment. The experiments 24,25 have not yet been of sufficientprecision to resolve these questions, but, disquietingly, in some cases all three curves seem to lie above their tangent. We shall shortly pose a related question about the surface tension of liquid helium near its lambda point.U \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ P ’Pc P ‘Pc 1 Figure 7 l4 N. Nagarajan, W. W. Webb, and B. Widom, J. Chem. Phys., 1982,77, 5771.’’I. L. Pegg and I. McLure: I. L. Pegg, Ph.D. thesis, University of Sheffield, 1982. Widom Another prediction of theory that experiment has so far been too imprecise to test is the difference in the way curves of oa,Bvus. Tat fixed non-critical composition or density come in to the boundary of the three-phase region according to whether the fixed composition is on one side or the other of critical. In Figure 7 we see again the curves oavand oaBof Figure 6a, and also the curve there called oa,Bv,now drawn dashed and labelled p = pc.The other dashed curves are also the tensions of the a,Py interface in the two-phase region as functions of temperature, each now at a fixed, off-critical composition or density.Because of (27), when all three phases are in equilibrium the phase p wets perfectly, or spreads at, the ay interface. Then when Py is still a single, homogeneous phase in equilibrium with a, with a fixed composition on the y side of critical (p < pc, say, as in Figure 7), as T decreases and the phase-separation point is approached a premonitory layer of p forms at the interface with a and gradually logarithmically slowly-grows to macroscopic thickness as p comes ever closer to being stable in bulk. 1-3*5,26-29 In the limit there is then an infinite adsorption at the ay interface, and so, according to the Gibbs adsorption equation, a (logarithmic) =divergence in (do/~?p)~ -(ap/a~),(aT/ap),(ao/aT)~(now p is again chemical potential, not the surface-tension exponent), and so also in (do/aT),, since away from the critical point the other two factors are both finite. Thus, the dashed curve marked p < pc in Figure 7 comes into the limiting curve oavwith infinite slope,26 as shown.On the /3 side of critical no premonitory layer of y forms at the interface with a because the ap interface is not wet by y, so the curve marked p > pcin Figure 7 comes into the limiting curve oaBwith finite slope, as also shown. Direct measurements 24 of oa, have so far failed to reveal the predicted divergence of the slopes of the curves p < pc.The phenomenon we just discussed is not specifically connected to the critical point. On the contrary, as we see in Figure 7, the amplitude of the predicted singularity becomes smaller as the critical point is approached: the curve of oa* us. Tat p = pcshares a common tangent there with the curve oavus. T,as in Figure 6. On the other hand, when too far from the critical point we may be in a regime 26 in which oar< oaB+ oBvinstead of (27); then P no longer wets the ay interface, a premonitory p layer is no longer formed, and we would no longer expect the curves p < pc(Figure 7) to be of infinite slope where they meet the curve oey. Measurements of the adsorption-in effect, direct measurements of the derivatives of o-would probe such singularities more sensitively than do measurements of o itself. Adsorption measurements 30-33 have been called on to 26 J.W. Cahn, J. Chem. Phys., 1977,66, 3667. 27 R. Pandit, M. Schick, and M. Wortis, Phys. Rev. B, 1982, 26, 5112. 28 H. Nakanishi and M. E. Fisher, Phys. Rev. Lett., 1982, 49, 1565. 29 A. Robledo, ‘Las Transiciones Interfaciales’, preprint (1984). 30 (a)D. Beaglehole, J. Chem. Phys.. 1980,73,3366; Phys. Left.A, 1982,91,237; (b)A. V. Mikhailov, V. L. Kuz’min, and A. I. Rusanov, Kolloid. Zhur., 1984, 46,481. 31 B. Heidel and G. H. Findenegg, J. Phys. Chem., 1984,88, 6575. 32 J. Specovius and G. H. Findenegg, Ber. Bunsenges. Phys. Chem., 1980, 84, 690 S. Bliimel and G. H. Findenegg, Phys. Rev. Lett., 1985, 54, 447.33 G. H. Findenegg and R. Loring, J. Chem. Phys., 1984,81,3270; R. Loring and G. H. Findenegg, J. Coll. Inter-Sci., 1981, 84, 355. Phase Equilibrium and Interfacial Structure test another prediction of theory, this one concerned more specifically with behaviour near the critical point. If we follow oa,8vas a function of Ap at T = T, we find,’ as the counterpart of (28), 0ol,87 0c +D*IAPl”8 (29) with J.I and P the critical-point exponents already defined and with D, two different constants according as Ap >0 or Ap <0. The singular terms IT -T,JMin (28) and in (29) are equivalent to each other via critical-point scaling [cf: equation 31, but refer to two different paths of approach to the critical point: the first to the path p = p, with Tvarying and the second to the path T = T, with p varying.Then the =(ac~/ap)~adsorption r =(ao/a~.~)~ p2x [from the identity x-’ =~~(dp/dp)~ referred to in Part 41 at the a,Py interface behaves as r -IAplcu-a-Y)/8 we along the critical isotherm (T = Tc),from (29), (3), and (4); and thus, again by critical-point scaling, r -IT -Tclu-8-von the path p =p,. By (23) this is r JT-TJ-(V-B) = IT -~~1-0.3 (30) a result first obtained by Fisher and de Genne~.~~ Here, too, experiment and theory are not yet in perfect accord: with pand y(or Py) the liquid phases of aniline-cyclohexane mixtures and with a their vapour, 13eaglehole30“ found an exponent close to 0, more nearly consistent with a logarithmic divergence than with IT -Tcl-0.3,while in analogous systems Mikhailov et al.30b found the exponent 0.35 & 0.15; with pand y (or Py)the fluid phases of pure ethylene, and with graphite as the a phase, Specovius and Findenegg 32 did indeed find the exponent -0.3; but these have been superseded by the more accurate measurements of Bliimel and Findenegg 32 with SF, in place of ethylene, and the result is an exponent closer to -0.5. For propane on two forms of graphite the exponents that are found are also in this range, -0.31 and -0.44 (Findenegg and Loring33). The theory may well be right, but it is not yet proved. The surface tension of liquid helium near its lambda point TAfalls in this class of phenomena, because that point is a critical point at which the superfluid order parameter yr-the analogue of our earlier Apfirst appears; then, as T falls below T,, Iwl increases from 0, as in (3).There do not appear two spatially separated liquid phases P and y, so there is no identifiable Pyinterface; also there are no separate aP and ay interfaces: there is only one liquid-vapour interface below as well as above TA. The appropriate theory3’ is one that recognizes explicitly both the density p, which differentiates liquid and vapour, and the superfluid order parameter yr, and is thus a two-component (or multi-component, depending on the dimensionality of w) generalization of the theory outlined in Part 3. The result is again as in (28). The helium lambda point is not exactly of the class of ordinary, two-phase critical points, so the value of the exponent p is slightly different from 1.26-now closer to 1.35-but that is a minor difference.The mean-field theory 35 of this liquid-vapour 34 M. E. Fisher and P. G. de Gennes, C. R. Hebd. Seances Acad. Sci., Ser. B, 1978, 287, 207. 35 P.Tavan and B. Widom, Phys. Rev. B, 1983, 27, 180. Widom interface again gives p = 3/2, but it gives a coefficient C in (28) that is non-zero only for T < T,; above T, there are only integer powers of T -TAin G. This is as in an earlier theory by S~byanin,~~ but Hohenberg 37 has argued that the singular term IT -T,lwshould appear in CJ on both sides of the 1-point. The question is still unresolved theoretically. E~perimentally,~~ it appears that the singular term is present on both sides, and that the data are well fit with the power p = 1.35.Both by theory 3s and experiment 38 the curve CJ us. Tlies entirely above the line that is tangent to it at T = T,, as shown schematically in Figure 8. This differs from the result in Figure 6. There we distinguished two curves, G,~and oay,for T < T,, while here for T < TA there is only one, but that does not account for the difference, which is still not clearly understood. The difference is not due only to there being here two densities, p and w, instead of one, for some of the calculations for the ordinary critical endpoint were also based on two-density versions of the theory.22 What may be the significant difference between the former and present problems is that in helium the vapour phase is symmetrically related to the superfluid ‘phases’ below T, while at an ordinary critical endpoint the distinct phase is unsymmetrically related to the two near-critical phases; but if this is the explanation it has yet to be convincingly demonstrated.t Q TA T-Figure 8 6 Long-range Forces Does it matter very much in the theory based on (9) or (12) whether the intermolecular potential q(r) vanishes proportionally to l/r6, say, at large r, or is 36 A. A. Sobyanin, Sou. Phys. JETP, 1972,34, 229. 37 P. C. Hohenberg, J. Low Temp. Phys., 1973, 13, 433. J. H. Magerlein and T. M. Sanders, Jr., Phys. Rev. Leu.,1976, 36,258. Phase Equilibrium and Interfacial Structure much shorter ranged than that? There have been some surprising answers.According to the mean-field theory it is only in an inhomogeneous system that it could matter. In a homogeneous fluid the theory reduces simply to the equation of state [cf: equation 61 with a given by (7). Thus, any two cp(r)that have the same integral outside hard cores of equal diameter b lead to the same equation of state; whether cp is long-ranged or short-ranged is irrelevant, as long as it is integrable. Even in an inhomogeneous system, if the square-gradient approximation (10)or (18) were sufficient, long-rangedness of cp(r) would again have no effect, provided only that cp were not so long-ranged that its second moment m failed to exist (equation 11); i.e.,provided that cp(r)vanished faster than l/r5(as l/r6,for example) as rd co.Then any two fluids in which the molecules had the same core diameter b and attractive potentials with the same integral and same second moment would behave identically, however different the ranges of those potentials. For an ordinary interface between two bulk phases, such as that whose profile is shown schematically in Figure 2, even when it is necessary for quantitative accuracy to go beyond the square-gradient approximation, neither the profile nor the tension is markedly affected by the range of cp as long as a and m are fi~ed.~~.~' To be sure, the rapidity of approach of p(z) to pe or p1as lzl-co does depend on the range of cp. When cp(r) is of strictly finite range [i.e., when cp(r) 3 0 for r greater than some fixed, microscopic distance], or when it vanishes exponentialy rapidly as r --+ 00, then p(z) -ps or p1 -p(z) vanishes exponentially rapidly as 121-co, proportionally to exp( -lzl/t), with 5 the correlation length we encountered earlier, which, near the critical point, is related to the second moment of cp and to the compressibility x by (20).When cp(r),instead, vanishes proportionally to -cpo/r"for large r, with some constant cpo > 0 and with n > 3, as lzl+ co,where g,l means either g or 1 consistently, and where pg,, and xg,, are the density and compressibility of the bulk phases.Thus, the decay is as l/lzJ"3 instead of exponential; but if n > 5 (so that cp has a second moment), by the time the power-law decay is perceptibly different from the exponential, p(z) is already so close to its asymptotic limit of ps or p1that the difference is of little consequence. It is entirely different for a slab with two interfaces, as when in Part 5, for example, we dealt with a premonitory layer of a wetting phase p, let us say of thickness 1, at an ay interface (Figure 9).The importance of long-range forces in such geometries has been emphasized by de Genne~;~' in particular, the range of the forces is found42-57 to be crucial in determining the order of the 'wetting j9 J. A. Barker and J. R. Henderson, J. Chem. Phys., 1982, 76, 6303. *O B. Q.Lu,R. Evans, and M. M. Telo da Gama, preprint (1985). 41 P. G. de Gennes, J. Phys. (Paris), Lett., 1981,42,L-377.Widom transition’ 26-29.58-60 which is a transition in the structure of the cry interface7 between microscopic and macroscopic thicknesses 1 (Figure 9). a Y Figure 9 Tarazona et ul.42-44 have made the important remark that when (9) (or its generalization for multi-component systems) is applied to the configuration of Figure 9, and the equilibrium profiles p(z) are found by solving (9) iteratively, the solution converges quickly, after only a few iterations, to its final form at the ap and Py faces of the p film, but only very slowly to the final equilibrium thickness 1. This then allows one to separate the short-range and long-range aspects of the film structure, and to define an effective film free-energy per unit area, o(Z), by first specifying an F(p) suitable to the problem at hand and then substituting for p(z) on the right-hand side of (12) a roughly correct form with variable thickness 1. The equilibrium 1 is then that which minimizes o(Z).In this way4244*59*60 problems such as that of the wetting transition are reduced to much simpler problems of one, or very few, degrees of freedom. It is related also (although it is more phenomenological and less microscopic) to the way one constructs an interface Hamiltonian,61 which may again be a function o(Z), or a functional o[l(x,y)]with x and y the co-ordinates in a plane parallel to the interfaces in Figure 9. *’ P. Tarazona and R. Evans, Mol. Phys., 1983, 48, 799. 43 P. Tarazona, M. M. Telo da Gama, and R.Evans, Mol. Phys., 1983, 49, 283. ** P. Tarazona, M. M. Telo da Gama, and R. Evans, Mol. Phys., 1983,49, 301. *’ P. G. de Gennes, C. R. Hebd. Seances Acad. Sci., (II), 1983, 297, 9. *6 E. H. Hauge and M. Schick, Phys. Rev. B, 1983, 27, 4288. 47 M. P. Nightingale, W. F. Saam, and M. Schick, Phys. Rev. Lett., 1983, 51, 1275. *8 M. P. Nightingale, W. F. Saam, and M. Schick, Phys. Rev. B, 1984, 30,3830. 49 R. Lipowsky and D. M. Kroll, Phys. Rev. Lett., 1984,52,2303; D. M. Kroll and T. F. Meister, Phys. Rev. B, 1985, 31, 392. ” G. F. Teletzke, L. E. Scriven, and H. T. Davis, J. Chem. Phys., 1982, 77, 5794. G. F. Teletzke, L. E. Scriven, and H. T. Davis, J. Chem. Phys., 1983, 78, 1431. ” R. E. Benner, Jr., G. F. Teletzke, L. E. Scriven, and H.T. Davis, J. Chem. Phys., 1984, 80, 589. ” V. Privman, J. Chem. Phys., 1984, 81, 2463. ’* M. P. Nightingale and J. 0.Indekeu, Phys. Rev. Lett., 1985, 54, 1824. ” S. Dietrich, M. P. Nightingale, and M. Schick, preprint (1984). ”S. Dietrich and M. Schick, Phys. Rev. B, 1985,31, 4718. ” C. Ebner, W. F. Saam, and A. K. Sen, preprint (1984).’* D. E. Sullivan, Faruduy Symp. Chem. Soc., 1981, 16, 191.’’ D. E. Sullivan and M. M. Telo da Gama, in ‘Fluid Interfacial Phenomena’, ed. C. A. Croxton, Wiley, 1985. 6o P. G. de Gennes, ‘Wetting: Statics and Dynamics’, preprint (1984). 61 D. A. Huse, W. van Saarloos, and J. D. Weeks, preprint (1984). Phase Equilibrium and Interfacial Structure We may illustrate this procedure by an example in which the p film of Figure 9 would not be quite stable as a bulk phase, there being some positive undersaturation-free-energy per unit volume, say f, which would vanish if the film were stable in bulk.We consider the example in which the mass density dBof the wetting film is greater than the density d, of the y phase,62 which is itself of macroscopic thickness H 9 I (Figure 10); then f= (do -dy)gH with g the acceleration of gravity. For simplicity, and just as a schematic illustration of the method, let us assume a common q(r) for all the interactions and then take the number densities pa, Po, and p, to be such that is the intermediate one, say pa < 06 < pr (unlike the mass densities, which are in the order da < dy< do).In this way the bulk phases are ordered a, y, p in the gravitational field and yet the interfacial energetics are such as to favour the order a, 0, y. This leads to the configuration in Figure 10 with a microscopic but positive I, as we shall see.a z=o B z=l z c Y z=I+H B Figure 10 We now take (12) to be o(l) =P + +Jm J cpWCP(Z') -P(Z)lP(Z)dT dz (33) -a, r>b with p(z) = pa for -a< z < 0, p(z) = for 0 < z < 1, and p(z) = p, for 1 < z < co,as is appropriate for the asymptotic limit of a macroscopically thick y phase (H-00). Then from (33), where j(r) --2~cpo/(n-2)(n -3)(n -4)1& (36) 62 OD. Kwon,D. Beaglehole, W. W. Webb, B. Widom, J. W. Schmidt, J. W. Cahn, M. R. Moldover, and B. Stephenson, Phys. Rev. Lett., 1982, 48, 185.Widom as I----, 00. The o(l)of (34), with thisj(l) and n = 6, is as given by de Gennes41 (See also Tarazona et ~l.,~~Sullivan and Telo da Gama," Dzyaloshinsky et ~l.,~~" and Kuni et ~1.~~~).Since p6 is here assumed intermediate between pa and pv,the coefficient ofj(l) in (34) is negative, hence l/r" appears in o(l)in the large-l limit with positive coefficient (because cpo > 0 and n > 4). Thus, whenfis small o(l)has its minimum at a large but microscopic 1, With n = 6 andf = (d6 -d,)gH this, too, is as found by de Genne~.~' When three bulk phases are in equilibrium and their tensions are related by (27), one of them, as we saw, spreads at the interface between the other two (the interface of highest tension), which, with 1 now macroscopic, is the configuration in Figure 9.But when the tensions are related by ouv< csa6 + 06vinstead of by (27), as already remarked, this is not so; instead, the three phases meet with non-zero (and non- 1800) contact angles at a common line of three-phase contact.'*64 (One of the contact angles becomes 180" in the limiting case in which one phase is a non- deformable solid.) De Gennes 6o has called attention to the potential importance of long-range forces for the structure and energetics of such a three-phase line. There are not yet any calculations comparable to those of Tarazona et al.,42-44say, to demonstrate that the properties of that line are qualitatively different for long- range forces from what they are for short-range forces (although, in principle, .~~calculations like those of Benner et ~1would be able to do that).The point, again, is not merely to show that different cps lead to different results but that long- rangedness per se has unique implications; i.e., that even if a short- and a long- ranged q(r) shared the same zeroth and second moments a and m they would still have qualitatively different consequences. That surprising result has been amply demonstrated for films and is thus plausible also for the three-phase line, but the question is still open. 7 Conclusion Although we have along the way seen some of the successes of current theories of phase equilibrium and interfaces, the primary purpose of this Lecture has been to call attention to discrepancies between theory and experiment, to key theoretical predictions that are still untested by experiment, and to important questions that are still unsettled even by theory.These have included the problem of the coefficient 00 in the surface tension near the critical point (defined by equation 1);the critical- point exponent for the vanishing of the interfacial tensions at a tricritical point; the signs of the singular terms in the tension of the non-critical interface near a critical endpoint, and the critical-point exponent for the divergence of the adsorption at that interface; the way in which a growing layer of a phase p at the interface between two other phases a and y manifests itself in the tension of the ay interface; the form 63 (a)I. E.Dzyaloshinsky, E. M. Lifshitz, and L. P. Pitaevsky, Adv. Phy., 1961, 10, 165; (6) F. M. Kuni, A. I. Rusanov, and E. N. Brodskaya, Colloid. J. USSR (Engl. Trunsl.), 1969, 31, 691. 64 R. E. Benner, Jr., L. E. Scriven, and H. T. Davis, Furuduy Symp. Chem. Sor., 1981, 16, 169. 139 Phase Equilibrium and Interfacial Structure of the singularity in the surface tension of liquid helium at its lambda point; and the effect of long-ranged interactions on the structure and energetics of the line of three-phase contact. This has left untouched two further vast areas of inquiry about interfaces, each with important unsolved problems of its own. One of these is the necessary modification 65-70 of the mean-field theory to take account of long-wavelength capillary waves, and includes the problem of whether there can be a unique deconvolution of an interfacial profile into capillary-wave and ‘intrinsic’ stru~ture.~’-~~The other is the problem of curved surfaces-the surfaces of drops and bubbles of finite size-on which there has been much recent progres~,~*~~-’~ but where there remain some paradoxes and uncertainties.Clearly our subject is still a lively one, and has enough unanswered questions to stay that way for a long time. Acknowledgement. This work was supported by the National Science Foundation and by the Cornell University Materials Science Center. The author thanks Dr. M. M. Telo da Gama for helpful conversations. 65 J. D. Weeks, J. Chem. Phys., 1977,67, 3106. 66 R.Evans, Adv. Phys., 1979,28, 143. 67 F. F.Abraham, Phys. Rep., 1979,53, 93. 68 J. D.Weeks, D. Bedeaux, and B. J. A. Zielinska, J. Chem. Phys., 1984,80, 3790. 69 D.Bedeaux and J. D. Weeks, J. Chem. Phys., 1985,82, 972. ’O J. K.Percus, in ‘Studies in Statistical Mechanics’, ed. E. W. Montroll and J. L. Lebowitz, North-Holland, 1985. D.Jasnow and J. Rudnick, Phys. Rev. Lett., 1978,41,698. 72 R. Evans, Mol. Phys., 1981,42, 1169. 73 D.B. Abraham, Phys. Rev. B, 1984,29, 525. 74 D.A. Huse, preprint (1984). l5 S. J. Hemingway, J. R. Henderson, and J. S. Rowlinson, Faruduy Symp. Chem. SOC..1981,16, 33. 76 J. S. Rowlinson, J. Chem. Soc., Faraday Trans. 2, 1983,79, 77. 77 J. S. Rowlinson, J. Phys. A, 1984, 17, L357. M. P. A. Fisher and M. Wortis, Phys. Rev. B, 1984,29, 6252. 79 P. Phillips, Mol. Phys., 1984,52, 805.

ISSN:0306-0012    

DOI:10.1039/CS9851400121

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

TILDEN LECTURE* The Collision Dynamics of Vibrationally Excited Molecules By Ian W. M. Smith DEPARTMENT OF PHYSICAL CHEMISTRY, UNIVERSITY CHEMICAL LABORATORIES, LENSFIELD ROAD, CAMBRIDGE CB2 1EP 1 Introduction Traditionally, chemical kinetics has been concerned with the rates of chemical change at defined temperatures. There is, however, at least one class of chemical reaction where the distribution of reagents over states can become markedly non- Boltzmann, even under quite ordinary conditions. At low total pressures, the states above and close to the threshold for unimolecular reaction become depleted as collisional redistribution of molecules into these levels fails to keep pace with their unimolecular rate of reaction. For bimolecular reactions, such effects are thought to be absent.At least for processes involving simple species at low and moderate temperatures, the reagents are predominantly in their lowest vibronic states. The variation of rate constant with temperature is generally believed to reflect the change of reaction cross-section with collision energy (and, possibly, rotational energy) and energy redistribution will occur more rapidly than reaction. Unfortunately, the dynamical details of reactive collisions are separated from the thermal rate constant by many ‘layers of averaging’. One result of this is that it is essentially impossible to recover the form of the excitation function (i.e.how the reaction cross-section varies with collision energy) by measuring how the rate constant depends upon temperature.The situation in regard to collisional transfer of vibrational energy is rather similar. Early experimental methods, using ultrasonic absorption or dispersion, the photoacoustic effect, and shock tubes, probed bulk relaxation: that is, the overall transfer of energy between vibrational (V)and translational, rotational (T,R)degrees of freedom. It soon became apparent that bulk relaxation is virtually always controlled by the rate at which molecules are transferred between their two lowest levels.’ At the other extreme, the rates of unimolecular dissociations (or the reverse association reactions) and their dependence on ‘third-body’ gas provide some measure of energy transfer rates at internal energies at the dissociation limit.Measurements on the transfer of molecules between levels of intermediate excitation are rare. Undoubtedly optical excitation provides the best, most controlled technique for * Delivered at a Symposium of the Faraday Division of the Royal Society of Chemistry on 9 February, 1984 at the Scientific Societies’ Lecture Theatre, London. 1. W. M. Smith, ‘Kinetics and Dynamics of Elementary Gas Reactions’, Butterworths, London, 1980. J. D. Lambert, ‘Vibrational and Rotational Relaxation in Gases’, Clarendon, Oxford, 1977; J. T. Yardley, ‘Introduction to Molecular Energy Transfer’, Academic, New York, 1980. The Collision Dynamics of Vibrationally Excited Molecules perturbing a Boltzmann distribution over molecular states.We are not concerned here with photochemical dissociation but rather with the promotion of a molecular species to a specific rovibronic state as the result of irradiation of the species with a frequency corresponding to one of its discrete absorptions. Where electronic excitation is involved, laser sources are not essential-electronic photochemistry has a long history-although the power, tunability, monochromaticity, coherence, short duration, and polarization of lasers have enormously extended the range and potency of such experiment^.^ Studies in oibrational photochemistry, however, rely on laser sources. The great power of some infrared sources-notably the CO, laser-has opened up an entirely new field of research: namely, that of the phenomenon of infrared multi-photon absorption and the isotopically selective dissociation that this process can bring about.4 I shall not deal at all with this topic: partly because my group have not been involved in this work and partly because the multiphoton excitation process is not state-selective; after the laser pulse, molecules are left in a distribution over many highly energized states.Absorption of single photons from a laser can lead to the immediate overpopulation of single rovibrational quantum states in small molecules. In principle, the evolution in time of the populations of the initially excited state and of neighbouring states can be observed spectroscopically, and state-specific rate constants for energy transfer and reaction can be deduced, if the number of levels involved is small and the number of independent observations sufficiently large.5 In practice, mapping the pathways for relaxation and reactive loss at the state-to-state level proves very difficult, and indirect reasoning has frequently to be used to explain the limited number of observations that are possible.In particular, the role of molecular rotations in reactions and in vibrational energy-transfer processes is difficult to assess-at least, from bulk experiments like our own, where rotational relaxation is rapid. Experiments that combine molecular beam and laser excitation or detection techniques can provide information at this level of detail. Two beautiful experiments that have been reported recently illustrate what can be done.Altkorn et a1.6 have used one laser (an optical parametric oscillator) to excite HF molecules to specific levels u = 1, J and a second laser to detect the u’S distribution of the CaF molecules that are formed in the reaction Ca + HF(u = 1,J) -CaF(v’,J) + H (1) under single collision conditions. In a different e~periment,~ beams of He and I, collide with relatively high and variable collision energy and the final states of I, are M. A. A. Clyne and I. S. McDermid in ‘Dynamics of the Excited State’, ed. K. P. Lawley, Ado. Chem. Phys., 1982, 50, 1. D. S. King in ‘Dynamics of the Excited State’, ed. K. P. Lawley, Adu. Chem. Phys., 1982, 50, 105. C. B. Moore and I. W. M. Smith, Faraday Discuss. Chem. SOC.,1979, 67, 146.R. Altkorn, F. E. Bartoszek, J. Dehaven, G. Hancock, D. S. Perry, and R. N. Zare, Chem. Phys. Lett., 1983, 98, 2 12. ’G. Hall, K. Liu, M. J. McAuliffe, C. F. Giese, and W. R. Gentry, J. Chem. Phys., 1983,78,5260. Smith probed by laser-induced fluorescence. This allows one to extract cross-sections for transfer between specified u,J states at defined collision energies. The observed result of any collection of molecular collisions is ultimately a reflection of the forces which act between the atoms during the collisions. Physical chemists prefer to think in terms of how the potential energy-the overall electronic binding energy-varies with the nuclear configuration.' In even the simplest system of interest-a total of three atoms labelled A, B, and C-the potential energy (V) depends on three spatial co-ordinates (frequently chosen as the internuclear separations rAB, rBC,and rCA)and V(rAB, rBc, rCA)is properly referred to as a potential energy hypersurface-though this rather unwieldy term is often shortened to potential energy surface.It is useful, not just within the context of this lecture, to distinguish three broad types of potential energy surface. The first arises when the two colliding species both have closed electronic shells. (I shall call such species molecules, even when they are monatomic, to distinguish them from radicals, which have one or more unpaired electrons.) Collisions between Ar atoms and N, atoms in their singlet ground states serve as an example.The interaction gives rise to only one singlet potential energy surface. There is only weak, Van der Waals', attraction (~/k2: 106 K), and the form of the potential reflects the strong bond between the N atoms, and the strong repulsion between N, and Ar as their electron clouds overlap. The potential energy surface in Figure 1 represents this case with the atoms collinear. V BC R X Figure 1 Typicalpotential energy surface for the collinear collision of a diatomic molecule BC (e.g., N,) with a noble gas atom A. Contours join points of equal potential energy; R is the distance from A to the centre-of-mass of BC; and the curve on the right indicates how V varies along the path of minimum energy, which is shown on the potential energy surface by the dashed line Clearly, no chemical reaction can occur in collisions of this kind at moderate energies.However, energy might be transferred. A simple (classical) argument can be used to explain qualitatively why the transfer of vibrational energy in N, + Ar collisions is extremely inefficient. Note that the approach of an Ar atom towards N, tends to compress the N, bond; but only very slightly until high energies are reached. Now imagine a collinear classical collision between a non-vibrating N, molecule and an Ar atom. At low and moderate collision energies, the incoming The Collision Dynamics of Vibrationally Excited Molecules and outgoing trajectory will scarcely deviate at all from the path of minimum energy, represented by the dashed line in Figure 1: there will be essentially no transfer of energy from relative translation to vibration.Only at very high collision energies, will inertia (or the ‘bobsleigh’ effect) cause much difference between the incoming and outgoing trajectories. In practice, diatomic molecules are only transferred between neighbouring vibrational levels with low probability, especially when the states are relatively widely separated and the collision energy (or temperature, in a thermal experiment) is low. Transfer is more probable the more impulsive the collision: consequently, the probabilities of energy transfer decrease along the sequence He to Xe. V ‘BC r X f~~ Figure 2 Typicalpotential energy surface for the collinear collision of a diatomic molecule BC (e.g., HF) with an atomic radical A (e.g., CI). As in Figure 1, the curve on the right shows the variation of V along the path of minimum energy, represented on the potential energy surface by the dashed line The second type of system whose collision dynamics I wish to consider has a potential energy surface like that shown in Figure 2.This form of surface arises when a radical and a molecule interact. Then if A is the radical, it is usually possible for an atom-transfer reaction, A + BC --* AB + C, to take place. Reaction occurs in an A + BC collision if the system surmounts the potential barrier on the surface. For reaction in the exothermic direction, e.g., F + HC1-HF + C1, this barrier is nearly always less than 10% of the energy of the bond being broken in the reaction-and is frequently very small.For systems where A and C are identical, reaction can only be experimentally observed by isotopic substitution, but vibrational relaxation may be enhanced substantially by the atom-transfer process. A third case arises when two radicals interacts8 Now it may be especially important to remember that more than one potential energy surface can correlate with the separated species, because of the existence of degenerate and near-degenerate states in species with open electronic shells. However, it is likely that at least one surface has a deep minimum or ‘well’, as electrons on the two radicals ‘pair up’ te form a chemical bond. Collisions on a potential energy surface like that shown in Figure 3 are likely to lead to the formation of collision complexescapable of surviving for at least several vibrations.The fate of these complexes can depend Smith on conditions as well as on the detailed form of the surface. When there are different pathways for dissociation, the most exothermic channel is likely to predominate. For example, the reaction: 0 + OH -O2 + H; AH,, = -70 kJ mob' (2) probably proceeds via an HO,? complex which dissociates preferentially to 0, + H rather than back to 0 + OH. On the other hand, the only simple dissociation channel for HONO, complexes from OH + NO, is redissociation; this process only competes against collisional stabilization of the adducts. However, when one of the radicals forming a collision complex is initially vibrationally excited, the association-dissociation process is very likely to be accompanied by relaxation, and consequently vibrational energy transfer is unusually rapid in radical-radical collision.8,9 V V X X Figure 3 Typical potential energy surface for the collinear collision of a diatomic radical BC (e.g., 0,)with an atomic radical A (e.g., H).The curve on the right shows how V varies along the path of minimum energy for a case where the (ABC)? complex can dissociate to A + BC, or to AB + C; the curve on the left shows the variation of V when only one such dissociation channel is accessible My main aims in this lecture are to describe how reactive and inelastic processes involving vibrationally excited molecules may be studied, and to see how the rates of such processes can be related to fundamental properties of the system in question-such as the form of the potential energy hypersurface.For each type represented by the potentials in Figure 1,2, and 3, I shall consider systems studied recently by my own group, but the general conclusions will be based on a much wider range of data. An experimental technique used in several of our measurements is that of infrared fluorescence or laser-induced vibrational fluorescence (LIVF).,.' O In a LIVF experiment, a laser pulse is used to excite a fraction of molecules in a * M. J. Howard and I. W. M. Smith, Prog. Reac. Kinet., 1983, 12, 55. M. Quack and J. Troe, Ber.Bunsenges. Phys. Chern., 1975,79,170 M. Quack and J. Troe, Ber. Bunsenges Phys. Chem., 1977, 81, 160. lo I. W. M. Smith in 'Lasers as Reactants and Probes in Chemistry', ed. W. M. Jackson, Howard University, 1984, in press. The Collision Dynamics of Vibrationally Excited Molecules gas sample to a defined vibrational level. (Usually the laser output contains only one or a few frequencies that coincide with molecular absorption frequencies, but any rotational state specificity is rapidly destroyed.) The lifetime of the vibrationally excited molecules is determined by observing the intensity of infrared fluorescence as a function of time: i.e., IF([).In the simplest case, an exponential decay of the fluorescence from the initially excited level is observed: ZF(t) = ZF(t = 0) exp( -kl,,t) where klst = k, + Ck,[Q] with k, being the effective first-order constant for removal by radiative decay and the ka[Q] are pseudo-first-order constants for removal of the molecules from the initial excited level in collision with Q, a component of the gas sample; e.g.AB(u) + Q -AB(d < u) + Q (3) Second-order constants can be found by studying how klst varies as the composition of the gas sample containing AB is systematically changed. In general, k, is ca. lo2 s-' and makes only a small contribution to kist. To exploit the LIVF method fully, one needs a powerful pulsed laser source which is tunable so that a variety of molecular levels can be accessed. Even with such a source it is difficult to promote adequate concentrations of molecules to other than low-lying levels.Most molecules in equilibrium samples are in the zeroth vibrational level and overtone and combination bands are weak. Even when higher excitations are possible it may be difficult to differentiate between fluorescence from the initially excited level and that from lower, but still excited, levels which are populated as the molecules relax down the manifold of levels. Finally, it has to be recognized that the method measures total loss from a particular level: it does not discriminate between reaction and relaxation in those cases where Q and the excited molecules can react, nor does the method usually provide direct evidence about the mechanism for relaxation, that is the states to which the excited molecules are transferred.2 Collisions Between Molecules Collisions between diatomic molecules and rare gas atoms are relatively simple to treat theoretically. Unfortunately, virtually all the experiments on vibrational relaxation in such systems probe transfer between the lowest levels of the diatomic, u = 0 and u = 1. As well as overcoming the difficulty of exciting molecules to levels above u = 1 conditions have to be reached so that relaxation is predominantly by V-T,R energy transfer in collision with the rare gas, rather than by the intrinsically much faster V-V processes like AB(u) + AB(u = 0)+AB(u -1) + AB(u = 1). (4) In this section, I shall give a brief review of experiments that we have carried out on the vibrational relaxation of HCN, which has interesting similarities to, and Smith differences from, the hydrogen halides.Furthermore, it is sufficiently simple that ab initio theory or theoretical analysis of the rovibrational spectra of HCN and its isotopic variants may provide a good intramolecular potential, and wavefunctions that accurately describe the nature of its vibrational eigenstates. From LIVF experiments we have determined rate constants for relaxation of HCN from four different vibrational levels: (Ool)," (Oo2),12 (01l),I3 and (l0l).l4 This has been made possible by a high energy opticazparametric oscillator (OPO). This device, pumped by 100 mJ pulses from a Nd:YAG laser system (J. K. Lasers Ltd.) generates pulses of 1-5 mJ energy, 10-20 ns in duration, with a linewidth of ca.0.2cm-'. The tuning range of our OPO is shown in Figure 4, where it is displayed alongside the vibrational energy levels of HCN. The solid lines on the energy diagram show which absorptions have been successfully used in our LIVF measurements. Photon energy / lo3cm-121 121 7, 300 111 090 111 1 3 5 Output energy per pulse / m J "1 v2 v3 Figure 4 Output of the optical parametric oscillator compared with the vibrational energy levels ofHCN. The vertical arrows show the absorption bands that have been directly excitedin our LIVF experiments;' '-14 the diagonal arrows show the emission bands that have been observed A common feature of the four HCN levels that have been excited is that they all contain at least one quantum of energy in the v3 mode, which is essentially the C-H stretch.The v3 fluorescence lies at ca. 3 pm, the best region for observing time- resolved infrared fluorescence. However, when an overtone (002) or combination (011 or 101) level is excited, the fluorescence at 3 pm may include contributions from molecules in levels other than that initially excited. Fortunately, judicious use l1 G. S. Arnold and I. W. M. Smith, J. Chem. SOC.,Faraday Trans. 2, 1981,77, 861; G. S. Arnold, R. P. Fernando, and I. W. M. Smith, J. Chem. Phys., 1980, 73, 2113. P. W. Hastings, M. K. Osborn, C. M. Sadowski, and I. W. M. Smith, J. Chem. Phys., 1983,78, 3893. l3 B. D. Cannon and I. W. M. Smith, Chem. Phys., 1984,83,429. l4 B.D. Cannon, J. S. Francisco, and I. W. M. Smith, Chem. Phys., 1984, 89, 141. The Collision Dynamics oj-Vibrationally Excited Molecules of a 'cold gas filter' (CGF)13 differentiates between v3 emissions arising from different states. The CGF consists simply of a small gas cel! containing HCN. It is placed between the fluorescence cell and the detector. The pressure of HCN in the filter can be adjusted either to absorb emissions from all levels with u3 = 1l2 or to remove just the (001,OOO) fluorescence.' Comparison of the infrared fluorescence signals with and without the CGF allows one to determine rates of relaxation both from the initially excited state and from (001). Moreover, in some cases, the route for relaxation can be deduced. Rate constants have been determined for the removal of energy from vibrationally excited HCN by a variety of collision partners.' O-' Table 1 summarizes only the data for collisions with noble gases.The numbers listed are the probability (P) of transfer of HCN from the specified level per collision. Two features of these results stand out: (i) the probabilities are, in all cases, small but show considerable variation depending on the HCN level; (ii) the dependence of P on rare gas (M) is different for different HCN levels. For example, for HCN(OO1) the values of P are almost independent of M, but for HCN(O11) P falls quite steeply from M = He to M = Xe. This is not the place to reproduce in detail our interpretation of the interesting data summarized in Table 1.However, the results merit some discussion, especially as they seem to represent the only case in which relaxation rates have been determined for all modes of a simple polyatomic molecule. Table 1 Probabilities" of vibrational energy transfer between HCN(rnn1) and the noble gases at 298 K He Ne Ar Kr Xe HCN(001) HCN(002) 1.9 (-5) 1.6 (-5) 2.9 (-5) 1.9 (-5) 2.9 (-5) 1.4 (-5) 3.2 (-5) 1.6 (-5) 2.8 (-5) - HCN(101) 0.95( -4) 1.32( -4) 1.27(-5) 1.24(-5) 1.43(-5) HCN(O11) 2.3 (-3) 7.8 (-4) 5.7 (-4) 3.4 (-4) - a 1.9 (-5) = 1.9 x lt5;thermally averaged probabilities are calculated from the 'hard-sphere' formula: P = k/(no2)cwhere k is the observed rate constant and (d)and c are thecollision cross-section and mean relative velocity, respectively.The following cross-sections were used HCN + He, Ne, Ar, Kr, and Xe: 2.84, 2.94, 3.26, 3.36, and 3.58 A It is appropriate to begin by considering relaxation of the v3 mode, which is a rather localized CH stretching vibration resembling the sole vibration of a hydrogen halide. Despite this parallel, it is clear that, in relaxation of its v3 mode by noble gases, HCN acts as a polyatomic." The evidence for this is that the probabilities listed in the first row of Table 1 are much larger than those for deactivation of HCl(u = 1)15 and HF(u = 1)16 by rare gases, and they show R. V. Steele, Jr. and C. B. Moore, J. Chem. Phys., 1974,60, 2794. l6 (a)J. F. Bott and N. Cohen, J.Chem. Phys., 1971,55,3698; (6) J. K. Hancock and W. H. Green, J. Chem. Phys., 1972, 57, 4515. Smith virtually no dependence on the nature of the rare gas. This is a clear indication that HCN(001) is not relaxed by direct transfer back to the (OOO) vibrational ground state, but by transfer to levels associated with the other, lower frequency, vibrations. In terms of the 'breathing sphere' modification17*18 of SSH theory," the insensitivity of P to the reduced mass of the colliding species means that the rate- determining step in the relaxation from the level being monitored is a resonant one in which the internal energy of the excited molecules changes by skBT that is, transfer is induced to a level with a similar energy but in other vibrational modes.For transfer from HCN(001), only the levels (120) and (050) appear close enough to fit the bill. Arnold and Smith" argued that the dominant process is probably HCN(OO1) + M -HCN(12'0) + M; AE = + 191 CIK' (5) on the basis that the (001) and (12'0) levels are both C+ vibrational states and that there was some evidence for mixing of these zeroth-order states in isolated HCN.*' The reverse process would occur with a somewhat larger rate constant than that for the process as written. However, HCN(12'0) will relax rapidly to lower levels, so the transfer back to (001) can be ignored. The proposed mechanism for relaxation of HCN(001) was supported when measurements were made on HCN(002). At first sight, it is surprising that (002) should relax more slowly than (001): calculations based on harmonic oscillator behaviour2' would predict an increase in relaxation rate by a factor of ca.2. However, the equivalent process to equation 5, i.e. HCN(OO2) + M -HCN(12'1) + M; AE = +247 cm-' (6) is now less resonant. This will reduce the term in the matrix element for collision- induced transfer which depends on the 'overlap' of translational wave-functions. In addition, the vibrational part of the matrix element will not increase as much as it might in the harmonic oscillator limit, because the increase in energy discrepancy reduces the mixing of the zeroth-order wavefunctions. The extent of the second factor is difficult to estimate. However, the way in which the ratios of the probabilities, Poo2/P0,,, vary with rare gas is reproduced extremely well by calculations based on the 'breathing sphere' model.'* These results highlight a major difficulty in trying to predict the rates and pathways for vibrational energy relaxation in polyatomic molecules: namely, our ignorance of the true form of the vibrational eigenfunctions.Even for quite low- lying levels, the labels that we attach to vibrational states give a misleading impression of their 'purity'. The eigenfunctions can be expressed as linear combinations of those for harmonic oscillators,-for example, for a triatomic F. I. Tanczos, J. Chem. Phys., 1956, 25, 439. Is J. L. Stretton, Trans. Faraday Soc., 1965, 61, 1053. l9 R. N. Schwartz, Z. I. Slawsky, and K.F. Herzfeld, J. Chem. Phys., 1952, 20, 1591; R. N. Schwartz and K. F. Herzfeld, J. Chem. Phys., 1954,22, 767. 2o V. K. Wang and J. Overend, Spectrochim. Acta, Part A, 1976,32, 1043. 21 D. Rapp and T. E. Sharp, J. Chem. Phys., 1963,38, 2641. The Collision Dynamics of Vibrationally Excited Molecules molecule, as CailqmqnqI),but it is formidably difficult to calculate the mixing coefficientsai accurately, either by ab initio methods or semi-empirically from an experimental knowledge of the energies of the vibrational levels. This is an area in which one might hope for progress, with measurements of radiatively and collisionally induced transition probabilities playing a role in establishing the true nature of vibrational states in polyatomic molecules.The experiments on relaxation of HCN from its (101) and (011) levels demonstrate that excitations in the v1 and v2 modes are lost more rapidly than excitation in v3. For HCN(101),14 the values of P are all 4-5 times greater than those for HCN(001). Again the lack of dependence on M is strongly suggestive of a near-resonant intramolecular V-V transfer mechanism. The most likely process is HCN(lO1) + M-HCN(03’1) + M; AE = -27cm-’ (7) which would be assisted either by direct mixing of the initial and final states via a Coriolis interaction, or by mixing between the zeroth-order (101) and (02O1) X’ states. HCN(O11) is relaxed very much faster than either HCN(101) or HCN(OO1) and the probabilities now depend rather strongly on M, the noble gas.Since the experimental observations demonstrate that the v3 excitation is retained in the first step of relaxation from (Oll), then that process must be HCN(Ol1) + M -HCN(001) + M; AE = -692 cm-’ (8) Now, AE, the energy which has to be transferred to the relative translational and rotational motions of HCN and M, is much greater than k,T, and a more marked dependence on the reduced mass of the collision pair is to be expected. However, the simple breathing sphere the~ry”~’~ would actually predict a much more marked variation of P along the sequence M = He to M = Kr. We have obtained13 much better agreement with experiment by adapting a theoretical model which includes the effects of molecular rotation in an approximate, classical manne~.~~,*~ The physical basis for these calculations is illustrated in Figure 5.HCN is treated as three ‘embedded hard-spheres’ and it is reasonably assumed that the most effective collisions in promoting energy transfer to and from the v2 bending mode are those in which M collides with the H atom perpendicular, or nearly perpendicular, to the HCN axis. As implied in the Introduction, the probability of energy transfer will be greatest for the most impulsive collisions. However, rotational motion of the HCN, as well as the translational motion, is assumed to contribute to the effective velocity at which the H and M atoms approach. The magnitude and direction of the relative velocity and the rotational momentum of the HCN are chosen by Monte Carlo methods.For each choice, a transition probability is calculated via first-order perturbation theory-subject to the constraints of energy and angular momentum ’’A. Micklavc and S. F. Fischer, Chem. Phys. Lett., 1976,44, 209. l3 A. Micklavc, J. Chem. Phys., 1978,69,281; A. Micklavc and S. F. Fischer, J. Chem. Phys., 1980,72,3805. Smith Figure 5 ‘Embedded hard-sphere’ model used in calculations on the V-R,T energy transfer from HCN(O11) to noble gas atoms. The approach shown defines Om,, the maximum angle between the molecular axis and the atom’s direction of approach for which the noble gas atom can impact on the H atom con~ervation.’~Calculations are carried out for a large sample to yield a thermally averaged collisional probability.For M = He, the value of P calculated by this method is less than that from the simple breathing-sphere calculations, largely because steric and mode-matching effects ’’are accounted for more correctly. For the heavier rare gases, the rotational contribution to the effective velocity of approach becomes dominant and raises the values of P by factors of 3.8 (Ne), 4.7 (Ar), and 6.5 (Kr) above the breathing-sphere values. Overall, the calculations parallel the variation of P with M extremely well. The absolute values calculated are approximately four times those observed experimentally, possibly reflecting the shortcomings of an approach based on first- order perturbation theory. In an excellent review on vibrational energy flow in polyatomic molecules, Weitz and FlynnZ4 postulate three ‘propensity rules’.The first is that molecules can be transferred rapidly between levels associated with the same single mode by near- resonant, intermolecular V-V exchange as represented in equation 4. Secondly, they proposed that energy transfer between modes is fast when their fundamental frequencies lie within ca. k,Tof one another, which is not the case in HCN. Finally, they suggested that transfer between levels associated with different modes is enhanced by mechanical Fermi mixing, or by large mechanical or electrical anharmonicity of the lowest frequency mode. Our results on relaxation in HCN lend support to this third propensity rule. In addition, our data indicate that molecular rotation can help to induce relaxation of energy in bending modes.’‘ E. Weitz and G. Flynn in ‘Photoselective Chemistry’, ed. J. Jortner, R. D. Levine, and S. A. Rice, Ah. Chem. Phys., 1981, 47(2), 185. The Collision Dynamics of Vibrationally Excited Molecules Such V--R,T energy transfer will be especially important for hydrides and deuterides. 3 Collisions Between Radicals and Molecules In this part of my review, I want to consider explicitly two examples of simple radical-molecule reactions: 0 + HCl(0) --+ OH + C1 and OH(v) + HC1-H,O + CI In the first of these two cases, the vibration which is excited is one along the bond that is broken in the reaction; in the second, it is along a bond that retains its integrity through the reaction.In both reactions, vibrational relaxation, i.e. 0 + HCl(v) -0 + HCl(u’ < v) and OH(v) + HCI -OH(u’ < u) + HCl can compete with reaction. Many experiments, including our own, measure the total rate of loss of the vibrationally excited species, but I hope to demonstrate that one can assess the relative importance of reaction and relaxation by comparing the rates for isotopically related processes. Moreover, it should be possible to relate the effects of vibrational excitation and isotopic substitution on reaction rates since both changes leave the underlying potential energy surface unaltered. Reaction (9a)is very nearly thermoneutral (A@ = +3.4 kJ mol-’) and has an HCI,DCL + 0 CI + OD,OH Figure 6 Energy level diagram for the reactions: 0 + HCI(u), DCl(v) -+OH(v’), OD(v’) + C1.The full lines represent vibrational levels of the hydrogenated species, the dashed lines those of the deuterated species Smith activation energy of ca. 25 kJ rn~l-’.’~*’~ Using these data and spectroscopic information, the approximate energy level diagram in Figure 6 can be constructed. The thermal rate constants exhibit a normal primary kinetic isotope effect (k,,, > kx,) which is consistent with AGO(the difference between zero-point levels in the transition state and in the reagents) being greater in the deuterated system.’ Our experiments” on the kinetics of HCl(u = 1) and DCl(u = 1) in the presence of 0-atoms used the LIVF method, steady-state concentrations of 0-atoms being created in a discharge-flow system.The rate constants which these experiments gave are compared in Table 2 with those for reaction of the (o = 0) molecules. Although DCl(u = 1) molecules have only just as much energy as the transition- state species, whereas HCl(u = 1) molecules have much more energy, the DCl(v = 1) molecules are removed more rapidly. This observation led us to conclude that the predominant channel for removal of the (u = 1) molecules is not reaction but vibrational relaxation-probably by an electronically non-adiabatic mechanism of the kind first proposed by Nikitin28 to explain the anomalously rapid self- relaxation of NO(u = 1). Table 2 Rate constants’for reaction and relaxation of HCI(u) and DCL(u) with 0-atoms at 298 K k/cm3 molecule-’ s-l Ref: HCI (U = 0) reaction 1.2 (-16) 25a reaction 1.4 (-16) 26 DCI (U = 0) reaction 2.7 (-17) 25b HCI (U = 1) DCI (U = 1) reaction + relaxation reaction + relaxation 1.0 (-12) 1.3 (-12) 27 27 HCI (U = 1) HCI (U = 2) reaction + relaxation reaction + relaxation 8.9 (-13) 5.2 (-12) 29 29 reaction 1.5 (-12) 29 HCl (U = 1) reaction 2 (-14) 30 HCI (U = 2) reaction 3 (-12) 30 HCI (U = 1) reaction + relaxation 1.0 (-12) 31 HCI (U = 2) reaction reaction + relaxation 6.4 (-14) 6.3 (-12) 31 31 a 1.2 (-16) = 1.2 x 25 (a)R.D. H. Brown and I. W. M. Smith, Inf.J. Chem. Kinet., 1975,7,301;(b)R. D. H. Brown and I. W. M. Smith, Int. J. Chem. Kinet., 1978, 10, 1. 26 D. L. Baulch, R.A. Cox, P. J. Crutzen, R. F. Hampson, Jr., J. A. Kerr, J. Troe, and R. T. Watson,J. Phys. Chem. Ref: Data, 1982, 11, 327. 27 R. D. H. Brown, G. P. Glass, and I. W. M. Smith, Chem. Phys. Lett., 1975, 32, 517. E. E. Nikitin, Opt. Spectrosk., 1960,9, 8 and 1961, 11,452; E. E. Nikitin and S. Ya Umanski, Faruday Discuss.Chem. SOC.,1972, 53, 1. 153 The Collision Dynamics of Vibrationally Excited Molecules This conclusion does not, of course, exclude the possibility that the 0 + HC1,DCl reactions are efficiently promoted by vibrational excitation. After all, the rate constant for removal of HCl(u = 1) is 104-times that for reaction of HCl(u = 0) so even though reaction makes only a small contribution to the overall removal it could represent a very significant rate increase.In fact, more detailed and elegant e~perirnents~~-~'than our own-in which the production of OH was observed- indicate that this is the situation. Table 2 gives rate constants that have been measured for reaction of HCl(u = 1,2) with 0-atoms. The observation that excitation of the vibration along the bond broken in an atom-transfer reaction enhances the reaction rate very appreciably has been confirmed in studies of some other nearly thermoneutral reaction^.'^-^^ Before considering how such observations might be interpreted, I shall summarize briefly our results 34,35 for reaction (10a) and its deuterated analogues. The rate of the thermal reaction between OH and HCI has now been measured directly in about eight independent studies.(The results are summarized most recently in ref: 36.) The activation energy is variously reported as between 2.4 and 4.4 kJ mol-', though the spread in rate constant values at 298 K is only about 15%. Our objective is to see whether vibrational excitation in the radical reagent in simple radical-molecule reactions can enhance their rates. A lot of the data in the literature have been derived from rather indirect experiments and provide no consistent answer. The OH + HC1 reaction seems a specially suitable case for study since isotopic substitutions can be made in both reagents, and the possibility also exists of observing the effect of vibrational excitation of the molecular (HC1) reagent. In the present experiments, OH or OD radicals are generated by pulsed photolysis of H20, HNO,, or D20, DNO,.At variable delays after the photolysis pulse, radiation from a tunable, pulsed, dye laser is used to induce A2Z+--X% fluorescence from the radicals. This LIF detection method is extremely sensitive and highly specific. (The laser excites a single line in either the (0,O)or (1,1) band of the A--X system of OH or OD, but rotational equilibration is extremely rapid, so no rotationally specific information is obtained.) The photodissociation processes give only low yields of vibrationally excited radicals, but this is a positive advantage since any effects of stepwise vibrational relaxation can be ignored. As with our LIVF experiments, these measurements yield rate constants for the sum of relaxation and reaction.In systems comprising two diatomics, a likely mechanism for rapid relaxation is intermolecular V-V transfer.2 However, the rates of V-V exchange will depend strongly on AEv--y, the discrepancy between 29 R. G. Macdonald and C. B.Moore, J. Chem. Phys., 1978, 68, 513. 30 J. E. Butler, J. W. Hudgens, M. C. Lin, and G. K. Smith, Chem. Phys. Lett., 1978, 58, 216. 31 M. Kneba and J. Wolfrum, Proc. 17th Int. Combust. Symp., The Combustion Institute, Pittsburgh, 1979, p. 497. 32 R. Zellner and W. Steinert, Chem. Phys. Lett., 1981, 81, 568. 33 V. B. Rozenshtein, Yu. M. Gershenzon, A. V. Ivanov, and S. I. Kucheryavii, Chem. Phys. Lett., 1984,105, 423. 34 B. D. Cannon, J. S. Robertshaw, I. W. M. Smith, and M.D. Williams, Chem. Phys. Lett., 1984,105,380. 35 I. W. M. Smith and M. D. Williams, to be published. 36 L. F. Keyser, J. Phys. Chem., 1984, 88, 4750. Smith the two vibrational transition energies, and that varies widely for the four different isotopic pairs from: OH,OD(v = 1) + HC1,DCl. Table 3 summarizes our rate data for members of the OH,OD + HC1,DCl family. The following points should be emphasized: (i) the rate constant for the OH + HCl reaction is in good agreement with values that have been determined previously; (ii) comparison of the rate constants for OH + HC1,DCl and for OD + HC1,DCl show that there is an appreciable primary kinetic isotope effect; (iii) comparison of the data for OH,OD + HC1 and for OH,OD + DC1 show that there is insignificant secondary isotope effect.(iv) the rate constants for removal of OH(v = 1) and OD(u = 1) are not much larger than the corresponding rate constants for reaction of radicals in (v = 0). Table 3 Rate constants at 298 K for total removal of OH,OD (v = 0,l) by HCI, DCI and energy discrepancies for V-V energy exchange 10' 3k/cm3molecule-' s-l Ev-v/cm-l OH (U = 0) + HCI OH (V = 0) + DCl OD (V = 0) + HC1 OD (U = 0) + DCl 6.8 f0.25 1.1 0.1 4.6 f 0.6 1.3 k 0.2 OH (Y = 1) + HCI OD (V = 1) + DCI 9.7 f 1.0 3.0 & 0.7 -684 -542 In experiments involving both deuterated and hydrogenated species, it is clear from our measurements that some isotopic scrambling occurs between the radical precursor and the molecular reagent.We believe that this might have caused earlier 37738measurements on the OH(u = 0) + DC1 reaction to yield rate constants that are too high. In our experiments using LIF, the radical precursor was present in much smaller concentrations and any isotope exchange would do little to degrade the isotopic purity of the molecular reagent. However, because the precursors do relax the vibrationally excited radicals rapidly, the effects of isotopic scrambling cause our measurements on OH(v = 1) + DC1 and OD(o = 1) + HC1 to be of relatively low accuracy, and rate constants are not reported here. It is evident from our results that the addition of 42.7 kJ mol-' of energy to OH or 31.5 kJ mol-' to OD does little to promote their reactions with HC1,DCl.These results, those for the 0 + HC1,DCl reactions, and others on the effects of reagent vibrational excitation can perhaps best be understood via vibrationally adiabatic transition-state theory.3940 Let us begin by considering a three-atom system, A + 37 I. W. M. Smith and R. Zellner, J. Chem. SOC.,Faraday Trans. 2, 1974, 70, 1045. 38 D. Husain, J. M. C. Plane, and N. K. H. Slater, J. Chem. SOC..Faraday Trans. 2, 1981, 77, 1949. 39 E. Pollak, J. Chern. Phys., 1981, 74, 5586. 40 A. D. Isaacson and D. G. Truhlar, J. Chem. Phys., 1982, 76, 1380. 155 The Collision Dynamics of Vibrationally Excited Molecules BC = AB + C, from this viewpoint. As Figure 2 shows, a path of minimum energy can be traced across the potential energy surface, and the variation of potential energy reduced to a one-dimensional representation, V(x).To apply the vibrationally adiabatic form of transition-state theory, 'q4' it is also necessary to examine how V varies with small displacements orthogonal to x. One then computes rate constants at various points along x and finds the minimum value, that being the best transition-state theory estimate. Each calculation depends importantly, but not only, on AEi which corresponds to the maximum energy on the u = 0 vibronic state curve. In general, this curve does not run parallel to V(x), since the zero-point energy changes as the interatomic forces alter. According to transition-state theory, it is the change in AE&, because of different zero-point effects, that is largely responsible for kinetic isotope effects.' In vibrationally adiabatic transition-state theory, these ideas are simply extended, the system being assumed to stay in the same vibrational state (but not retain the same vibrational energy) as it progresses from separated reagents to the transition state.The rate of reaction now crucially depends on AG, the height of the barrier on the vibrationally adiabatic curve for state u. The extent to which vibrational excitation promotes reaction depends on how AE, is reduced as u is increased. It is clearly vital to ask how vibrationally adiabatic the collisions are likely to be, up to the point where trajectories reach the transition state. The answer to this depends crucially on the location of the transition state.The earlier the position along the reaction path which the transition stage occupies, the less is the coupling between relative translational and vibrational motions up to that point, and vibrational adiabaticity is likely to be a good appro~imation.~~ However, because the V-T coupling is weak the vibrationally adiabatic or vibronic curves are likely to run almost parallel, the values of A@ only change slowly with u, and effects of vibrational excitation are slight. This case is illustrated in Figure 7a. The opposite limit is reached when the barrier is 'late'; i.e., as the products separate.'^^^ This is typical for a strongly endothermic reaction. Now there is very strong curvature in the reaction path and strong V-Tcoupling before the barrier is reached on the potential: at least the simplest versions of vibrationally adiabatic theory cannot be applied.In intermediate cases, vibrationally adiabatic transition-state theory is more appropriate than one might immediately suppose-especially for those reactions, which have been comparatively widely studied, in which an H-atom is transferred between two heavier species. Now, as u is increased, A@ falls appreciably and, in addition, the position of the barriers on the vibrationally adiabatic curves move to progressively earlier positions along the reaction path, thereby improving the approximation of vibrational adiabaticity. Figure 7b illustrates this effect for a supposedly thermoneutral reaction, such as (9a).So far I have implicitly considered systems in which vibrational excitation is in 41 P. Pechukas, Ann. Rev. Phys. Chem., 1981, 32, 159. 42 I. W. M. Smith, J. Chem. SOC.,Faraday Trans. 2, 1981, 77, 747. 43 J. C. Polanyi, Acc. Chem. Res., 1972, 5, 161. 156 Smith -x--x-Figure 7 Variations of potential energy (v) and vibronic energies for: (a) an exothermic atom- transfer reaction with a low and early potential energy barrier, so that the barriers on the vibrationally adiabatic curves vary only siightly with v; and (b) a thermoneutral reaction with a high and symmetrically located barrier where the heights and positions of the vibrationally adiabatic barriers depend strongly on v the bond which 'disappears' in the reaction.However, the ideas of vibrationally adiabatic transition-state theory are even more appropriate to excitation of other reagent modes. For example, Isaacson and Truhlar4' have considered the effects of both OH and H, excitations on the reaction OH(v1) + H2(~2)+H,O + H (1 1) which is, of course, closely related to reaction 10a. They calculated normal mode frequencies at all points along the reaction path. For excitation of the H, molecule, the vibronic curves are not parallel to one another or to V(x);consequently, AEz decreases appreciably as u2 is raised. On the other hand, the OH vibration which correlates through to the v1stretching mode in H20 is almost unchanged along x. Hence the u2 vibronic curves are almost parallel to V(x),the values of AE are almost independent of ul, and there is very little change of the rate constant with OH excitation.This is entirely consistent with experiments on reaction (11)32 and with our own results on the OH,OD(u, = 1) + HC1,DCl systems. 4 Collisions Between Free Radicals As was pointed out in the Introduction, the result of collisions between two radicals will usually be dominated by the motion of the system on a potential energy surface (or surfaces) which has a deep 'well' or minimum.' Collision complexes are formed easily and it is their rate of formation which is rate-determining for many radical- radical processes. In trying to estimate what this rate might be, several factors must be borne in mind. One is that V(x),the potential energy along the path of minimum energy, will probably show no maximum.The transition state can be chosen by variationa141 or maximum free-energy criteriag-or the concept of a defined The Collision Dynamics of Vibrationally Excited Molecules transition-state for canonical reaction can be abandoned alt~gether.~~ Whichever approach is taken, it is clear that the critical configurations are 'loose' or 'early'. So the vibrationally adiabatic curves for different states will be very nearly parallel up to this region and the rates of formation of collision complexes should be virtually independent of initial vibrational state. A second factor to be remembered is that interactions of two radicals lead to more than one potential energy surface with statistics determining what fraction of collisions occur on any particular surface. It may be necessary to take account of the fact that more than one such surface leads to bound states45 and hence to recombination, reaction, or vibrational relaxation.Here, I shall concentrate on the dynamics of radical-radical collisions when one of the radicals is vibrationally excited. In the two studies which I shall specifically refer to, the experimental difficulties were reduced by making one radical NO or NO,. In the first set of experiment^,^^ which we did a few years ago, NO was vibrationally excited in 'indirect' LIVF experiments. It is not easy to find a pulsed i.r. laser to excite NO directly. However, it can be excited by V-V energy transfer from HCl: HCl(u = 1) + NO(u = 0) -HCI(u = 0) + NO(u = 1); AE = -1009 cm-' (12) when the HCl is excited by radiation from a pulsed HCl chemical laser.These experiments gave the results for NO(u = 1) listed in Table 4. The data for OH(u = 1) and OD(u = 1) were found47 by the same experimental method that was used for studying the kinetics of these radicals with HCl and DCl. The rate constants shown in Table 4 are all within an order-of-magnitude of rate constants for the total number of 'hard-sphere' collisions. In contrast, relaxation at Table 4 Rate constants (lo-" cm3 molecule-' s-l) at 298 K for relaxation ofN0 (u = 1) by 0,C1 and Br atoms and of OH, OD (u = 1) by NO and NO2" kiax kZc kest NO (U = 1) + 0 6.5 f0.7 -2.8 5.0 NO (U = 1) + C1 3.5 f 0.5 -5.8 6.9 NO (u = 1) + Br 2.0 f0.3 -3.3 5.4 OH (U = 1) + NO 3.8 & 0.6 -2 1.4 OD (U = 1) + NO 2.7 f 0.3 -1.4 OH (U = 1) + NO2 4.8 f 0.8 23 2.0 OD (U = 1) + NO2 4.3 f0.3 -2.1 Experimental data from ref: 46 and 47, values ofk,,, and kcstare taken from Howard and Smith's review,8 which lists the original references 44 M.Quack and J.Troe, Ber. Bunsenges. Phys. Chem., 1974, 78, 240. 45 I. W. M. Smith, Int. J. Chem. Kinet., 1984, 16, 423. 46 R. F. Fernando and I. W. M. Smith, Chem. Phys. Lett., 1979,66,218; R. P. Fernando and I. W. M. Smith, J. Chem. Sor., Faraday Trans. 2, 1981, 77, 459. *' I. W. M. Smith and M. D. Williams, to be published. Smith room temperature of NO(v = 1) and OH(u = 1) by argon requires more than 2 x lo748 and more than lo547 collisions, respectively.In addition, it should be remembered that, in all these systems, more than 50%of collisions between radicals occur on potential energy surfaces without significant minima on them. The most reasonable explanation for the unusually facile vibrational relaxation found in these radical-radical collisions, is that the initial specific vibrational excitation is randomized during the lifetime of the collision complexes that form. At the same total energy, a much greater volume of phase space is associated with separated radicals when they are not vibrationally excited, than when they are. As a result, when the complexes redissociate there is only a very low probability that they do so to yield the radicals in excited vibrational states.If this explanation is correct, and if the rate of formation of collision complexes is independent of whether or not the colliding radicals are vibrationally excited, the relaxation rate constants (krelax) should correspond quite closely to the rate constants for recombination of the same pair of radicals in thermal collisions and the limit of high pressure (kzc). Unfortunately, for these small systems, the latter rate constants can be determined only after a long, and rather uncertain, extrapolation. Within the errors of estimating kz,,the values are in reasonable agreement with those of krelax. The experimental values of both rate constants can be compared with estimated values based on Quack and Troe’s maximum free-energy interpolative meth~d.~ Again, fair agreement is found, given that the theoretical model is quite sensitive to the form assumed for the long-range potential.However, uncertainties about the real form of potential energy surfaces in the regions where chemical bonds just start to influence the potential energy is a real stumbling block to making accurate calculations on radical-radical processes. 5 Conclusions In this review, an attempt has been made to sketch out the connection between the results of elementary bimolecular processes involving vibrationally excited species, as revealed by measured rate constants, and the form of the potential energy hypersurface which controls the dynamics of collisions between any particular pair of species.Our knowledge and understanding is steadily improving, not least because of the stimulating interplay between experiment and theory. However, I am afraid that my simple and short account may leave the reader with the impression that, in this field, few problems remain. This is certainly not so. The state of our present understanding is fragile, based as it is on the results of rather few definitive studies and these often limited; for example, to the behaviour of molecules in low- lying states of vibrational excitation. Some very interesting problems relate to systems which cannot easily be placed into the categories that I have identified and considered. For example, we remain largely ignorant about the dynamics of isolated molecules at intermediate levels of excitation: that is, between the regime of discrete levels associated with normal mode motions and investigated by infrared absorption spectroscopy, and the ergodic ** J.C. Stephenson,J. Chem. Phys., 1973,59, 1523. 159 The Collision Dynamics of Vibrationally Excited Molecules region near dissociation limits where rapid intramolecular vibrational relaxation and statistically based unimolecular rate theories are the order of the day. In bimolecular processes, there is a similar region of uncertain behaviour. Thus, it is not yet clear when transitions occur between direct dynamics and collisions involving complexes, as either the collision energy or the strength of intermolecular attraction is increased.One would like to know how deep the attraction between colliding species must be for complexes to form in bimolecular collisions, and whether other factors have any influence on complex f~rmation.~'.~' In summary, this field of research endeavour is vital and challenging, and it promises to remain so for many years yet. Acknowledgements. I should like to express my thanks to those who have supported the research of my group over the years, particularly SERC, Shell Research Ltd., USAFOSR, NATO, and the Department of the Environment. I should also like to thank all my present and past research colleagues for their contributions to our joint efforts. 49 Ch. Schlier in 'Energy Storage and Redistribution in Molecules', ed. J. Hinze, Plenum Press, New York, 1983, p. 585. so M. K. Osborn and I. W. M. Smith, Chem. Phys., 1984,91, 13.

ISSN:0306-0012    

DOI:10.1039/CS9851400141

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

The B,,dependent Isomerase Enzymes; How the Protein Controls the Active Site By John M.Pratt * DEPARTMENT OF CHEMISTRY, UNIVERSITY OF WITWATERSRAND, 1 JAN SMUTS AVENUE, JOHANNESBURG, 2001 SOUTH AFRICA 1 Introduction Metalloenzymes and metalloproteins can be considered as a partnership between the metal ion or complex, which provides the active site, and the protein, which enhances and controls the activity of the metal through, for example, changes in conformation which are somehow coupled to changes on the metal. The role of the protein and the mechanism of coupling are now understood in the case of the non-enzymatic haemoglobins; see the review by Perutz.’ The aim of this article is to summarize present knowledge concerning the B ,-dependent isomerases, which are the first family of redox-active metalloenzymes where we can probably claim to understand the role of the protein and, at least in outline, the mechanism of coupling; the peroxidases may provide the second such family, together with an instructive contrast in the mechanism of coupling (see below).Good sources of background information are available on the B, ,-dependent enzymes and on the co-ordination ~hemistry,~ redox potential^,^ and organometallic chemistry 2*3 of the B ,(Co corrinoid) complexes. 2 The B12Coenzymes and the B,,dependent Isomerase Enzymes The structure of vitamin B,, (cyanocobalamin) was determined by Hodgkin and co-worker~;~,~the molecular structure is shown in Figure 1.It is a diamagnetic, six- co-ordinate complex containing the & Co”’ ion.The naturally occurring Co corrinoids comprise a family of complexes which all possess the same conjugated ‘corrin’ ring as in B,,, but may differ in the nature of the side-chains and nucleotide base as well as other axial ligands; those that possess the same side-chains as B12 itself are called ‘cobalamins’. The B ,-dependent isomerases catalyse the isomerization of various organic substrates, as exemplified by the reversible C-skeleton rearrangement shown in Figure 2a (one of the B ,-dependent reactions occurring in man) and the irreversible rearrangement of diols to aldehydes shown Present address: Department of Chemistry, University of Surrey, Guildford GU2 5XH M. F. PerutZ Ann. Rev. Biochem., 1979,48, 327.‘BIZ’(2 vols.), ed. D. Dolphin, Wiley, New York, 1982. J. M. Pratt, ‘Inorganic Chemistry of Vitamin B,,’, Academic Press, London, 1972. * D. Lexa and J. Saveant, Acc. Chem. Res., 1983, 16, 235. D. C. Hodgkin, J. Lindsey, R. A. Sparks, K. N. Trueblood, and J. G. White, Proc. R. SOC. London, Ser. A., 1962,266,494. C. Brink-Shoemaker, D. W. J. Cruickshank, D. C. Hodgkin, M. J. Kamper, and D. Pilling, Proc. R. SOC. London, Ser. A, 1964, 278, 1. B,2-dependent Isomerase Enzymes CONH2 CH20H-H Figure 1 Molecular structure of B, (Reproduced by permission from re$ 3) in Figure 2b. All these enzymatic reactions involve the interchange of H with C, N, or 0 between neighbouring atoms; each enzyme is reasonably, but not completely, specific for one substrate.The corrinoid is present in a so-called 'coenzyme' form, in which the upper (p) axial co-ordination site is occupied by the extraordinary ligand 5'-deoxyadenosyl (hereafter denoted by R)which is shown in Figure 3, while the lower (a) ligand may be the heterocyclic base (denoted by B) present in the cobalamins or another base, or may even be absent (to leave a five-co-ordinate Co-atom). The structure of the red six-co-ordinate cobalamin coenzyme has been determined by Hodgkin and Lenhert; the atomic positions are shown in Figure 4. Evidence that the yellow alkyl-corrinoids (including the acidified form of the coenzyme) are five-co-ordinate is indirect,' but supported by the very high rate of reaction (k2 2 10' dm3 mol-I s-') of methylcobinamide with ~yanide.~ Both the normal red and acidified yellow forms of the coenzyme are diamagnetic lo and can formally be regarded as complexes of Co"' with a carbanion.Other families of B12-dependent enzymes are known; here we consider only the isomerases. The (a)P. G. Lenhert and D.C. Hodgkin, Nature, 1961,192,937; (6)D. C. Hodgkin, Proc. R. SOC.London,Ser. A, 1965,288, 294; (c) P. G. Lenhert, Proc. R. SOC.London, Ser. A., 1968,303,45. J. M. Pratt, in ref: 2., Vol. 1, p. 325. D. A. Baldwin, E. A. Betterton, and J. M. Pratt, S. Afr. J. Chem., 1982, 35, 173. lo J. A. Hill, J. M. Pratt, and R. J. P. Williams, J. Chem. Soc., 1964, 5149. Pratt CO. SR CO -SR I I CIS-CH CHz-CHz I I COOH COOH (a1 OH OHI RCH-CH2 +[?RC%-CH+ ’I RCHz-CCHO + H20 OHI hi COOR COOR COOR I I I Co-CHZ-CR’ + CH,-CR’ + CH2-CHR’ I 1 I COOR COOR COOR (C1 Figure 2 Examples of B, ,-dependent enzymatic rearrangements, (a) (RSH = coenzyme A)and(b)(R = H, Me, HOCH,), andprotein-freemodelreactions,(c) (R = Me, Et; R’ = H, Me) kHTHOH01 \,qCHOH Figure 3 The 5’-deoxyadenosyl ligand present in the B, ,coenzymes (Reproduced bypermission from ref: 3) mammalian enzymes accept only the cobalamin coenzyme, although a range of coenzymes is found in bacterial enzymes; here we discuss only the cobalamin coenzyme.The cobalamin coenzyme is bound by the apoenzyme (protein) with no significant change in its absorption spectrum; this suggests that no change has occurred in the axial ligands, although the possibility that B is replaced by another nitrogenous base (lysine, histidine) cannot be rigorously excluded.The cobalamin coenzyme is, therefore, the only metallo-coenzyme (excluding other B, coenzymes) whose chemistry can be directly compared in the presence and absence B, ,-&pendent Isomerase Enzymes Figure 4 Structure of the cobalamin coenzyme (Reproduced by permission from rej 7b) of protein; in addition, the physical properties of the Co corrinoids (especially their u.v.-visible spectra) provide an excellent probe for following changes at the active site during the enzymatic reaction. The enzymes require no additional cofactors for the C-skeleton rearrangements and only an alkali-metal cation for the diols; their reactions involve only the single substrate.In spite of the lack of any X-ray data on these enzymes, they provide the ideal opportunity for investigating the basis of the partnership between a metal complex and its protein. 3 The Main Steps in the Enzymatic Reaction It is now generally agreed,2 though not proven (see below), that the first step involves homolytic fission of the C0-C bond according to reaction 1 (axial ligands only given) and that the (protein-bound) radical R derived from the ligand then abstracts a hydrogen atom from the substrate (SH)according to reaction 2. The Co" complex (BIZ,)is assumed to be five-co-ordinate and this is supported by the very high rate of reaction with methyl radicals to form the Co-Me complex; but l1 J.F. Endicott and G. J. Ferraudi, J. Am. Chern. SOC.,1977,!29, 243. Pratt an X-ray determination of the structure would be desirable. There is, however, no agreement on the nature of the subsequent steps, i.e. on the form in which the substrate undergoes rearrangement; candidates include radicals, fully or partially formed carbocations (by electron transfer to Co"), and organo-cobalt derivatives.2 The product is formed, and the coenzyme re-formed, by the reverse of these steps. KO-R B-Co" + Re k, R* + SH RH + S- (2) In the absence of protein the Co-C bond of the coenzyme is surprisingly stable in neutral aqueous solution, although decomposed by light, by heating in acid, and by reagents such as cyanide; we shall see below that in the presence of protein and substrate the Co-C bond can be labilized by a factor of 10 'I! The abstraction of a hydrogen atom according to reaction 2 is, of course, well known but C-skeleton rearrangements of the type shown in Figure 2a were unprecedented in organic chemistry until 1975 when they were modelled (see Figure 2c)with various protein- free alkyl-cobalt complexes (cobalamins, cobaloximes), in which the radical was generated by photolysis or thermolysis of the Co-C bond; 12-' it was subsequently shown that the radical does not undergo rearrangement except in the presence of the Co" ion.I4 The protein therefore plays a major role in step (1) but a relatively minor role in the subsequent steps, perhaps confined to preventing the various species from diffusing away from each other.This review focuses exclusively on step (1) of the enzymatic reaction, where the main theme is the use (by the protein) of steric distortion to labilize the Co-C bond. It should be mentioned that many groups throughout the world are interested in mechanistic studies on the B ,-dependent isomerases and relevant protein-free models (e.g.those of B. T. Golding, H. A. 0.Hill, A. W. Johnson, M. D. Johnson, and R. J. P.Williams in the U.K., Arigoni in Switzerland, Retey in Germany, Yurkevich in Russia, Fukui and Toraya in Japan, Dolphin in Canada, Abeles, Babior, Barker, Halpern, Hogenkamp, Marzilli, Schrauzer and Stadtman in the U.S.), but that most of this work is aimed at unravelling the nature of the steps after fission of the Co-C bond; reference 2 provides a good introduction to this area.The greater instability of the Co-C bond to a secondary, compared to a primary, alkyl ligand was first noted (with Co-Pr') in 1968.'' This led to the suggestion that the enzyme might distort the coenzyme such that the state of the Co-C bond approached that of the transition state for heterolytic fission to give the carbonium ion.16 The first e.s.r. evidence for homolytic fission of the Co-C bond in the P. Dowd, M. Shapiro, and K. Kang, J. Am. Chem. SOC.,1975,97,4754. G.Bidlingmaier,H. Flohr, U.M. Kempe, T. Krebs, and J. Retey, Angew. Chem., 1nt. Ed. Engl., 1975,14, 822. 14 J. Retey, in 'Vitamin BIZ', ed.B. Zagalak and W. Friedrich, de Gruyter, Berlin, 1979, p. 439. Is R. A. Firth, H. A. 0.Hill, B. E. Mann, J. M. Pratt, R. G. Thorp, and R. J. P. Williams, J. Chem. SOC.A, 1968, 2419. l6 H. A. 0.Hill, J. M. Pratt, and R. J. P. Williams, Chem. Brit.,1969, 5, 156. B, ,-dependent Isomerase Enzymes isomerases and related ribonucleotide reductases was reported in 1969-197 1,' '-19 and it was suggested that the protein somehow served to stabilize Co" and the organic radical.lg It was later suggested more specifically that the protein used steric distortion of the cobalt co-ordination sphere (most probably of the Co-C-C bond angle), caused by a substrate-induced change in conformation, to displace equilibrium (1) to the right." Others have stressed distortion of the corrin ring,' which may lead to pressure on the adenine group of the ligand 22 or on the C-atoms closer to the Systematic studies over the last five years have provided a fairly comprehensive picture of steric effects in protein-free alkyl-corrinoids, against which one can assess the behaviour of the protein-bound coenzyme.4 Steric Distortion in Protein-free Alkyl-corrinoids The effects of increasing distortion around the co-ordinated carbon atom (C,) have been studied systematically *,-" by varying the alkyl ligand in the series -CH, to -CHMe2 (the Co-CMe, complex is too unstable to study), -CH,CH, to -CH,CMe,, and cyclopropyl to cyclohexyl (here denoted by c3-c6) with particular reference to equilibrium (1) and to the equilibrium between the red six-co-ordinate alkylcobalamin and its yellow five-co-ordinate form (with B remaining unco-ordinated at the end of the side-chain).Neopentylcobalamin (B-Co-CH,CMe,) undergoes reversible homolytic fission at room temperature and provides an unambiguous protein-free model for homolytic fission in the ~oenzyme.~~~~~~Alkyl-corrinoids with H on the C,-atom usually decompose by overall p-elimination (to give Co' and an ~lefin),~.~~ but recent evidence indicates that p-elimination occurs uia an initial homolytic fission to give the caged (Co" + radical) pair.26 The results form a very simple pattern, in which all three series of alkyl ligands can be placed in virtually a single order of increasing steric compression, uiz.Me -C, -R (the ligand in the coenzymes) <Et -Pr" < Bu' -C, < Pr', neopentyl, C,, c6.8'26 If we start with methyl-cobalamin and increase the degree of steric distortion around C,, then we observe an increase in the lability of the Co-C bond in the six-co-ordinate form and an increasing stability of the five- over the six- co-ordinate form. If we next study the five-co-ordinate forms alone (by protonating B in the cobalamins or by removing the side-chain to form the 'cobinamides'), we find a parallel increase in lability with increasing distortion, although the absolute rates are lower (by about lo3) than in the six-co-ordinate forms; in spite of increasing conversion into the more inert five-co-ordinate form, the net result of B.M. Babior and D. C. Gouid, Biochem. Biophys. Res. Commun., 1969.34, 441. J. A. Hamilton and R. L. Blakley, Biochim. Biophys. Acta, 1969, 184, 224. l9 M. A. Foster, H. A. 0.Hill, and R. J. P. Williams, Biochem. SOC.Symp., 1971, 31, 187. 2o J. M. Pratt, in 'Techniques and Topics in Bioinorganic Chemistry', ed. C. A. McAuliffe, Macmillan, London, 1975, p. 109. R.H. Abeles and D. Dolphin, Acc. Chem. Res., 1976, 9, 114. 22 J. S. Krouwer, B. Holmquist, R. S. Kipnes, and B. M. Babior, Biochim. Biophys. Acra, 1980,612, 153. 23 J. H. Grate and G. N. Schrauzer, J. Am. Chem. Soc., 1979, 101, 4601. 24 G. N. Schrauzer and J. H. Grate, J. Am. Chem. Soc., 1981, 103, 541. 25 S. M. Chemaly and J. M. Pratt, J. Chem. Soc., Dalton Trans., 1980, (a) 2259, (b)2267, (c) 2274.26 J. M. Pratt, J. Mol. Catal., 1984, 23, 187. Pratt increasing distortion in the cobalamins is always a marked increase in lability, e.g. by a factor of 2lo7 on replacing -Et (t+ -6 months) by -CHEt, (t+ -3 s).’~ It has been shown27 that at 95 “C, t+ for the coenzyme (180 min) is greater than that for ethylcobalamin (9 min), as expected from the steric order; by assuming that the ratio of rates is not too temperature-sensitive, one can derive a ‘working’ t+ of ca. 10 years for the coenzyme at room temperature. Increasing distortion also produces a regular change in the spectrum of the five-co-ordinate forms, such that the maximum shifts from 458 nm in the five-co-ordinate form of the coenzyme to -440nm in the strained complexes;* the spectrum serves as a molecular strain- gauge! Finally, we can use the alkyl-cobinamides to study the co-ordination of imidazole and show that the same pattern (of equilibria and Co-C bond lability) is observed, irrespective of whether the heterocyclic base is imidazole or the much bulkier ligand B present in the cobalamin side-chain.8926 These studies on alkyl- corrinoids in solution are complemented by structural determinations of various alkyl-cobaloximes of formula [R-Co(dmgH),py], where dmgH -is the dimethylglyoximate monoanion, which reveal remarkable distortions; cf: the lengthening of the C0-c bond from 1.998 (R = Me) to 2.085 A (R = Pr’) and the surprisingly large ‘tetrahedral’ Co-C-C bond-angle of 130” when R = -CH2CMe3.” 5 Co-C Bond Fission in the Protein-bound Coenzyme The main evidence relating to the first step in the enzymatic reaction can be summarized as follows.(i) The use of coenzyme analogues has shown that none of the 0-atoms in the ribose moiety are essential for a~tivity,~’-~~ i.e. one can at this stage treat the ligand as a simple alkyl ligand. (ii)The u.v.-visible spectrum of the enzyme in the resting state (max. ca. 520 nm) is virtually identical to that of the protein-free coenzyme, but the addition of substrate causes partial and reversible conversion into CO” (max. ca. 475 nm), probably with the base B still co- ~rdinated,’~and often the appearance of an anomalous band at ca.440 nm.19*33-36 Since protein-free corrinoids do not exhibit a prominent band ca.440 nm either in aqueous solution or in organic solvents such as CH2C12,37 this probably indicates the presence of some additional species. (iii) Several isomerases show an unusual and characteristic e.s.r. signal in the presence of substrate (i.e.when frozen during the enzymatic steady-state), which indicates the presence of both paramagnetic Co” and an organic radical with overlap between their atomic orbitals at a distance of l7 J. Aron, J. M. Pratt, and M. S. Shaikjee, unpublished work. ”L. Randaccio, N. Bresciani-Pahor, P. J. Toscano, and L. G. Marzilli, J. Am. Chem. SOC., 1981,103,6347. 29 H. P. C. Hogenkamp and T. G. Oikawa, J. Biol. Chem, 1964, 239, 1911. 30 S. S. Kerwar, T. A. Smith, and R. H. Abeles, J. Biol. Chem., 1970, 245, 1169.31 T. Toraya, K. Ushio, S. Fukui, and H. P. C. Hogenkamp, J. Biol. Chem., 1977,252, 963. 32 H. P. C. Hogenkamp, in ‘Vitamin BIZ’, ed. B. Zagalak and W. Friedrich, de Gruyter, Berlin, 1979, p. 489. 33 B. M. Babior, T. H. Moss, W. H. Orme-Johnson, and H. Beinert, J. Biol. Chem., 1974,249, 4537. 34 Y. Tamao and R. L. Blakley, Biochemistry, 1973, 12, 24. 35 T. Toraya, E. Krodel, A. S. Mildvan, and R. H. Abeles, Biochemistry, 1979, 18, 417. 36 M. R. Holloway, H. A. White, K. N. Joblin, A. W. Johnson, M. F. Lappert, and 0. C. Wallis, in ‘Vitamin BIZ’, ed. B. Zagalak and W. Friedrich, de Gruyter, Berlin, 1979, p. 471. 37 Y. Murakami, Y. Hisaeda, and A. Kajihara, Bull. Chem. SOC. Jpn., 1983,56, 3642. 167 BIZ-dependent Isomerase Enzymes 2 10 In both cases where the radical has been identified, it is derived from the substrate or a substrate analog~e.~~.~' Even though the occurrence of step (1) is accepted by almost all workers in the B,, field, failure to detect the radical derived from the ligand leaves a slight question mark hanging over any discussion of step (1) and of the role of the protein. There is a real need to find a substrate analogue (perhaps a diol in which the C-H bonds are replaced by C-F bonds) which can trigger step (1) of the reaction without itself reacting further with the radical produced.The 'working' value oft+ -10 years for homolytic fission of the Co-C bond in the protein-free coenzyme at room temperature can be combined 26 with the value of k, -2 x lo9 dm3 mol-' s-' (obtained by picosecond flash photolysis) 11*42 to give a value of K,, = ([B-Co**][R*])/[B-Co-R] = k,/k, x lo-'* dm3 mol-' for the equilibrium constant of reaction (1); the Co-C bond is clearly very stable towards homolytic fission.By contrast, turn-over numbers of up to 370 s-' (i.e. t+ < 2.5 ms) have been reported for one enzyme43 and similar values may be expected for others; this corresponds to labilization of the Co-C bond by 2 10''in the presence of substrate. In the same enzyme the Co-C bond is slowly decomposed by 0, (presumably via an initial homolytic fission)42 even in the absence of substrate with t+ = 20 min at 30 0C,44which corresponds to labilization by ca. 3 x lo5.The sensitivity of 0, can, however, be reduced by small changes in the ligand and side-chains of the coenzyme 35*38 and also appears to be lower in many other isomerases; one can therefore assume a working figure of ,<lo5 for labilization by the protein alone. The main role of the protein is therefore to displace equilibrium (1) to the right by a massive overall factor of 2 10' l, which is achieved in two stages, viz.a smaller step of < lo5 when the coenzyme is bound by the protein and a larger step of 2 lo6 when the substrate is added; further discussion will be confined to the latter. Since k, (in free solution) is virtually diffusion controlled, there is no scope for the protein to labilize the Co-C bond by increasing both the rate constants k, and k, without significantly altering the equilibrium constant K,, (i.e.by a kinetic effect) and the protein must labilize the Co-C bond by changing K,, (i.e.by a thermodynamic effect).8*26 Little is known about the associated changes in protein conformation, though in one enzyme the difference in conformation (between the forms with intact and broken Co-C bond) is sufficient to produce significant differences in their susceptibility to denaturing agent^.^' Co-C bond labilization therefore appears to be associated with the binding of substrate and with a change of conformation; it j8 K.L. Schepler, W. R. Dunham, R. H. Sands, J. A. Fee, and R. H. Abeles, Biochim. Biophys. Acra, 1975, 397, 510. jgG. R. Buettner and R. E. Coffman, Biochim. Biophys. Acta, 1977, 480,495.40 J. F. Boas, P. R. Hicks, J. R. Pillbrow, and T. D. Smith, J. Chem. SOC.,Faraday Trans. 2, 1978, 417. 41 J. E. Valinsky, R. H. Abeles, and A. S. Mildvan, J. Biol. Chem., 1974, 249, 2751. 42 J. F. Endicott and T. L. Netzel, J. Am. Chem. Soc., 1979, 101,4000. *' T. Toraya, T. Shirakashi, T. Kosuga, and S. Fukui, Biochem. Biophys. Res. Commun., 1976,69,475. 44 0.W. Wagner, H. A, Lee, P. A. Frey, and R. H. Abeles, J. Biol. Chem., 1966, 241, 1715. 45 B. M. Babior, in ref: 2, Vol. 2, 1982, p. 263. 168 Pratt can be re-written as in equation (3), where P, and P, represent the two conformations of the protein and the square brackets enclose the protein and protein-bound species. Equation (3) serves to emphasise that a large change in equilibrium (1) can readily be accommodated, even with a low binding constant for the substrate, provided it is accompanied by a large compensating change in free energy due to the change in protein conformation.20 [P,*B-Co-R] + SH [P,-B-Co" + R-SH] (3) The protein must also ensure that the products of step (1) are not destroyed by, for example, irreversible reaction with 0,.Kinetic studies on the protein-free Co" indicate that autoxidation can be suppressed simply by preventing interaction between two Co" complexes and by excluding amino-acids such as cysteine and tyrosine, without the need to exclude 02.46Free radicals, on the other hand, react avidly with 0,;protection is much more difficult and, as shown by decomposition even in the absence of substrate, only partly effective.Perhaps the protein provides a (?H-bonded) environment which reduces the 'solubility' of 0,in the vicinity of the radicals. 6 Synthesis and Summary The B ,-dependent isomerase enzymes exploit the organometallic chemistry of an unusual organocobalt corrinoid coenzyme to catalyse some highly unorthodox reactions involving the intermediate formation of radicals. These B ,coenzymes offer an apparently unique opportunity for directly comparing the properties of a metallocoenzyme in the presence and absence of protein. Studies on the protein- free corrinoids have shown that increasing steric distortion around the co-ordinated ca-atom (a)can cause a dramatic labilization of the Co-C bond through the displacement of equilibrium (l), and (b) will eventually convert a relatively stable red six-co-ordinate form into a much more labile yellow five-co-ordinate form with a band at ca.440 nm, while studies on the enzymes have shown that the addition of substrate (i) produces Co" through the displacement of equilibrium (1) by a factor of b lo6and (ii) often causes the appearance of an additional band at ca. 440 nm which does not belong to Co". Point (a) alone provides experimental support for the postulate that the protein (on the binding of the substrate) displaces equilibrium (1) to the right by steric distortion around C,. The coincidence between (ii) and (b) provides supporting evidence and, in addition, suggests that the single step of reaction (1) should be expanded into two distinct steps as follows.26 The coenzyme is present in the resting state as the relatively unstrained red form with the protein in a strained conformation (PT).The binding of the substrate is accompanied by a change to the relaxed conformation (PR) which (by, for example, moving the adenine part of the ligand relative to the corrin ring) distorts the co- ordination sphere around C, (and also strains most of the coenzyme-protein contacts) and converts the coenzyme into a strained yellow form with its absorption band at ca. 440 nm (first step). The strained form is then in labile 46 E. W. Abel, J. M. Pratt, R. Whelan, and P. J. Wilkinson, S. Afr. J. Chem., 1977, 30, 1. B, ,-dependent Isomerase Enzymes equilibrium with 'base-on' Co" and the free radical with minimal change in the protein conformation (second step); the steady-state concentration of the strained intermediate (and whether it can be 'seen' or not) will obviously vary from one enzyme to another, depending on the relative rates of formation and removal.In other words, the protein acts as a 'molecular switch' to ensure that the dangerously active radical is generated only when the substrate is present, with the red form and strained yellow form of the coenzyme corresponding to the 'off and 'on' positions respectively. [N-Fe"] + 0,,--'.[N-Fe"-O,] (4) The B,,-dependent isomerases have helped to crystallize some of the emerging themes of metalloenzyme chemistry such as; metal-ligand specificity (why Co for organometallic chemistry or Mo for N,-fixation?); sophisticated redox reactions (of a type hardly yet studied by the co-ordination chemist); and, perhaps most important of all, the mechanism of coupling between changes involving the protein and changes involving the metal, which form the basis of the partnership between metal and protein.Where no new bonds are formed between the protein and the metal complex during the reaction, steric and coulombic interactions provide the main method of coupling. There are interesting parallels between the isomerases and the non-enzymatic haemoglobins, both of which use steric effects. In the haemoglobins the Fe" porphyrin functions by the reversible co-ordination of 0, according to equation (4), where only the axial ligands are given and N represents the co-ordinated (proximal) histidine.In both the isomerases and haemoglobins the protein controls a key equilibrium of the active site [uiz. (1) and (4)] by using a change in conformation to link the equilibrium on the metal with some other equilibrium (binding of substrate and, for example, changes in the other haemoglobin sub-units) and couples the changes involving the metal and the protein through steric distortion of one of the axial ligands (5-deoxyadenosyl and histidine). Coulombic effects are less well known but are obviously involved in the proton-coupled reactions (reduction, co-ordination of monobasic anions and probably of HO, -) of the Fe"' ion in peroxidases; our understanding of these effects will be aided by the recent structural determination of a peroxidase47 and by the availability of protein-free models for proton-coupled red~ction.~~.~' In the field of purely organic enzymes, stress is usually placed on the role of the protein in 'activating' the substrate and increasing a rate constant by lowering the energy of the transition state; the metalloenzymes and proteins, on the other hand, emphasize the ability of the protein (on interaction with substrate, another protein, etc.) to 'activate' the metal (i.e.the active site)50 by changing an equilibrium constant.2o 47 T. L. Poulos, S. T. Freer, R. A. Alden, S. L. Edwards, U. Skogland, K. Takio, B. Eriksson, N. Xuong,T. Yonetani, and J. Kraut, J. Biol. Chem., 1980, 225, 575. 48 D. A. Baldwin, V. M. Campbell, L. A. Carleo, H. M. Marques, and J. M. Pratt, J. Am. Chem. SOC.,1981, 103, 186. 49 D. A. Baldwin, V. M. Campbell, H. M. Marques, and J. M. Pratt, FEBS Lett., 1984, 167, 339. B. L. Vallee and R. J. P. Williams, Proc. Natl. Acad. Sci. USA, 1968, 59, 498. 170

ISSN:0306-0012    

DOI:10.1039/CS9851400161

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

S-Nitrosation and the Reactions of S-Nitroso Compounds D. Lyn H. Williams DEPARTMENT OF CHEMISTRY, DURHAM UNIVERSITY, SOUTH ROAD, DURHAM DH1 3LE 1 Introduction Nitrosation reactions generally have been well known in chemistry for a long time. The reactions have been much used synthetically and many aspects have been examined mechanistically. By far the largest literature refers to N-nitrosation, which has been important synthetically, industrially, and biologically, as well as mechanistically. The nitrosation and diazotization of amines has been much studied and it has been possible to identify a range of specific nitrosating agents, as well as to describe their reactivity quantitatively for a wide range of substrates. Standard texts in organic chemistry usually discuss the chemistry of N-nitrosation in some detail.Azo dyes result from the nitrosation of aromatic and heterocyclic amines, and there is much concern regarding the formation of carcinogenic nitrosamines from secondary (and tertiary) amines. In addition, C-nitrosation of both aliphatic and aromatic systems is well described, the synthesis of alkyl nitrites from alcohols represents an example of 0-nitrosation, and there is a large literature referring to nitrosation at metal centres. Much less is known about S-nitrosation processes (although there is an obvious formal similarity to 0-nitrosation) due at least in part to the relative instability of the initially formed S-nitroso species. It is to be expected that sulphur-containing compounds generally would be susceptible to electrophilic nitrosation since the more polarizable sulphur atom is known (in other reactions) to be more nucleophilic than a corresponding oxygen atom.In recent years however there has been a significant interest in the area of S-nitrosation, both from the synthetic and mechanistic viewpoints, derived from better handling techniques for the unstable compounds and the availability of fast reaction methods to measure rate constants of the rapid reactions involved. Further there is at least one important biological aspect involving the use of vasodilatory drugs such as alkyl nitrites, alkyl nitrates, and the pentacyano- nitrosylferrate anion (nitroprusside) where it is believed that reactions with -SH sites are involved.It seems an appropriate time to review this area of S-nitrosation generally, to include both synthetic and mechanistic aspects, and to draw comparisons and contrasts, where appropriate, with the corresponding 0-and N-nitrosation reactions. 2 Nitrosation of Tbiols The reaction of thiols, both aliphatic and aromatic, with nitrosating agents NOX (NOCl, RONO, N,O,, NO,, N,O,, HNO, etc.) to form S-nitrosothiols or J. H.Ridd, Quart. Reo., 1961, 15, 418. S-Nitrosation and the Reactions of S-Nitroso Compounds thionitrites (equation 1) probably represents the best-known example of a S-RSH + NOX -RSNO + HX (1) nitrosation process. The products are coloured yellow or red or in some cases green, and have been known in solution for some time, indeed this reaction has been used as a quantitative test for nitrosyl sulphuric acid using thioglycolic acid as the reagent.2 The reaction appears to be quite general from the point of view of the thiol and also from the range of conventional nitrosating agents, including the reaction of alkyl nitrites with thiol~.~ In contrast to the corresponding reaction of alcohols4 the formation of nitrosothiols is essentially irreversible.There is a recent comprehensive review of the chemistry of nitrosothiols written from the synthetic angle and in terms of the use of these compounds as synthetic reagents. Nitrosothiols are mostly unstable (particularly when compared with alkyl nitrites), decomposing to give the disulphide and nitric oxide (equation 2), 2RSNO-RSSR + 2N0 (2) presumably by a homolytic mechanism, although other non-radical pathways are possible.The most stable compounds appear to be those with bulky substituents at the carbon atom attached to the sulphur, e.g. t-butyl nitrosothiol,6 triphenylmethyl nitrosothiol,’ and the nitrosothiol derived from N-acetylpenicillamine which (as a deep green solid with violet reflections) is indefinitely stable as the solid at room temperature, but which decomposes slowly in solution. Until recently it was believed that the best yields of nitrosothiols were obtained by reaction of the thiol with dinitrogen tetroxide N204in equimolar amounts at ca. -10°C in an inert solvent such as hexane, ether, carbon tetrachloride, or acetonitrile.’ Dinitrogen tetroxide can act as an excellent nitrosating agent (as well as being a source of NO2 radicals) and can be thought of as NOfN03- in a number of inorganic reactions lo as well as organic reactions where, for example, alkyl nitrites can be made from alcohols,’ nitrosamines from secondary amines,12 nitrosamides from amides,13 and the nitroso nitrate adduct from an alkene.14 Interestingly, a number of thiols can be converted into nitrosothiols directly M.Ziegler and 0.Glemser, Z. analyt. Chem., 1955, 144, 187; G. Robisch and E. Ludwig, Z. Chem., 1974, 14, 103. H. Lecher and W. Siefken, Ber., 1926, 59, 1314, 2594. P. A. S. Smith, ‘The Chemistry of Open-Chain Organic Nitrogen Compounds’, Benjamin, New York, 1966, Vol.11, p. 498.’S. Oae and K. Shinhama, Org. Prep. Proced. Int., 1983, 15, 165. G. Kresze and U. Uhlich, Chem. Ber., 1959, 92, 1048. ’H. Rheinboldt, Ber., 1926, 59, 1311. L. Field, R. V. Dilts, R. Ramanthan, P. G. Lenhert, and G. E. Carnahan,J. Chem. SOC.,Chem. Commun., 1978, 249. S. Oae, Y. H. Kim, D. Fukushima, and K. Shinhama, J. Chem. SOC.,Perkin Trans. 1, 1978, 913. lo C. C. Addison, Angew. Chem., 1960,72, 193. l1 P. Gray and A. D. Yoffe, Chem. Rev., 1955, 55, 1069. l2 E. H. White and W. R. Feldman, J. Am. Chem. SOC.,1957, 79, 5832. l3 E. H. White, J. Am. Chem. SOC.,1955, 77, 6008, 6011, 6014. l4 L. Parts and J. T. Miller, J. Phys. Chem., 1969, 73, 3088. WiIIiams using nitric oxide under conditions when oxygen is completely eliminated.’ ’vl These include’ cysteine, N-acetylcysteine, glutathione, P-D-thioglucose etc.On standing for 20 h however, the reaction mixture from cysteine gave the disulphide; l6 it is believed, by decomposition of the intermediate nitrosothiol. Earlier workers” had also reported that thiols reacted with nitric oxide in the presence of base to give either the nitrosothiols or the disulphides. Recent work l8 has suggested that nitrosation and subsequent decomposition to yield the disulphide occurs only by a base-catalysed reaction (see Scheme 1) and that reactions were much inhibited by reduction of the pH below 4. RSH + B-eRS-+ BH RS--f-NO --+ RS-N-0-5RS-I~OH 2RS-N-OH -RSN(0H)-N(0H)SR 1 RSSR + HzNzO, (-Nz + N,O) Scheme 1 If oxygen is not completely eliminated then a reaction via NO, is observed, again leading to the disulphide uia, it is thought, the intermediate nitrosothiol.’* This work suggests that NO and NO, could be potent inhibitors of thiol-dependent enzymes; such effects have been brought about by cigarette smoke,lg which is known to contain nitrogen oxides.A similar reaction scheme (see Scheme 2) has been proposed 2o for the reaction of thiolate anion with two simple nitrosamines, where minute quantities of a free radical, assigned as R,NN(O*)SR’ were detected by e.s.r. spectroscopy. /O’R‘S’ + RZNNO ,T RZN-N ‘SRI l5 L. J. Ignarro, B. K. Barry, D. Y. Gruetter, J. C. Edwards, E. H. Ohlstein, C. A. Gruetter, and W. H. Baricos, Biochem. Biophys. Res. Commun., 1980, 94, 93.l6 U. Schultz and D. R. McCalla, Can. J. Chem., 1969, 47, 2021. l7 H. Reihlen, A. Friedolsheim, and W. Oswald, Justus Liebigs Ann. Chem., 1928,465,72; R. Longhi, R. 0. Ragsdale, and R. S. Drago, Znorg. Chem., 1962, 1, 768. W. A. Pryor, D. F. Church, C. K. Govindan, and G. Crank, J. Org. Chem., 1982,47, 156. l9 M. L. Fenner and J. Braven, Br. J. Cancer, 1968, 22, 474. 2o W. A. Waters, J. Chem. SOC.,Chem. Commun., 1978, 741. S-Nitrosation and the Reactions of S-Nitroso Compounds A similar situation with regard to possible reactions by nitric oxide exists for amines. There, in the complete absence of air and catalysts (such as metal halides and iodine) no reaction occurs either in organic solvents 21 or in the gas phase,22 whereas rapid nitrosation occurs when air is admitted, no doubt via nitrogen dioxide.There is no report of nitrosation of an amine by nitric oxide in strongly basic media, though perhaps this is not a reaction one would expect, given that secondary amines are much less acidic than are the thiols. Cysteine is also converted almost quantitatively into cystine by a nitrosamide in aqueous ethanol at pH 7 in an atmosphere of nitrogen. A mechanism was suggested based on the formation and subsequent decomposition of S-nitrosocysteine, though it was not established whether it was formed by direct attack of the nitrosamide or whether some free nitrosating species was first formed. It has recently been suggested 23 that convenient syntheses of nitrosothiols can be achieved using an alkyl nitrite, specifically t-butyl nitrite, as the reagent in a solvent such as chloroform.Quantitative yields were observed from t-butyl thiol and benzyl thiol. All the experiments so far described on the nitrosation of amino thiols (cysteine etc.) indicate that S-rather than N-nitrosation occurs. However it does seem that under certain conditions the product derived from N-nitrosation can be formed in reasonable yield. Thus 24 from cysteine itself 55% of the thiirancarboxylic acid (1) was recovered, probably by way of the diazonium intermediate (2) by nucleophilic 72" HS-CH2-CF displacement of nitrogen by the thiol group. Similar results were obtained for the methyl ester of cysteine and (in rather lower yield) for penicillamine.It is possible that the N-nitroso intermediate derives from the initially formed S-nitroso species; an example of such a rearrangement will be discussed in Section 4. Some rate measurements have been carried out on the nitrosation of thiols. Three groups of workers have independently established the rate law given in equation 3 for the reactions of t-butyl thiol (in 50% aqueous dio~an),~~ cysteine (in Rate = k[Thiol][H+][HNO,] (3) 21 B. C. Challis and S. A. Kyrtopoulos, J. Chem. SOC.,Perkin Trans. I, 1979, 299. 22 G. B. Neurath, M. Dunger, and F. G. Pein in 'Environmental N-Nitroso Compounds, Analysis, and Formation', ed. E. A. Walker, P. Bogovski, and L. Griciute, IARC Scientific Publication No. 14, International Agency for Research on Cancer, Lyon, 1976, pp.215-225; D.Spincer and D. T. Westcott, ibid., pp. 133- 139. 23 M. P. Doyle, J. W. Terpstra, R. A. Pickering, and D. M. LePoire, J. Org. Chem., 1983, 48,3379. 24 C. D. Maycock and R. J. Stoodley, J. Chem. SOC.,Chem. Commun., 1976, 234. 25 G.Kresze and J. Winkler, Chem. Ber., 1963, %, 1203. 174 Williams water) 26 and a number of mercapto-carboxylic acids (in water).27 This is a familiar rate equation 28 which applies under certain experimental conditions to the nitrosation of a wide variety of substrates including primary and secondary amines (aromatic and aliphatic), alcohols, hydrazine, hydrazoic acid, ureas etc., and is generally interpreted in terms of a mechanism involving rate-limiting attack by a positively charged species H2N02+ or NO+.At the acidities employed the extent of protonation of the thiol is negligible so there is no complication here of the kind generally found for basic substrates. Values of the third-order rate constant k equation 3) for these thiols together with some unpublished results are included in Table 1,together with a few selected values for some other substrate for comparison Table 1 Values of k (equation 3) for acid-catalysed nitrosation in water at 25 "C Substrate k/I2 molP s-l Ref: CO(NH2)Z 0.89 29 2P-Dinitroaniline 2.5 29 p 3 160 29 NH3NH2 620, 611 29, 30 MeOH 700 31 (NHz)zCS 6 960 26 Sulphanilic acid 7 300 29 Me3CSH 47 * 25 Cysteine methyl ester 213 32 C ysteine 456,443 26,27 N-Acet ylpenicillamine 840t 31 Glutathione 1080 32 Mercaptosuccinic acid 1334 27 Thioglycolic acid 2 630 32 Mercaptopropanoic acid 4 764 27 In 50% dioxan-water.t At 31 "C purposes. It is clear that S-nitrosation is a very reactive process, and requires a fast reaction technique such as stopped-flow spectrophotometry to enable rate constants to be measured. Most of the thiols studied, for example are more reactive than HN,, +NH,NH2, and CH,OH. It has been argued,,, on the basis of the relative constancy of k values for very reactive species (e.g. thioureas and aniline derivatives), that these reactions occur at the encounter limit of the substrate with H2N02+ or NO+. This limit appears to be around 7 OOO l2 mol-2 s-I for neutral 26 P.Collings, K. Al-Mallah, and G. Stedman, J. Chem. SOC.,Perkin Trans. 2, 1975, 1734. 27 L. R. Dix and D. L. H. Williams, J. Chem. SOC.,Perkin Trans. 2, 1984, 109. 28 D. L. H. Williams, Ah. Phys. Org. Chem., 1983, 19, 381. 29 J. Fitzpatrick, T. A. Meyer, M. E. O'Neill, and D. L. H. Williams, J. Chem. SOC.,Perkin Trans. 2, 1984, 927. 30 J. R. Perrott, G. Stedman, and N. Uysal, J. Chem. Soc., Dalton Trans., 1976, 2058. 31 S. E. Aldred, D. L. H. Williams, and M. Garley, J. Chern. Soc., Perkin Trans. 2, 1982, 777. 32 P. A. Morns and D. L. H. Williams to be published. 33 J. H. Ridd, Ado. Phys. Org. Chem., 1978, 16, 1. S-Nitrosation and the Reactions of S-Nitroso Compounds substrates. A number of the thiols studied (although containing deactivating substituents e.g.-CO,H) have kvalues approaching this limit, which suggests that here also reaction occurs at or close to the diffusion limit. In one case" the measured experimental activation energy is 56 kJ mol-' which is close to that expected for such a process because of the pre-equilibria involved and the temperature dependence of the viscosity of the solvent. Nucleophilic catalysis of nitrosation of amines and other species is well known and is believed to arise by reaction of an equilibrium concentration of the corresponding nitrosyl compound NOX (equation 4). The extent of catalysis is HNO, + Hf+ X-eNOX + H,O (4) principally governed 28 by the size of the equilibrium constant for NOX formation + which increases along the series ONCl < ONBr < ONSCN < ONSC(NH,), as expected from the nucleophilicities of the anions concerned, even though the rate constants for attack by NOX decrease along the same series.Such catalysis does not occur for amides 34 nor for some amines containing electron-withdrawing groups,29 and this has been attributed 35 to the importance of the reversibility of attack by the nitrosating species, in these systems. Thiols however all show catalysis by halide ion and thiocyanate ion and it is possible to extract the second-order rate constants k, for attack by the NOX species. The data are presented in Table 2 and Table 2 Second-order rate constants k2 for the reactions of NOCl, NOBr and NOSCN with thiols in water at 25 "C k2 values/l mol-' s-l NOCl NOBr NOSCN Cysteine methyl ester 1.0 x lo6 4.9 x lo4 7.0 x 10, Cysteine 1.2 x lo6 5.8 x lo4 7.0 x 10, Glutathione 5.7 x lo6 2.9 x 10' 1.9 x lo3 Mercaptosuccinic acid -2.6 x 104 -Thioglycolic acid 1.4 x 107 1.1 x 106 2.5 x 104 Mercaptopropanoic acid -4.5 x 10' -are taken from references 27 and 32.The by now well-established28 trend of reactivity ONCl > ONBr > ONSCN is now extended to their reaction with thiols. Surprisingly the most reactive thiols towards ONCl have k, values which are 2-3 powers of ten less than the calculated value33 expected for a diffusion-controlled process. It is not clear why this should be so. The high reactivity of thiols in nitrosation suggests that in general they would make excellent scavengers for nitrous acid, particularly those with a high solubility in water, such as the mercaptocarboxylic acids.There are a number of situations jq C. N. Berry and B. C. Challis,J. Chem. SOC.,Perkin Trans.2,1974,1638; M. Yamamoto, T. Yamada, and A. Tanimura, J. Food Hygiene SOC.Jpn., 1976, 17, 363. 35 G. Hallett and D. L. H. Williams,J. Chem. SOC.,Perkin Trans. 2, 1980, 1372. 176 Williams which call for the removal of nitrite or nitrous acid from solution, e.g. in some nitration reactions in strong nitric acid and in some biologically related systems to prevent nitrosamine formation. The efficiency of mercaptopropanoic acid has been examined quantitatively27 by noting its ability to suppress the reversibility of a denitrosation reaction of a nitrosamine.As expected, the thiol was very efficient in this respect and was significantly more so than azide, a well-known nitrous acid scavenger. Similarly it is to be expected that added thiols would suppress amine nitrosation by a direct competition reaction. This is indeed the case36 for the nitrosation of an aniline derivative where the extent of N-nitrosation is markedly re- duced by added cysteine or N-acetylpenicillamine, and it was possible in both cases to inhibit completely the N-nitrosation reaction when sufficient thiol was present. A major difference between 0-and S-nitrosation is that whilst the former is significantly reversible the latter is effectively not. For e~ample,~~?~~ equilibrium constants for the nitrosation of a number of alcohols (and carbohydrates) are in the approximate range 0.5-2.0 1 mol-’, whereas values are too large to measure reasonably for the corresponding sulphur case.The denitrosation of nitrosothiols can be effected at high acidity if steps are taken to remove the free nitrous acid as it is formed.38 An explanation has been suggested 31 based on the differences between the nucleophilicity and basicity of the corresponding oxygen and sulphur sites. For the forward reaction in equation 5, the important factor in the comparison of the H20 RSH + NOX Rs’ *> RSNO + H30’ t X’ (5) ‘NO reactivities of ROH and RSH,is the nucleophilicity difference between the oxygen and sulphur atoms. There is much independent evidence which shows that the sulphur is the more nucleophilic.This is borne out in nitrosation studies, where it has been shown 31 that N-acetylpenicillamine (3) (a reasonably good model for t-butyl thiol) is several orders of magnitude more reactive than t-butyl alcohol (4). For the reverse reaction, however, an important factor (see equation 5) is the HSCMe,CH(NHAc)CO,H HOCMe, (3) (4) relative basicities of the sulphur (in RSNO) and the corresponding oxygen (in RONO) atoms. Again there is much evidence to suggest that oxygen is the more basic by ca. 10’. This corresponds quite well with the estimated38 difference of 2 x lo6 in the reactivity of nitrosothiols and corresponding alkyl nitrites, with the latter being the more reactive. This is an example of where the relative basicity and nucleophilicity of oxygen and sulphur operates in opposite directions for the forward and reverse reactions.It is not possible to compare exactly the reactivities in S-,0-,and N-nitrosations, 36 D. L. H. Williams and S. E. Aldred, Fd. Chem. TOX.,1982, 20, 79. 37 J. Casado, F. M. Lorenzo, M. Mosquera, and M. F. R. Prieto, Can. J. Chem., 1984,62, 136. S. S. Al-Kaabi, D. L. H. Williams, R.Bonnett, and S. L. Ooi, J. Chem. Soc., Perkin Trans. 2, 1982, 227. S-Nitrosation and the Reactions of S-Nitroso Compounds because the relevant compounds have not been subject to a kinetic study under the same conditions, but it is clear that thiols are generally more reactive than corresponding alcohols, and are at least comparable in reactivity with amines.More results are needed in this area. The reactivity of thiols towards nitrosamines has been measured quantitatively using the denitrosation reaction carried out in the presence of a sufficient excess of hydrazine to suppress the reverse reaction 39 (equation 6). Kinetic measurements PhN(Me)NO + Y -!% PhNH(Me) + NOY' (6)1Removed with a range of nucleophiles Y gave the relative reactivity order shown in Table 3. Table 3 Relative reactivities of some nucleophiles towards N-nitroso-N-methylaniline in aqueous acid solution Nucleophile Relative Reactivity Chloride ion 1 Cysteine 2 Glutathione 3 S-Methylcysteine 35 Bromide ion 55 Methionine 65 Thiocyanate ion 5 500 Thiourea 13 OOO Iodide ion 15 750 Clearly both cysteine and glutathione are about as reactive as chloride ion whereas, as expected, the two sulphides are somewhat more reactive whilst falling well short of that of thiourea.Thiols are also oxidized to disulphides by reaction with the pentacyanonitrosyl- ferrate ion in alkaline solution4' as shown in equation 7. Reactions are 2[Fe(CN),NO] 2-+ 2SR--RSSR + 2[Fe(CN),NO] 3-(7) characterized by a very rapid formation of an intense colour, usually red, which then fades gradually. This has been interpreted as a rapid adduct formation (equation 8) which is the coloured species, and the rate of this process has been [Fe(CN),NO] 2-+ SR-e[Fe(CN),N0(SR)l3-(8) measured by stopped-flow spectr~photometry.~~ The bonding in such adducts is not described, but since the nitroso group is believed 42 to act in the NO+ sense in 39 G.Hallett and D. L. H. Williams, J. Chem. SOC.,Perkin Trans. 2, 1980, 624. 40 D. Mulvey and W. A. Waters, J. Chem.SOC.,Dalton Trans.,1975,951; P. J. Morando, E.B. Borghi, L.M. de Schteingart, and M. A. Blesa, ibid., 1981, 435.'' P. A. Rock and J. H. Swinehart, Inorg. Chem., 1966,5, 1078. 42 N. G. Connelly, Inorg. Chim. Acfa, 1972, 6, 47. Williams this iron complex, it could well be that the sulphur is bonded to the nitrogen of the nitroso group. An equivalent interaction has been proposed in the deamination of primary aliphatic amines by the pentacyanonitrosylferrate ion.43 The de-composition of the adduct, in the case of the thiol reaction, has not been established unequivocally. Two suggestions have been made, equations 9 and 10, the first [Fe(CN),NO(SR)] 3-+ RS--[Fe(CN),NO] 4-+ RSSR (9) [Fe(CN),NO(SR)] 3--[Fe(CN),fiO] 3-+ RS' (10)1 RSSR involving attack by RS- and the other a one-electron transfer process.An alternative might well be a unimolecular decomposition of the adduct, eliminating the nitrosothiol which then loses nitric oxide and forms the disulphide. The physical properties of nitrosothiols are well documented in the review by Oae and Shinhamas and will not be reported in detail here. In general they are coloured red or green as the pure compounds but often are yellow in solution in common with many S-nitroso species. Dipole moments, infrared spectra, and U.V. and visible spectra have been reported and are much as expected, and compare reasonably with those of the more widely studied alkyl nitrites, given the expected changes consequent upon the greater electronegativity of oxygen than sulphur."N-N.m.r. chemical shifts have been and have been used to identify S-compounds from model peptides such as N-acetylcysteine, N-acetylpenicillamine, and thioacetic acid, using triphenylmethylthionitrite as a reference. In addition 14N-n.m.r. shifts have also been reported45 for CF3SN0 and EtSNO. Finally the crystal structure of the stable nitrosothiol (3) derived from N-acetylpenicillamine has been determined.* 3 Reactions of S-Nitrosothiols (Thionitrites) Nitrosothiols readily decompose thermally to give disulphides and nitric oxide (equation 11); the same products can be obtained photochemically.Both clearly 2RSNO -RSSR + 2N0 (1 1) involve homolysis of the S-N bond. Oxidation can give the thionitrate, and reduction the thiol or the disulphide. Reaction with thiols gives the unsymmetrical disulphides, and with sulphinic acid the thiolsulphonates. Nitrosothiols have also been used to convert secondary amines into N-nitrosamines, to effect deamination of arylamines, and to convert alcohols into alkyl nitrites; these reactions are all described in reference 5 and in references therein. The remainder of this section will concentrate on the more recent work and in particular on the limited amount of mechanistic work undertaken. 43 A. R. Butler, C. Glidewell, J.Reglinski, and A. Waddon, J. Chern. Res. (S),1984, 279. '* R. Bonnett, R. Holleyhead, B. L. Johnson, and E. W. Randall, J. Chern. Sor., Perkin Trans. 1,1975,2261. 45 L. 0.Andersson, J. B. Mason, and W. van Bonswijk, J. Chern. SOC.(A), 1970, 296. 179 S-Nitrosation and the Reactions of S-Nitroso Compounds The acid-catalysed hydrolysis of nitrosothiols (equation 12) is many orders of RSNO + H,O&RSH + HNO, (12) magnitude slower than that of alkyl nitrites in general; this has been attributed to the greater basicity of the oxygen atom.38 The reaction has only been studied at relatively high acidities and in the presence of a sufficient excess of added sodium azide (or some other nitrous acid trap) to ensure irreversibility. The method is entirely analogous to that used to study the denitrosation of nitrosamines, where again the equilibrium greatly favours the formation of nitro~amines.~~ As for the nitrosamines the reaction of nitrosothiols is catalysed by halide ion and other nucle~philes,~~in the sequence Cl- -c Br-< SCN--SC(NH,), (equation 13).R&H(NO) + Hal--RSH + NOHal (13)1Removed It has been found47 that the mercuric ion also catalyses the hydrolysis (again in the presence of a nitrous acid scavenger). This was used as the basis of an analytical procedure for the quantitative analysis of thiols. No kinetic data are available, but it has been suggested47 that a S-bound mercury complex (5) is formed which undergoes hydrolysis as in equation 14.In a study relating to the possible in uiuo formation of carcinogenic nitrosamines the trans nitrosation reaction (nitrosothiol + amine -nitrosamine) has been studied using S-nitrosocysteine, S-nitrosoglutathione, and a protein-bound nitrite model sy~tem.~~,~~ All brought about nitrosation of N-methylaniline and other amines in acid and alkaline solution, thus reinforcing an earlier suggestion 50 that S-nitrosothiols can act as transfer-nitrosating agents (equation 15), although it is RSNO + R’R”NH-R’R”NN0 + RSH (15) not yet established whether this is a direct one-stage process or not. In contrast there was no observable catalysis of nitrosation of morpholine by ~ysteine,~~ although S-methylcysteine did show a small catalytic effect. 46 I.D. Biggs and D. L. H. Williams, J. Chern. Soc.. Perkin Trans. 2, 1975, 107. 47 B. Saville, Analyst, 1958, 83, 670. ‘it M. J. Dennis, R. Davies, and D. J. McWeeny, J. Sci. Food Agric., 1979, 30, 639. 49 M. J. Dennis, R. C. Massey, and D. J. McWeeny, J. Sci. Food Agric., 1980, 31, 1195. A. Mirna and K. K. Hofmann, Fleischwirfschaf,1969, 49, 1361. T. A. Meyer and D. L. H. Williams, J. Chern. Soc., Perkin Trans. 2, 1981, 361. 180 Williams It has been known for some time that many potential nitrosating agents can act as vasodepressor agents and a number have been widely used in the treatment of angina, heart failure, and hypertensive emergencies. The substances used include organic nitrites (and nitrates), nitric oxide, sodium nitrite, and the pentacy- anonitrosylferrate (nitroprusside) anion.Little is known about the mechanism of these compounds but it is thought 52 that the relaxation in vascular smooth muscle is dependent on the presence of tissue-bound SH groups. More recently it has been shown ’’that the vasodilator action of alkyl nitrites etc. can be attributed at least in part to the formation of unstable S-nitrosothiols as intermediates. The detailed mode of action is still not resolved but it is believed that S-nitrosothiols markedly activate the enzyme guanylate cyclase which brings about relaxation of vascular smooth muscle and so decreases the systemic arterial pressure. S-Nitrosothiols also bring about inhibition of human platelet aggregati~n.~~ More work remains to be done before a complete mechanistic picture emerges but it is clear that S-nitrosothiols play an important part.4 Nitrosation of Sulphides Although it is to be expected that the sulphur atom in a sulphide is at least as nucleophilic as that in a thiol, there is not such a convenient leaving goup (H+) in the case of the sulphides so S-nitrosothiol formation is not to be expected, and indeed is not found from simple sulphides RSR.However there are reports of S-nitrosothiol formation from disulphides. Thus 55 the reactions of disulphides with N20, give products compatible with the intermediacy of RSNO and RSO’ (equation 16). Similarly the photolysis of disulphides in the presence of nitric oxide RSSR + N,O,-RSNO + RSO+ (16)1 1 Products Products yields the S-nitrosothi~l,~~ presumably according to equation 17.There are also CH3SSCH332CH3S’-% 2CH,SNO (17) reports of ring opening reactions of cyclic sulphides which are interpreted in terms of a S-nitrosation. It has also been known for a long time that simple alkyl sulphides react with alkyl nitrites (and tetranitromethane) to give coloured solutions which fade on tand ding.'^ As far as is known, intermediates and products of such reactions have not been identified positively but it does seem likely that the coloured intermediates are in fact S-nitroso ions. s2 P. Needleman, B. Jakschik, and E. M. Johnson, J. Pharmacol. Exp. Ther., 1973, 187, 324. ” L. J. Ignarro, H. Lippton, J. C. Edwards,W. H. Baricos, A. L. Hyman, P.J. Kadowitz, and C. A. Gruetter, J. Pharmacol. Exp. Ther., 1981, 218, 739. 54 B. T. Mellion, L. J. Ignarro, C. B. Myers, E. H. Ohlstein, B. A. Ballot, A. L. Hyman, and P. J. Kadowitz, Mol. Pharmacol., 1983, 23, 653. ”S. Oae, D. Fukushima, and Y. H. Kim, Chem. Lett., 1978, 279. s6 P. M. Rao, J. A. Copeck, and A. R. Knight, Can. J. Chem., 1967, 45, 1369.’’E. M. Harper and A. K. Macbeth, Proc. Chem. Soc., 1914,30, 15; A. K. Macbeth and D. D. Pratt, J. Chem. Soc., 1921, 119, 354. S-Nitrosation and the Reactions of S-Nitroso Compounds 15-L -Methi? 10 -c I v) \s“ m 5-s lo2[ Substrate 1 / M Figure 1 Variation of thejrst-order rate constant with [Substrate] for the deamination of methionine (l), S-methylcysteine (2) and alanine (3) The deamination of both methionine and S-methylcysteine proceed normally to give the expected alcohol products,58 but at a much enhanced rate (see Figure l), compared with a similar amine without the -SR This has been interpreted as first involving a S-nitroso intermediate which undergoes an intramolecular S-to N-rearrangement, leading to the alcohol product as shown in equation 18 for S-methylcysteine. In this case a favourable five-membered ring transition-state would be involved, whereas for methionine a six-membered ring interaction is proposed. There are a number of examples in the literature where S-to N-rearrangement of this kind occurs, although it is not established whether it occurs intramolecularly or intermolecularly.For example,60 acylation of thioamides and thioureas give initially the S-bonded isomer which rearranges on heating to the N-acylated product.It is possible that S-nitroso ions of the type (6) could also act intermolecularly to effect nitrosation of a suitable species. In practice this would be observable kinetically as catalysis of nitrosation (or diazotization) by added RSR,in the same way as catalysis is brought about by added C1-, Br-, SCN-etc. Such experiments have not yet been described, although there is one report 51 of a measure of catalysis of diazotization of aniline by added S-methylcysteine, but this example is complicated by the possibility of an intramolecular reaction. ’’G. A. Maw and C. M. Coyne, Arch. Biochem. Biophys., 1966, 117, 499.59 T. A. Meyer and D. L. H. Williams, J. Chem. Soc., Chem. Commun., 1983, 1067. 6o M. L. Moore and P. S. Crossley, J. Am. Chem. SOC.,1940,62, 3273. Williams There is strong kinetic evidence for S-nitrosation of sulphides using nitrosamines on the source of the nitroso The denitrosation of N-methyl- N-nitrosoaniline in the presence of a sufficient excess of hydrazine to ensure irreversibility is catalysed by both methionine and glutathione to approximately the same extent. Both are significantly more reactive than the thiol cysteine as expected and are approximately as reactive as bromide ion (see Table 3, Section 2). This reaction scheme (Scheme 3) implies also that the nitrososulphonium ion (7) PhN(Me)NO + H+ Ph&H(Me)NO Ph&H(Me)NO + RSR'"OW.PhNHMe + RiR' INO RBR' + NH,NH, -@+ RSR' + decomposition products INO Scheme 3 can itself nitrosate hydrazine. More work is needed to develop this idea further. Sulphides can also undergo another kind of reaction with the nitrosonium ion in a one-electron transfer reaction leading to radical cation formation. This is not in the normal sense a S-nitrosation reaction but, since it is closely related, it is included in this section. The cyclic disulphides 1,5-dithiocyclo-octane and 13- dithiocyclononane are oxidized by one equivalent of NO+ (added as the tetrafluoroborate) in acetonitrile or propionitrile to the long-lived cation radical 61 as shown in equation 19. The ion which can be isolated as the hexafluorophosphate salt, shows sulphur-sulphur transannular interaction as evidenced by the e.s.r.spectrum. Similarly, two equivalents of NO + yield the dication (equation 20) again with the S-S transannular bridge, and an analogous N-S bond is formed when one of the sulphur atoms in the cyclo-octane is replaced by -NMe.62 S,S-Acetals (which can be thought of as sulphides) undergo reaction with nitrous 61 W. K. Musker, T. L. Wolford, and P. B. Roush, J. Am. Chem. SOC.,1978, 100, 6416. 62 W. K. Musker, A. S. Hirschon, and J. T.Doi, J. Am. Chem. SOC.,1978, 100, 7754. 183 S-Nitrosation and the Reactions of S-Nitroso Compounds acid (and with other electrophiles) to give the corresponding carbonyl compound.63 Reaction is believed to involve S-nitrosation followed by loss of RSNO and subsequently loss of RSH as outlined in Scheme 4.This has a certain x:,” Xi,”HN02’HJ I NO H20 -RSNO1 \ -RSH r0-OH Scheme 4 similarity to the reaction discussed in Section 5 whereby thiocarbonyl compounds generally are converted into the carbonyl compounds by reaction with nitrous acid. 5 Nitrosation of Tbiocarbonyl Compounds The sulphur atom in thiocarbonyl compounds has a pronounced nucleophilic reactivity (e.g. thiourea is as reactive as iodide ion as measured by the Pearson nucleophilicity parameter 64) so it is to be expected that electrophilic S-nitrosation occurs. This is indeed the case and all reactions studied in this area involve the formation of the S-nitrososulphonium ion as shown in equation 21.Most of the work reported has been concerned with the nitrosation of thiourea and its alkyl derivatives no doubt because of the stability of these species compared with other thiocarbonyl compounds e.g. thioketones. It has been known for some time by the work of Werner 65 that thiourea can undergo two reactions with nitrous acid, one leading to nitrogen and thiocyanate ion (equation 22) and the other to a disulphide HNO, + (NH,),CS -H+ + SCN-+ N, + 2H20 (22) cation (C,C-dithiodiformamidinium),equation 23. The former, which predominates ++ 2HN02 + 2H’ + 2(NH,),CS -(NH2),CSSC(NH2), + 2N0 + 2H20 (23) at low acidities, can readily be rationalized in terms of N-nitrosation whilst the 63 M. T. M. El-Wassimy, K. A. Jsrgensen, and S.0.Lawesson, J. Chem. Soc., Perkin Trans. I, 1983,2201. 64 R. G. Pearson, H. Sobel, and J. Songstad, J. Am. Chem. Soc., 1968,90, 319. A. E. Werner, J. Chem. SOC.,1912,101,2180; M. E. Coade and A. E. Werner, J. Chem. Sor., 1913, 102, 1221. Williams latter, which can be explained by S-nitrosation, takes over at higher acidity. Again, in common with all S-nitrosation reactions, a transient yellow or red colour is observed here and seems to be a property of the S-nitroso ion. The structure of the dication has been established by crystal structure analysis of its salts.66 The same product can be obtained using other oxidizing agents such as hydrogen peroxide, the halogens and peracids, and also using nitro~amines.~~ The reaction is very general and not only thioureas, but also thiocarbonates and thioketones (including heterocyclic systems) have been converted into stable dications containing the -S-S-bond by a variety of chemical and electrochemical oxidation procedures.66 The equilibrium constant for the formation of the S-nitrosothiouronium ion (equation 24) has been measured6' as has the rate constant for S-nitrosation, by H+ + HNO, + (NH2),CS e(NH,),C&NO + H20 (24) noting the increasing absorbance due to the yellow S-nitroso ion in a stopped-flow spectrophotometer.The value of the equilibrium constant is 5 OOO l2 moF2 at 25 "C, which means that at suitable concentrations of the reagents substantial conversion of the nitrous acid into the S-nitroso ion can occur.This contrasts with the corresponding situation for the reaction with halide ion where only very small equilibrium quantities of the nitrosyl halides are formed. The rate constant k (equation 25) for the nitrosation of thiourea was found 26 to Rate = k[H+][HNO,][(NH,),CS) (25) be 6 960 l2 moF2 s-' at 25 "C which is only slightly greater than the values reported 70 for aniline, o-toluidine, and o-chloroaniline, and very close to that found for sulphanilic acid,29 all of which suggests that in each case these reactions occur by diffusion-controlled reaction between the substrate and the positively charged nitrosating agent. Further the value of k is little changed by N-methyl substitution26 and the activation energy of 65 kJ is close to that expected for such a diffusion-controlled process.A similar reaction occurs with nitrosamines and thiourea in acid solution,67 yielding initially the yellow colour characteristic of the S-nitroso species which then gives the disulphide salt as before. Quantitative kinetic measurements were carried out under slightly different conditions, in the presence of an excess of a suitable nitrous acid trap (hydrazine etc.) which destroyed the S-nitroso ion rapidly, so that the direct S-nitrosation of thiourea could then be examined without any reversibility problems. That the reaction is a direct one (equation 26) is shown by 66 0.Foss, J. Johnsen, and 0.Tvedten, Acfa Chem. Scand., 1958, 12, 1782, and references therein. 67 D. L. H. Williams, J. Chem.SOC.,Perkin Trans. 2, 1977, 128. 68 R. L. Blankespoor, M. P. Doyle, D. M. Hedstrand, W. H. Tamblyn, and D. A. Van Dyke, J. Am. Chem. SOC.,1981, 103, 7096. 69 K. Al-Mallah, P. Collings, and G. Stedman, J. Chem. SOC.,Dalton Trans., 1974, 2469. 'O H. Schrnid and C. Essler, Monatsh., 1960,91,484. S-Nitrosation and the Reactions of S-Nitroso Compounds the fact that it occurs in the presence of nitrous acid traps so that hydrolysis to free nitrous acid can be ruled out. It was possible to establish the reactivity of thiourea (and alkyl thioureas) in this reaction for two nitr~samines,~~,~’ along with that for a number of other nucleophiles. The data (in Section 2, Table 3 for R = Ph, R’ = Me) shows clearly that thiourea is indeed very reactive in this reaction and is comparable with iodide in this respect.The data for the nucleophiles fit well the Pearson nucleophilicity relation~hip~~ particularly for R = Ph, R’ = Me, but rather less well for R = R’ = Ph, in the case of the larger nucleophiles I-and SC(NH&, where steric effects might operate. The Fischer-Hepp rearrangement of aromatic nitrosamines occurs in parallel with a normally reversible denitrosation reaction72 (Scheme 5). By the addition of powerful nucleophiles such as thiourea (and a nitrous acid trap) it is possible to divert the reaction pathway towards denitrosation at the expense of rearrangement.73 MeNNO MeNH MeNH Scheme 5 Alkyl nitrites can also apparently directly bring about S-nitrosation of thiourea as deduced from the results of kinetic experiments using propyl nitrite in propanol ~olvent.’~Reactions were first-order in halide ion, thiocyanate ion, and thiourea.S-Nitrosothiols (or thionitrites) behave in the same way.38 It seems that the formation of the S-nitrosothiouronium ion from thiourea is a general one, occurring readily with any of the conventional nitrosating agents. An apparent exception is the reaction of thiourea with photolysed pentacyanonitrosylferrate ion,75 where nitric oxide is generated and the iron complex product contains a bonded thiourea group. ” J. T. Thompson and D. L. H. Williams, J. Chem. SOC.,Perkin Trans.2, 1977, 1932. ’’ D. L. H. Williams, Tetrahedron, 1975, 1343. 73 D. L. H. Williams, J. Chem.SOC..Perkin Trans.2, 1982, 801. 74 S. E. Aldred and D. L. H. Williams, J. Chem. Soc., Perkin Trans.2, 1981, 1021. 7s P. A. Stoeri and D. X. West, J. Inorg. Nucl. Chem., 1974, 36,3883. Williams For reactions of thiourea with nitrous acid at lower acidities where products are derived from N-nitrosation, it had been suggested 69 that the initial attack might be at sulphur followed by a Sto N rearrangement. The evidence for this is kinetic and derives from the absence of a term in [H'] in the rate equation for reaction leading to thiocyanic acid. However the results of "N-n.m.r. experiment^^^ have been interpreted in terms of a direct N-nitrosation under these conditions. At low acidities it was possible to isolate the N-nitrosothiourea.The acid-catalysed hydrolysis of N-nitrosothioureas do appear to involve the reverse N to S migration. The final product of the nitrosation of thiourea has been described as the disulphide cation, but it appears that urea itself can also be formed (Scheme 6). SC(NH,), + HNO, ONk(NH,), (NH2),6SS6(NH,), OC(NHd2 Scheme 6 Thus urea is claimed to be the major product from the S-nitrosation at higher acidity," and is also formed in the hydrolysis of N-nitros~thiourea~~ (after N to S migration). It is believed that this occurs by nucleophilic attack of water or by the elimination of HSNO giving a carbodiimide which subsequently undergoes hydration. The transformation of thiocarbonyl compounds to carbonyl compounds (equation 27) can be achieved using a variety of other reagents including interestingly an alkyl nitrite,79 a reaction which could also involve S-nitrosation.It is not clear under what conditions the two possible alternative products (the dication and the carbonyl compound) are formed. The kinetics of the carbonyl-forming reaction were studied using N-methyl-2-thiopyrrolidoneas the substrate. As expected the reaction is acid-catalysed and also catalysed by thiocyanate ion. On the synthetic side it has been showns1 that a range of thiocarbonyl compounds (8) with R = Ph, Me and R' and R2 Ph, Me and also various cyclic structures, can be converted smoothly and generally in good yield, into the corresponding amide structures, by treatment with excess nitrous acid in 4M-HCl. Similarly a range of thiono compounds (9) where X and Y = 0,S etc., are also converted into their carbonyl analogues by the same treatment.These results were discusseds1 in terms of the HSAB principle where the soft (borderline) acid NO' (or H2N02+) attacks the soft sulphur atom of the thiocarbonyl compound, but it is to be expected for the primary and secondary '6 J. W. Lown and S. M. S. Chauhan, J. Chem. SOC.,Chem. Commun., 1981,675. 77 J. W. Lown and S. M. S. Chauhan,J. Org. Chem., 1983,48,3901. "J. W. Lown and S. M.S. Chauhan, J. Org. Chem., 1983,4%, 507. 79 K. A. Petrov and L. N. Andrew, Russ. Chem. Rev., 1971, 40,505. K. A. Jsrgensen and S. 0.Lawesson, Chem. Scr., 1982,20, 227. K. A. Jsrgensen, A. B. A. G. Ghattas, and S. 0.Lawesson, Tetrahedron, 1982, 38, 1163.S-Nitrosation and the Reactions of S-Nitroso Compounds thioamides that the nitrogen atom is extensively protonated, and indeed no N-nitrosamides are recovered. The thiocarbonyl-carbonyl transformation can also be brought about using nitrosamines as the source of NO+.82A large range of thioamides (8) with R, R', and R2 being various combinations of H, Ph, Me and other groups, yield the corresponding amides by treatment with N-methyl-N-nitrosoaniline or N-nitrosopiperidine in acid solution containing potassium iodide. In the presence of a nitrous acid trap (ascorbic acid) the reaction is extensively inhibited and so a likely mechanism involves the denitrosation of the nitrosamine (see reJ:46),followed by S-nitrosation of the thioamide by the free nitrosating species, as outlined in Scheme 7. The likely rate-limiting step is the denitrosation step with X-as nucleophile PhN(Me)NO + H+ Ph&H(Me)NO PhiH(Me)NO + X' + PhNHMe + NOX \+ NOX +>=S ,C=S-NO which should then lead to a zero-order dependence upon [>C=S]. This was not examined experimentally however.On the synthetic side, it has been shown83 that nitrous acid (and alkyl nitrites) react with primary thiobenzamides to give the disubstituted thiodiazole (equation 28). It is not immediately clear whether this involves S-nitrosation. N-C-Ph PhCSNH2 + HN02 +Ph-C' II (28) 'S-N 82 K. A. Jsrgensen, M. T. M. El-Wassimy, and S. 0.Lawesson, Tetrahedron,1983, 39, 469. 83 M. W. Cronyn and T.W. Nakagawa, J. Am. Chem. SOC.,1952, 74, 3693. Williams 6 Reactions of S-Nitrosothiouronium Ions The further reactions of the S-nitrosothiouronium ions derived from thiourea and related compounds, have already been mentioned in Section 5, see Scheme 6. The reaction leading to the disulphide cation has been examined in more mechanistic detaiLE4 The reaction was found to be inhibited by nitric oxide (in an oxygen-free atmosphere) and the rate equation given by equation 29 was established from Rate = kl[(NH2)2CS][(NH2),CSNO+]+ kz[(NH2)2CSNO+]2 (29) initial rate measurements. The results were consistent with a mechanism involving two parallel pathways, the first a reversible formation of a radical intermediate (10) from thiourea and the S-nitrosothiouronium ion, and the second involving a bimolecular reaction between two molecules of the S-nitroso ion, as outlined in equations 30 and 31.The subsequent fate of (10) is open to speculation but could (NH,),CS + [(NH,),CSNO]+ e[(NH,),CSSC(NH,),]+' + NO (30) (10) 2[(NH2)2CSNO]++(NH,),&SS&(NH,), + 2N0 (31) involve a bimolecular disproportionation, or a unimolecular radical cation breakdown, or possibly further oxidation of (10) by another molecule of the S-nitrosothiouronium ion. A similar rate equation was found for the decomposition of the tetramethylthiourea derivative, but an additional term now appears in the rate equation involving catalysis of the decomposition by nitrous acid, and this is not easy to interpret mechanistically.The possibility arises that S-nitrosothiouronium ions generally might act as nitrosating agents in their own right. This was first suggested by some results of a kinetic study on the denitrosation of nitrosamines using thiourea as a nucleophilic catalyst6' The reaction was found to be reversed by added secondary amine product suggesting the outline mechanisn in Scheme 8. This was later substan- PhN(Me)NO + H+ PhNH(Me)NO PhkH(Me)NO + (NH,),CS ePhNHMe + (NH,),C$NO (NH2),CSN0 + Nitrite trap -Various products Scheme 8 tiated by the observation of marked catalysis of nitrosation and diazotization by added thio~rea.~ 1*85 The well-known catalysis of nitrosation by added halide ion and thiocyanate ion has been interpreted in terms of intermediate formation of the corresponding nitrosyl halide or thiocyanate which effects nitrosation. It appears that the same is true for thiourea.The substantial catalytic effect of thiourea is shown in Figure 2 where catalysis of nitrosation of morpholineS1 by thiocyanate 84 P. Collings, M. Garley, and G. Stedman, J. Chem. Soc., Dalton Trans., 1981, 331; M. S. Garley, G. Stedman, and H. Miller, J. Chem. Soc., Dalton Trans., 1984, 1959. ''M. Masui, C. Ueda, T. Yasuoka, and H. Ohmori, Chem. Pharm. Bull., 1979, 27, 1274. S-Nitrosation and the Reactions of S-Nitroso Compounds 400c Isc( H2)2 Nitrosation ofmorpholine 3001 i' -'"1; 0 5 10 15 20 25 lo3 [x-] /M Figure 2 Comparison of the catalytic eficiencies of Br-, SCN-, and SC(NH2), in the nitrosation of morpholine ion and bromide ion are also shown.The same effect is observed in diazotization of aniline. This makes thiourea one of the best catalysts known for nitrosation processes. The extent of catalysis by X-depends on the magnitude of the equilibrium constant KNoxfor NOX formation and also upon the magnitude of its rate constant for reaction wth substrate S (see Scheme 9). KNoxcan be separately HNO, +X-(or X) +H+ %NOX (or NOX') +H,O NOX (or NOX') +S k,product Scheme 9 measured, and is known for C1-, Br-, SCN-, and SC(NH,),, so that k can readily be extracted. The extent of catalysis seems to depend more on the value of KNox than on the value of the bimolecular rate constant. Some data are collected in Table 4 for the diazotization of aniline and also 4-aminobenzoic acid.The rate constants Table 4 Values of KNoX together with values of k (firom Scheme 9) for the diazotization of aniline and 4-aminobenzoic acid at 25 "C in water k/l mol-' s-' k/l mol-' s-l NOX ~~~x/1~mol-2 (Aniline) (4-aminobenzoic acid) NOCl 1.1 1 ~ 3 2.2 x109 1.1 x109 NOBr NOFCN 5.1 x1C2 30 1.7 x109 1.9 xlo8 4.3 x 108 1.4 xlo6 NOSC(NH2)z 5 OOO 1.3 xlo6 1.8 x104 are taken from references 86 and 87 and the equilibrium constants from references 88 (NOCI), 89 (NOBr), 90 (NOSCN), and 69 (NOk(NH,),). A reasonable Bronsted plot was obtained for both NOSCN and NOk(NH2)2,86whereas k values tend towards the diffusion-controlled limit for both NOBr and NOClg7 190 Williams as the pK, of thf aniline is increased.The reactivity trend NOCl > NOBr > NOSCN > NOSC(NH,), is now well-established for a range of substrates. Catalysis of nitrosation of aliphatic amines by thiourea and alkyl thioureas also occurs,85 and again the overall effect is greater than that of added thiocyanate ion. 7 Nitrosation of Thiocyanate Ion Nitrosyl thiocyanate is known only as a blood-red species, stable only in solution,g1 which decomposes in high concentration at room temperature, and is readily synthesized from nitrous acid and thiocyanic acid, nitrosyl chloride and silver thiocyanate, or ethyl nitrite and thiocyanic acid.', Its structure has never been established, presumably as a result of its instability (it decomposes to give nitric oxide and thiocyanogeng3), but it is generally believed that the nitroso group is bound to sulphur and not to nitrogen.It thus represents an example of S-nitrosation. Arguments based on Hard- Acid-Soft-Base theory favour bonding to sulphur and recent ab initio molecular orbital calculationsg4 reveal that nitrosyl thiocyanate should be significantly more stable than the isomeric nitrosyl isot hiocyanate. There are now many examples of substantial degrees of catalysis by thiocyanate ion of nitrosation in the literature, and this is generally interpreted in terms of intermediate formation of nitrosyl thiocyanate which then effects nitrosation. Among substrates studied which show such catalysis are hydroxylamine and its methyl derivative^,'^ aniline derivative^,^^^^^ morph~line,~ hydrazoic acid,97 alcohol^,^' thi~ls,,~etc.For most cases the rate-limiting step is the attack of NOSCN with the substrate (Scheme 10).The equilibrium constant KNosCN has been HNO, + H+eH,NO,+ H,NO,+ + SCN-eNOSCN + H,O NOSCN + Substrate -nitrosation product Scheme 10 determined separatelyg0 as 30 1, rnol-, at 25 "C and so the bimolecular rate constant for NOSCN attack can be determined. Such values are always less than for the corresponding NOCl and NOBr reactions, and this difference has been discussed theoretically by molecular orbital calculationsg4 using the concept of L. R. Dix and D. L. H. Williams, J. Chem. Res. (S),1984,97. M. R. Crampton, J. T.Thompson, and D. L. H. Williams, J. Chem. SOC.,Perkin Trans. 2, 1979, 18. H. Schmid and E. Hallaba, Monatsh. Chem., 1956,87, 560. 89 H. Schmid and M. G. Fouad, Monatsh. Chem., 1957,88,631. 90 G. Stedman and P. A. E. Whincup, J. Chem. SOC.,1963, 5796. "C. C. Addison and J. Lewis, Quart. Rev., 1955,9, 115. 92 E. Siiderback, Annalen, 1919,419, 217; H. Lecher and F. Graf, Ber., 1926, 59, 2601. 93 F. See1 and D. Wesemann, Chem. Ber., 1953,86, 1107. 9* K. A. Jergensen and S. 0.Lawesson, J. Am. Chem. SOC.,1984, 106,4687. 95 M. N. Hughes, G. Stedman, and T. D. B. Morgan, J. Chem. SOC.(B), 1968, 344. 96 C. A. Bunton, D. R. Llewellyn, and G. Stedman, J. Chem. SOC.,1959, 568. 97 G. Stedrnan, J. Chem. SOC.,1959, 2949. 191 S-Nitrosation and the Reactions of S-Nitroso Compounds charge- and frontier-controlled reactions. Such calculations predict that NOSCN should be less reactive than NOCl as is found experimentally.For some very reactive substrates however (e.g. aniline96 and azide ion97) it was found that the reaction became zero-order in substrate. This has been interpreted as a change in rate-limiting step to NOSCN formation. Such behaviour has also been found for other nitrosyl species. This enables the third-order rate constant k in equation 32 to be evaluated as 1 46097 and 1 50096 l2 moF2 s-' at 0 "C. Values Rate = k[H+][HNO,][SCN-] rather close to these have been observed for a whole range of anions which has prompted the explanation' that these reactions occur at the encounter-controlled limit.Reaction with very reactive neutral substrates produces the same effect except that the limit is somewhat lower than for the anions, which is to be expected by electrostatic considerations. More recently the same effect has been noted for reaction at 25 "C for both hydrazoic acid29 and also thioglycolic acid,32 where the values of k are 11 700 and 11 OOO l2 moF2 s-' respectively. This corresponds to an activation energy of ca. 56 kJ, which is in the region expected for an encounter- controlled reaction between a positively charged nitrosating species and an anion. Catalysis of nitrosation by thiocyanate ion has implications in the in uiuo formation of carcinogenic nitrosamines from naturally occurring secondary amines and sources of nitrous acid such as sodium nitrite (used widely as a food preservative particularly of cured meats) and nitrate ion in water supplies (which is readily reduced to nitrite in the saliva).Thiocyanate is secreted in the saliva and so will catalyse the formation of nitrosamines. This is particularly of concern for smokers, where the thiocyanate concentration is three or four times that of non- smoker~.~* have shown that, in the presence Experiments with N-meth~laniline~~ of thiocyanate, reaction proceeds much more rapidly in acid conditions such as gastric juice (between pH 1 and 2), whereas the catalytic action is less at higher pH values. Kinetic studies on the denitrosation of nitro~amines~~ strongly suggest that in the presence of thiocyanate ion a direct S-nitrosation occurs to give nitrosyl thiocyanate (equation 33). In this case it is necessary to include a nitrous acid trap Ph&H(Me)NO + SCN--PhNHMe + NOSCN (33)1Removed to drive the reaction to the right.As expected thiocyanate is more reactive than both chloride ion and bromide ion. Similarly the observation of substantial thiocyanate catalysis in the denitrosation of alkyl nitrites3I and also S-98 P. M. Densen, B. Davidow, H. E. Bass, and E. W. Jones, Archs. Envir. Hlth., 1967, 14, 865. 99 E. Boyland and S. A. Walker, Nalure (London), 1974, 248, 601. Williams nitrosothiolsJ8 has been interpreted as intermediate formation of nitrosyl t hiocyanate. Use is made of thiocyanate catalysis in the trans-nitrosation reaction observed (equation 34) where a nitrosamine transfers its nitroso group to another amine.R’R”NN0 + R”’R””NH sc“;l-* R’R”NH + R”’R””NN0 (34) This is believed to occur by a denitrosation reaction (thiocyanate ion catalysed) and subsequent nitrosation of the secondary amine. loo Similarly in the nitrosation of amines by propyl nitrite in propanol (equation 35),74 virtually no reaction occurs PrONO + PhN(H)Me s,!!i-9 PrOH + PhN(Me)NO (35) in the absence of a nucleophile such as halide ion or thiocyanate ion, but in their presence N-nitrosation occurs smoothly, again by way of nitrosyl thiocyanate for the thiocyanate ion catalysed reactions. 8 Nitrosation of Thiosulphate Ion Yellow solutions are readily formed from nitrous acid and thiosulphate ion. Early kinetic measurements’ O’ yielded a rate equation which included a non-integral dependence upon the thiosulphate concentration and the interpretation included rate-determining formation of the nitrosonium ion NO However a more recent + kinetic investigation,’ O2 using the stopped-flow technique, resulted in the rate equation given in equation 36.These measurements were made in aqueous Rate = k,[H+][HN02][S20,2-] + k2[HNO2I2 (36) perchloric acid at 25 “C. The interpretation involves two pathways (a) rate- determining attack by NO+ (or H2N02+)at the thiosulphate ion and (6) rate-determining formation of N20, which then effects the nitrosation of the thiosulphate ion more rapidly than its hydrolysis to nitrous acid.The second pathway has been observed on many occasions for many amine substrates whilst the first is also well-known. The value of k, is 18 OOO l2 moF2 s-’, which is taken to represent the diffusion-controlled limit. This is somewhat higher than that found for neutral (ca.7 OOO l2 moF2s-l) and singly negatively charged (ca. 11 OOO l2 mol-2 s-’) substrates, but the trend is to be expected on consideration of the effect of coulombic interactions on the diffusion rate. The activation energy of 50 kJ mob’ is close to that found for diffusion-controlled reactions involving singly charged anions. The product (yellow in solution) is believed to be the S-nitrosated anion [O,SSNO] -which decomposes fairly rapidly in solution. Its decomposition products have not been investigated.Nevertheless it was possible to measure the loo S. S. Singer,J. Org.Chem., 1978,43,4612; S. S. Singer, G.M. Singer, and B. B. Cole, J. Org. Chem., 1980, 45, 4931; S. S. Singer and B. B. Cole, J. Org. Chem., 1981, 46, 3461. lo’ J. 0.Edwards, Science, 1951, 113, 392. lo2 M.S. Garley and G. Stedman, J. Inorg. Nucl. Chem., 1981,43, 2863. S-Nitrosation and the Reactions of S-Nitroso Compounds equilibrium constant for its formation at 1.66 x lo71' mol-' at 25 "C. The visible absorption spectrum is very similar to those of other S-nitroso species including S-nitrosothiols, with an extinction coefficient in the same region. The equilibrium constant is also constant over a range of concentrations of the starting materials, for measurements made in acetate buffers.This then appears to represent another example of a S-nitrosation. The possibility arises that if the S-nitroso ion [O,SSNO]- can itself act as a nitrosating agent, then it would represent an unusual example of a negatively charged electrophilic nitrosating species. Intermediates capable of nitrosation have been detected kinetically, by the observed catalysis of nitrosation by added nucleophiles, and include the nitrosyl halides, ONSCN, ON$C(NH,),, ONSR, and possibly ON$(RR'). The extent of catalysis is governed principally by the size of the equilibrium constants for the formation of these intermediates, and not by the rate constants of their subsequent reactions. On these grounds it is perhaps reasonable to expect [O,SSNO]- to behave similarly. We have looked, in a preliminary study,' O3 for catalysis by thiosulphate ion in nitrosation.For the comparatively slow nitrosation of morpholine, no catalysis was detected, but the much more rapid nitrosation of N-methylaniline was markedly subject to catalysis. Further work in this area is necessary but the preliminary findings do support the suggestion that [O,SSNO]-can act as a nitrosating agent. 9 Nitrosation of Sulphinic Acids Sulphinic acids, particularly in their anion form (11) are well-known S-nucleophiles, reacting with a range of conventional systems including alkyl halides, carbonyl compounds, alkenes etc.' O4 Reaction with alkyl halides invariably yields (11) (12) the sulphone derivative.It has been suggested'05 that structure (12) contributes to the overall sulphinate ion, no doubt because of the clear indication of S-rather than 0-bonded products formed. It is not surprising therefore that sulphinic acids undergo S-nitrosation, but the observed products, the alkane- or aryl-sulphonyl hydroxylamine derivatives (equation 37), are somewhat unexpected in the context 2RS0,H + HNO, e(RSO,),NOH + H,O (37) of nitrosation chemistry, in that two molecules of reactant are used up and apparently two nitrosation steps are involved. The reaction, which appears to be Io3 T. Bryant and D. L. H. Williams, unpublished work. Io4 C. J. M. Stirling, Inf. J. Surfur Chem. B, 1971,6, 280 S. Oae and N. Kunieda, 'Organic Chemistry of Sulfur', ed.S. Oae, Plenum, 1977, p. 603. E.N. Guryanova and Y.A. Syrkin, Zh. Fiz. Khim., 1949, 23, 105. Williams quite general, has been known for some time'06 and has been used in the characterization of a long chain aliphatic sulphinic acid."' The reaction has also been used'O* in the analytical determination of sulphinic acids by a simple titration with a solution of sodium nitrite in dilute acid. More recently a number of such hydroxylamine derivatives have been prepared using this reaction,'09 in an attempt to generate by oxidation of the hydroxylamines, neutral analogues of Fremy's radical ion. The kinetics of the reaction of benzenesulphinic acid with nitrous acid have been determined very recently."' The reaction is very rapid and rate constants were determined by stopped-flow spectrophotometry. Reaction was first-order in nitrous acid and also in total stoicheiometric concentration of the sulphinic acid the rate constants were the same, within experimental error, for the measurement of the appearance of the product, as for the disappearance of the reactant.Acid catalysis was observed but is not strictly first-order. A plot of the first-order rate constant us. [H'] (when [sulphinic acid] % [HNO,]) is linear above ca. 0.06 M, but with a substantial positive intercept, and is curved at lower acidities. This type of behaviour is typical of a reaction which occurs via both the neutral substrate and its anionic form. Application of the protonation equilibrium (pK, 1.84'") leads to the expression given in equation 38 for the first order rate constant, where k,is the third order rate constant (Rate = k,[H+][HNO,][Substrate]) for reaction of the neutral form and k, that for the anion reaction (Scheme 11).The total substrate RSO,HeRSO,-+ H+1 1HNO, k, HNO, k, Product Product Scheme 11 concentration is [HA], and K, the acid dissociation constant for benzenesulphinic acid. This equation fits the experimental data well and values of k, and k, of 820 and 11 800 1, moF2 s-' are readily obtained. Thus both the acid and its anion are very reactive towards nitrosation, with the anion being significantly the more reactive. Its rate constant is very similar to that2' of the reaction of SCN-implying that this is the diffusion limit for an anion reaction generally with H,NO,+ or R.Otto and H. Ostrop, Annulen, 1867, 141, 370; W. Koenigs, Ber., 1878, 11, 615. lo' C. S. Marvel and R. S. Johnson, J. Org. Chem., 1948, 13, 822. ''* B. Lindberg, Actu Chem. Scund., 1963,17,383 and references therein; J. P. Danehy and V. J. Elia, Anal. Chem., 1972,44, 1281. Io9 J. D. Birchall and C. Glidewell, J. Chem. SOC.,Dalton Trans., 1977, 10. T. Bryant and D. L. H. Williams, J. Chem. SOC.,Perkin Trans. 2, 1985 in the press. ''' R. K. Burkhard, D. E. Sellers, F. De Cou, and J. L. Lambert, J. Org. Chem., 1959, 24, 767. S-Nitrosation and the Reactions of S-Nitroso Compounds PhS _I) 8-+ H,NO; 3 b-+ ( Ph SO,) ,NOH Scheme 12 NO+. Since the reaction is clearly first-order in substrate, the mechanism is likely to involve rate-limiting attack by the nitrosating species (designated H2N02 in-+ Scheme 12), yielding the nitrososulphinate intermediate, which rapidly effects another S-nitrosation with a molecule of reactant (here written as the anion), finally forming the hydroxylamine derivative as product, by proton transfer from the solvent.Nitrososulphinates (or sulphonyl nitrites) have previously been suggested as intermediates in this reaction,"' and have recently been isolated112 as rather unstable brown crystals from the reaction of sulphinic acids with Nz04at -20 "C in ether. They appear to be amongst the most powerful nitrosating agents known (although no quantitative comparison data are available), reacting with amines, ' alcohol^,^ and thi~ls,~ often yielding some of the hydroxylamine derivative as well, by nitrosation of the product sulphinic acid.The chemistry of nitrososulphinates is described fully in reference 5. Sulphinic acids also react with alkyl nitrites' l4 yielding the same hydroxylamine products as for the nitrous acid reaction, again it is believed by the intermediate formation of the reactive nitrososulphinates. In common with many (but not all) nitrosation reactions in aqueous acid solution, the reaction of benzenesulphinic acid is also catalysed by added nucleophiles C1-, Br-, SCN-, and SC(NH2)2,110involving the appropriate nitrosyl halide etc. Again, reaction of both the neutral acid and the anion occurs, with the latter again being the more reactive. The usual sequence of reactivity of the various nitrosyl species was observed, with the reaction of nitrosyl chloride with the sulphinate ion being close to the calculated diffusion limit. Acknowledgement. I would like to thank Dr. G. Stedman for many valuable discussions in the area of S-nitrosation, and for his comments on this manuscript. 'I2 S. Oae, K. Shinhama, K. Fujimori, and Y. H. Kim, Bull. Chem. SOC.Jpn., 1980, 53, 775. 11' S. Oae, K. Shinhama, and Y. H. Kim, Bull. Chem. SOC.Jpn., 1980,53, 1065. 114 G. Kresze and W. Kort, Chem. Ber., 1961,94, 2624.

ISSN:0306-0012    

DOI:10.1039/CS9851400171

出版商:RSC    

年代:1985

数据来源: RSC