摘要:

J. Chem. SOC., Faraday Trans. I, 1985, 81, 1375-1388 Analysis and Application of Linear Dichroism on Membranes Description of a Linear-dichroism Spectrometer BY LENNART B-A. JOHANSSON* Department of Physical Chemistry, University of Umei, S-901 87 Umei, Sweden AND AKE DAVIDSSON Department of Inorganic Chemistry, University of Lund Chemical Centre, S-220 07 Lund, Sweden Received 9th July, 1984 Linear dichroism (LD) refers to anisotropic absorbance of a system of macroscopically ordered chromophores. LD measurements based on light-modulation techniques are more sensitive and accurate than the conventional two-spectra technique. The set-up and an analysis of an LD spectrometer are reported. A method for the calibration of the instrumental LD scale has been developed. The method is based on the difference in Fresnel reflections of a tilted quartz plate for different linear polarizations of light.Lyotropic liquid-crystalline phases are frequently used as models for biological membranes. Many lamellar liquid crystals align spontaneously in thin layers between glass or quartz plates and form a macroscopically uniaxial system. LD of chromophores solubilized in such systems can be measured if the plates are tilted with respect to the analysing light beam. The interpretation of LD data on these systems may be complicated by the birefringence of the anisotropic phase and by the reflections at the various interfaces. A model taking these complications into account has been derived and is compared with a simplified model commonly used; both these models are tested experimentally.Linear dichroism (LD) refers to the differential absorption by an anisotropic system of chromophores which is obtained with two mutually perpendicular linear polarizations of light. The simplest experimental equipment for LD studies is an absorption spectrometer and a pair of linear polarizers. This method is for many purposes excellent,l but becomes inexact for small values of the dichroism. A more accurate and up-to-date method is offered by the light-modulation techniques. The commercial circular-dichroism (CD) spectrometers are based on such techniques and could in principle easily be transformed to linear-dichroism spectrometers.2 Although many excellent CD instruments are commercially available the manufacturers have hitherto paid LD spectrometers very little attention.In this work we describe an LD spectrometer built with optical and electronic components that are easily available. An optical analysis of the instrument is presented. A general problem with both CD and LD spectrometers is calibration. Here a new method for LD calibration is presented. The physical basis of the method is Fresnel reflection. The reflection of a beam of light impinging on a tilted quartz plate depends on the polarization state of the light, the angle of incidence and the refractive index of the plate. From LD measurements of a tilted quartz plate and its calculated reflection losses the calibration constant of the instrument can be determined. The orientation of chromophores in anisotropic systems can be investigated with 13751376 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES linear-dichroism spectro~copy.~ In this work the applicability of the technique is focused on the investigation of chromophoric molecules in natural and model membranes.We have previously shown that the orientation of chromophores solubilized in lamellar liquid-crystalline phases can be studied with LD.3 Many lamellar liquid-crystalline phases align spontaneously in thin layers between glass or quartz plates and form macroscopically uniaxial systems. The optical axis of these systems is perpendicular to the plane of the plates. The linear dichroism of such a specimen can only be measured if the optical axis is tilted relative to the incident light beam. However, the optical treatment of light passing through this kind of system is complicated by the birefringence of the aligned phase and by the multiple reflection.This paper presents a detailed theoretical and experimental investigation of the significance of these phenomena in lyotropic liquid crystals. EXPERIMENTAL The light source of the LD equipment was a 450 W Xe lamp mounted in a home-built water-cooled lamp house. A power supply (M 30 1) from Photochemical Research Associates Inc. (Ontario, Canada) was used. The monochromator was a Jobin Yvon HRS 2 instrument with a grating (M.29) that has 1200 grooves mm-l. All samples in this work were studied with a spectral bandwidth of 0.6 nm. The long pass filters were low fluorescing and belong to the KV series made by Schott (Mainz, West Germany).The linear polarizer, a Glan prism, had an aperture of 19.5 mm, a useful spectral range of 215-2700 nm and an extinction ratio of 10-5 (Bernhard Halle, West Berlin). The photoelastic modulator, from Jobin Yvon (Longjumeau, France) operated at a modulation frequency of 18.5 kHz. The Hanle wedge depolarizer was from B. Halle. The photomultiplier tube was a Hamamatsu R 376 and the lock-in amplifier a Princeton Applied Research 129A. The regulation circuit of the direct current of the photomultiplier was home built essentially in accordance with ref. (4). For the calibration we used slides made of fused quartz (Synsil, Zeiss AB, Stockholm). The refractive indices at the wavelengths of calibration were 1.4940 (280.4 nm), 1.4667 (435.8 nm), 1.4619 (508.6 nm), 1.4585 (587.6 nm), 1.4564 (656.3 nm) and 1.4552 (706.5 nm).The stress birefringence of the plates was 5 nm cm-l. The quartz plates were cleaned as follows. The plates were kept in a warm (5G60 "C) potassium bichromate+sulphuric acid solution (containing a mixture of 0.40 mol dmP3 K,Cr,O, in concentrated H,SO,, 1 : 10 by volume) for ca. 5 min and washed with a large excess of distilled water and ethanol (spectroscopic grade). After a final washing with distilled water the plates were dried in an oven (at ca. 80 "C). In the calibration the linear dichroism was recorded at various angles of incidence (a) between the beam of light and the normal of the plate. In order to reduce errors caused by misalignments of the plate, LD was recorded both at + and - R and the mean value of the two LD values was used.Thefluorophore indocarbocyanine DiICl,(3) was from Molecular Probes Inc. (Texas, U.S.A.). The dye was solubilized in a lamellar liquid-crystalline phase composed of 14 wt% sodium octanoate, 36 wt% decan-1-01 and 50 wt% water. Sodium octanoate (B.D.H., Poole) was purified by filtering a solution of the amphiphile in methanol and active coal. Decan-1-01 (Serva) was of analytical grade and the water was distilled twice. The lamellar phases were macroscopically aligned between quartz slides (synsil). The alignment was checked by observing the sample when placed between two crossed polarizers. The chromophore concentrations were 5.0 x lop4 and 1.0 x mol dm-3 and the absorbances of the samples were varied between 0.2 and 1.0.The absorption spectra were recorded on a Varian Cary model 219 U.V. spectrometer. The ordinary and extraordinary refractive indices of the lamellar liquid-crystalline phase were determined with an Abbe refractometer (model A, Zeiss) at 20 and 25 "C. The numerical calculations were performed using the Cyber computer, UMDAC, at the University of Umeb.L. B-A. JOHANSSON AND A. DAVIDSSON 1377 X t propagat ion \ of light Fig. 1. (a) Schematic set-up of the LD spectrometer. A, light source; B, monochromator; C, long-pass filters; D, linear polarizer ; E, photoelastic modulator (p.e.m.); F, dichroic sample; C;, depolarizer; H, photomultiplier; I, regulation unit, keeping the direct current level constant; J, lock-in amplifier; K, control unit for p.e.m.; L, wavelength program; M, recorder.(b) Optical elements of an LD spectrometer considered in the optical analysis. A beam of linearly polarized light (along y ) , so, is passing through the p.e.m. (E), the dichroic sample (F) and the depolarizer (G). So, R, F and D are Stokes-Mueller-Go matrices (see text). RESULTS AND DISCUSSION THE LINEAR-DICHROISM SPECTROMETER The optics and the electronics of the linear-dichroism spectrometer are illustrated schematically in fig. 1. Light from a Xe lamp is passed through a monochromator and long-pass filters placed after the exit of the monochromator reduce stray light.1378 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES A linearly polarized beam is produced by a polarizer and passes through the photoelastic modulator (p.e.m.).The state of polarization of the beam leaving the p.e.m. varies periodically with frequency, w. The depolarizer placed in front of the photomultiplier tube eliminates the effect of the polarization of light on its detection. Electronically the photomultiplier direct current is kept at a constant level with a home-built regulation circuit. The lock-in amplifier selects the 2wt Fourier component of the alternating current. OPTICAL ANALYSIS AND CALIBRATION The formalism5-’ of Stokes, Mueller and Go is very convenient for describing the intensity and the state of polarization of a light beam passing various optical elements. In the LD spectrometer linearly polarized light passes through a photoelastic modulator which causes a change in the state of polarization dependent upon both the induced birefringence of the p.e.m. and the angle @) between the polarization vector of light and the induced optical axis of the crystal.The beam of light emerging from the p.e.m. passes through the dichroic sample where both the intensity and the polarization of the beam may be changed. In the Stokes-Mueller-Go formalism the linearly polarized light, incident on the photoelastic modulator, is represented by a 4 x 1 Stokes matrix, so, and the p.e.m. by a 4 x 4 Mueller matrix. These matrices are given in Appendix A. The sample, f?, and the depolarizer, D, are also described by 4 x 4 matrices. The Stokes matrix for the light beam, s, reaching the photomultiplier (1) is given by Three of the Stokes elements represent the state of polarization, while the fourth describes the intensity.6 Since a depolarizer eliminates the former elements only the intensity element, I, is left.The photomultiplier current i( t ) is directly proportional to the intensity, which is given by s = DPR@, S) so. I(t) = I, 10-A[cosh( LD In 10 )-c0s22B sinh( LD In 10 ) + sin2 2/3 cos S(t) sinh (“D;lO)] ( 2 ) as is shown in Appendix A, where p denotes the angle between the polarization vector of the light and the optical axis of the p.e.m., d(t) is the induced phase lag of the p.e.m., LD = A , - A,, 2 = ( A , + A , ) / 2 and I, is the intensity of the linearly polarized light beam 9,. A Fourier-series expansion of cos d(t) = cos [So sin (wt)] yields i(t) = k[cosh( LD In 10 )-cos22psinh( LD In 10 ) a, LD In 10 +sin2 2p sinh ( ) ( J0(S0)+2 E J2,(S,) cos 2nwt n - i where J,, J,, are the Bessel coefficients of the first kind, 6, is the maximum phase lag and k is a constant. If the d.c.level of the photomultiplier current is electronically kept constant, eqn (3) can be written as sin2 2p tanh (LD f: lo) $ J2,(d,) cos 2nwt \ L / n - l LDln 10 iac(t) = k’ 1 - tanh ( ) [J,(d,) sin2 2p- cos2 2fl (4)L. B-A. JOHANSSON AND A. DAVIDSSON 1379 10.0 > E --- 0 2 5 .O 0 I I 1 I I I I 1 2 3 4 5 6 7 lo2 LD Fig. 2. Reflection and transmission of a light beam impinging on a tilted (90" - w) quartz plate. The two linear polarizations are denoted w and 1. Plot of the recorded signal against the calculated LD corresponding different angles of tilt. The slope K is the instrumental constant at 656.3 nm.where k' is a new constant. A phase lag of 6 = 142" (2.49 rad) is preferable since the Bessel function J0(2.49) equals zero and J2(2.49) is 91.6% of its maximum value. By means of the lock-in amplifier the 2at component of iac(t) is extracted and eqn (4) then becomes sin2 2p tanh(LD lo) \ L / LDln 10 ' i,, = k"J2(SO) l+cos22ptanh( ) If p is further adjusted to 45", i,, will reach its maximum value, i.e. i,, = k"J2(d0) tanh (6) A misalignment of p by 1" for LD values < 0.1 will cause an error which is < 0.1 % . The recorded signal, Urec, is directly proportional to iac. For values of LD < 0.1 we obtain which is a very good approximation of eqn (6). U,,, = KLD (7) In order to obtain LD the instrumental constant K must be determined. For this1380 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES 210 I 1 I I I I I 170 1 I I 1 I J 200 300 LOO 500 600 700 800 wavelength/nm Fig.3. Wavelength dependence of the instrumental calibration constant, K. calibration we have used the fact that the Fresnel reflections of light passing a tilted quartz plate depend on the state of polarization. As shown in Appendix B the linear dichroism due to the different reflections is where R, and RI are the reflection coefficients of the light beams having the polarizations set as w and 90" (I) with respect to the normal of the plate (fig. 2). R, and RI have been calculated for different angles of incidence and at the different wavelengths of calibration where the values of the refractive indices of the quartz plate are known.According to eqn (7) the measured signal U,,, will be directly proportional to the calculated reflection L D which we also find experimentally (cf. fig. 2). The proportionality constant, which is equivalent to the instrumental constant K, has been determined at different wavelengths. Ideally Kshould be independent of the wavelength but this is not found, as can be seen in fig. 3. One explanation is a possible wavelength dependence of the phase lag 6, and thereby a variation in the Bessel function Jz(6,) with [cf. eqn (6)]. For our equipment the variation of K over a wavelength region of 100 nm is ca. 3%. For an instrument calibrated at only one wavelength, which is the case for most commercial CD and LD spectrometers, considerable errors may be introduced at other wavelengths. LINEAR DICHROISM OF LYOTROPIC LIQUID CRYSTALS In the study of the orientation of chromophores in membrane systems linear- dichroism spectroscopy is one of the very few techniques available.Lamellar liquid-crystalline phases composed of detergents and/or lipids in water are frequently used as model membrane^.^ These systems may spontaneously align in thin layers between a pair of glass or quartz slides, so that the director of the lamellae coincides with the normal of the sides. The linear dichroism of such a sample can only be measured if the slides are tilted with respect to the impinging beam of light (cf. fig. 2 and 4). From measurements of LD and the absorbance A I , the order parameter, S , of a chromophore can be determined through the relation3 LD 3s cos2 w Al (1-S) n2 - (9)L.B-A. JOHANSSON AND A. DAVIDSSON refractive index intensity from liqht source intensity Ip t o detector 1381 (4 (bl Fig. 4. (a) Multilayer system air/quartz/anisotropic medium with refractive indices equal to I , rn and vi, respectively, together with the transmitted light components which are considered. MI and I denote the two linear polarizations. (h) Boundary conditions at the first quartz/ anisotropic sample interface for w-polarized light. The Oi denote the angles of incidence, reflection and refraction, and a corresponds to the deviation from 7c/2 of the electric-field vector from the refractive wave vector i. where S = i(3 (cose/3) - 1) describes the average orientation of the transition dipole moment with respect to the director @) (for a perfect orientation of the transition dipoles parallel or perpendicular to this director, S becomes 1 and -i, respectively), cv is the angle of tilt and n is the refractive index of the phase.Eqn (9) is based on the assumption that the sample is optically isotropic. This is not strictly true since the system is uniaxially anisotropic. However, for sufficiently small values of the birefringence one can expect this idealization to be a very good approximation. The purpose here is to investigate the usefulness of this simple model through comparisons with a more accurate model and their correlation with experimental data. Let us therefore consider a uniaxially anisotropic medium in optical contact with two isotropic media, i.e.the slides. The optical path and the reflections at the different interfaces will depend on the polarization of the impinging ray and on the refractive indices of the media. The isotropic media with refractive index rn and the uniaxial anisotropic medium with the real refractive indices n, and n, are illustrated in fig. 4, where the various reflections considered in this work are also shown. The derivation of a formula relating the intensity of the impinging ray to that having passed through the system is, from a quantum-electrodynamic view, a hopeless task.s Rigorous classical treatments of optically anisotropic systems have been published by Rama- chandran and Ramaseshans and Szivessy . lo Their investigations include many general cases of wave propagation between anisotropic media but not the case which is of particular interest here.Our derivation, which is condensed in Appendix C , is based on the following assumptions. ( a ) The absorbing medium is considered as homogeneous. This medium interacts with the electromagnetic wave through its real (n) and imaginary ( k ) parts of the complex refractive index (13). The imaginary part then corresponds to the absorption coefficients of the chromophores and the real part is the refractive index of the transparent matrix. (b) Incident light beams reflected more than twice are neglected. All beams illustrated in fig. 4 are taken into account. Phase correlation between these components are neglected.1382 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES 0.4 0.3 T+ 5 0.2 4 1 0.1 0 I I I I I 1 1 I 0.5 cos* w 1.0 Fig.5. LD/A, as a function of cos2 w. 0, Experimental data, (-) extended model [eqn (lo)] and (---) simple model [eqn (9)]. The derivation yields a rather complicated expression of L D / A , as a function of the refractive indices (m, n,, n,), the angle of incidence (w) and the order parameter 6): The explicit expressions of the absorbances A,(n,, n,, w, S ) , A,(n,, w, S ) and q4(m,nZ,n,, w, S ) are given by eqn (C 6)-(C 14). Since n,,n,,rn and w are known or can be determined experimentally, S can be obtained from a fitting procedure to experimental L D / A I values. The approximate formula of (LD/AJ (n, w, S ) , i.e. eqn (9) and the more accurate description eqn (lo), have been tested experimentally. We used the dye indocarbocyanine [DiICl,(3)] solubilized in a lamellar liquid- crystalline phase of sodium octanoate, decan- 1-01 and water.ll The w-dependence of L D / A , as a function of cos2 w was measured and is shown in fig. 5.L D / A l is linear with cos2 w for angles > 40" but deviates at smaller angles. From the linear part, which is predicted by eqn (9), an order parameter S = -0.280 can be determined with n = (nZ+2n,)/3. The extended model predicts both the linear and the non-linear dependence with cos2 w and gives an order parameter S = - 0.274. Since the deviation between the two order parameters is only ca. 2% we conclude that the simplified model for this system is very good. In fig. 6 a comparison of the two models is presented. The order parameters have been calculated for different possible experimental L D / A , values at w = 45" and 60".The refractive indices chosen are typical for lyotropic lamellar liquid crystals. A deviation of > 10% is found for -8 x < LD/A, < 2 x 10-l at 45" and for - 3 x < LD/A, < 8 x lop2 at 60". It is obvious that the simple model must be used with care, as has been pointed out previ~usly.~ For smaller LD/Al values the extended model developed here is necessary.L. B-A. JOHANSSON AND A. DAVIDSSON 1383 Fig. 6. Comparison of the order parameters calculated with (---) the simple [eqn (9)] and (-) the extended [eqn (lo)] models for different LD/AI values. The refractive index of the quartz plates is 1.460 and the refractive indices n, and n, of the anisotropic medium are 1.39 and 1.40, respectively.The isotropic absorbance (A, + 2A,)/3 is chosen as 0.5. w denotes the angle of incidence of the light beam. CONCLUSIONS An alternative to the method of LD calibration described here would be to use a quarter-wave retarder (A/4) and a circular-dichroism standard. By passing the beam of light that leaves the photoelastic modulator through a A/4 plate the polarization becomes circular. This beam then passes through a sample of known CD before it reaches the detector. An analysis analogous to that given in Appendix A yields U,,, = KCD (1 1) where the instrumental constant Kis the same as in eqn (7). This method of calibration would require A/4 retarders of good optical quality for each wavelength of calibration as well as a set of CD standards. In comparison, our method of calibration is more simple and precise.The LD due to all the reflections of a quartz plate are easily obtained from the Fresnel equations, and the adjustments of the angles of tilt are rather easily set within 0.5". However, one disadvantage is the limitation in calibration to < LD < 10-l. Of course, with materials of lower refractive index an extension to smaller LD values can be made. Another important point is the optical quality of1384 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES the quartz plate. We have used several different plates in the calibration but always obtained the same result. The plates are therefore assumed to be optically homogeneous. In the LD model of an absorbing and optically anisotropic system presented here it is assumed that the imaginary part of the refractive index, k , is much smaller than the real part, n.For a typical system the thickness of the absorbing medium is ca. 100 pm and the absorbance ca. 1. k is then ca. and hence k2 is much smaller than n2. Furthermore, the refractive indices at 598 nm of a lamellar liquid-crystalline phase of sodium octanoate, decan-1-01 and water are 1.402 (n,) and 1.393 (n,) when the chromophore [DiIC,,(3)] is solubilized and 1.402 (n,) and 1.392 (n,) in its absence. Thus the assumption that the refractive index of the absorbing medium is dominated by the refractive index of the transparent matrix is correct here. The extended model predicts a non-vanishing LD even if the order parameter S = 0.In lyotropic liquid-crystalline systems where the birefringence, n, - n,, is in the order of f one then expect LD/A = lo+ even if S = 0. When small LD values are studied a careful optical analysis is therefore necessary in the interpretation. The optical model presented here may easily be extended to biaxial systems such as a Couette ce11.12 APPENDIX A OPTICAL ANALYSIS OF THE LINEAR-DICHROISM SPECTROMETER The Stokes matrix of the linearly polarized light beam entering the optical elements of the LD spectrometer shown in fig. 1 (6) is given by I,n this work the convention of element order used by Troxell So passes a photoelastic modulator which can be regarded as a and Scheraga5 has been used. plate having a time-dependent retardation, S(t), and whose optical axis forms an angle Q with the x axis.The Muellei matrix of this retarder2 is / 1 0 0 0 \ 0 sin2 2p+ cos2 2p cos 6 - cos 28 sin 6 cos 2p sin 6 cos 6 -sin 2p sin 6 sin 2Q cos 2Q( 1 - cos S) w,4 = \ o cos 2~ sin 2p( 1 - cos 6 ) The dichroic sample may be represented according to Go7 by sin 2p sin 6 cos2 2 ~ + sin2 2~ cos 6 / P = exp( - yfI) where 0 0 b, f I = ( i b, 0 0 ;, I) 0 In 10 2n - yb, = - ( A , - A J , ya, = - (n, -nu) d and A = +(A, + Au). 2 1 A , and A , are the polarized absorbances with respect to the x and y axes, n, and nu are the corresponding refractive indices and d is the optical path length.L. B-A. JOHANSSON AND A. DAVIDSSON 1385 The depolarizer is represented by 1 0 0 0 0 0 0 0 D = ( O 0 0 0 0 O O The Stokes matrix of the beam entering the photomultiplier then becomes s = DPR(J?,d)S, which with eqn (A 1)-(A 6) yields s = I, 10-K i" + sinh ( 1) LD In 10 [sin2 0 0 0 The intensity of the beam is thus APPENDIX B LINEAR DICHROISM AND REFLECTION A light beam passing from air through a quartz plate with refractive index n undergoes reflection losses at the two interfaces.For the directions of polarization denoted by w and I and illustrated in fig. 2, the reflection coefficient^^^ are R w = [ n sin w + ( l -n-2 cos2 w): 1 R , = [ sin w + n( 1 - n-2 cos2 w): where the refractive angles have been substituted using Snell's law. The intensities of the transmitted light beams are (B 2) sin w - n( 1 - n-2 cos2 w): 2 ] (B 1 ) n sin w-(1 -np2 cos2 w): I, = Ioy [(l - R,)2+ Rt(1 - R,)2+ R:(1 - R,)4+...I, v = I, W . For an infinite number of reflections I, 1-R, - I,, 1 + R,' The linear dichroism due to the reflections is then which with eqn (B 3) Bves LD = l o g [ ( % ) ( 3 ) ] . 1-R, 1 + R , Given the refractive index of the quartz plate at a certain wavelength the linear dichroism is easily calculated for various angles of the impinging light beam. 46 F A R 11386 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES APPENDIX C PROPAGATION OF LIGHT IN A UNIAXIAL ACHIRAL SYSTEM Consider a system of a homogeneous and absorbing dielectric embedded between two plane parallel transparent isotropic plates of refractive index m (cf. fig. 4). The thicknesses of the absorbing medium and the plates are much larger than the wavelength of light. The absorbing medium is characterized by the components ri, = n, - ik, of the complex refractive-index tensor, which is symmetric for an achiral medium, and where v denotes the Cartesian components and a denotes a complex quantity. A coordinate system is located with its origin at the boundary between the first plate and the anisotropic medium.The z axis is parallel with the normal to the plates. Define the plane of incidence as the y = 0 plane. Furthermore, let the polarization of the incident light be linear with its electric-field vector either in y = 0 or along y ; i.e. we only regard light propagation in the system when the electric-field vectors are in the symmetry planes of the anisotropic medium, which indeed are the solutions of the wave propagation in an anisotropic medium for a general state of polarization.At the boundary z = 0 the tangential components of the electric- and magnetic-field vectors must be continuous. l 3 This requirement and the relation between electric ( E ) and magnetic (H) fields, H,, = ri,,(s x E), (C 1) allow us to derive the Fresnel coefficients of refraction, t , and reflection, r , at the boundary, where s is the unit vector of propagation of the electromagnetic wave. The derivation of the t and r coefficients is here shown only for the second boundary. Because of induced polarization the electric vector of the light in the anisotropic medium will not in general be orthogonal to s. Let the direction of the w-polarized xz electric-field vector in the anisotropic medium be (71/2) - a relative to s.For w-polarized light with angle of incidence O2 at the boundary z = 0 we get the following relations between the electric-field amplitudes at this boundary : W W (C 2a) Eeflected = i,, p i c i d e n t fpfrarted = f2w p i c i d e n t where 8, denotes the angle between the z axis and s. The refractive index for the w-polarized light is fi; = fi;ri;[ri; cos2(8,-a)+fi; sinZ(8,-a)]-1. Since a = 0 for y-polarized beam, corresponding relations are easily obtained. Proceeding along these lines for the system as a whole, we obtain the amplitude relations between the incident beam, E,, and the amplitude of the main beam that has travelled through the system.14 (C 4) where R is the position vector of the light in the anisotropic medium and ti with i = 1-4 are the Fresnel coefficients for transmission at the four boundaries.The intensity of light is given as the product EE*, where * denotes the complex conjugate of E. If we consider those components of I, illustrated in fig. 4(a) that contribute to the output intensity, write down expressions analogous to eqn (C 4) for them and finally consider the symmetry in the system, we get the intensity of p-polarized light: E, = E,,(t, i2 i3 t4), exp (imt) exp (im/c) ri,(Rs),], p = I, w where co is the circular frequexy of the light and c its velocity, d is the thickness of the anisotropic medium along z and d, is its complex phase factor.14 Im denotes the imaginary part. The Fresnel coefficients in eqn (C 5 ) are all substituted with the analogous expressions to eqnL. B-A. JOHANSSON AND A.DAVIDSSON 1387 (C 26). By hell's law and eqn (C 3) the reflective and refractive angles can be eliminated. After a lot of tedious algebra involving complex conjugation, the output intensity is obtained as a formula depending on w, n,, n,, rn, d, k, and k,. The absorbances, A,, are related to the k, as 2mk, d A,=-----. C The final equations for the linear dichroism LD = Aw--A, are then A , = Yw--og4wrw A , := y,-iog4,r, mON, - rn-lmBn, 16WX CB2n;(m-'M-rn sin w ) ~ + 10-2yw[(rnc. 2Nz + rn-lMBn,) + (rncN, + rn-'MBn,)4 (rn-'M+ rn sin w)2 1 >'I 4MN,( M - sin w) (M+ NJ2 (M+ sin w) ( M - N,) ( M - sin w) ( (M+ N,) (M+ sin w) 4, := 1 +2 2 5 6 M 4 x CB2ni (rn sin ~+m-~n/1)~(m~N,+rn-'MBn,)~ rw = 256M4W, sin2 w (M+ sin w ) ~ (M+ N,)4 r, := M == (m2 - cos2 w): C = ni W, + n4, cos2 w N , =(~;-cos~w): B=n;T+n;cos2w N , := (n," - cos2 w): where n, and a, are the real refractive index and absorption, respectively, of the absorbing medium when the polarization of light is directed along the optical axis and n, and a, are corresponding quantities with polarization orthogonal to the optical axis. These expressions cover the linear dichroism of an uniaxial system at any angle of incidence in a symmetry plane of the anisotropic medium. For w near n / 2 , i.e. close to normal incidence, a term which takes multiple-beam interference into account should be added.14 We thank Miss Lilly Sjolander for technical assistance and the Swedish Science Research Council for financial support. E. W. Thulstrup and J. Michl, J. Am. Chem. Soc., 1982, 104, 5594. A. Davidsson and B. Norden, Chem. Scr., 1976, 9, 49. L. B-A. Johansson, A. Davidsson, G. Lindblom and B. Norden, J. Phys. Chem., 1978, 82, 2604; L. B-A. Johansson and G. Lindblom, Q. Rev. Biophys., 1980, 13, 63. F. A. Modine, Rev. Sci. Instrum., 1979, 50, 386. T. C . Troxell and H. A. Scheraga, Macromolecules, 1971, 4, 519. P. S. Theocaris and E. E. Gdoutos, in Matrix Theory ofPhotoelasticity, ed. D. L. MacAdam (Springer- Verlag, Berlin, 1979). ' N. Go, J. Phys. Soc. Jpn, 1967, 23, 88. J. M. Schurr, Chem. Phys., 1976, 15, 1 . 46-21388 ANALYSIS AND APPLICATION OF LINEAR DICHROISM ON MEMBRANES G. N. Ramachandran and S. Ramaseshan, in Handbuch der Physik, Band X X V / I , Kristalloptik- Beugnung (Springer-Verlag, Berlin, 1961). lo G. Szivessy, in Handbuch der Physik, Band XX Licht als Wellenbewegung (Springer-Verlag, Berlin, 1928). P. Ekwall, Adu. Liq. Cryst., 1975, 1, 1. l2 B. NordCn, A. Davidsson and L. B-A. Johansson, Nature (London), 1976, 261, 460. l3 M. Born and E. Wolf, in Principles of Optics (Pergamon Press, Oxford, 6th edn, 1980). l4 A Davidsson, Chem. Phys. Lett., 1981, 83, 346. (PAPER 41 1 188)

ISSN:0300-9599    

DOI:10.1039/F19858101375

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1985,81, 1389-1400 Polarized-light Spectroscopic Study of Indocarbocyanine Dyes Solubilized in Amphiphile Aggregates BY LENNART B-A. JOHANSSON, * TOMMY VALLMARK AND GORAN LINDBLOM Department of Physical Chemistry, University of Umei, S-901 87 Umei, Sweden Received 9th July, 1984 The solubilization and the fluorescence properties of indocarbocyanines in ionic and non- ionic micelles and cubic liquid crystals have been studied. The indocarbocyanines are two dialkyl derivates (C2H5- and C14H29-), commonly denoted as DiIC2(3) and DiICl,(3). The steady-state anisotropy and the quantum yield of fluorescence have been measured, and the radiative lifetime of both dyes determined (2.2 ns). In the aggregates of ionic amphiphiles such as sodium octanoate, octylammonium chloride and dodecyltrimethylammonium chloride the rotational correlation time of both dyes was found to increase with increasing concentra- tion of the detergent.In micelles of the non-ionic pentaethyleneglycol mono-n-dodecylether no such dependence on the concentration is found. A reasonable explanation is that the large indocarbocyanine molecules perturb the relatively small ionic aggregates while the large non-ionic ones are much less affected. From linear-dichroism studies of the dyes in lamellar liquid crystals it has been concluded that both these chromophores are oriented with their long axis preferentially parallel to the plane of a bilayer. The reciprocal rate of the non-radiative processes is found to increase linearly with increasing rotational correlation time of the fluorophore.Very similar behaviour has been found for ordinary solutions of alcohols, which suggests that the orientational distribution of the dye molecules in the amphiphile aggregates has a negligible effect on the non-radiative processes of the indocarbocyanines. Dyes are frequently used as probe molecules in studies of various systems containing aggregates of amphiphiles or macromolecules such as proteins and synthetic polymers. Thus carbocyanines have been used to probe lipid phase transitions1 and membrane potentials2 and to measure translational diffusion coefficients in membranes1 These investigations were based on measurements of the fluorescence of the chromophore. The strength of such a method is undoubtedly the high sensitivity; also, no other method is often available for studies of biological systems. Note, however, that the results obtained from such a probe method may depend on the probe molecule used.Therefore it seems highly desirable to investigate the behaviour of such fluorescence probe molecules in various simple model systems. To our knowledge very few such studies have been reported for probe molecules used for studies of biological membranes or even whole cells. Several different systems have been used as model systems3 for biological membranes but for these optical studies we find it convenient to use isotropic phases of simple surfactants such as micellar solutions and cubic liquid-crystalline phases. This work is part of a systematic investigation of the solubilization of indocarbocyanine molecules in such model systems aimed at obtaining a detailed picture of the location and orientation of this probe molecule in aggregates of ionic and non-ionic amphiphiles.The molecular motion of the indocarbocyanines and the effect of structure and/or dynamics on the photophysics of the dye molecules have also been briefly studied. 13891390 INDOCARBOCYANINE DYES JN AMPHIPHILE AGGREGATES EXPERIMENTAL MATERIALS The fluorophores, indocarbocyanine DiIC2(3) and DiIC,,(3), were obtained from Molecular Probes Inc. (Texas, U.S.A.) and used as received. Rhodamine B of analytical grade was purchased from Merck. The detergents sodium octanoate (NaC,) (B.D.H.), n-octylamine hydrochloride (OAC) (Alfa Products) and dodecyltrimethyl ammonium chloride (DOTAC) (Eastman Kodak Co.) were all purified by filtering a solution of the amphiphile in methanol and active coal.The tetra-, penta- and hexa-ethyleneglycol mono-n-dodecyl ethers (CI2EO4, C12E05, C12E06) were obtained from Nikko Chemicals Ltd, Tokyo, Japan and used without further purification. Octan- 1 -oyl glycerol (C,G) was synthesized at ' Syntestjanst ', Chemical Centre, Lund and decan-1-01 (C,,OH) (Serva) and n-octane (C,) (Fluka), both of analytical grade, were used as received. The water was distilled twice. The linear dichroism (LD) was recorded on a homebuilt equipment, described el~ewhere,~ and the absorption spectrum was recorded on a Varian Cary model 219 U.V. spectrometer. The refractive indices were determined with an Abbe refractometer (model A, Zeiss). When the ordinary (n,) and the extraordinary (n,) refractive indices of the lamellar liquid-crystalline phases were measured the ocular of the Abbe refractometer was supplemented with a sheet polarizer (HNPB Polarizer, U.K.).The fluorescence emission was studied with a Spex Fluorolog 112 instrument (Spex Ind., New Jersey, U.S.A.) equipped with Glan Thompson polarizers. The spectral bandwidth, in all fluorescence experiments, was 4.1 and 2.7 nm for the excitation and the emission monochro- mators, respectively. The samples were thermostatted to within & 0.2 K. METHODS The orientation of the indocarbocyanine dyes, DiIC,(3) (n = 2 and 14), solubilized in lamellar liquid-crystalline phases was studied at 293 & 1 K. The lamellar phases were composed of sodium octanoate + decan- 1-01 + water, octylammonium chloride + decan- 1-01 + water, pentaethylene glycol mono-n-dodecyl ether + water and octan- 1 -0y1 glycerol + water. The lamellae align spontaneously in thin layers between quartz plates.The orientation was checked using a microscope supplemented with polarizers. For a macroscopically ordered sample placed between two crossed polarizers no birefringence was visible for light passing at normal incidence to the plates. The order parameter, which is a measure of the average orientation (p) of the transition dipole moment with respect to the normal of the lamellae,5 is given by S = 2(3 cos2p- l), where ( ) stands for the space average. To a good approximation [see ref. (4)] the linear dichroism depends on S as A , 3 s cos2 w LD = (1 - q n 2 where A , is the absorbance of the ordinary beam, w is the angle of tilt of the sample as illustrated in fig.1 and the refractive index n = $(n, + 2n,). A more rigorous treatment that accounts for the birefringence of the uniaxial medium and the reflection losses at the various interfaces is given in ref. (4). This treatment shows that LD depends on S, w, n, and n, in a complicated way, which is best handled with a computer. In this work the LD data were analysed using both the approximate and the extended models. The fluorescence steady-state anisotropy, t,, is defined by FVV-GFVH FVV + 2GFVH rs = where Fvv and FVH are the fluorescence intensities measured with the excitation and the emission polarizers set vertical-vertical and vertical-horizontal, respectively.6 The fluorescence emission is monitored at 90" with respect to the propagation of the excitation beam.The instrumental correction factor G = FHV/FHH. FHV and F H H are defined by analogy with Fvv and FVH, and G was determined with a solution of Rhodamine B in ethanol (1.5 x lop7 mol dm-3). TheL. B-A. JOHANSSON, T. VALLMARK AND G. LINDBLOM 1391 KN-R rl incident beams Fig. 1. Upper left: structural formula of indocarbocyanine, R = C,H, or C14H2,. Upper right: schematic diagram of the linear-dichroism (LD) experiment showing an aligned lamella and the impinging linearly polarized light beams. The absorption ( A 3 and the LD spectra (recorded at w = 45") of DiIC14(3) in a lamellar phase of NaC,+n-Cl,0H+H20 is also shown.steady-state anisotropy of DiIC,(3) (n = 2 and 14) was calculated from the excitation spectra measured at 568 nm when the excitation wavelength was varied between 500 and 600 nm. r, shows no significant dependence on dye concentration between 1.0 x and 1.0 x mol dm-3. The reproducibility of Y, is within k0.003. The fluorescence quantum yield of DiIC,(3) (n = 2 and 14) was measured using Rhodamine B as ~ t a n d a r d . ~ Rhodamine B dissolved in ethanol (1.5 x lo-' mol dm-3) was degassed by the freeze-pumpthaw method. The emission spectra were recorded with vertically polarized excitation light and the emission polarizer set at the 'magic' angle, 54.7". The wavelength of excitation was 550 nm. The quantum yield, @, can be calculated from where @RhB is the quantum yield of Rhodamine B (0.65 at 298 K) and and A are the1392 INDOCARBOCYANINE DYES IN AMPHIPHILE AGGREGATES 10 Fig.2. Uncorrected emission spectra of (- - - -) Rhodamine B and (-) indocarbocyanine. The excitation wavelength is 550 nm. absorbances at the wavelength of excitation of the Rhodamine B standard and the sample, respectively. Since the emission spectra of Rhodamine B and the indocarbocyanine dyes are overlapping and similar in shape (cf. fig. 2), the areas of the emission spectra, in the region 520-700nm, are taken as FRhB and F. The variation in quantum yields found for dye concentrations between lo-' and is < 10%. The corrections due to differences in refractive indicesa of the samples and the standard are found to be within these variations.RESULTS The linear dichroism (LD) was measured according to the method described previously5 (see also fig. 1). The four systems investigated were of sodium octanoate (NaC,) + decan- 1-01 + water, octylammonium hydrochloride (OAC) + decan-1-01 + water, pentaethylene glycol mono-n-dodecylether (C,,EO,) +water and octan- 1 -oyl glycerol (CSG) + water. The order parameters, S, were determined for the indocarbo- cyanine dyes DiIC2(3) and DiIC,,(3) solubilized in lamellar liquid-crystalline phases. The lamellar phases were macroscopically aligned between quartz plates. Table 1 shows the order parameters together with the ordinary (n,) and extraordinary (n,) refractive indices for the various lamellar phases. The fluorescence steady-state anisotropy (rs) of the dyes solubilized in micellar solutions and cubic liquid crystals at different compositions and temperatures was studied in amphiphile systems containing dodecyltrimethylammonium chloride (DOTAC), OAC, NaC, and C,,EO, (x = 4, 5 and 6).Typical anisotropy data are shown in fig. 3. The peak at ca. 570 nm is due to light scattering. The rs values obtained are summarized in table 2, where some calculated values are included (vide infra). As can be inferred from this table, rs varies between 0.2 and 0.3 for all the systems studied. Furthermore, the trends in rs upon changes in temperature or amphiphile concentration are similar for all systems studied. The emission lifetime (zf) was calculated from the measured quantum yield of fluorescence (@) (see table 2).The radiative lifetime (T:) was calculated from theTable 1. The ratio of the linear dichroism (LD) and the absorbance (A3 at 550 nm of DiIC,(3) and DiIC,,(3) solubilized in various aligned lamellar liquid crystals. The lamellae were tilted an angle w = 45" with respect to the impinging light beams (cf. fig. 1). n, and n, are the ordinary and the extraordinary refractive indices of the lamellar systems measured at 589 nm. The order parameter calculated from eqn (1) is denoted by S, and S is the value obtained with an extended model., The compositions of the samples are given as wt o/, . NaC, + C,,OH + H,O -0.174 1.392 1.401 -0.292 -0.289 -0.163 1.392 1.400 -0.269 -0.263 (14: 36: 50) ( 1 7 : 33 : 50) (70: 30) (70: 30) OAC + C,,OH + H,O -0.149 1.396 1.405 -0.241 -0.233 -0.147 1.393 1.405 -0,235 -0.228 Cl,EO, + H,O -0.095 1.424 1.429 -0.148 -0.134 -0.128 1.424 1.429 -0.209 -0.196 C,G + H,O -0.117 1.413 1.424 -0.186 -0.176 -0.052 1.413 1.424 -0.075 -0.0601394 INDOCARBOCYANINE DYES IN AMPHIPHILE AGGREGATES 0 1 500 550 600 wavelength/nm Fig.3. Fluorescence anisotropy (rS), calculated from the excitation spectra and recorded at the maximum of fluorescence (568 nm) for (a) DiICl,(3) solubilized in the cubic phase of Cl,EO, at 282 K and (b) DiIC,(3) in a micellar solution of NaC, (10%) at 293 K. A typical excitation spectrum is also shown. absorption spectra of DiIC,(3) and DiICl,(3) dissolved in ethanol. It was found to be equal to 2.2 ns for both derivatives. The rate constant of the non-radiative processes (knr) calculated from is presented in table 2.The fluorescence lifetime of DiIC,(3) based on the quantum-yield measurement (zp) has been compared with that obtained in picosecond absorption measurementsg (zp) of the ground-state recovery. The following values of zq/zfs are found in decan- 1-01 solutions at different temperatures (in ps): 429/415 (298 K), 367/360 (303 K), 315/ 310 (308 K), 271/260 (313 K) and 204/190 (318 K). z: is equal to zrs within the experimental accuracy. Thus the calculated radiative lifetime is in excellent agreement with experiment. DISCUSSION From the LD experiments it can be concluded that both the dye molecules are preferentially oriented with the long axis of the chromophore perpendicular to the director (the normal) of the lamellae. Furthermore, since the absolute value of the order parameters is large and it has about the same value for both DiIC, and DiIC,,, it is concluded that the fraction of the chromophore at the surface of the lamellar aggregate is close to unity, i.e.all the chromophores are associated to the surface. This is an important experimental finding and it is assumed that this high degree of association also occurs for aggregates in micellar solution. Note that the I S I value of DiIC, is slightly higher than it is for DiIC,, in the systems containing amphiphiles with short alkyl chains (NaC,, OAC and C,G), while the opposite situation is found for the Cl,EO, system. This observation can be explained by the fact that incorporation of the relatively long C,, alkyl chain into the thin lamellae of short alkyl chains willL.B-A. JOHANSSON, T. VALLMARK AND G . LINDBLOM 1395 cause a perturbation in the bilayer ordering leading to a decrease in the order parameter. For the C,G system this effect gives rise to a difference in the order parameter which is as large as a factor of three. For the C12E0, system the solubilization of the dye causes a much smaller pertllrbation of the bilayer since the length of the alkyl chain of the detergent and the dye differ by only two carbons. Note that the difference between the order parameters calculated with the approxi- mate model, i.e. eqn (l), and with the extended model increases with decreasing values of I LD/A I. For LD/AI z -0.05 the approximate model yields a value of S which is ca. 25% too small.The data shown in table 2 clearly demonstrate the need for the extended model when analysing small I LD/AI I values. The fluorescence steady-state anisotropy, r,, of DiIC2(3) and DiIC14(3) solubilized in non-ionic and ionic micelles and cubic liquid crystals are presented in table 2. For all systems studied rs increases with increasing concentration of the amphiphile and decreasing temperature. In the analysis of these data we will use a model based on the following assumptions: (a) the rotational motion of the dye molecules at the aggregate surface can be described with a strong-collision model, (b) the absorption and the emission transition dipole moments are parallel and (c) the dependence of non-radiative processes on the orientation of the dye molecule with respect to the amphiphile aggregate can be neglected. One can then showlo'l1 that for spherical aggregates 2 / 1-As2 )"( where dc is the local rotational correlation time of the fluorophore and the correlation time Tc is related to correlation times of the rotational diffusion of the aggregate (&) and the translational diffusion of the dye molecules (&) as Thus, r, of micellar solutions and isotropic liquid crystals, which are microscopic anisotropic systems, depends on both molecular order and dynamics. By assuming that the micelles rotate as hydrodynamic spheres, the rotational correlation time can be calculated from where 47 is the viscosity of the medium and r the radius of the aggregate.For sodium octanoate micelles we obtain dR = 2.2 ns when the radius r = 10 A.This value is the fastest rotational correlation time of any micelle studied here. The correlation time of the translational diffusion for the dye molecules along the aggregate surface can be estimated from (8) r2 4 n = E where D is the translational diffusion coefficient. D for DiIC14(3) solubilized in a lipid bilayer typically12 is of the order of 10-l' m2 s-l, which for sodium octanoate micelles yields #D = 30 ns. Since the fluorescence lifetimes are 5 1 ns the contribution of dD to r, [cf. eqn (5)] can safely be neglected in all the systems containing DiIC14(3). The order parameter appearing in eqn ( 5 ) can be estimated from the LD measurement on the lamellar phases. It is then assumed that S does not depend on1396 INDOCARBOCYANINE DYES IN AMPHIPHILE AGGREGATES Table 2.Anisotropy (r,) and the quantum yield of fluorescence (@) of DiIC,(3) and DiICl,(3) in different micellar and cubic liquid crystals at different temperatures and compositions. The compositions are given in wt ”/, . zf and knr denote the calculated fluorescence lifetime and rate of the non-radiative processes, respectively. T : ~ ~ is an effective rotational correlation time of the dye molecules and +br is the local rotational correlation time of the fluorophores. C,,EO, + H,O (28.3: 71.7) (10:90) C12E05 + H 2 0 C,,E05 + H,O C12EO5 + H2O” (50: 50) (65.5 : 34.5) C,,EO, + H,O (25 : 75) NaC, + H,O ( 10 : 90) NaC, + H,O (38: 62) NaC, + C, + H,Oa (40: 4: 56) OAC + H,O (10: 90) OAC + H,O (40 : 60) DOTAC + H 2 0 (10: 90) DOTAC + H,O (38: 62) DOTAC + H,O” (50 : 50) 289 299 293 289 282 299 293 289 282 299 282 289 298 293 298 293 298 29 3 289 298 293 289 298 293 289 298 293 289 298 293 289 298 293 289 0.2903 0.2739 0.2763 0.2825 0.2870 0.2747 0.2789 0.2823 0.2887 0.2788 0.3050 0.2845 0.206 1 0.2109 0.2267 0.2308 0.2246 0.2297 0.23 14 0.2034 0.2050 0.2064 0.2052 0.2077 0.21 13 0.2480 0.25 10 0.2522 0.2572 0.2594 0.2632 0.2645 0.2679 0.2724 DiIC,(3) 0.160 0.1 13 0.143 0.I68 0.208 0.1 10 0.136 0.165 0.206 0.1 12 0.240 0.177 0.079 0.086 0.1 10 0.126 0.134 0.157 0.181 0.072 0.100 0.1 15 0. I37 0.154 0.176 0.079 0.086 0.102 0.092 0.094 0.1 10 0.102 0.1 18 0.141 352 249 315 370 458 242 299 363 453 246 528 39 1 173 190 242 276 294 345 397 160 22 1 252 30 1 339 387 173 190 224 20 1 208 242 224 260 31 I 2.39 3.56 2.72 2.25 1.73 3.68 2.89 2.30 1.75 3.61 1.44 2.10 5.33 4.8 1 3.68 3.17 2.95 2.44 2.06 5.80 4.07 3.51 2.87 2.49 2.13 5.33 4.8 1 4.0 1 4.52 4.35 3.68 4.0 1 3.39 2.76 803 803 46 1 46 1 60 1 60 1 762 762 1000 1000 453 453 590 590 747 747 1013 1013 485 48 5 1473 1473 826 826 149 161b 172 26 1 300b 318 37gb 309 365b 385 477b 45 1 582b 135 138 184 190 208 215 255 266 296 311 352 374 237 24 1 270 275 322 329 306 313 323 330 395 406 372 382 448 462 567 590L. B-A.JOHANSSON, T. VALLMARK AND G. LINDBLOM Table 1. (cont.) 1397 system C,,EO, + H,O (28.3:71.7) C,,EO, + H 2 0 (10:90) C,,EO, + H 2 0 (50 : 50) C,,EO, + H,Oa (65.5: 34.5) C,,EO, + H,O (25: 75) NaC, + H 2 0 ( 10 : 90) NaC,+ H,O (38: 62) NaC, + C, + H20a (40:4:56) OAC + H 2 0 (10:90) OAC + H,O (40: 60) DOTAC + H,O (1 0: 90) DOTAC + H,O (38 : 62) DOTAC + H,Oa (50: 50) 289 299 293 289 282 299 29 3 289 282 299 282 289 298 293 289 298 293 289 298 293 289 298 293 289 298 293 289 298 293 289 298 293 289 298 29 3 289 0.2924 0.2795 0.2835 0.2885 0.293 1 0.28 17 0.2873 0.2888 0.2946 0.2947 0.3005 0.2916 0.2324 0.2347 0.2383 0.2376 0.2413 0.2484 0.2371 0.2462 0.2476 0.23 15 0.2340 0.2338 0.2347 0.2355 0.2386 0.2532 0.2538 0.2544 0.2534 0.2538 0.2577 0.2588 0.2625 0.2654 DiIC,,(3) 0.245 0.175 0.21 1 0.246 0.299 0.174 0.215 0.258 0.300 0.175 0.377 0.268 0.094 0.134 0.149 0.126 0.149 0.173 0.157 0.173 0.204 0.105 0.1 11 0.124 0.136 0.152 0.178 0.1 10 0.126 0.141 0.157 0.181 0.204 0.189 0.220 0.244 539 385 464 54 1 658 383 473 568 660 385 834 590 208 294 329 276 329 38 1 345 38 1 450 23 1 244 272 298 334 39 1 242 276 31 1 345 397 450 415 484 536 1.40 2.14 1.70 1.39 1.06 2.16 1.66 1.31 1.06 2.14 0.744 1.24 4.35 2.95 2.58 3.17 2.58 2.17 2.44 2.17 1.77 3.87 3.64 3.22 2.90 2.54 2.10 3.68 3.17 2.76 2.44 2.06 1.77 1.95 1.61 1.41 1264 765 970 1207 1558 78 1 1037 1222 1595 818 2185 1368 238 347 403 339 41 7 523 418 51 1 61 3 264 285 315 34 1 397 48 1 35 1 405 459 493 582 689 642 783 898 1264 765 970 1207 1558 78 1 1037 1222 1595 818 2185 1368 270 420 505 408 527 708 528 686 884 276 299 332 362 425 523 360 417 474 510 606 723 672 828 957 a Cubic liquid-crystalline phase.May contain contribution from &.1398 INDOCARBOCYANINE DYES IN AMPHIPHILE AGGREGATES DOTAC W DiIC,(3) OAC W Fig. 4. Schematic diagram showing the relative dimensions of the various amphiphile aggregates and the indocarbocyanines DiIC,(3) (n = 2 and 14).The circles represent the cross-section of the aggregates. the geometry of the amphiphile aggregate.l3?l4 Here we have used a fixed value of S = - 0.3 in the evaluations of all the data. A change in the region I S I < 0.3 will only slightly affect the value of r, owing to its S2 dependence. Therefore the Tc dependence is dropped in the second term on the right-hand side of eqn (5) and the dynamics are calculated from (9) rs=( 2 1-S2 zf zf)+E. 5 l+-+- rc d c Since zf is determined from a separate experiment l/zEff = l / r c + l/dc can be calculated from rs (see table 2). For the DiICI4(3) where dD is negligible qJc can be obtained by subtracting the contribution from a calculated &.The correlation times dc of DiIC2(3) presented in table 2 are also corrected for the rotation of the micelles but note that the values for the C, micelles may contain contributions from the translational diffusion. It is found that dc increases significantly with increasing concentration of the amphiphile except for the systems containing C,,EO,, where it is almost independent of the amphiphile concentration. Thus the rapid local molecular motion of the dye molecule is different in ionic and non-ionic amphiphiles. In order to investigate whether this observation is due to the electrostatics of the system the dependence on dc was investigated for micellar solutions. Varying the NaCl concentration up to ca. 35 mmol dm-3 gave no significant effects.It is therefore likely that the charges on the micellar surface are of no importance and we must search for other explanations. It is well known that non-ionic amphiphiles form very large aggregates,15 while the ionic amphiphiles form1 s 1 0 0 5 a= O 2 : --. 01 06 0 4 0 2 L. B-A. JOHANSSON, T. VALLMARK AND G. LINDBLOM 1399 1 1 I I I 1 500 1000 1500 2000 2500 3000 0 1 I C 2 ( 3 ) 0 2 00 400 600 ” 200 400 600 800 1000 4Jc IPS Fig. 5. Dependence of the reciprocal rate of the non-radiative processes (1 /knr) on the rotational correlation time (#c) for DiIC,(3) (n = 2 and 14) in various systems. In the upper plots the systems are: and A, decan-1-01; 0 and A, micelles and cubic liquid crystals of pentaethyleneglycol mono-n-dodecylether C,,EO, ; 0, vesicles of dipalmitoylphosphatidyl choline obtained at 323, 313, 298, 289 and 277 K.#c increases with decreasing temperature. In the lower plots the systems contain the amphiphiles 0, octylammonium chloride, OAC; 0, sodium octanoate, NaC, ; a, dodecyltrimethyl ammonium chloride, DOTAC. The compositions are given in table 2. comparatively small ones. n reasoname interpretation or me ooservea aata is Inen that the dye molecules are easily solubilized in the ordinary large aggregates of C,,EO, and a further increase in the size of the aggregate upon an increase in surfactant concentration has no effect on the rotational correlation time of the dye. However, for the small ionic micelles the large dye molecules cannot be completely incorporated in the normal aggregates so the micelles have to change size in order to fully solubilize the chromophore (see fig.4). This interpretation is supported by the observation that the order parameter for DiIC,, is smaller for the amphiphiles having shorter alkyl chains than for DiIC,. Thus, the micelles containing the chromophore are different from the ordinary micelles built up of detergents only. Consequently the local environment of the dye molecules change with concentration, leading to the variation in correlation time. This increase in q5c continues into the cubic phases. The temperature dependence of the correlation time is quite normal, i.e. q5c increases with1400 INDOCARBOCYANINE DYES IN AMPHIPHILE AGGREGATES decreasing temperature because of a decrease in the molecular motion. From a comparison between the correlation times obtained for the two dyes DiIC, and DiIC,, it can be concluded that the dye with the shorter alkyl chain moves more rapidly, which is also expected.The largest difference is found for the C,,EO, system where the correlation times for the two dyes differ by more than a factor of two. An interesting result of this study is the observation that the rate constants for the non-radiative processes depend on the rotational correlation time of the dyes (see fig. 5). Evidently both systems containing amphiphile aggregates and ordinary solutions show a very similar dependence of l/knr on $c. This suggests that the orientational distribution of the dyes with respect to the amphiphile aggregate has a negligible effect on the non-radiative processes, which is an assumption made in the model above [eqn (5)].This indicates that a time-resolved fluorescence measurement of DiIC,(3) solubilized in amphiphile aggregates should give one lifetime, i.e. no effects of microheterogeneity. We thank Miss Lilly Sjolander for technical assistance and the Swedish Natural Science Research Council for financial support. R. D. Klausner and D. E. Wolf, Biochemistry, 1980, 19, 6199; M. F. Ethier, D. E. Wolf and D. L. Melchoir, Biochemistry, 1983,22,1178 ; K. Kurihara, K. Onuki, Y. Toyoshimaand M. Sukigara, Mol. Cryst. Liq. Cryst., 1981, 68, 69. P. J. Sims, A. S. Waggoner, C-H. Wang and J. F. Hoffman, Biochemistry, 1974, 13, 3315; G. L. Bashford, Biosci. Rep., 1981, 1, 183. R. N. Robertson, The Lively Membrane (Cambridge University Press, Cambridge, 1983); V. Luzzati, in Biological Membranes, ed. D. Chapman (Academic Press, New York, 1968), vol. 1, p. 7 1. L. B-A. Johansson and A. Davidsson, J. Chem. SOC., Faraday Trans. 1, 1985,81, 1373. L. B-A. Johansson and G. Lindblom, Q. Rev. Biophys., 1980, 13, 63. J. R. Lakowicz, Principles of Fluorescence Spectroscopy (Plenum Press, New York, 1983). ' R. F. Kubin and A. N. Fletcher, J. Lumin., 1982, 27, 455. * J. B. Birks, Photophysics of Aromatic Molecules (Wiley-Interscience, London, 1970); F. R. Lipsett, Prog. Dielectr., 1968, 7 , 217. V. Sundstrom and E. Akesson, to be published. lo L. B-A. Johansson and G. Lindblom, J. Chem. Phys., 1983, 78, 1519. l1 P-0. Eriksson, L. B-A. Johansson and G. Lindblom, in Surfactants in Solution, ed. K. L. Mittal and la Z. Derzko and K. Jacobson, Biochemistry, 1980, 19, 6050. l3 L. B-A. Johansson, 0. Soderman, K. Fontell and G. Lindblom, J. Phys. Chem., 1981, 85, 3694. l4 J. Charvolin and P. Rigny, J. Chem. Phys., 1973, 58, 3999. l5 C. Tanford, The Hydrophobic Eflect (Wiley-Interscience, New York, 1980). B. Lindman (Plenum Press, New York, 1984), vol. 1, p. 219. (PAPER 4/ 1 189)

ISSN:0300-9599    

DOI:10.1039/F19858101389

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985,81, 1401-1414 Separation of the Effect of Solvent Structure on the Kinetics of Substitution Reactions into Contributions to the Initial and Transition States using Free Energies of Transfer Kinetics of the Solvolysis of 1,2-Chlorothiocyanatobis( 1,2-diarninoethane)- cobalt(II1) Ions in Water and Water + Propan-2-01 Mixtures BY ALI E. EID AND CECIL F. WELLS* Department of Chemistry, University of Birmingham, Edgbaston, P.O. Box 363, Birmingham B15 2TT Received 16th July, 1984 Rates of the first-order solvolysis of 1,2-~hlorothiocyanatobis( 1,2-diaminoethane)cobalt(111) ions have been measured for a range of temperatures in water and in water+propan-2-01. A plot of log (rate constant) against the Y factor is linear, but a plot of log (rate constant) against the reciprocal of the dielectric constant is curved. From a comparison of the variation of the enthalpy and entropy of activation with the physical properties of the solvent mixture it is concluded that solvent structure is an important factor determining solvolytic reactivity.The evidence currently available concerning the nature of the transition state in substitution reactions and, in particular, the solvolytic transition states of ColI' complexes is assessed. The influence of values for the free energy of transfer of ions on the assignment of the contributions of solvent structure to the initial and transition states for substitution reactions is discussed. .As a result of the application of a free-energy cycle relating the process initial state + transition state in water to that in the mixture, it is concluded that the influence of solvent structure predominates on the pentacoordinated cation in the transition state over its influence on the hexacoordinated cation in the initial state.Following the suggestionl9 that solvent structure is important in controlling rates of reactions involving ions in water and in water + cosolvent mixtures, a survey of the solvolysis reactions of neutral molecules and of ions revealed3 that it is rare for these to obey the relationships predicted on the basis of point charges in a dielectric continuum, viz. that logk (k is the rate constant) should vary linearly with the reciprocal of the dielectric constant, Oil, independent of the molecular structure of the cosolvent. Subsequently it has been shown that, for the solvolysis of complex ions, ;a good correlation is obtained between the extrema in the enthalpy AH* and entropy 14S* of activation and the extrema in the physical properties4 of the water+cosolvent mixtures which are related to changes in solvent structure, both for complexes such as I ,6-[C0py,Cl,]+,~ 1,6-[C0en,Cl,]+~ and 1,6-[C0(4Me-py),Cl,]+,~ which show no linear relationship between log k and D;l, and for complexes such as 1,2-[Coen,N3C1]+ and 1,6-[C0en,N,Cl]+,~ which do show3 a linear relationship between logk and Ill1.In general, we have used propan-2-01 as the cosolvent, as the physical properties of water + propan-2-01 mixture^,^ such as the decrease in the partial molar volume of the alcohol, E- V i , the ultrasonic absorption, the excess contribution to the change in the temperature of maximum density, ATE, and the excess enthalpy of mixing, AH:, 14011402 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION show that this branched-chain alcohol produces more structure when added in low concentrations to water.However, with 1 ,6-[Copy4C1,]+ methanol, propan-2-01 and 2-methylpropan-2-01, and with 1 ,6-[Co(4Me-py),C12]+ propan-2-01 and 2-methylpro- pan-2-01, have been used as cosolvents: it was found that the extrema in AH* and AS* move to lower mole fractions of alcohol, x,, along that series, in a similar manner to the movement4 of the extrema in 6- V i and the ultrasonic absorption. As the correlationlo of volumes of activation, AV*, with overall volume changes, AVO, for the solvolysis of complexes of this type suggests that the rate-controlling process for a hexacoordinated complex CZc is the stretching of the bond between the metal MZm and the leaving ligand Xzx (where 2 represents the overall charge on the species), with the extent of the stretching in the transition state approximating to full dissociation, a free-energy cycle has been constructed3 relating the free energies of activation in water, AG& and in the mixture, AG;, by the free energies of transfer of the individual species i from water into the mixture, AG,"(i), AG; C zw' M",m + where subscripts w and s refer to water and the mixture, respectively.The equation 2.303RT log (k,/k,) - AG:(XZx) = AG,"(MZm) - AG:(CZrn) (1) can be derived from this cycle.3 It has been found independently that, usually, for water+cosolvent mixtures, AG,"(i) is negative when i is a cation.ll Consequently, if the left-hand side of eqn (1) is negative, - AG,"(MZm) > -AG:(Czc), and the influence of solvent structure on the cation is greater in the transition state than in the initial state; or, if the left-hand side is positive, - AG,"(CZc) > - AG,"(M'm), and the influence of solvent structure on the cation is greater in the initial state than in the transition state.It was found for a wide selection of complex cations CZc, based on the metal ion Co"* and when XZx is a halide ion, that usually the left-hand side of eqn (1) is negative, showing the dominance of the effect of solvent structure on the cation in the transition state for these solv~lyses.~~ 5-9 The relevance of such free-energy cycles to the assignment of the effects of solvent structure to the initial and transition states will be discussed again later.We now examine the kinetics of the solvolysis in water + propan-2-01 for another complex cation, 1,2-[Coen2SCNCl]+, where the plots of log k against are linear and coincident for the cosolvents acetone, methanol, ethane- 1,2-diol and sucrose.12 EXPERIMENTAL MATERIALS 1,2-Chlorothiocyanatobis( 1,2-diarninoethane)cobalt(m) perchlorate was prepared by the method of Werner.I3 AnalaR perchloric acid and AnalaR propan-2-01 were used. A solution of sodium perchlorate was prepared as described previo~sly.~ Water was distilled once from an all-glass still.A. E. EID AND C.F. WELLS 1403 Table 1. Variation of rate constant k with additions of HC10, and NaC10, at a constant temperature of ca. 50 "C concentration addition /mol dm-3 k / s-l - 2.64 HClO, 2.0 x 10-5 2.59 HClO, 2.0 x 10-4 2.47 HClO, 2.0 x 10-2 2.38 NaClO, 2.0 x 10-2 2.4 1 - PROCEDURE Rates of solvolysis were measured in the thermostatted cell compartment of a Unicam SP6-500 spectrophotometer with thermostatting using water circulation from a Pye Unicam temperature controller. Optical densities were either read directly from the digital display or recorded on a Pye Unicam DR 16 digital printer. Changes in optical density with time were followed at 490 nm using an initial concentration for the complex of 0.01 mol dm-3. Spectra were measured using a Unicam SP 800 spectrophotometer.RESULTS AND DISCUSSION PRELIMINARY OBSERVATIONS The spectrum of the 1 ,2-[Coen2SCNC1]+ ion in water +propan-2-01 was found to have a maximum in the visible region at 500 nm. On solvolysis, this maximum absorption increased with the product, having a maximum at 490nm. A sharp isosbestic point was found at 532 nm and there may be another one in the region of 400 nm. Optical densities, O.D.,, at 490 nm were measured at varying times, t , and the final optical density, O.D.,, was recorded. Plots of log(O.D., -0.D.J against time were always linear. RATE OF SOLVOLYSIS IN WATER Table 1 shows the effect of various inorganic additions on the rate constant at a constant temperature of ca. 50 "C. This shows that an added acid, HCIO,, has no effect beyond a possible small ionic-strength effect.Accordingly, no acid was added to kinetic determinations in water or in the mixtures. The decrease in optical density was followed in pure water at 40, 50, 55 and 60 "C and values for the first-order rate constant were calculated from the slopes of the linear plots of log (O.D., -O.D.,) against time. Two determinations were made at each temperature and were always in close agreement : the mean values of the rate constants are given in table 2 and the deviation from the mean never exceeded 1 % . A plot of log k against the reciprocal of the absolute temperature was linear, and values for AH * and AS* determined using the least-squares procedure and a computer with all the individual values for k are given in table 3. The mean value found for k at 50.0 "C of 2.81 x lo-, s-l can be compared with k = 1.48 x s-l found by Panasyuk and ArkharoP using a potentiometric method.VARIATION OF RATE CONSTANT WITH SOLVENT COMPOSITION The same kinetic procedure as described for pure water was used for a range of concentrations of propan-2-01 up to 50 vol% for the same range of temperatures as1404 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION Table 2. Variation of the first-order rate constant k temperature s-I) with solvent composition and [propan-2-01] T/"C ~ . _ _ _ _ _ _ _ _ _ mole wt ;4 fraction 40.0 50.0 55.0 60.0 0 3.98 8.04 12.19 16.44 18.60 20.78 25.22 29.76 34.41 39.17 44.04 0 0.0 123 0.0255 0.0400 0.0557 0.0641 0.0729 0.09 18 0.1 13 0.136 0.162 0.191 0.85 0.80 0.7 1 0.69 0.65 0.65 0.61 0.55 0.476 0.457 0.441 0.420 2.8 1 2.72 2.40 2.1 1 1.94 1.96 1.84 1.65 1.53 1.47 1.43 1.36 4.90 4.55 3.91 3.58 3.29 3.37 3.10 2.88 2.95 2.78 2.69 2.61 8.7 7.9 6.9 6.5 6.2 6.2 5.9 5.6 5.3 5.0 4.65 4.55 Table 3.Variation of transition-state parameters with solvent composition" [propan-2-01] mole AH* AS* at 25 "C AG* at 25 "C wt% fraction /kJ mol-l /J K-' mol-1 /kJ mol-l 0 3.98 8.04 12.19 16.44 18.60 20.78 25.22 29.76 34.4 1 39.17 44.04 0 0.0 123 0.0255 0.0400 0.0557 0.0641 0.0729 0.09 18 0.1 13 0.136 0.162 0.191 98+ 1 97+ 1 96k 1 94+ 1 94+2 94+2 95+2 98+2 102 + 2 101 & 1 loo+ 1 101 f 1 -9.5+ 1.6 -15f3 - 19+4 -25+4 -25+6 -25+5 -25+6 -18&7 - 1.6k5.1 - 4.7 f 4.3 - 9.3 + 3.1 - 5.8 f 4.3 101 + 1 101 + 2 101 + 3 101 + 3 102 & 4 102+3 102 f 4 102f.5 103 f 3 103 & 3 103f2 103 + 3 a & indicates the standard error.used with water. Linear plots were also obtained for log (O.D., - O.D.,) against time, and the first-order rate constants calculated from the slopes of these plots were always in good agreement for duplicate measurements for the same set of conditions of concentration of propan-2-01 and of temperature. The mean values of these duplicate runs for each set of conditions are collected in table 2, and the deviation from the mean never exceeded 2.2%. Plots of log k against reciprocal of absolute temperature were linear for each concentration of propan-2-01. Values for AH*, A S * and AG* together with their standard errors were determined by applying the least-squares procedureA. E. EID AND C . F. WELLS 1405 0.5 r 0.1 1 I I I I I 1 1.4 1.6 1.8 2.0 22 2.4 2.6 102 0;' Fig.1. Plot of logk against the reciprocal of the dielectric constant at 50 "C for the solvolysis of 1 ,2-[Coen2SCNC1]+ in water + propan-2-01. to all individual values of the rate constants using a computer and they are collected in table 3. Laidler et aI.I4 considered the application of electrostatics to the separation of charges involved in the extension of the metal-ligand bond, MZm.*.XZx, in the transition state for the solvolysis of a complex CZr in a changing medium, resulting finally in the Laidler-Landskroener equation based on Born and Kirkwood effects on the ions ( r is the radius) and dipoles (G) in a dielectric continuum. This equation was later modified3 to allow for deviations from a dielectric continuum due to changing solvent structure by including terms involving the free energy of transfer of each species from water into the mixture, resulting in where D represents dielectric constant and subscript n indicates AG:(i) for the transfer process water + mixture, excluding the electrostatic effects to which the earlier terms on the right-hand side of eqn (2) relate.If AG;(C), z AG;(Mn)+AG:(X),, i.e. considering the original Laidler-Landskroener equation containing only the electro- static terms, logk should have an inverse relationship with D,; other electrostatic approa~hes~~3 l6 to this type of dissociative mechanism which assume a continuity in the medium also require a relationship of the same type. Although the kinetic data of Panasyuk and Arkharov12 for mixtures of water with one cosolvent chosen from sucrose, ethane-l,2-diol, glycerol, ethanol, methanol and acetone all lie on the same linear plot for log k against Dkl at 50 "C, fig.1, drawn using our data at 50 "C against D;' for mixtures of water with a cosolvent which induces much greater4 structural changes in water than the cosolvents used by Panasyuk and Arkharov, is a pronounced curve. We conclude, therefore, that when a cosolvent is chosen for its1406 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION c I I I I 0.0 5 0.10 0.15 0.20 x2 (mole fraction) 901 Fig. 2. Plot of AH* against mole fraction of propan-2-01 for the solvolysis of 1,2-[Coen2SCNC1]+ in water + propan-2-01. T x2 (mole fraction) Fig. 3. Plot of AS* against mole fraction of propan-2-01 for the solvolysis of 1,2-[Coen2SCNC1]+ in water + propan-2-01.structure-enhancing properties, the simple Laidler-Landskroener equation1* without the terms in AG:(i)n (or with them cancelling out) does not apply. Fig. 1 shows that there is a differential effect of solvent structure on the initial state, CZc, and the transition state, (M'm.. exz,), for the solvolysis of 1 ,2-[Coen,SCNCl]+. This differential effect of solvent structure on the initial and transition states is also shown by the variation of AH* and AS* with solvent composition in fig. 2 and 3. These plots both show a minimum in the region of x, = 0.05 well outside the standard deviations, followed by an approximate flattening off in the region of x, z 0.1. This is comparable with similar extrema or changes in AH* and AS* found for theA.E. EID AND C. F. WELLS 1407 solvolysis in water + propan-2-01 of 1,2-[C0en,N,Cl]+,~ 1,6-[C0(4Mepy),Cl,]+,~ 1,6- [C~en,Cl,]+~ and 1,6-[C0en,N,Cl]+,~ and AH* and A S * for the solvolysis of 1 ,6-[Copy4C1,]+ in water + propan-2-o15 give broad maxima covering both regions. The extrema at x, z 0.05 correspond closely to the minimum in the decrease in the partial molar volume of propan-2-01, - V i , found17 in its mixtures with water, which is thought'? l1 to correspond to maximum structure formation within the 'flickering icebergs' of water1* induced by the pressure imposed by the alkyl chains of the alcohol in the intervening cavities. The flattening off in fig. 2 and 3 corresponds approximately to the maximum in the ultrasonic absorption foundlg in water + propan-2-01, where the increased presence of alkyl groups starts1? l1 the break-up of the water structure.l8 As with the other solvolyses of complex ions in water +propan-2-01,~-~ a plot of AH* against AS* for the solvolysis of 1 ,2-[Coen,SCNCl]+ is linear.EXTENSION OF THE MZm-XZx BOND IN THE TRANSITION STATE We now wish to determine whether this large effect of solvent structure on the rate of solvolysis is due to a dominating effect on the initial state or the transition state. As discussed above, for the solvolysis of a variety of complexes in mixtures of water with a range of cosolvents, we have used a free-energy cycle. However, the free-energy cycle requires an assumption that the extension of the MZ~.-.XZx bond in CZc in the transition state is sufficiently long that it is a reasonable approximation to consider MZm and XZx as separate entities:3t5-9 as this assumption has been criticised20 we should, perhaps, assess critically the evidence in its favour.Moreover, as our interpretation21 of the kinetics of a related process, the substitution of ligands into the aquanickel(I1) ion, where we have also used a free-energy cycle based on the assumption that the extension of the Ni2+---H,0 bond is long enough in the transition state to approximate to full separation, has also been criticised without evidence being supplied against it,22 it would seem appropriate to examine this here also. It can be regarded as a special case of the solvolysis cycle with Xzx = H,O and the incoming ligand YzY in close proximity to CZc.As stated above, much of the evidence usually presented in favour of full dissociation MZm..*XZx in the solvolysis transition state rests on the comparison of A I/* with the overall volume change A Vo.lo A criticism levelled at the interpretation of A I/* is that these often restrict the application of A V* to one process, the extension of MZm...XZx, neglecting contributions to AV* from the movement of other species in the mixture. However, this was obviated by Palmer and KelmlO in their application of Pm = E+AV*-Px (3) where 6 is the partial molar volume of species I, to the solvolysis of CZc = [CO(NH~)~X](~-~X)+. With XZx varying over C1-, Br-, SO:-, NO;, Me,SO and H,O, they obtain rm = 54.9 0.9 cm3 mol-l, which convincingly suggests a common intermediate cation with Xzx fully removed in the transition state.10923 The same authors have now applied eqn (3) to complexes of the type [CoN,X,]+, covering complexes [Coen,Cl,]+ used by us, with a comparable result: for a range of [CoN,Cl,]+, with the mean AV* = 1.4f 1 cm3 mol-1 and the mean A V" = - 13.9 f 1.1 cm3 mol-l, the difference, 12.5 2.1 cm3 mol-l, corresponds closely24 to the volume of 14 cm3 mol-l for the coordinated H,O added subsequently to the full extension of the CO~+-C~- bond in the transition state.The fact that more positive values for AV* have been for the solvolysis of complexes of [CoN,Cl,]+ by other workers adds support to the conclusion that the Co3+.-C1- bond is fully broken in the transition state.1408 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION Table 4.Grunwald-Winstein m values for the solvolysis of [CoL,XCl]+ complexes in water + cosolventa c o m p 1 ex cosolvent m ref. 1,6-[COPY,C1,1+ 1,6-[COPY,C1,1+ ~,6-~COPY,C~,1+ 1 ,6-[Coen2C1,]+ 1 ,6-[Coen,Cl2]+ 1,2-[Coen,N3C1]+ 1,6-[Coen,N3C1]+ 1,2-[Coen2SCNC1]+ 1 ,6-[Co(4Mepy),C12]+ 1 ,6-[Co(4Mepy),C12]+ methanol propan-2-01 t-butyl alcohol methanol propan-2-01 propan-2-01 propan-2-01 propan-2-01 propan-2-01 t-butyl alcohol -0.13 -0.14 -0.13 0.25 0.17 0.17 0.08 0.20 non-linear plot non-linear plot 5 5 5 45 6 8; this work 9 this work 7 7 a py, pyridine; 4Mepy, 4-methylpyridine ; en, 1,2-diaminoethane. Strong additional support for this is available from an alternative experimental source. The constancy of the ratio of stereochemical forms in the solvolytic products from complexes [CoN,LX]"+ with varying X indicates26 that common penta- coordinated intermediates exist for either the cis or trans forms.This strongly suggests that full dissociation CO~+--X- is a good approximation for the transition state.26 This result can be compared with the experimental kinetic evidence on the substitution of ligands Y with the aquanickel(I1) ions, Niii,. where the independen~e~~? 28 of the derived first-order rate constant for varying identity of Y for the substitution of Y"Y within the intermediate complex NiiiYZy over a wide range of Y in size and charge supports the conclusionz1 that the Ni2+.-.H20 bond is completely broken to a first approximation in the transition state.When allowance is made for the volume change in achieving the intermediate complex from the separated Niit and YzY, A V* for the first-order substitution28p29 for a range of YZy agree closely with that found in the water exchange, AV* = 7.1 cm3 m~l-l,~O as expected from the Eigen-Wilkins mechanism.27 This is less than AV* z 14-15 cm3 m01-131*32 expected for a complete transfer of an H20 from within the coordination shell of Niii to a situation outside the complex l$iiYzy, and this has been inter~reted~~ as corresponding to 20-50% complete extension of the Ni2+*-.H20 bond in the transition state. However, it has been pointed that this assessment makes no allowance for the expected contraction in the Ni2+-. .H20 bond distances remaining in the penta-aquanickel(I1) ion, based on the shorter Ni2+-H20 bond distance found in a pentacoordinated trigonal-bipyramidal situation compared with that in the hexacoordinated octahedral situation;33 moreover, this contraction is supported by the observation that Ni-Cl bond distances are shorter in the trigonal bipyramid than in the o c t a h e d r ~ n .~ ~ It has been estimated31 that this will cause a contraction of ca. 4 cm3 mol-l, which would make the contribution to AV* derived from the Ni2+*-.H,0 extension ca. 11 cm3 mol-l. However, this is based on bond distances found in the solid phase:31 it seems highly likely that in solution the solvent molecules will also be pulled in closer to the Ni2+ in the more open trigonal bipyramid compared with the closely packed octahedron, and this contraction could easily account for the disparity between 11 cm3 mol-l and the expected 15 cm3 mol-l. Thus the estimate for the Ni2+..-H20 extension of Caldin et ~ 1 .~ ~ must be taken as a lower estimate on two counts: the observed AV* z 7 cm3 mol-1 does not contradict the expectation from the zeroA. E. EID AND C. F. WELLS 1.1 1.0- .y 1409 - 3 0.9 0.8 I 1 I I 4 0.71 1 2 3 Y Fig. 4. Plot of logk against the Y factor at 25 "C for the solvolysis of 1,2-[Coen,SCNCl]+ in water+propan-2-01: Y values calculated6 from the rate constants of 0, Akhtar and Begum46 and 0, Robertson and Sugarn01-i.~' order2' in YzY in the rate-determining step that incipient bond formation of Y with Ni2+ is absent in the transition state, i.e. that the Ni2+.--H20 bond is completely extended before Y becomes involved with the Ni2+ ion.To a first approximation, the A V* values, the stereochemical ratios of products for solvolysis and the zero order in YzY of the rate-determining step for substitution indicate that the M%-XZx bond is completely extended to its breaking point in the transition state when MZm is Co3+ and Xzx is C1- for solvolysis and MZm is Ni2+ and Xzx is H,O for substitution. One further line of evidence supports this for solvolysis. Linear plots have been found for log (rate constant) against the Grunwald-Winstein Y factor34 for the solvolysis of many of these complexes involving the loss of a chloride ion. With such a linear plot, the effect of changing the solvent on the solvolysis of M-Cl is proportional to the effect the same change of solvent has on the solvolysis of t-butyl chloride relative to the reference solvent: this suggests that the solvent inolecules are involved mechanistically in the same way in the two cases.Specifically, the effect of the change in solvent on the difference in the free energy, AG'*, between the transition state and the initial state for the solvolysis of MCl is proportional to the effect of the same change in solvent on the difference in free energy, AG*, between the same two states for the solvolysis of t-butyl chloride, expressed as AG2 - AGk* = m(AGg - AGZ) (4) where subscript R refers to the reference solvent and m is a proportionality constant. The extent of charge development in the transition state for the solvolysis of t-butyl chloride is practically complete.35 If the charge development is less than complete in the transition state of MCl, the extent of M+..-Cl- would be expected to vary with change in the cosolvent or with change in the other substituents X in ML4XCl.However, table 4 suggests that this variation with cosolvent does not occur, as the slope m for the linear plots of log k against Y for the solvolysis5 of 1,6-[Copy4C1,]+ are in close agreement over a range of cosolvents. Fig. 4 shows that an approximately linear1410 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION 1.6 - 1.4- - EY + m 1.2 - 1 2 3 4 1 .o Y Fig. 5. Plot of logk against the Y factor at 25 "C for the solvolysis of 1,2-[Coen,N,Cl]+ in water+propan-2-01: Y values calculated6 from the rate constants of 0, Akhtar and Begum46 and 0, Robertson and Sugamori.*' plot is obtained for log k against Y for the solvolysis of 1,2-[Coen,SCNCl]+ reported here and fig.5 shows that a linear plot is also obtained for the solvolysisa of 1 ,2-[Coen2N3Cl]+ in water + propan-2-01 : the values of Y in water + propan-2-01 are taken from Groves and Wells.' The values of rn taken from the slopes are given in table 4 together with the values of rn for the solvolysis of other complexes [CoL,XCl]+ in water +propan-2-01:~~ this shows that rn remains relatively constant for varying X with the same cosolvent, The invariance of rn with changing cosolvent for 1,6-[Copy4C1,]+ and the invariance of rn with varying X for complexes [CoL4XCl]+ with the same cosolvent support the above evidence for complete extension of the C O ~ + - C - bond in the transition state for the solvolysis of this type of complex.In view of the change in sign of rn between [Copy4C12]+ and [Coen,XCl]+, the absence of linearity for plots of log k against Y for 1,6-[Co(4Mepy),C12]+ with propan-2-01 and t-butyl alcohol as cosolvents imply that rn z 0 in these cases. SEPARATION OF SOLVENT STRUCTURE EFFECTS FOR INITIAL AND TRANSITION STATES FREE ENERGIES OF TRANSFER The criticism of our work by Blandamer et aZ.20v22 includes both a misconception and an error. The former lies in the different intentions of their work and of ours. They use Leffler's and Grunwald's solvent operator, 8m,36 to assess the effect of solvation in the whole of the transition state via 6, AG* = 6, GGs - 6, GYs ( 5 ) where subscripts IS and TS indicate initial and transition states, whereas we use free-energy cycles to explore the influence of solvation and solvent structure on individual species within the transition state.In either case, free energies of transfer are required. In our analysis of substitution21 and solvolysis3~ 5-9 reactions we introduce values of AG:(i) directly determined for i itself, and cycles such as eqn (1) give a difference of AG:(i) for the individual species i in the transition and initial states: AG:(i) for i in the transition state is only obtainable if AG:(i) in the initial state can be determined. However, Blandamer and Burgess have used AGY(i') for AG:(i) in theirA. E. EID AND C. F. WELLS 141 1 Table 5. Values for the free energies of transfer AG,"(i) (kJ mol-l) at 25 "C for transferring i from water into water + methanol [methanol] substitution of bipy into Niii mole wt% fraction AG,"(ReCli-) AG,"(Fe*+) 6, GI, 6, GTS 8.1 8.1 8.1 16.5 16.5 16.5 25.2 25.2 25.2 34.4 34.4 34.4 47.7 54.2 0.0472 1 .80" 0.0472 - 0.0472 - 0.100 4.73" 0.100 - 0.100 0.159 6.8" 0.159 0.159 0.228 10.1" 0.228 - 0.228 0.313 - 0.400 - - - - - - 3.90"' - 3.6"' - 7.0"' - 7.5".-9.8" - 9.2e - 6.6"' - 6.4"~ - 1 1.0"p - 9.6"' - - - - - - - - - - - - - - - - - 6. la*d - 16.4"*' - 21.6" - 20Se - 14.8",' - - - 8 . 2 " ~ ~ - - - - - 8. la*d -9.1"qd - - a Mole fraction scale; using data for Fephen,ReCl, of Blandamer and Burgess; using using data for Fe(phen),(ClO,), of data for Fe(phen),(ClO,), of Blandamer and Burgess; Sengupta et al. ; molar scale using data of Blandamer and Burgess.calculations where i # i' but i is similar to i': they have 37 AG,"(Fe2+) and AG,"(Cu2+) as a measure of AG,"(Ni2+) in their analysis of the substitution of 2,2'-bipyridine into Niih. This can, however, introduce large errors. Their values for AG,"(Fe2+) were from Fe2+ + 3 phen Fe(phen)i+ (6) using AG,"(Fe2+) = AG,"[Fe(phen),] - SAG,"(phen) - AAGO, (7) where phen is 1,lO-phenanthroline and AAGO, is the change in AG;, the standard free energy of formation of Fe(phen)i+, in transferring the equilibrium from water into water + cosolvent : Blandamer and however, use + AAGO, in eqn (7) and find a wide variation in AG,"(Fe2+) depending on the source of data used for AG:[Fe(phen)i+]. We have calculated AG,"(Fe2+) in water+methanol using their source of data for AG&39 their solubilities of [Fe(phen),][ReCl,] and [Fe(phen)3](C10,)2,38 their values for AG,"(~hen),~~ values for AG,"(ReCli-) calculated using their solubilities for CS2[ReC16]40 and our values for AG,"(Cs+),ll together with our values for AG,"(C10;) :11 all corrections have been applied to our values used for AG,"(i).ll Table 5 shows that we obtain good agreement for AG,"(Fe2+) using the two alternative solubilities, the hexafluororhenate(1v) or the perchlorate, together with eqn (7).Table 5 also con- tains our values for AG:(Fe2+) calculated using eqn (7) with the solubilities of [Fe(phen),](ClO,), and values for AG,"(phen) of Sengupta et al.,41 with the same values for AG",g and AG,"(CIO,)ll used above: these values differ from those calculated using the data of Blandamer and but have, nevertheless, a substantial negative magnitude, as expected for a M2+ ion.ll Our values calculated for AG,"(ReCli-) in1412 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION I I I I 0.05 0.10 0.15 020 -1oL x2 (mole fraction) Fig.6. Plot of the left-hand side of eqn (1) at 25 "C against mole fraction of propan-2-01 for the solvolysis of 1,2-[Coen2SCNCl]+ in water + propan-2-01. water+methanol are also included in table 5 : these positive values are higher than those found for singly charged negative ions in water + methano1,ll comparable with the situation found in water + The determination of AG,"(Cu2+) assumes43 that the liquid-junction potential Ej x 0 for the cell M1(CIO,), 1 NEt,Pic 1 AgClO, M1 1 0.01 mol dm-3 0.1 mol dm-3 0.01 mol dm-3 1 Ag and, although this appears approximately correct when M, is Ag+ and n = 1 in pure organic liquids, Ej # 0 in pure water.,, It would seem unlikely, therefore, that Ej = 0 when M, is Cu2+ and n = 2 in pure water and its water-rich mixtures with methanol: this doubt is supported by the positive values for AG,"(Cu2+) derived assuming Ej = 0 compared with the expectedllv 42 negative values found above for AG,"(Fe2+).Using the new values found for AG,"(Fe2+) converted to the molar scale, we have repeated the analysis of Blandamer et a1.22 on the substitution of 2,2'-bipyridine into NiEt using the solvent operator 6, and eqn ( 5 ) to find the effect of solvation on the whole of the transition state. The values for S, GTS and 6, GIs relative to pure water in table 5 show that the effect of solvent structure is greater on the whole of the initial state than on the whole of the transition state. This can be compared with our finding2, for the cation alone in water+methanol and, for this substitution, a similar comparison appears to hold for the whole states22 and the cation alonez1 in water+DMSO: for this substitution the effect of solvent structure on the cation dominates in the initial state for other cosolvents mixed with water.21 This appears to be true for the substitution of other neutral ligands into Niii for x2 > 0.03 (with approximately equal effects for NH,), but the reverse effect can occur for the substitution of charged ligands into Nii;.21A.E. EID AND C.F. WELLS 1413 APPLICATION OF THE FREE-ENERGY CYCLE TO THE SOLVOLYSIS OF 1 ,2-[Coen2SCNCl]+ As the evidence discussed above for solvolysis reactions supports the view that the Co-Cl bond for complexes of this type approximates to full extension in the transition state, the free-energy cycle and eqn (1) can be applied here to apportion the effect of solvent structure on the cation between the initial and transition states. Values for logk,/k, at 25 "C were calculated using the values of AH* and AS* in table 3 and the appropriate values for AG:(Cl-) were interpolated from it's variation against mole fraction of propan-2-01.'~ Fig. 6 shows that the variation of the left-hand side of eqn (1) against mole fraction of propan-2-01, x,, is negative. This shows that the influence of solvent structure on the cation is greater in the transition state than in the initial state, comparable with the findings from the kinetics of solvolysis of a range of similar ColI1 complexes in water + cosolvent mixtures3* 5-9 I C.F. Wells, Discuss. Faraday SOC., 1960,29, 219; Trans. Faraday SOC., 1961,57, 1719; J. Chem. SOC., 1962, 3100; Trans. Faraday SOC., 1970, 66, 204. H. P. Bennett0 and E. F. Caldin, J. Chem. SOC. A, 1971, 2191; 2198; 2207; J. Solution Chem., 1973, 2, 217. :' C. F. Wells, J. Chem. Soc., Faraday Trans. I , 1977, 73, 1851. F. Franks and D. J. G. Ives, Q. Rev. Chem. SOC., 1966, 20, 1. C. N. Elgy and C. F. Wells, J. Chem. SOC., Dalton Trans., 1980, 2405; 1982, 1617; J. Chem. Soc., Faraday Trans. I , 1983, 79, 2367. G. S. Groves and C.F. Wells, J. Chem. Soc., Faraday Trans. I , 1982, 78, 619. I. M. Sidahmed and C. F. Wells, J. Chem. SOC., Dalton Trans., 1983, 1035; 1984, 1969. A. E. Eid and C. F. Wells, J. Chem. SOC., Faraday Trans. I , 1981,77, 1621. A. E. Eid and C. F. Wells, J. Chem. Soc., Faraday Trans. 1, 1983, 79, 253. l o W. E. Jones, L. R. Carey and J. W. Swaddle, Can. J . Chem., 1972, 50, 2739; D. A. Palmer and H. Kelm, Coord. Chem. Rezl., 1981, 36, 89. C. F. Wells, J. Chem. SOC., Faraday Trans. I . 1973, 69,984; 1974, 70, 694; 1975,71, 1868; 1976,72, 601; 1978, 74, 636, 1569; Aust. J. Chem., 1983, 36, 1739; G. S. Groves, I. M. Sidahmed and C. F. Wells, unpublished work. l 2 V. D. Panasyuk and A. V. Arkharov, Russ. J. Inorg. Chem., 1968. 13, 406. l 3 A. Werner, Ann., 1912, 386, 140.l 4 K. J . Laidler and H. Eyring, Ann. N . Y . Acad. Sci., 1940, 39, 30; S. Glastone, K. J. Laidler and H. Eyring, The Theory ofRate Processes (McGraw-Hill, New York, 1941), chap. 8; K. J. Laidler and P. A. Landskroener, Trans. Faraday SOC., 1956, 52, 200; K. J. Laidler, Suom. Kemistil. A , 1960, 33, 44; K. J. Laidler, Chemical Kinetics (McGraw-Hill, New York, 2nd edn, 1965), chap. 5. l 5 E. A. Moelwyn-Hughes, Proc. R. SOC. London, Ser. A, 1936, 155, 308; 1936, 157, 667; The Kinetics of Reuctions in Solution (Oxford University Press, London, 2nd edn, 1947), chap. 4, 5 and 7; Physical Chemistry (Pergamon Press, Oxford, 2nd edn, 1961), chap. 7-9 and 24. l 6 E. S. Amis, Kinetics of Chemical Change in Solution (Macmillan, New York, 1949), chap. 5 and 9; Solvent Efleects on Reuction Rates and Mechanisms (Academic Press, New York, 1966), chap.1-3; Soltrent Eflects on Chemical Phenomena (Academic Press, New York, 1973), vol. 1, chap. 5. H. S. Frank and M. W. Evans, J. Chem. Phys., 1945, 13. 507; H. S . Frank and W.-Y. Wen, Discuss. Faraday SOC., 1957, 24, 133; G. Nemethy and H. A. Sheraga, J. Chem. Phys., 1962,36, 3382; 3401; N. Laiden and G. Nemethy, J. Phys. Chem., 1970, 74, 3501. l i K. Nakanishi, Bull. Chem. Soc. Jpn, 1960, 33, 793. l 9 M. J. Blandamer, Introduction to Chemical Ultrasonics (Academic Press, London, 1973), chap. 1 1. 2n M. J. Blandamer and J. Burgess, Coord. Chem. Rev., 1980, 31, 93. 21 C. F. Wells, J. Chem. SOC., Dalton Trans., 1980, 1494; C. N. Elgy and C. F. Wells, J. Chem. Soc., Faraday Trans. I , 1981, 77, 2529; 1983, 79, 2439.M. J. Blandamer, J. Burgess, P. P. Duce, N. Gosal, R. Sherry, P. Guardado and F. Sanchez, Transition Met. Chem., 1984, 9, 3. 23 D. A. Palmer and H. Klem, Inorg. Chem., 1977, 16, 3139. 24 G . Daffner, D. A. Palmer and H. Kelm, Inorg. Chim. Acta, 1982, 61, 57. '' G. A. Lawrance, Inorg. Chim. Acta, 1980, 45, L275; G . A. Lawrance and S. Suvachittanont, Aust. 26 W. G. Jackson and A. M. Sargeson, Inorg. Chem., 1978.17, 1348; W. G. Jackson and C . M. Begbie, J. Chem., 1980, 33, 273. Inorg. Chim. Acta, 1982, 60, 115.1414 SOLVENT EFFECT IN THE KINETICS OF SUBSTITUTION 27 M. Eigen and R. G. Wilkins, Adu. Chem. Ser., 1965,49,55; F. Basolo and R. G. Pearson, Mechanism 28 K. Ishihara, S. Funahashi and M. Tanaka, Inorg. Chem., 1983, 22, 2564. 2s E. F. Caldin, M. W. Grant and B. B. Hasinoff, J . Chem. Soc., Faraday Trans. I , 1972, 68, 2247; M. W. Grant, J. Chem. SOC., Faraday Trans. I , 1973, 69, 560; E. F. Caldin and R. C. Greenwood, J. Chem. SOC., Faraday Trans. I , 1981,77,773; T. Inone, K. Kojima and R. Shimozawa, Inorg. Chem., 1983, 22, 3972. of Inorganic Reactions (Wiley, New York, 2nd edn, 1967), p. 156. 30 H. Vanni and A. E. Merbach, Inorg. Chem., 1979, 18, 2758. 31 C. H. Langford, Inorg. Chem., 1979, 18, 3288. 32 D. R. Stranks, Pure Appl. Chem., 1974, 38, 303. 33 A. F. Wells, Structural Inorganic Chemistry (Oxford University Press, London, 4th edn, 1975), pp. 34 E. Grunwald and S. Winstein, J. Am. Chem. SOC., 1948, 70, 846. 35 M. H. Abraham, Prog. Phys. Org. Chem., 1974, 11, 1. 36 J. E. Leffer and E. Grunwald, Rates and Equilibria of Organic Reactions (Wiley, New York, 1963), 37 M. J. Blandamer and J. Burgess, Pure Appl. Chem., 1983, 55, 55. 38 M. J. Blandamer and J. Burgess, Inorg. Chim. Acta, 1982, 64, L113. 3s G. Biswas, S. Aditya, A. Battacharyya and S. C. Lahiri, J . Indian Chem. SOC., 1977, 54, 1137. 40 J. Burgess, N. Morton and J. C. McGowan, J. Chem. SOC., Dalton Trans., 1977, 1775. 41 D. Sengupta, A. Pal and S. C. Lahiri, J . Chem. SOC., Dalton Trans., 1983, 2685. 42 C. F. Wells, J. Chem. Soc., Faraday Trans. I , 1984, 80, 2445. 43 J. F. Coetzee and W. K. Istone, Anal. Chem., 1980, 52, 53. 44 R. Alexander, A. J. Parker, J. H. Sharp and W. E. Waghorne, J . Am. Chem. SOC., 1972,94, 1148. 45 C. H. Langford, Inorg. Chem., 1964, 3, 228. 46 F. Akhtar and R. A. Begum, J . Bangladesh Acad. Sci., 1978, 2, 9. 47 R. E. Robertson and S. E. Sugamori, J. Am. Chem. SOC., 1969, 91, 7254. 971-972. p. 48 and chap. 8. (PAPER 4/ 1226)

ISSN:0300-9599    

DOI:10.1039/F19858101401

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

.I. Chem. SOC., Faraday Trans. 1, 1985,81, 1415-1424 Kinetics of the Gas-phase Thermal Decompositions of 1 -Methoxy- 1 -methylcyclopropane and cis- and trans- 1 -Met hoxy-2-me thylc yclopropane BY IFTIKHAR A. AWAN AND MICHAEL C . FLOWERS* Department of Chemistry, The University, Southampton SO9 5NH Received 25th July, 1984 In the temperature range 665-737 K the thermal decomposition of 1 -methoxy- 1 - k/s-l = 1014.76*0.81 exp (-252f 10 kJ mol-'/RT). 'The presence of the 1 -methyl substituent destabilises the transition state for reaction. Secondary decomposition of the initially formed isomeric products precludes the determination of their individual rates of formation. Cis- and trans- 1 -methoxy-2-methylcyclopropane undergo first-order, reversible, geometric isomerisation in competition with structural isomerisation to give cis- and trans- 1 -methoxybut- 1 -ene and 1 -methoxy-2-methylpropene in the temperature range 597-689 K: methylcyclopropane follows first-order kinetics with a rate constant given by the equation kcis + t r a n s / ~ - l = 1015.25*0.23 exp (- 235.4 f 2.8 kJ mol-l/RT) ktrans --* cis/s-l = 1014.99*0.80 exp (-233.7 k9.9 kJ mol-l/RT) kcis -rcis.l.methoxybut.l.ene/S-l = 1013.79*0.29 eXp (-233.1 & 3.6 kJ mOl-l/RT) kcis + trans.l.methoxybut.l-ene/S-l = 1013.29k0.66 eXp (- 234.4 f 8.1 kJ mOl-'/RT) kcis -+ 1-methoxy-2-methylpropene /s-l = 1012.4*1.0 exp(-225f 12 kJ rnol-'/RT) ktrans + trans-l-methoxybut-l-ene/s-' '= 1013'0*0'9 exp (- 233 & 1 1 kJ mol-l/RT) ktrans -t l-metho~y-2-methylpropane/~-~ = 1013'5 "*O exp(-235+ 13 kJ mol-l/RT).ktrczns -+ cis-l-methoxybut-l-ene/s-l = 1014'05'0.62 exp (- 243.7 & 7.8 kJ mol-l/RT) On the basis of a biradical mechanism the results provide evidence for the formation of distinguishable biradicals on opening the cis- and trans- 1 -methoxy-2-methylcyclopropane ring. Estimates are made of the relative rates of ring closing, internal rotation and hydrogen-atom transfer of the biradicals. We have previously reported a study of the gas-phase thermal decomposition of methoxycyclopropane. The methoxy group increased the reaction rate and lowered the activation energy for structural isomerisation by ca. 45 kJ mol-1 compared with the value for cyclopropane. The transition state and biradical formed were envisaged to be resonance stabilised through participation of the oxygen-atom lone pair of electrons and to have some resultant ionic character.Structural isomerisation occurred by hydrogen-atom transfer in the biradical. Using scheme 1, the measured rate constant can be equated to k, kh/(k-, + k,). It was thought that for methoxy- cyclopropane k, $ k-,, whereas for alkylcyclopropanes it is believed that k, 4 k-,. Scheme 1. 14151416 DECOMPOSITION OF METHOXYMETHYLCYCLOPROPANE In order to investigate further the mechanism of this reaction we have studied the kinetics of the gas-phase reactions of 1 -methoxy-1 -methylcyclopropane (1 MMC) and cis- and trans- 1 -methoxy-2-methylcyclopropane (cMMC and tMMC). EXPERIMENTAL All thermal kinetic studies were carried out as previously described2 in a conventional 'static' system using Pyrex reaction vessels aged by pyrolysis of ca.20 Torrt hexamethyldisiloxane at 798 K for 24-48 h. Two reaction vessels were employed: one was packed with short lengths of Pyrex tubing to give a surface-to-volume ratio ( S / V ) of ca. 12 cm-l and the other was of similar external dimensions but was not packed ( S / V = 1 cm-l, volume ca. 150 cm3). Young's PTFE barrel, greaseless stopcocks were used in all parts of the vacuum system in contact with pyrolysed material. Analyses were carried out on a Perkin-Elmer 452 gas chromatograph using a 50 m squalane SCOT column at ca. 15 "C with helium carrier gas and flame ionisation detection (f.i.d.). F.i.d. sensitivity coefficients for all compounds were assumed to be directly proportional to carbon number.Peak areas were determined using an LDC 308 computing integrator. Products were identified by gas-chromatographic retention time and by g.l.c./mass spectrometry (Kratos MS 30 mass spectrometer). IMMC, cMMC and tMMC were prepared by the reaction of methylene iodide with the appropriate methoxypropene in the presence of a zinc+opper couple3 and 1,2-dimetho~yethane.~ In the absence of 1,2-dimethoxyethane low yields of the required cyclopropane were obtained. The products were purified, first by fractional distillation on a 0.6 m Podbielniak column and then by preparative gas chromatography. Final product purities were 1 MMC (b.p. 60 "C) 99.7 % (0.2% 2-methoxypropene), cMMC (b.p. 69 "C) 99.4% (0.2% tMMC, 0.3 % cis-l-methoxy- prop-1-ene) and tMMC (b.p.65 "C) 94.4% (0.4% cMMC, 0.2% diethyl ether, 4.9% an unidentified C, vinyl or ally1 ether not formed as a reaction product in kinetic studies). The identities of the products were confirmed by n.m.r. 3-Methoxybut-1-ene was prepared by the action of methyl iodide on the sodium salt of 3-hydroxybut- 1-ene.5 Other methoxypropenes and methoxybutenes used in the study for kinetic, synthetic or identification purposes were prepared by the elimination of methanol from the appropriate acetal.6$ RESULTS 1 -METHOXY- 1 -METHYLCYCLOPROPANE The kinetics of the thermal decomposition of lMMC were investigated in the temperature range 665-737 K at up to 20% decomposition. The results showed that 3-methoxybut-1-ene and (a- and (Z)-2-methoxybut-2-ene, the expected isomeric reaction products, were formed in rather small amounts together with very much larger amounts of other molecules that had clearly been formed as the result of fragmentation reactions.The extent of the fragmentation was such that at 18% decomposition the isomeric products constituted ca. 1.3 % of that total, i.e. they accounted for only ca. 7% of the reactant decomposed. Methane, ethane, propane, but-1-ene and cis- and trans-but-2-ene formed ca. 75 % of the fragmentation products. Separate pyrolyses of 20 Torr 3-methoxybut-1-ene and a mixture of (a- and (Z)-2-methoxybut-2-ene with 2-methoxybut- 1 -ene at 702 K for 3 1 min (this time would have resulted in ca. 25% decomposition of IMMC) produced ca. 95% decomposition of these molecules to mainly hydrocarbon products (methanol, carbon monoxide, water etc.would not have been detected). Scheme 2 summarises the observed reactions. Because of the rapid decomposition of the initial reaction products it was clear that t 1 Torr = 133 N rn+.I . A. AWAN AND M. C. FLOWERS 1417 ---w (E)- and (2)-2-rnethoxybut-2-ene+ - 3-methoxybut-1 -ene fragmentation products Scheme 2. Table 1. First-order rate constants for the thermal decomposition of 1 -methoxy- 1 -methylcyclopropane ~~ 737.1 83 729.1 55 717.2 27 709. 1 18 693.4 5.8 666.1 0.95 649.7 0.37 accurate determination of the rate constants for the formation of these products from lMMC would not be possible and it was therefore decided only to determine the overall rate of loss of IMMC; this proved to be a first-order process on the basis of the time dependence.First-order rate constants were determined at 7 temperatures using 28 Torr initial reactant and they are listed in table 1. In Arrhenius form they are given by the equation k/s-l = 1 014.76 exp ( - 25 1.7 10.4 kJ mol-l/RT). All error limits quoted in this paper are statistical 95% certainty limits. Cis- AND trans- I -METHOXY-2-METHYLCYCLOPROPANE The kinetics of the thermal decomposition of cMMC and tMMC were investigated in the temperature range 597-689 K. The initial reactant underwent a reversible, first-order (on the basis of time and pressure dependence), geometric isomerisation in competition with structural isomerisation and some fragmentation reactions. The structural isomerisation products were identified as cis- and trans- 1 -methoxybut- 1 -ene and 1 -methoxy-2-methylpropene.No 3-methoxy-2-methylpropene or 1 -methoxybut- 2-ene were detected; however, if these allylic ethers were formed in the same ratio to the vinylic ethers as observed from methoxycyclopropanel then, under the conditions used, they would not have been detected. Methane, ethane, propane, butenes and pentenes were identified as the major fragmentation products although they were not formed in quite the same proportions from cMMC and tMMC. The fragmentation products constituted < 9% of the total decomposition products under the conditions used. No structural isomers were detected that would have resulted from the cleavage of the carbon-carbon bond of the cyclopropane ring opposite to the methoxy substituent, i.e.the C(2)-C(3) bond. Scheme 3 summarises the various channels for the reaction of cMMC and tMMC. Variation of the initial pressure of reactant between 5 and 88 Torr at 690 K produced no change in the fractional rates of formation of any isomerisation products. Secondary decomposition of the initially formed structural isomers proved significant at high percentage conversions and hence the majority of runs used to determine rate 47 FAR 18 0 fs Table 2. Rate constants for reactions of cis- and trans- 1 -methoxy-2-methylcyclopropane for 28 Torr initial reactant pressurea 689.8 674.4 657.8 643.3 625.8 613.2 597.3 8 E rate constants/ s-l 7 8 2630 1900 500 140 (4) 38.0 (3) 19.0 (4) 220 45 (3) 25 (3) 40 (1) E 1070 768 200 55 (3) 15.0 (3) 8.50 (4) 83 14.3 (5) 9.5 (2) 20 (2) 2 273 65 17.8 (4) 3.86 (4) 2.52 (1) 24 4.5 (3) 3.0 (2) 7.0 (3) B 140 107 27 8.25 (4) 1.95 ( 5 ) 1.15 ( 5 ) 12 1.8 (3) 1.0 (2) 2.8 (1) 4 39.6 28.9 6.7 2.17 (4) 0.53 (3) - 0.83b 0.41 (3) 1.3b Ei - - 0.096 (1) 2 kl k2 xk1i kl 1 kl2 kl3 xk2i k21 k22 k23 3 50 - 15.9 1 5.6b 2.9 0.89 (5) 0.18 (2) - 1.7 0.22 (2) 0.15 (2) 0.25 (2) 3 4.6 4.9b 0.66 0.24 (2) 0.07 (3) 0.59 0.053 (1) cj r a Weighting factors used for the determination of Arrhenius parameters are shown in parentheses.These data deviated considerably from 8 ? the best straight line through the data points and were not used in determining the quoted Arrhenius parameters. 3I. A. AWAN AND M. C. FLOWERS 1419 - cis-CH3CH2CH=CHOCH3 trans€H3CH2CH=CHOCH3 - other products 4 Scheme 3. Table 3. Arrhenius parameters for the reactions of cis- and trans- 1 -methoxy-2-methylcyclopropane reactant product rate constant log A / s - l EJkJ molP1 cMMC tMMC cMMC cMMC cMMC tMMC tMMC tMMC tMMC cMMC cis-1-methoxybut-1-ene trans- 1 -methoxybut-1 -ene 1 -methoxy-2-methylpropene cis- 1 -methoxybut- 1 -ene trans-1 -methoxybut-1 -ene 1 -methoxy-2-methylpropene 15.25f0.23 14.99 f.0.80 13.79f0.29 13.29 f. 0.66 12.4f 1.0 14.05 f 0.62 13.0 f 0.9 13.5+ 1.0 235.4f2.8 233.7 f 9.9 233.1 & 3.6 234.4 f. 8.1 225 f. 12 243.7 & 7.8 2 3 3 k l l 235k 13 constants were limited to < 15% reaction. At seven temperatures, with 28 Torr initial reactant, estimates of k,+Zk,, and k,+Xk,i were obtained from the slopes of the first-order rate plots for loss of initial reactant. Estimates of Zkli and ZkZi were deduced from the initial rates of formation of all products other than the geometric isomer.These values were then used as the input values for a Simplex non-linear least-squares program,8 which utilised the experimental data from both cMMC and tMMC, as initial reactants at the same temperature, to obtain the best estimates for k,, k,, Zkli and CkZi. Partition of CkIi and CkZi into the values for the individual rate constants was carried out using the initial slopes of plots of percentage of the relevant reaction product against percentage total reaction. The rate constants so determined are listed in table 2 and the Arrhenius parameters derived from them are reported in table 3. The Arrhenius parameters quoted for the structural isomerisation used a weighted least-squares procedure in which the weighting factors were based on the number of data points available to determine the separate rate constants.Some runs were undertaken at 657, 674 and 689 K in an attempt to determine directly the equilibrium constant for the geometric isomerisation, by pyrolysing cMMC for long times. The ratio of tMMC to cMMC passed through a rather flat maximum of ca. 1.60 at ca. 40% decomposition of cMMC+ tMMC to structural isomers and fragmentation products, instead of reaching a constant value. It seems likely that a secondary reaction product was contributing to the cMMC peak at high percentage reaction and so further attempts to determine the equilibrium constant directly were not made and these data were not used in the determination of the quoted rate constants.47-21420 DECOMPOSITION OF METHOXYMETHYLCYCLOPROPANE Studies of the rates of reaction of cMMC at 675 K in the packed reaction vessel gave results identical to those that had been obtained in the unpacked vessel. Because of lack of reactant only one run was carried out in the presence of nitric oxide. Addition of 10% nitric oxide to cMMC at 690 K did not affect the rate of geometric isomerisation but reduced the rates of formation of fragmentation products by ca. 50% and also slightly reduced the rates of formation of structural isomerisation products. It is thus possible that our measured rates for structural isomerisation do have some contribution from radical pathways, although it is unlikely that this contribution is significant.DISCUSSION A biradical mechanism has been used to explain the thermal unimolecular rearrangements of cyclopropanes and this mechanism proved capable of explaining the reaction products and the observed parameters for the thermal decomposition of methoxycyclopropane. The studies reported here on cMMC and tMMC indicate that the major processes are homogeneous and probably unimolecular and so also might be expected to proceed via a similar mechanism. Data for lMMC indicate a first-order reaction but the measured rate constants can only be considered to represent the maximum values for the unimolecular structural isomerisation. Table 4 lists the rate constants for structural isomerisation of some cyclopropanes bearing alkyl and/or methoxy substituents.For alkyl-substituted cyclopropanes the reaction rate increases slightly from cyclopropane to methylcyclopropane to 1,l- dimethylcyclopropane, whereas comparison of the total reaction rates for methoxy- cyclopropane and lMMC shows that the addition of the 1-methyl substituent reduces the reaction rate significantly. This effect is entirely in accord with the biradical mechanism above, as the consequence of attaching the electron-donating methyl group to the partially negatively charged carbon atom is to reduce the resonance stabilisation caused by the methoxy group in the transition state for hydrogen-atom transfer and hence to slow the reaction rate. The observationg that the activation energy for the isomerisation of 1 -methoxy- 1 -vinylcyclopropane to 1 -methoxycyclopentene is only ca.2 1 kJ mob1 below the activation energy for the corresponding reaction of vinylcyclo- propane, whereas the reduction in activation energy is ca. 46 kJ mol-l for the isomerisation of 1 -methoxy-2-vinylcyclopropane, is presumably the consequence of similar counter-stabilising effects of the vinyl and methoxy groups when they are attached to the same carbon atom. Analogously, in studies of the differences in the heats of formation of various vinyl ethers, Taskinen7v10 has observed that a methyl group p to the methoxy group has a less stabilising effect by ca. 7 kJ mol-1 than is observed in the equivalent alkenes. Comparison of the rates of structural isomerisation of 1,2-disubstituted cyclopro- panes given in table 4 shows that the presence of a methyl group at the 2 position has very similar effects in most instances for both methyl- and methoxy-substituted cyclopropanes. In both systems the rates of geometric isomerisation are much faster than the rates of structural isomerisation and the A factors for geometric isomerisation are identical within experimental error.It is clear therefore that our previous contention1 that k, 3 k-, is not tenable for 1 -methoxy-2-methylcyclopropane (MMC) and estimates of k-,/k, are made in the Appendix. Thus the stabilisation energies for the methoxy substituent in our earlier paper refer to the transition states for hydrogen-atom transfer rather than to the initial ring opening. Comparison of the Arrhenius parameters for geometric isomerisation of MMC and 1,2-dimethylcyclo-Table 4.Rate constants at 665 K for the gas-phase thermal decomposition of some substituted cyclopropanes [C( 1)-C(2) or C( 1)-C(3) bond fission as marked] c yclopropanes met hoxycyclopropanes 7 k(s truc tural k(structura1 ? isomerisation) 9 isomerisation) k( tot al) reactant / 10-6 s-' ref. reactant /lo+ s / 10-6 s-1 ref. 5 z a % 41 33 1 W 10 - this work F s s E do"' VOMe Id v 0.8 d 1.2 b 1.3, C( 1)-C(2) C e: 110 39, C( 1)-C(2) this work P r 5, C(l)-C(3) 3 .O 0.8, C( 1)-C(3) 0.7, C( 1)-C(2) Id 0.8, C( 1)-C(3) this work 14, C( 1)-C(2) 1 1 , C(l)-C(3) 47 d d If v a W. E. Falconer, T. F. Hunter and A. F. Trotman-Dickenson, J . Chern. SOC., 1961, 609. Phys., 1965, 69, 2141; D. W. Setser and B. S. Rabinovitch, J . Am. Chern. SOC., 1964, 86, 564.1959, 3953. D. W. Placzek and B. S . Rabinovitch, J. Chern. M. C. Flowers and H. M. Frey, J . Chern. Soc., M. C. Flowers and H. M. Frey, Proc. R. SOC. London, Ser. A , 1960, 257, 122; 1961, 260; 424.1422 DECOMPOSITION OF METHOXYMETHYLCYCLOPROPANE propane suggests that stabilisation energy from the methoxy substituent in respect of the ring-opening reaction may be no more than 13 kJ mol-l. The error limits of the Arrhenius parameters do not allow an assessment of whether or not it is energetically easier to break the C(l)-C(2) or the C(1)-C(3) bond of MMC. However, it is clear that the rates of formation of products resulting from C( 1)-C(2) rupture is greater than for C( 1)-C(3) rupture and the rates are greatest from the cis isomer; these observations are all also true for 1,2-dimethylcyclopropane.Estimates of the relative rates of breaking C( 1)-C(2) and C( 1)-C(3) bonds can be obtained using the detailed mechanism shown in the Appendix. For C( 1)-C(3) bond rupture tMMC reacts faster than cMMC, in contrast to 1,2-dimethylcyclopropane for which the rates are the same from both isomers. This difference is presumably a consequence of the details of the relative rates of all the elementary reactions involved, i.e. the relative rates of ring opening and ring closing, internal rotation and hydrogen-atom transfers in the biradicals. Lustgarten and Richey,ll in a study of the isomerisation of 7-alkoxybicyclo[2. 1 . 13- heptadienes to cycloheptatrienes, also found that the alkoxy group had large accelerating effects on the reaction rates.They considered that the most likely reaction path involved a biradical and considered a zwitterionic intermediate unlikely in the light of a small solvent effect. Although we do find evidence, from the effect of a 1 -methyl substituent, to support a partial negative charge residing on the radical centre immediately bonded to the oxygen, the lack of effect of the 2-methyl substituent also suggests that the other radical centre is not charged. Kirmse and Zeppenfeld12 determined the rate constants and Arrhenius parameters for the interconversion of the geometric isomers of I ,2-dimethoxy-3-methyIcyclopro- pane and a rate constant of 3.6 x s-l at 288 "C for the conversion of cis- 1 -methoxy-cis-2,3-dimethylcyclopropane into its geometric isomers.At 288 "C we calculate that the rate constant for conversion of cMMC into tMMC would have been 2.3 x s-l, i.e. essentially the same value bearing in mind thecombined uncertainties. Kirmse and Zeppenfeld deduced from their data that there was no preference for a synchronous double rotation, rather than a single rotation, mechanism. Our data also have information pertinent to this as a consequence of an important difference that exists between reactions of 1,2-dimethylcyclopropane and MMC. The ratios of the rates of formation of 2-methylbut-2-ene to 2-methylbut- 1 -ene and of cis-pent-2-ene to trans-pent-2-ene are essentially the same from both cis- and trans- 1,2-dimethylcyclopropane. Thus for this reaction no conclusions can be drawn as to whether or not distinct and different biradicals are formed when opening the cyclopropane ring of the two different geometric isomers.However, in the current study the ratio of cis- to trans-1-methoxybut-1-ene is ca. 4.1 from cMMC and ca. 1.6 from tMMC. Hence if a biradical mechanism is assumed, different biradicals must be formed from the two isomers on breaking the C(l)-C(2) bond. A concerted, conrotatory (or disrotatory) ring-opening reaction would be consistent with this observation, though it is not a requirement. Concerted ring opening and ring closing would necessitate that the geometric isomerisation occurred as the consequence of breaking the C( 1)-C(3) bond as no geometric isomerisation can occur for a concerted opening at the C(1)-C(2) bond. Thus the geometric isomerisation would be in competition with isomerisation to 1 -methoxy-2-methylpropene.However, based on the work of Kirmse and Zeppenfeld12 and on other cyclopropane studies (e.g. the geometric isomerisation and racemisation of 1 -ethyl-2-methylcyclopropane13), we feel that the predominant ring-opening reaction is not concerted and that the geometric isomerisation is a consequence of a biradical mechanism in which either the C( 1)-C(2) or the C(l)-C(3) bond has been ruptured. From our data, estimates of the variousI. A. AWAN AND M. C. FLOWERS 1423 elementary processes involved may then be obtained based on some reasonable assumptions (see the Appendix). The general conclusions are that (a) the cMMC ring opens at C( 1)-C(2) up to three times faster than tMMC, (b) hydrogen-atom transfer in the biradicals is significantly slower (> 30 times) than either internal rotation or ring closure and (c) the ratio of kcyclisation to kinternal rotation is ca. 1.These are very similar conclusions to those that may be drawn from alkylcyclopropane studies13 with perhaps a slight reduction in the rates of internal rotation relative to cyclisation (cf. key = 0.2 for 1 -ethyl-2-methylcyclopropane13) from that expected, bearing in mind the similar relative size of the rotors. We thank Prof. H. M. Frey for the Simplex computer programme used to derive the rate constants from the experimental data and I. A. A. thanks the Government of Pakistan for the award of a scholarship. APPENDIX Scheme 4 represents the detailed biradical mechanism for the geometric isomerisation of, and for the formation of cis- and trans-1-methoxybut-1-ene from, M M C .All rate constants with lower case subscripts are rate constants for elementary processes. Rate constants k , and k , represent the overall rates of geometric isomerisation via C( 1)-C(3) bond rupture. Biradicals A and B are formed by C( 1)-C(2) bond rupture in cMMC and t M M C , respectively. “ biradical A cis-1-methoxybut-1-ene cMMC% k:I k* 1 tkB k: krotl IkIOtX tMMC 7 biradical B kht * trans- 1-methoxybut- 1-ene k-I Scheme 4. In order to derive expressions for the relative rate constants for the various elementary processes occurring, an initial simplifying assumption is necessary. We have chosen to assume that the rates of internal rotation of the two biradicals are the same.Using this, and the steady-state assumption for biradical concentrations, leads to the following equations for relative rate constants in terms of the measured rate constants defined in scheme 3.1424 DECOMPOSITION OF METHOXYMETHYLCYCLOPROPANE Table 5. Calculated rate-constant ratios based on rate-constant data at 643 K {! assumed values 0 0 0.2 0.2 0.2 0 0.2 0 0 0.2 0.2 1 0.2 0 0 1 0 1 1 1 1 kFr/kgc 57 28 63 32 73 33 28 k t r 1% 76 38 374 188 156 38 195 kgclkkt 2.8 2.8 5.9 6.0 4.1 2.6 6.4 k5-lkrot 0.9 0.9 2.3 2.3 5.4 1 .o 2.1 k t r lkrot 0.4 0.4 2.3 2.3 2.8 0.5 2.2 k,clkt, 2.8 2.8 1.3 I .3 2.5 3.0 1.2 6 represents the rate of geometric isomerisation via C(l)-C(3) bond rupture relative to a rate via C( 1)-C(2) bond rupture equal to unity. Using rate constants calculated at 643 K from the Arrhenius parameters, the relative rate constants for the elementary processes are given in table 5 for assumed values of a, p and 6. For consistency with the experimental data a must be < 0.25 and p < 1.45. I. A. Awan and M. C . Flowers, J . Chem. SOC., Faraday Trans. I, 1983, 79, 1413. M. C. Flowers and T. h i i r k , J . Chem. SOC., Faraday Trans. I , 1975, 71, 1509. E. Le Goff, J . Org. Chem., 1964, 29, 2048. H. E. Simmons, E. P. Blanchard and R. D. Smith, J. Am. Chem. SOC., 1964, 86, 1347. K. Wiberg, J. Am. Chem. SOC., 1952,74, 3891. ti W. L. Howard, E. C . Jacobson and R. A. Newton, J. Org. Chem., 1961, 26, 3574. ’ E. Taskinen, J. Chem. Thermodyn., 1974,6, 345. J. A. Nelder and R. Mead, Comput. J., 1964-65, 7, 308; M. J. Box, Comput. J., 1965-66, 8, 42. J. M. Simpson and H. G. Richey Jr, Tetrahedron Lett., 1973, 2545. R. K. Lustgarten and H. G. Richey Jr, J. Chem. SOC., 1974, 96, 6393. lo E. Taskinen, J, Chem. Thermodyn., 1973, 5, 783. l2 W. Kirmse and M. Zeppenfeld, J . Chem. Soc., Chem. Commun., 1977, 124. l3 W. L. Carter and R. G. Bergman, J. Am. Chem. SOC., 1968, 90, 7344; 1969,91, 741 1 . (PAPER 4/ 1304)

ISSN:0300-9599    

DOI:10.1039/F19858101415

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

S. Chem. SOC., Faraday Trans. I, 1985, 81, 1425-1434 Broad Single-minimum Proton Potential and Proton Polarizability of the Hydrogen Bonds in Trifluoroacetic Acid + Pyridine-N-Oxide Systems as a Function of Donor and Acceptor Properties and Environment Infrared Studies BY ULRICH B~HNER AND GEORG ZUNDEL* Institut fur Physikalische Chemie der Universitat Miinchen, Theresienstrasse 41, D-8000 Miinchen 2, Federal Republic of Germany Received 26th July, 1984 Trifluoroacetic acid + pyridine-N-oxide (1 : 1) systems have been studied in CC1, and aceto- nitrile as a function of the basicity of the N-oxide. In CC1, acid-base complexes are formed almost completely, whereas in acetonitrile the degree of complex formation increases with increasing basicity of the N-oxide, i.e. with increasing ApK, (pK, of the protonated N-oxide minus pK, of the acid).Log& increases in proportion to ApK,. In the O-...H+...ON bonds a single-minimum proton potential is present, and if the degree of symmetry of this potential is sufficiently large continua indicate that these hydrogen bonds show a large proton polarizability caused by the fluctuating proton. From the shift of v(C=O) and from the intensity of the continuum, both as a function of ApK,, the following conclusion is drawn: the broad single-minimum potential well shifts with increasing basicity of the N-oxide, i.e. with increasing ApK, from the donor to the acceptor. This shift is larger in acetonitrile solutions than in CCl, solutions since in the polar solvent the interaction of the hydrogen-bond dipole with the reaction field induced by it in the solvent is larger.Heteroconjugated hydrogen bonds AH---B (I) eA-..-H+B (11), if present in the gas phase, usually have asymmetrical proton potentials with a deep single minimum at the donor, i.e. under gas-phase conditions the equilibrium is completely shifted to the In solution the interaction of the dipole of these hydrogen bonds with the reaction field induced by it in the ~ o l v e n t ~ - ~ together with other specific interactions6 change the proton potential fundamentally and shift this equilibrium in favour of the polar proton-limiting structure(I1). Thus, in the case of long hydrogen bonds these interactions may cause double-minimum proton potentials to occur. This is the case if the pKa values of the donor, AH, and of the protonated acceptor, BH+, are relatively high,7y whereas short hydrogen bonds with broad single-minimum potentials may be realized if the pKa values are l ~ w .~ - l l Hydrogen bonds in which the proton may fluctuate show a large proton polarizability caused by proton motion.l2? l3 Owing to these large polarizabilities the hydrogen bonds interact strongly with their environment. As a result of these interaction effects continuous absorptions are observed in the i.r. spectra.7$ 8* 1 3 9 l4 In this studyO-...H+---ON hydrogen bondsin trifluoroaceticacid (TFA)+pyridine- N-oxide (R-PyNO where R is a substituent) systems (with CC1, and acetonitrile as solvents) have been examined. With regard to their pK, values short hydrogen bonds are expected.This type of hydrogen bond has been studied by various a u t h o r ~ l ~ - ~ ~ 14251426 I.R. STUDIES OF HYDROGEN BONDS using different methods (u.v., i.r. and n.m.r. spectroscopies and vapour-pressure osmometry). HadZi et aZ.l1? 1 5 7 l6 have shown that these hydrogen bonds are short, with single-minimum proton potentials. In solution not only the above-mentioned hydrogen-bonded complex but also other species may be present [see reaction (2) later]. These equilibria have been studied extensively as a function of various parameters by Szafran’s In a recent p ~ b l i c a t i o n ~ ~ these authors also concluded that preferentially 1 : 1 complexes with single-minimum proton potentials are present. The same conclusion was drawn from U.V. and i.r. measurements by Kreevoy’s 28 However, Dega-Szafran and Szafran26 have shown in a study of the osmotic coefficients that in benzene these 1 : 1 complexes associate strongly.Brycki and S ~ a f r a n ~ ~ studied such systems in chloroform. They found that the n.m.r. signal of this hydrogen-bonded proton is shifted strongly toward lower fields and shows a sharp maximum as a function of ApK, (ie. pK, of the protonated N-oxide minus pK, of the acid). The position of the centre of gravity of the intensity of the continuum25 shows an analogous shift as a function of ApK, and has a sharp maximum. In the case of lH n.m.r. spectroscopy they found the maximum at ApK, = 1.13; in the case of i.r. spectroscopy the maximum was at ApK, = 1.17. From this result Brycki and S ~ a f r a n ~ ~ concluded that these hydrogen bonds are strongest if they are on average symmetrical.EXPERIMENTAL The materials were purchased from Fluka (Switzerland), EGA and Merck (Federal Republic of Germany) and were of the highest degree of purity available. Acetonitrile, [2H,]acetonitrile and carbon tetrachloride (for spectroscopy) from Merck were used and dried over 3 A molecular sieves. Trifluoroacetic acid was dried with P,O, and then distilled twice under reduced pressure in a nitrogen atmosphere. The solid pyridine-N-oxides were dried over P,O, at ca. 30 K below their melting points under reduced pressure. Substances with minor impurities were sublimed at lo-, Torr. 2-Chloropyridine-N-oxide is commercially available as its hydrochloride (2-C1-PyNO.HC1). The free base was prepared by reaction with a concentrated Na,CO, solution, extracted with CHCl,, recrystallized from dry benzene and dried in a vacuum.4-CO2CH,-PyNO was prepared from isonicotinic acid-N-oxide by methylation with CH,OH and SOCl,.29 All preparations and transfers of solutions were carried out in a water-free glove-box under a nitrogen atmosphere. For the i.r. investigations cells with silicon windows were used. Because of the high reflectivity of this material, wedge-shaped layers were applied to avoid an interference pattern superposed on the spectra.’ The mean layer thickness was found to be 96 or 528 pm. The solvent bands were compensated for by an adjustable layer of solvent in the reference beam. In the regions of the strongest solvent bands the energy loss was too high to obtain any information.In the spectra these regions are indicated as dotted lines. The spectra were plotted with a model 325 spectrophotometer from Bodenseewerk Perkin-Elmer, Uberlingen, Federal Republic of Germany. It was flushed with dry and C0,-free air and the sample temperature was 298 K. The absorbances obtained from the spectra were corrected as outlined in ref. (7), taking into account the fact that with wedge-shaped layers the Lambert-Beer law is no longer valid. With the aid of this corrected scale the integral absorbances of the bands and the absorbance of the continua used for the evaluations were determined. The association equilibrium constants Ka were determined by evaluating the carbonyl bands corresponding to the donor and to the hydrogen-bonded complex as described in the next section.The pKa values were taken from ref. (30)-(33). The absorbance of the continuum was evaluated at 900 cm-l in CH,CN and at 1060 cm-l in CCl,, ie. in regions where no bands are present. The absorptivity values, E , were referred to 1 mol dmp3 of absorbing hydrogen bonds in a reference layer of 1 cm thickness.u. BOHNER AND G. ZUNDEL 1427 RESULTS AND DISCUSSION Examples of i.r. spectra of TFA + R-PyNO (1 : 1) systems in CC1, are shown in fig. 1 ( a ) and (b); those of the respective systems in acetonitrile are shown in fig. 1 (ct(e). In fig. 1 (a) and (c) the spectra of the pure donor solutions are shown for comparison and in all parts of fig. 1 the spectra of the pure acceptor solutions are also given.To obtain information on the hydrogen bonds in these acid-base complexes it is necessary to investigate the appearance of all species which could be present in solution and measure their concentrations. Hence the following equilibria must be taken into account: Ka Kd OH + ON O-*.*H+**.ON 0- + H+ON Kh [OH. * -0- -0- * - HO] + [NO+H * *ON NO. - -H+ON]. (2) OH+NO To estimate whether charged species are present, the conductivities were measured. In the case of the CCl, solutions no noticeable change in conductivity occurs in comparison with the pure solvent. The conductivity values of the acetonitrile solutions are given in table 1 . In the case of these solutions the concentrations of the various charged species were estimated from the conductivity measurements. For this purpose spectra of solutions were obtained in which the different types of pure charged species were present; examples are given in fig. 2.In the case of the acetonitrile solutions the specific conductivities increase by up to two orders of magnitude for the acid-base solutions in comparison with solutions of the pure components. This increase in conductivity may be caused either by free anions and cations or by complexes formed by homoconjugated hydrogen bonds. The dashed line in fig. 2 shows that v,,(CO,) of free trifluoroacetate ions is observed at 1694 cm-l. This band is not observed in the spectra of the acid-base solutions shown in fig. 1. The solid line in fig. 2 shows that an intense band characteristic of the homoconjugated TFA- trifluoroacetate group is found at 1745 cm-l.No indication of such a band is found in the spectra of fig. 1 . Thus the concentration of the charged species causing the conductivity is below our spectroscopic accuracy. Nevertheless, to obtain information on the amount of charged species present conductivity measurements and i.r. spectra of solutions with an excess of base were obtained. In the system TFA + 4-CH3-PyN0 (1 : 5 ) in acetonitrile (acid concentration 0.1 mol dm-,) a conductivity of 0.59 x lo-, R-l cm-l was found. The trifluoroacetate band at 1694 cm-l yields an amount of free anions that is ca. 9% of the initial acid concentration. A comparison of this conductivity value with those in table 1 shows that this value is one order of magnitude larger than in the 1 : 1 acid-base systems.This demonstrates that < 2% charged species are present in the 1 : 1 systems, and thus can be neglected. FORMATION OF THE HYDROGEN-BONDED COMPLEX The spectra of pure TFA solution in CC1, and in acetonitrile are shown by the dot-dashed lines in fig. l(a) and (c), respectively. In CCl, v(0H) of the monomeric acid is at 3508 cm-l and v(0H) of the dimeric acid is observed as a broad band with several maxima in the range 320&2900 cm- l. In acetonitrile only one broad absorption is found with a maximum at cu. 3000 cm-l due to v(0H) of the acid-acetonitrile hydrogen bonds. The v(C=O) bands of the acid in CCl, are found at 1805 cm-l (monomer) and 1789 cm-l (dimer), whereas in acetonitrile only one v(C=O) band is found at 1792 cm-'.1428 1.R.STUDIES OF HYDROGEN BONDS 0.7 1.0 - 0.2 0.4- 0.7 - 1.0 - - 1 . 5 - r t I I I I ' I ' I r - - - - 0.01 I I I I I I 1 I I I 1 I I I I I I 1 I I I 1 I I 1 1 1 . 1 0.2 0.4 0.7 0.4 - 0.7 - ('1 1.0 - ;. 0 ~ k 0 0 3 5 0 0 3000 2500 2000 1800 1600 1400 1200 1000 800 600 400 200 I I I l I 1 1 1 - 0.01 I I I 1 I I I I I I I I I I 1 I 1 1 1 I I I -.h 0.2 0.4 - - (elu . BOHNER AND G. ZUNDEL 1429 Table 1. Specific conductivities li R-' cm-l) at 298 K of the systems 0.1 mol dm-, TFA + R-PyNO in CH,CNav acceptor: R 1, L H + B 4-N02 2-c1 4-C02CH3 H 4-C6H5 3-CH3 4-CH3 4-CH30 2,6-(CH,), 1.32 1.68 1.64 1.54 1.70 1.37 1.36 5.48 1.77 7.5 36.0 61.1 90.2 82.5 95.0 94.5 83.5 83.7 a Specific conductivity of pure CH,CN = 0.5 x R-l cm-l. Donor: CF,COOH, I,, = 2.75 x lop6 Q-l cm-l.O . O . I I 1 I l l l l l 1 1 1 I l l l l l l 1 1 - C. 0.7 - 1.0 - - 1 . 5 - ~ 1 I I I I 1 I 1 1 1 I 1 I 1 I 1 1 1 1 1 I GOOO~SOO 3ooaz500 2000 iaoo 1600 1400 1200 ioao 800 600 w avenum ber/cm -' Fig. 2. Infrared spectra of (-.-.-) CF,COOH, (---) Bu,N+CF,CO, and (-) Bu,N+ [(CF,CO,),H]- in CD,CN; concentration 0.1 mol drn-,, layer thickness 96 pm. In the CCl, solutions of the 1 : 1 mixtures no bands of the free acid are found. Thus almost complete complex formation occurs in this system, i.e. K, is very high. In acetonitrile solutions of the 1 : I mixtures the amount of free acid has been calculated from the v(C-0) band at 1792 cm-l found in all spectra of the mixtures, at least as a weak shoulder. For calibration the band of the pure acid solution was used. From these values and the initial concentrations the association constants, K,, are obtained and are given in table 2.In fig. 3 log K, is plotted as a function of ApK,. Log K , increases almost in proportion to ApK,. Analogous linear relations between K, and ApK, have been found with other families of systems.$ THE 0 - * H+* * *ON HYDROGEN BOND IN THE ACID-BASE COMPLEXES In the i.r. spectra (fig. 1) of solutions of the 1 : 1 acid-base mixtures only one band is observed in the carbonykarboxylate stretching-vibration region which is caused by the acid-base complex. This is true for both the CCl, and acetonitrile solutions. The values of the absorption maxima of this band are given in columns 3 and 4 of table 3. The result that only one band is observed demonstrates that with all systems in both solvents a single-minimum proton potential is present in these hydrogen bonds.1430 I.R.STUDIES OF EIYDROGEN BONDS Table 2. Association data of the systems 0.1 mol dmP3 TFA + R-PyNO in CH,CN at 298 Ka 4-N02-PyN0 2-Cl-Py NO 4-C0,CH3-PyN0 4-C6H5-PyN0 3-CH3-PyN0 4-CH3-PyN0 2,6-(CH3),-PyN0 4-CH30-PyN0 PyNO - 1.75b -0.81b -0.41' 0.69d 0.83b 0.92d 1 .26d 1 .37d 2.05b - 2.27 - 1.33 - 0.93 0.17 0.3 1 0.40 0.74 0.85 1.53 2 11 16 21 1 187 205 317 3.76 5.08 a Donor: CF,COOH, pK, = 0.52 [ref. (30)]. The pK, values were taken from: ref. (31); ref. (32); ref. (33). - 3 - 2 - 1 0 1 2 3 A PK, Fig. 3. Log K, plotted against ApK, for the systems TFA + R-PyNO in CH,CN (concentration 0.1 mol dm-,): 0, experimental values; (-) best fit. R = (1) 4-N02, (2) 2-C1, (3) 4-C0,CH3, (4) H, (5) 4-C,H5, (6) 3-CH3, (7) 4-CH3, (8) 2,6-(CH,), and (9) 4-CH30.In the pure acid solution v(C=O) is found in the case of acetonitrile at 1792 cm-l. In the case of the CCl, solution this band is found at 1805 cm-l (monomer) and at 1789 cm-l (dimer). Table 3 shows that this band shifts with increasing ApK, from 1770 to 1755 cm-l in the acetonitrile solutions and from 1780 to 1768 cm-I in CCI, solutions. This result demonstrates that the proton is less strongly attached to the carboxylic group with increasing ApK,, i.e. the single-minimum well for the hydrogen bond shifts from the hydrogen-bond donor in the direction of the acceptor site because of the increasing basicity of the acceptors. Hence within this series of compounds the dipole of the hydrogen bond increases.This shift is stronger in acetonitrile than in CC1, (table 3, columns 3 and 4). This is caused by the stronger interaction of the dipole of the hydrogen bond with theU. BOHNER AND G . ZUNDEL 1431 Table 3. Absorption maxima of v(C=O) of the hydrogen-bonded complex in the systems TFA + R-PyNO v(C=O)/cm-l system CC1, CH,CN APK, solution solution TFA + 4-N02-PyN0 TFA + 2-C1-PyNO TFA + 4-C02CH3-PyN0 TFA + 4-C6H5-PyN0 TFA + 3-CH3-PyN0 TFA + 4-CH3-PyN0 TFA+2,6-(CH3),-PyN0 TFA + 4-CH30-PyN0 TFA + PyNO - 2.27 - 1.33 - 0.93 0.17 0.3 1 0.40 0.74 0.85 1.53 - - 1780 1770 - 1770 1776 1762 - 1762 1770 1762 - 1757 1768 1760 - 1755 reaction field induced by this dipole in the solvent., This reaction field is determined by the permittivity E ; it increases with the Onsager parameter (E - 1)/(2~+ 1) and hence is much stronger with acetonitrile than with cc1,.34 The enthalpy of this interaction is negative and relatively large [ref.(6) and references therein], and thus this interaction changes the proton potential considerably. Hence this interaction effect loosens the proton from the donor, shifting the well of the proton potential from the donor in the direction of the acceptor, and in addition the well is lowered. CONTINUA AND PROTON POLARIZABILITY OF THESE HYDROGEN BONDS In the spectra of the 1 : 1 complexes in fig. 1 continua are observed, demonstrating that these hydrogen bonds show a large proton polarizability due to the proton hctuation in the broad single-minimum potentia1.12-14 Fig. 1 (c)-(e) show that with the acetonitrile solutions these continua are observed in the region 1600-700 cm-l and are most intense at ca.900 cm-l. This wavenumber-dependent intensity distribution of the continua is characteristic of short hydrogen bonds with a single minimum showing large proton polarizabilities.l49 35* 36 Fig. 1 (a) and (b) show that in the case of CC1, solutions the maximum in the intensity of the continuum is at slightly higher wavenumbers, and in addition in the 2750-2250 cm-l region a band-like stucture is observed. This wavenumber-dependent intensity distribution is characteristic of slightly longer hydrogen bonds showing large proton polarizability. The band-like structure in the region 275&2250 cm-l is probably caused by the 0-2 proton transition [see fig.5 in ref. (3711. This difference in the continua observed in acetonitrile and in CCl, solutions may be explained as follows. The reaction field at the O--.-H+-..ON bonds is stronger with acetonitrile ( E = 37.5) than with CCl, (E = 2.24). As already discussed, the enthalpy of this interaction effect is negative and much larger with acetonitrile than with CCl,. Hence in the case of the acetonitrile solutions the proton potential well becomes deeper owing to the influence of this interaction effect, and hence comparable hydrogen bonds are slightly shorter in acetonitrile than in CCl,. In fig. 4 the absorbance of the continua is plotted as a function of ApK,. In the case of the acetonitrile solutions the intensity of the continua increases and then decreases with increasing ApK,, i.e.the intensity of the continua shows a broad flat1432 I.R. STUDIES OF HYDROGEN BONDS 5 A PK, Fig. 4. Absorptivity of the continuous absorption in the systems TFA + R-PyNO plotted as a function of ApK,: (a) at 900 cm-l for the systems in CH,CN; (b) at 1060 cm-l for the systems in CCl,. The numbers on the points refer to the same systems as in fig. 3. maximum at ApK, z 0.5. Thus the proton polarizability of the O-...H+*..ON bonds is largest at approximately this ApK, value. With the CCl, solutions the intensity of the continua increases with the systems that could be studied (solubility), but this increase in intensity occurs at higher ApK, values than for the acetonitrile solutions. This result is understandable since the influence of the reaction field on the proton potential is much weaker with CCl, and hence it does not increase the polarity of the hydrogen bond as strongly as is the case for acetonitrile.This result is consistent with the trend in the shift of the v(C=O) band. As discussed earlier, this band shifts less strongly toward smaller wavenumbers in the case of CCl, as compared with acetonitrile solutions. In chloroform the largest lH chemical shift of the hydrogen-bonded proton and the largest shift of the centre of gravity of the continuum intensity are found at ApK, z 1.1 5.25 A comparison of this value with the value ApK, z 0.5 for the systems in acetonitrile solution shows that the interaction of the hydrogen-bond dipole with the reaction field is stronger with acetonitrile than with the chloroform, as expected from inspection of the permittivity constants (for chloroform E = 4.81, while for acetonitrile E = 37.5).However, in addition to the non-specific interaction chloroform causes a specific interaction effect,6 since it forms a hydrogen bond to the second oxygen atom of the carboxylic acid group. Our results show that in acetonitrile the non-specific interaction of the hydrogen bond with the reaction field also over- compensates for the additional specific interaction effect of the complexes with chloroform molecules. CONCLUSIONS In the case of trifluoroacetic acid + pyridine-N-oxide (1 : 1) systems in CCl, almost complete formation of acid-base complexes occurs, whereas in acetonitrile the degree of complex formation increases with increasing basicity of.the N-oxide. Log K,U. BOHN'ER AND G. ZUNDEL 1433 increases in proportion to ApK,, i.e. the pK, value of the protonated base minus that of the acid. Only one v(C=O) vibration is observed in the carbonyl-carboxylate stretching- vibration region, demonstrating that a single-minimum proton potential is present in the O-..-H+...ON hydrogen bonds. However, v(C=O) shifts with increasing basicity of the N-oxide toward smaller wavenumbers, indicating that the proton is loosened from the acid. This effect is more pronounced with acetonitrile than with CCl, solutions. This occurs because the interaction of the dipole of the hydrogen bond with the reaction field induced by it in the solvent is much stronger in acetonitrile than in Intense continua indicate that these hydrogen bonds may show a large proton polarizability.In the case of acetonitrile solutions this polarizability is largest at ApK, z 0.5, indicating that the proton potential is on average symmetrical at this ApK, value. In the case of the systems in CCl, the intensity of the continuum increases at slightly higher ApK, values with increasing ApK, owing to the weaker interaction of the dipole of the hydrogen bond with the reaction field induced by it in the solvent. In summary, all these results show that the minimum of the broad single-minimum proton potential well shifts from the donor to the acceptor; this shift increases with increasing ApK,, i.e. the greater the basicity of the N-oxide. Furthermore, this shift is larger in more polar solvents since under this condition the interaction of the dipole of the hydrogen bond with the reaction field induced by it in the solvent is larger.Thus, the ApK, value and the polarity of the solvent determine the position of the minimum of the proton potential in the O-...H+-..ON bonds formed. CCl,. We thank the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie for providing facilities for this work. A. I. Kulbida and V. M. Schreiber, J . Mol. Struct., 1978, 47, 323. N. S. Golubev and G. S. Denisov, Chem. Phys. (USSR), 1982, 563. A. Beyer, A. Karpfen and P. Schuster, in Topics in Current Chemistry, ed. F. L. Boschke (Springer, Berlin, 1984), vol. 120, pp. 1-40. J. Fritsch and G. Zundel, J . Phys. Chem., 1981, 85, 556. G.Zundel and J. Fritsch, in Chemical Physics of Solvation, ed. R. R. Dogonadze, E. Kalman, A. A. Kornyshev and J. Ulstrup (North Holland, Amsterdam, 1985), vol. 11, chap. 3. R. Lindemann and G. Zundel, J . Chem. Soc., Faraday Trans. 2, 1977,73, 788. G. Albrecht and G. Zundel, J . Chem. SOC., Faraday Trans. 1, 1984,80, 553. I. Olovsson and P. E. Jonsson, in The Hydrogen Bond - Recent Developments in Theory and Experi- ments, ed. P. Schuster, G. Zundel and C . Sandorfy (North Holland, Amsterdam, 1976), vol. 11, chap. 8. fi G. Zundel and J. Fritsch, J . Phys. Chem., 1984,88, 6295. l o G. C. Pimentel and A. L. McClellan, Annu. Rev. Phys. Chem., 1971, 22, 347. l 1 D. Hadii and S. Bratos, in The Hydrogen Bond- Recent Developments in Theory and Experiments, ed. P. Schuster, G.Zundel and C . Sandorfy (North Holland, Amsterdam, 1976), vol. 11, chap. 12. l 2 E. G. Weidemann and G. Zundel, Z . Naturforsch. Teil A , 1970, 25, 627. I B R. Janoschek, E. G. Weidemann, H. Pfeiffer and G. Zundel, J . Am. Chem. SOC., 1972,94, 2387. G. Zundel, in The Hydrogen Bond - Recent Developments in Theory and Experiments, ed. P. Schuster, G. Zundel and C . Sandorfy (North Holland, Amsterdam, 1976), vol. 11, chap. 15. l 5 D. Hadii, J , Chem. SOC., 1962, 5128. Ifi D. Had5 and N. Kobilarov, J . Chem. SOC. A , 1966, 439. I' D. Hadii, J . Chem. SOC. A , 1970, 418. I n Z. Dega-Szafran, E. Grech and M. Szafran, J. Chem. SOC., Perkin Trans. 2, 1972, 1839. Z. Dega-Szafran, E. Grech, M. Z. Naskret and M. Szafran, Roczn. Chem., 1973, 47, 1849. 2o Z. Dega-Szafran and M. Szafran, J . Mol. Struct., 1978, 45, 33. 2 1 Z. Dega-Szafran, M. Szafran and J. Rychlewski, J . Chem. SOC., Perkins Trans. 2, 1978, 536. 22 B. Brycki, Z. Dega-Szafran and M. Szafran, Adv. Mol. Relax. Interact. Proc., 1979, 15, 71.1434 I.R. STUDIES OF HYDROGEN BONDS 23 (a) B. Brycki, Z. Dega-Szafran and M. Szafran, Pol. J. Chem., 1980, 54, 221; (6) B. Brycki and 24 M. Szafran and Z. Dega-Szafran, J. Mol. Struct., 1984, 99, 189. 25 B. Brycki and M. Szafran, J. Chem. SOC., Perkin Trans. 2, 1984, 223. 26 Z. Dega-Szafran and M. Szafran, J. Mol. Liquids, 1983, 25, 109. 27 M. M. Kreevoy, K. Chang, J. Phys. Chem., 1976, 80, 259. 28 K. Chang, Thesis (University of Minnesota, 1975). 29 B. Brzezinski, personal communication. 30 E. P. Serjeant and B. Dempsey, Zonisation Constants of Organic Acidr in Aqueous Solution, IUPAC 31 D. D. Perrin, Dissociation Constants of Organic Bases in Aqueous Solution (Butterworths, London, 32 J. H. Nelson, R. G. Garvey and R. 0. Ragsdale, J . Heterocycl. Chem., 1967, 4, 591. 33 C. Klofutar, S. Paljk and D. Kremser, Spectrochim. Acta, Part A, 1973, 29, 139. 34 L. Onsager, J. Am. Chem. SOC., 1936, 58, 1486. 35 A. Hayd, E. G. Weidemann and G. Zundel, J. Chem. Phys., 1979,70, 86. 36 B. Brzezinski and G. Zundel, J. Mol. Struct., 1981, 72, 1 101. 37 R. Janoschek, E. G. Weidemann and G. Zundel, J. Chem. SOC., Faraday Trans. 2, 1973,69, 505. M. Szafran, J. Chem. SOC., Perkin Trans. 2, 1982, 1333. Data Series no. 23 (Pergamon Press, Oxford, 1978). 1965, Supplement 1972). (PAPER 4/ 13 16)

ISSN:0300-9599    

DOI:10.1039/F19858101425

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 81, 1435-1440 Silicon-29 Magic-angle-spinning Nuclear Magnetic Resonance Study of the Crystalline-Amorphous Transition of Zeolite A Containing Trapped Krypton BY JACEK KLINOWSKI* AND XINSHENG LIU Department of Physical Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EP AND RALF-DIETER PENZHORN AND PETER SCHUSTER Kernforschungszentrum Karlsruhe, Institut fur Radiochemie, Postfach 3640, D-7500 Karlsruhe, Federal Republic of Germany AND NIGEL J. CLAYDEN AND CHRISTOPHER M. DOBSON Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR Received 27th July, 1984 High-resolution solid-state 29Si n.m.r. and i.r. spectroscopies have been used to monitor the crystalline-amorphous transition of (Ca, Na)- and (Sr, Na)-exchanged zeolites A hydrothermally treated in the presence of densified krypton gas.Both techniques are capable of monitoring the transition on a smaller scale than X-ray diffraction and demonstrate that thermally stable gas fixation takes place in a truly amorphous glassy matrix. Zeolites are of considerable interest as host matrices for troublesome by-products of nuclear fission. Some species, such as 137Cs, are most conveniently scavenged by the zeolite by ion exchange.l Appreciable amounts of gas can be introduced into a zeolite at elcvated temperatures and pressures and trapped on cooling if the kinetic diameter of'the gas molecule is larger than the effective size of the channel The drawback of this method of gas encapsulation is its essential reversibility: despite the fact that activation energy is required for the intracrystalline diffusion, leakage of trapped gas is often unacceptably high, especially when long-term storage of radioactive gas with high heat of decay is desired.A novel method of irreversible trapping of gases has recently been developed, in which the gas-bearing crystalline zeolite is hydrothermally vitrified in the presence of a pressurised gas.* The method has been extensively examined as a means of encapsulation of noble gases in zeolite A exchanged with alkaline-earth-metal We have used three techniques for the characterization of the gas-trapping matrix: X-ray diffraction, i.r. spectroscopy and 29Si magic-angle-spinning n.m.r. (m.a.s.n.m.r.). The advantage of this three-pronged approach is that each technique is sensitive to a different level of crystallinity. EXPERIMENTAL Partially exchanged (Ca, Na) and (Sr, Na) forms of zeolite A (Na-A 70% exchanged with (:a2+ and 77% exchanged with Sr2+) were used in the form of pellets ca.2 mm in diameter. While 14351436 29Si N.M.R. STUDY OF ZEOLITE A Sr- and Ca-exchanged zeolites A show little difTerence in their physical properties and crystal structure, the strontium form appears to be less stable towards hydrothermal treatment. Krypton encapsulation was performed in Nimonic autoclaves filled with pellets in such a way that dead space was minimized, as follows. (i) The water content of the zeolite was reduced to a preselected level by heating to 400 "C under vacuum for 16 h.(ii) Krypton gas was admitted to the autoclave at room temperature and a pressure up to 1000 bar. Krypton was pressurized by freezing with liquid nitrogen and vaporization in a Cu/Be trap. (iii) The autoclave was then closed and heated up to the preselected encapsulation temperature (625-675 "C) over a period of ca. 4 h. Increasing the residence time at elevated temperature caused no further change in the degree of crystallinity. During this treatment the zeolite became partially or totally amorphous, depending on the water content and the temperature. (iv) Finally the autoclave was cooled and the supernatant krypton pumped out. The efficiency of krypton encapsulation was determined by heating a fraction of the sample to 1100 "C. The amount of liberated gas was measured volumetrically as a function of temperature and analysed by gas chromatography.The relative thermal stability of the samples was determined by temperature-programmed heating at 8 "C min-l in the range 500-850 "C. 29Si m.a.s.n.m.r. measurements were carried out at 39.90 MHz using a Bruker CXP-200 spectrometer equipped with an Andrew-Beams probehead. Powdered samples were spun in Delrin spinners at ca. 3.5 kHz without cross-polarization or decoupling. Ca. 5000 scans were accumulated per sample and all chemicals shifts are quoted in ppm from tetramethylsilane (TMS). Powder X-ray diffractior! patterns were obtained on a Philips vertical goniometer using Cu Ka radiation selected by a curved graphite monochromator in the diffracted beam. The degree of crystallinity, pi, has been calculated from the relative intensity of diffraction peaks in the range 3-70" using the formula Eli urn p .= - where li and I , are the intensities (peak areas) in the sample and a zeolite A standard (taken to be 100% crystalline), respectively. 1.r. measurements were obtained from a zeolite/KBr disc using a Nicolet lOMX Fourier- transform spectrometer. RESULTS AND DISCUSSION Zeolites may display partial or total loss of long-range crystalline order and become amorphous to X-rays while at the same time retaining ion-exchange capacity and catalytic activity. This indicates that porosity is present on the unit-cell level. Such ' amorphous zeolites ' have been characterized by high-resolution electron micro- copy*-^^ and by i.r. l2 Our results indicate that the crystalline- amorphous transition in zeolites can also be monitored readily by high-resolution solid-state 29Si n.m.r.with magic-angle spinning,l3, l4 a technique primarily sensitive to short-range order. Highly crystalline (Ca, Na)-exchanged zeolite A gives a single n.m.r. signal at - 90.5 ppm [fig. 1 (a)], which is characteristic of this unique structure.li33 l5 The f.w.h.m. of this line is 3.3 ppm. The spectrum of (Sr, Na)-exchanged zeolite A is identical. The i.r. spectrum [see fig. 3(a) below] is typical of highly crystalline zeolite A. We have found that when (Ca, Na)-exchanged zeolite A is heated gradually in an open quartz vessel, the crystal structure breaks down at 775 "C, resulting in the formation of an amorphous solid. Since isochoric heating causes loss of crystallinity at temperatures much lower than 775 "C, it is clear that water plays an important role in the process.Water contents of the amorphous samples are given in ref. (7). When phase transition occurs in the presence of pressurized gas, a large fraction of the gas is immobilized in the vitreous material. When (Ca, Na)-exchanged zeolite A containingJ. KLINOWSKI et al. -90.5 1437 pprn from TMS Fig. 1. High-resolution 29Si m.a.s.n.m.r. spectra at 39.90 MHz. Chemical shifts are given in ppm from tetramethylsilane (TMS). (a) Untreated (Ca, Na)-exchanged zeolite A. (b) (Ca, Na)- exchanged zeolite A loaded with Kr at 625 "C. The degree of crystallinity (by X-ray diffraction) is 0.46 and the fraction of Kr released on heating is relatively large.(c) (Sr, Na)-exchanged zeolite A loaded with Kr at 650 "C. The sample is totally amorphous to X-rays and Kr is released only at very high temperatures. ( d ) (Ca, Na)-exchanged zeolite A loaded with Kr at 675 "C. The sample has recrystallized to nepheline + anorthite +hexagonal CaAl,Si,O, and contains little trapped Kr. < 0.8 wt % of residual water was loaded with krypton at 625 "C and 900 bar pressure the content of krypton in the product was 47 cm3 s.t.p. g-l. The sample showed a substantial loss of X-ray crystallinity [the degree of crystallinity estimated from the diffractogram in fig. 2(b) was 0.46 of that in the parent material] and a partial loss of i.r. crystallinity, although D4R vibrations, characteristic of zeolite A, were still present [fig.3(b)]. However, the 29Si m.a.s.n.m.r. spectrum [fig. 1 (b)] is completely unchanged. Evidently, the crystalline framework has undergone fragmentation in the course of heat treatment, each fragment retaining its identity as zeolite A, but having insufficient long-range order to be detected by X-ray techniques. The thermal behaviour of the sample supports this conclusion. While the largest fraction of the immobilized gas is liberated at temperatures > 850 "C, some gas leakage is observed in the 500-850 "C temperature range. This behaviour is typical of partially 'vitrified' samples (as judged by X-ray methods).1438 29Si N.M.R. STUDY OF ZEOLITE A I L I t u J 3 8 13 18 23 28 33 38 43 48 53 58 zeio Fig. 2. Powder X-ray diffraction patterns of zeolite samples; (u)-(d) as in fig.1. (Ca, Na)-exchanged zeolite A loaded with krypton under the same conditions as above but at 650 and 660 "C yields a product with the same gas content, but which is completely amorphous to X-rays and i.r. [fig. 2(c) and 3 (C), respectively]. The m.a.s. spectra of both samples were identical [the spectrum of the zeolite treated at 650 "C is given in fig. l(c)] and very different from that of the parent zeolite and from the sample treated at 625 "C. The chemical shift is now -85.0 ppm, characteristic of Si surrounded tetrahedrally by four Al-centred tetrahedra, while the f.w.h.m. of the (symmetrical) line is 16.7 ppm, i.e. five times greater than in the parent material. This suggests a wide distribution of Si-0-T angles, where T is Si or A1.16 The large change of chemical shift shows that structural identity of zeolite A, even on the unit-cell level, has been lost.The unusual value of the 29Si chemical shift in zeolite A permits us toJ. KLINOWSKI et al. 1439 a E 0 + i 0 c c Fig. 3. F.t.i.r. spectra of zeolite samples; (a)-(d) as in fig. 1. D4R denotes double four-membered ring vibrations and TO denotes T-0-T bending (T is Si or Al). observe this effect directly in a manner entirely analogous to that described by Engelhardt et aZ.,17 who monitored by 29Si m.a.s.n.m.r. the emergence of zeolite A from its amorphous-gel precursor. The samples prepared at 650 and 660 "C liberate the encapsulated gas only at temperatures > 850 "C. Their i.r. spectra [see fig. 3(c)] are typical of amorphous aluminosilicates, such as glasses.It has been shown7 that amorphous samples containing a noble gas can also be obtained at much lower temperatures (between 350 and 650 "C). Once trapping has been accomplished recrystallization into some anorthite-like species takes place at the same transformation temperature (850 "C), regardless of the parameter combination selected for gas fixation. In principle, gas can be trapped in zeolite pellets in one of the following three ways : (i) in the intercrystalline space inside a sintered pellet, (ii) in structurally perfect fragments of zeolite which are embedded in an amorphous mass or (iii) within a truly1440 29Si N.M.R. STUDY OF ZEOLITE A amorphous glassy matrix. Method (i) clearly does not apply, since essentially no gas is released when a loaded pellet is finely ground.As shown by i.r. and m.a.s.n.m.r., both short-range techniques, method (ii) occurs under non-optimized fixation con- ditions. With proper selection ofparameters (activity ofwater, duration and temperature of fixation) krypton is trapped in a thermally stable way within a truly amorphous solid [method (iii)]. When (Ca, Na)-exchanged zeolite A is in contact with krypton at still higher temperatures, e.g. 675 "C, the sample recrystallizes, producing an intimate but perfectly crystalline mixture of nepheline, anorthite and hexagonal CaAl,Si,O,, all of which have been identified by X-ray powder diffraction. All the peaks in the diffractogram could be assigned, and the spacings correspond exactly to those reported by Flank et al.,1s~19 who prepared a 'low-density ceramic foam' by high-temperature treatment of Ca-exchanged zeolite A.Clearly, the compact alumino- silicates such as nepheline, anorthite and CaAl,Si,O, cannot contain gases in their intracrystalline space and so the sample contains very little encapsulated gas. Its n.m.r. spectrum, given in fig. 1 (d), again shows a single line with an f.w.h.m. of 9.3 ppm, i.e. much narrower than in the amorphous material but, in view of the number of different Si-0-T angles present, not as narrow as the pure parent zeolite. We thank University of Cambridge and B.P. Research Centre, Sunbury for support and Prof. J. M. Thomas, F.R.S. for discussions. We also thank the Commission of the European Communities for support under its programme on radioactive waste and storage.L. A. Bray and H. T. Fullam, in Molecular Sieve Zeolites I, ACS Adv. Chem. Ser., 1971, 101, 450. W. J. Sesny and L. H. Shaffer, U S . Patent, 3 316 691, 1967. R. M. Barrer and D. E. W. Vaughan, J. Phys. Chem. Solids, 1971,32,73 1 ; Trans. Faraday SOC., 1971, 67, 2129. * R-D. Penzhorn, P. Schuster, H. Leitzig and H. E. Noppel, Ber. Bunsenges. Phys. Chem., 1982, 86, 1077. R-D. Penzhorn, H. Leitzig, K. Gunther, P. Schuster and H. E. Noppel, 17th DOE Nuclear Air Cleaning Conference, Denver, Colorado, 1982. E. F. Vansant, A. Thijs, G. Peeters, P. de Bikvre, R-D. Penzhorn, A. Dorea and P. Schuster, Zeolites, 1984, 4, 35. ' R-D. Penzhorn and W. Mertin, J. Solid State Chem., 1984, 54, 235. * J. M. Thomas and L. A. Bursill, Angew. Chem., Znt. Ed. Engl., 1980, 19, 745. lo L. A. Bursill, J. M. Thomas and K. J. Rao, Nature (London), 1981, 289, 157. l1 P. A. Jacobs, E. G. Derouane and J. Weitkamp, J . Chem. Soc., Chem. Commun., 1981, 591. l2 G. Coudurier, C. Naccache and J. C. Vedrine, J. Chem. Soc., Chem. Commun., 1982, 1413. l3 E. Lippmaa, M. Magi, A. Samoson, G. Engelhardt and A-R. Grimmer, J. Am. Chem. SOC., 1980,102, l4 C. A. Fyfe, J. M. Thomas, J. Klinowski and G. C. Gobbi, Angew. Chem., Int. Ed. Engl., 1983,22,259. l5 J. M. Thomas, C. A. Fyfe, S. Ramdas, J. Klinowski and G. C. Gobbi, J. Phys. Chem., 1982,86,3061. l6 J. M. Thomas, J. Klinowski, S. Ramdas, B. K. Hunter and D. T. B. Tennakoon, Chem. Phys. Lett., L. A. Bursill and J. M. Thomas, J. Phys. Chem., 1981,85, 3007. 4889. 1983, 102, 158. G. Engelhardt, B. Fahlke, M. Magi and E. Lippmaa, Zeolites, 1983, 3, 292. l8 W. H. Flank, J. E. McEvoy and J. R. Stuart, U S . Patent, 3 574 647, 1971. l9 W. H. Flank, J. E. McEvoy and J. R. Stuart, U S . Patent 3 775 136, 1973. (PAPER 4/ 1328)

ISSN:0300-9599    

DOI:10.1039/F19858101435

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

.I. Chem. Soc., Faraduy Trans. 1, 1985,81, 1441-1445 Substituent Effects on Proton Tunnelling Reaction between 2,4,6-Trinitrotoluene and 1 -Substituted Piperidines in Acetonitrile BY NAOKI SUGIMOTO AND MUNEO SASAKI* Department of Chemistry, Faculty of Science, Kyoto University, Kyoto 606, Japan Received 3 1 st July, 1984 The proton/deuteron-transfer reaction of 2,4,6-trinitrotoluene (TNT) with l-methylpiperi- dine, 1 -ethylpiperidine or 1 -phenethylpiperidine in acetonitrile is a simple reaction forming a 1 : I ion pair. The proton- and deuteron-transfer reactions have been followed at 5-35 "C by a stopped-flow method. The isotope rate ratio kH/kD at 25 "C is 14.7-16.2, the activation- energy difference Ea(D) - Ea(H) is 9.2-10.9 kJ mol-' and the ratio of Arrhenius pre-exponential factors A(D)/A(H) is 2.5-4.0 for the forward reaction.Similar values of kH/k", Ea(D)-Ea(H) and A(D)/A(H) are also found for the reverse reaction. All these values signify a tunnelling contribution to the proton-transfer reaction. An analysis based on an unsymmetri- cal parabolic potential barrier shows that as the bulkiness of the substituted group in the piperi- dines increases the top of the barrier becomes much rounder, and so tunnelling decreases. It is an important problem as to how tunnelling is affected by geometrical factors of the reactants.l*, We should consider two kinds of effects concerning the size of the substituents near the reactive site and also the effect of the s01vent:~~~ One is the effect of forming a cage-like structure in which the proton-transfer process is much less coupled to solvent motion; the other is the effect of increasing acid-base distance.In comparison with the reaction of 4-nitrophenylnitromethane (4NPNM) with 1,8- diazabicyclo[ 5 . 4 . Olundec-7-ene (DBU),5 a larger tunnelling contribution was observed for 2,4,6-trinitrotoluene (TNT),6-8 in which the reactive site is partly concealed by the adjacent NO, groups. This fact was considered in terms of the former influence. A similar effect was found in that pyridines having bulky substituents in the 2- and 6-positions enhanced the isotope rate ratio (and thus tunnelling) in proton transfer from 2-nitropr0pane.~ On the other hand, there have been few studies confirming the second view described above, although in order to evaluate the degree of tunnelling it would be informative to investigate systematically the influence of substituents bonding directly to an active base centre.In this paper we deal with the proton- and deuteron-transfer reactions of TNT with piperidine (P), 1 -methylpiperidine (MP), I -ethylpiperidine (EP) and 1 -phenethylpiperidine (PP) in acetonitrile (see scheme 1). Scheme 1. Q AH : TNT B : P MP EP PP 14411442 SUBSTITUENT EFFECTS ON PROTON TUNNELLING EXPERIMENTAL MATERIALS 2,4,6-Trini tro toluene (TNT) and deuterated 2,4,6- trini tro toluene ([2H,]TNT) were treated as described lo Piperidine (P), 1 -methylpiperidhe (MP), 1 -ethylpiperidine (EP) (Nakarai Chemical Col. Ltd) and 1 -phenethylpiperidine (PP) (Aldrich Chemical Co.) were dried with potassium hydroxide and then distilled. Acetonitrile was purified by a standard method.' APPARATUS AND PROCEDURES Absorption spectra were recorded with a Shimadzu UV-200s spectrophotometer. Kinetic measurements were carried out under an N, atmosphere with a Union Giken RA-401 stopped-flow apparatus.The temperature of the solution was regulated to within kO.1 "C by the circulation of thermostatted water. All solutions were freshly prepared just before use at the following concentrations (in mmol dmP3): TNT, [,H,]TNT, 0.09-0.6; P, 12-57; MP, 3-84; EP, 3-57 and PP, 9-80. We determined the proton-transfer equilibrium constants at 25 "C according to the Benesi-Hildebrand equation." We also obtained the rate of the protonldeuteron-transfer reaction by monitoring change in the absorbance ( A ) with time at 520 nm.The increase in absorbance obeyed first-order kinetics when the base was present in a large excess over TNT or [2H,]TNT. The observed first-order rate constant kobs was obtained as the slope of either plots of ln(A,-A,) against time or the Guggenheim plot. The rate constants determined by both calculations were in accord with each other. RESULTS The reaction of TNT or [2H3]TNT with piperidine gave a purple solution within the range of the present experimental concentrations. The absorption maxima of the product were around 378, 520 and 635 nm in acetonitrile, but the peak at 378 nm was concealed when the concentration of the acid exceeded 0.2 mmol dm-3. These absorption maxima were quite similar to those observed for the systems TNT+ DBU69 and TNT + tetramethylguanidine (TMG).l07 l2, l3 This indicates that the present reaction produces a 1 : 1 proton-transferred ion l2 The rate of growth of the absorbance was the same for all peak maxima, showing that they are due to a single species.In a kinetic measurement the stopped-flow trace obeyed a first-order kinetic equation when the base concentration was in large excess. Only the [2H3]TNT+P system did not obey first-order kinetics exactly. This behaviour may be due to isotope exchange,14 but the observed rate constant, kobs, at 25 "C was determined by a single-exponential function as for other systems. Plots of kobs against the base concentration, [B], were linear, in agreement with eqn (1) (see fig. 1): In this equation kF and kk are the forward and backward rate constants of the proton- and deuteron-transfer reactions (L = H or D).We give the values of k,L and kk at 25 "C in table 1, while table 2 lists the values of the equilibrium constant K at 25 "C obtained from k,/kb and those of Keg from equilibrium measurements. From the temperature dependence of the rate constants, the activation energies, E,, and the Arrhenius pre-exponential factors, A , of the forward and backward reactions were determined by a least-squares method (table 3). The tunnelling factor, Q, and curvature, v of the unsymmetrical parabolic potential barrier were computed on the basis of Bell's equation.15 In these calculations we used the values of (kf"/k?)* obtained from Arrhenius plots over the whole range of experimental temperatures.N .SUGIMOTO AND M. SASAKI 1443 5 4 3 - I v) --. Q 4 2 1 0 20 40 60 80 [B]/mmol dm3 Fig. 1. Dependence of kobs on [PP] for the TNT+PP reaction in acetonitrile; T = @, 35; 0, 25; A, 15 and A, 5 "C. Table 1. Rate constants kp, kp, k,D and k; for the proton- and deuteron-transfer reactions from TNT to 1-substituted piperidines at 25 "C piperidine ky/dm3 mol-1 s-' k,H/s-l kp/dm-3 mol-l s-l k;/ lo-' s-l P 24.0 & 0.5 1.74 & 0.4 2.03 f 0.10 1.74 & 0.09 MP 49.4& 1.5 1.94 f 0.06 3.10 f 0.12 1.22 f 0.05 EP 39.2 f 0.4 0.95 f 0.01 2.32 & 0.05 0.60 f 0.01 PP 19.8 f 0.8 0.88 f0.04 1.50 f 0.06 0.62 k0.02 Table 2. Equilibrium constants K&, K H and K D (in dm3 mol-l) for the proton- and deuteron-transfer reactions from TNT to 1 -substituted piperidines at 25 "C piperidine K&" KH K D b P 16.1 1.3 13.8 f 0.6 11.7f 1.2 MP 23.1 1.8 25.5 f 1.6 25.4 + 2.0 EP 32.7 & 2.9 41.3 f 0.9 38.7 f 1.5 PP 28.3 f 2.5 22.5 & 1.9 24.2 f 1.7 a, Values from equilibrium measurements.Values from rate-constants ratios ( K = kf/kb).1444 SUBSTITUENT EFFECTS ON PROTON TUNNELLING Table 3. Activation parameters for the proton- and deuteron-transfer reactions from TNT to 1-substituted piperidines at 25 "C parameter MP EP PP 39.3 f 1.3 8.6 f 0.1 50.2 f 0.4 9.2 k 0.1 29.7 +_ 0.4 5.5 f 0.1 39.8 f 0.8 6.0 & 0.1 48.1 f 0.8 10.0 +o. 1 58.6 f 0.8 10.6 f 0.1 41.4 f 0.4 7.2 f 0.1 51.9f 1.7 7.8fO.I 33.5 f0.8 7.3k0.1 42.7 f 0.4 7.7 & 0.1 53.1 f 0 . 4 9.3f0.1 62.8 f 0.8 9.7 f 0.1 DISCUSSION The kinetic isotope effect is anomalously large (kH/kD = 14-17 at 25 "C), although the isotope effect on the equilibrium constant is small.The large value of kH/kD suggests the involvement of nuclear tunnelling in the reaction coordinate. Tunnelling is favoured for a smaller effective mass of a species in motion. It will be diminished if the effective mass of the transferred particle increases with strong coupling of the solvent molecules. As is well known, tunnelling in a proton/deuteron-transfer reaction is revealed by the following characteristics;l? 7 * l5 an isotope rate ratio kH/kD > 7-1 1 at 25 "C, an activation-energy difference Ea(D) - Ea(H) > 4.6-5.8 kJ mol-l and a ratio of the Arrhenius pre-exponential factors A(D)/A(H) > 1-1.4. Values of (kH/kD)* at 25 "C, Ea(D) - &(H) and A(D)/A(H) for both the forward and backward reactions in acetonitrile are shown in table 4.These values are all larger than the semiclassical limits as described above, even after correction for a secondary isotope effect.6y l3 These facts lead to the conclusion that tunnelling occurs in these proton-transfer reactions even in a polar solvent, so that this reaction system is a good model for studying how the microenvironment near a reactive site affects the degree of tunnelling in terms of both solvent coupling and steric effects of the reactants. The values of kH/kD decrease in the order MP z EP > PP, and these are smaller than those of TNT + DBU (19. 1)6 and TNT + TMG (1 8.5)12 in the same solvent. With regard to the effect of the 1-substituent of piperidine, the values of v(H), which correspond to the curvature of the proton-transfer barrier, are 877 cm-l in MP, 868 cm-l in EP and 806 cm-l in PP.The proton-tunnelling factors Q(H) at 25 "C are 2.44, 2.39 and 2.07, respectively. The different behaviour of these bases is related to the difference in orbital hybridization and geometrical variations near the reactive site.l2? l6 The larger isotope effect and tunnelling contribution for TMG and DBU may be attributed to the narrower barrier width caused by the more compact sp2 hybridized orbitals of the nitrogen atoms in comparison with the sp3 orbitals of the piperidines.17 With regard to the effect of substituents on tunnelling we must consider two different interactions; one is that between a transferred particle and the solvent molecules and the other is that between acid and base molecules. First, the steric group of a base may exclude solvent molecules from the reactive site and reduce their coupled motions to the transferred proton. As the effective mass of the proton decreases or the barrier height increases, e.g.by energy of repulsion, the first substituent effect increases the tunnellingN. SUGIMOTO AND M. SASAKI 1445 Table 4. Characteristic observations of tunnelling for the TNT + 1 -substituted piperidines parameter MP EP PP (kfH/kfD>A" (25 O C ) 16.2f0.7 16.2f0.3 14.7f0.6 Af(D)/A f(H) 4.0 & 1.6 4.0 f 1.6 2.5 & 1.6 (kbH/kbD>A" (25 O C ) 16.2f0.3 16.4 k 0.4 14.2 f 0.4 Ea, b(D)-E,, mol-' 10.1 * 1.2 10.5 & 2.1 9.74 1.2 3.2 1.6 4.0 f 1.6 2.5 & 1.6 E,, f(D) - E,, ,(H)/kJ n-Io1-l 10.9f 1.7 10.5 & 1.6 9:2 f 1.2 A ,(D)/A b(H) u Values of the rate constants were taken from Arrhenius plots over the whole range of experimental temperatures.contribution. On the other hand, the steric hindrance of substituents in a base should prevent the approach of an acid to the nitrogen base centre and increase the barrier width. Thus the second effect decreases the degree of tunnelling. As a result of the sum of these two opposing contributions to tunnelling, the curvature v(H) of the barrier decreases with increasing bulkiness of the substituent bonding directly to the nitrogen atom, and consequently the tunnelling factor Q(H) decreases progressively. Thus the bulky substituents bonding directly to the reactive base centre operate so as to prevent the approach of an acid but not that of a solvent molecule.CONCLUSIONS The reaction of TNT with MP, EP or PP in acetonitrile forms a proton-transferred ion pair. While the isotope effect on equilibrium is small, the kinetic isotope effect is very large owing to the tunnelling contribution. The curvature v(H) of the barrier becomes much smaller and so the tunnelling factor Q(H) decreases with increasing size of the substituted group bonding directly to the base centre. E. F. Caldin and V. Gold, The Proton-transfer Reaction (Chapman and Hall, London, 1975). Faraday Discuss. Chem. Soc., 1982, 74. (a) M. Sasaki, N. Sugmoto and J. Osugi, Chem. Lett., 1980, 887; (6) N. Sugmoto, M. Sasaki and J. Osugi, Bull. Inst. Chem. Rex, Kyoto Uniu., 1981, 59, 63. N. Sugimoto, M. Sasaki and J. Osugi, Bull. Chem. Soc. Jpn, 1984, 57, 366. E. F. Caldin and 0. Rogne, J. Chem. SOC., Faraday Trans. I , 1978, 74, 2065. N. Sugimoto, M. Sasaki and J. Osugi, J. Phys. Chem., 1982, 86, 3418. ' N. Sugimoto, M. Sasaki and J. Osugi, J. Am. Chem. Soc., 1983, 105, 7676. + N. Sugimoto, M. Sasaki and J. Osugi, J. Chem. Soc., Perkin Trans. 2, 1984, 655. E. S. Lewis and L. H. Funderburk, J. Am. Chem. Soc., 1967, 89, 2322. I " N. Sugimoto, M. Sasaki and J. Osugi, Bull. Chem. Soc. Jpn, 1981, 54, 2598. I I H. Benesi and J. H. Hildebrand, J. Am. Chem. SOC., 1949, 71, 2703. I ! P. Pruszynski and A. Jarczewski, Roczn. Chem., 1977, 51, 2171. A. Jarczewski. P. Pruszynski and K. T. Leffek, Can. J. Chem.. 1979, 57, 669. (a) J. H. Blanch and 0. Rogne, J. Chem. Soc., Faraday Trans. I, 1978, 74, 1254; (h) H. Bulska and A. Chodowska, J. Am. Chem. SOC., 1980, 102, 3259. I " R. P. Bell, The Tunneling EfSect in Chemistry (Chapman and Hall, London, 1980). I " E. F. Caldin and S. Mateo. J. Chem. Soc., Faraday Trans. I , 1975, 71, 1876. B. K. Carpenter, J . Am. Chem. Soc., 1983, 106. 809. (PAPER 4/ 1358)

ISSN:0300-9599    

DOI:10.1039/F19858101441

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 81, 1447-1454 Isomerization of Alkanes on Epitaxially Oriented (1 1 1) Pd-Cu and Pd-Ag Alloy Films BY ZBIGNIEW KARPINSKI, * WOJCIECH JUSZCZYK AND JAROSLAW STACHURSKI Institute of Physical Chemistry of the Polish Academy of Sciences, Kasprzaka 44/52, 0 1-224 Warszawa, Poland Received 6th August, 1984 The reactions of neopentane and n-butane in the presence of an excess of dihydrogen have been studied on films of Pd-Cu and Pd-Ag evaporated onto mica. The characterization of films showed that they were predominantly (1 1 1) oriented. The catalytic activity for isomerization of neopentane and n-butane was higher for ca. 10 at % Cu than for pure Pd, resembling the catalytic behaviour of Pd-Au( 1 1 1) films. This does not, however, confirm the existence of the ligand effect in alkane isomerization over Pd-IB metal alloys.The catalytic activity of Pd-Ag( 1 1 1) alloys decreases with increasing silver content, but on the basis of the inferred surface composition of these alloys this agrees with the conclusion drawn for the Pd-Cu system, i.e. the possible absence of the ligand effect. Recently it has been discovered that neopentane undergoes selective isomerization over (1 11) oriented palladium' and Pd-Au(ll1) alloy films.2 On the basis of the relationship between the catalytic activity and the alloy composition (the maximum for ca. 10 at% Au) and additional experiments with CH,/D, and neopentane/D, exchange over Pd-Au( 1 1 1) alloys (almost exclusively the stepwise mode of exchange at higher temperatures), the hypothesis was advanced2 that neopentane isomerizes via an adsorbed alkyl species (not by surface carbenes) at one metal site according to the mechanism of Rooney and coworker~.~ On the other hand it has been suggested that the electronic ligand effect operates for the bond-shift isomerization, mainly because some Pd-Au alloys were more active than pure palladium.Even though it has been realized that the maximum on the activity against alloy composition curve is not evidence of the ligand e f f e ~ t , ~ so the relative importance of ensemble size as against electronic effects in the Pd-Au alloys is still unsettled. Sachtler and Somorjai5 have recently found that the ligand effect is absent for CO adsorption on Pt-Au( 1 1 1) single-crystal surfaces.However, the enhancement of n-hexane isomerization by these alloys could follow from electronic as well as from geometric changes, and they5 suggested the use of other Group IB elements (Cu or Ag) to prove the relative importance of the ligand effect. Replacing Au by Cu would influence the course of the isomerization for one or more of the following reasons. (1) Changing the ligand may introduce considerable geometric changes: gold has a larger atomic radius (1.44 A) than copper (1.28 A) and so the distances between atoms are larger in Pd-Au than in Pd-Cu alloys, which may influence the catalytic behaviour of the Pd alloy. (2) Replacing Au by Cu has a great efyect on the catalytic behaviour of platinum alloys in alkane conver~ion~-~ (more hydrogenoiysis occurs with copper) and other Group VIII metals (Ir, Ni and Pd) when highly diluted in copper also show increased activity for the hydrogenolysis of 14471448 ISOMERIZATION OF ALKANES ON Pd-Cu, Ag FILMS neopentane.lo ( 3 ) Qualitative differences have been found in the ESCA" valence- band-electron spectra of the Pd-Cu and Pd-Ag systems. A tail tending toward the bottom of the valence band for the Cu-based alloys and an absence of such a tail for Pd-Ag suggests that the Pd 4d states are more involved in bonding in the Cu alloys than they are in the Ag alloy. Also a larger chemical shift of the Pd 3d5,, line occurs in the copper host (0.8 eV) than in the silver host (0.5 eV). For the palladium-rich alloys, the copper d band hybridizes strongly with the host d band and loses its identity; a similar result was observed for the Pd-Au system where the 5d5/2 part (but not the 5d3,, part) of the Au d band could not be identified in the spectrum for high palladium concentrations.12 On the other hand, Hiifner et aZ.13 showed that in the valence band of the Pd-Ag system the d bands of Pd and Ag are well separated. It was decided to investigate whether these differences in the valence-band structures (and different chemical shifts) influence the catalytic behaviour of palladium atom centres.EXPERIMENTAL The (1 1 1) oriented films of palladium-copper and palladium-silver alloys were prepared by epitaxial growth of an evaporated layer of metals on a sheet of mica in the manner described in ref. (2). Briefly, both alloy components, palladium and copper (or silver), were evaporated simultaneously from two independent sources on the inner wall of a cylindrical reactor.The entire side wall of the reactor was covered by the freshly cleaved mica sheet. During evaporation (vacuum ca. 5 x Torr, 1 Torr = 133.3 Pa) and subsequent annealing in hydrogen for 2 h ( 5 Torr) the reactor was maintained at 500 "C. Spectrographically standardized palladium, copper and silver wires used for evaporation were purchased from Johnson Matthey. As previously,2 the lateral homogeneity of our films was poor. This was checked by analysing various sections of the alloy films by X-ray diffractometry (Rigaku-Denki X-ray diffractometer with Cu Ka radiation and a scintillation counter). It was estimated from the lattice parameter that the top section of the film of nominal composition 11.3 at% Cu consisted of 10 at% Cu, whereas the middle contained 14 at% Cu.For the films of 20.0, 30.8 and 35.3 at% Cu the atomic percentages of copper in the top and middle sections were: 20 and 29, 37 and 30, and 41 and 35, respectively. Large percentage differences were also observed for the Pd-Ag alloys. The X-ray diffraction study confirmed the presence of adequate phase homogeneity and gave important information as to the extent of the epitaxy of our films. The investigations of the reactions of neopentane (Fluka, puriss) or n-butane (Merck, purity of ca. 99.5%) with dihydrogen were conducted in a static circulation system. The whole system was made of Pyrex glass with Young's stopcocks. The main part of the apparatus was the cylindrical reactor connected to the rest of the system by two joints with W o n 0 rings.The progress of the reaction was followed by g.1.c. The hydrocarbons were dried and outgassed before use. RESULTS FILM CHARACTERIZATION X-ray diffractometry indicated that both the Pd-Ag and Pd-Cu films had a very high degree of (I 11) preferred orientation with respect to the mica base. The degree of orientation was expressed as height of (1 11) peak x 1 0 0 % height of (1 1 1 ) peak + 3 x height of (200) peak' degree of (1 11) orientation = The last columns in tables 1-4 show average degrees of orientation of films after kinetic runs. It follows from tables 3 and 4 that all Pd-Ag alloys are well oriented, whereas for Pd-Cu films (tables 1 and 2) the degree of (1 11) orientation diminishes with the copper content, and, as shown in tables 1 and 2, for 56.0 and 54.4 at% Cu the existence of the intermetallic compound PdCu was established.However, since both alloys werez . KARPINSKI, w. JUSZCZYK AND J. STACHURSKI 1449 Table 1. Reaction of neopentane on (1 1 1) oriented Pd-Cu films turnover initial product distribution (% )* degree of (1 11) film reaction frequencya composition, temperature /molecule orientation' Cu (at%) /"C atom-'s-' CH4 C,H, C,H, i-C,H,, i-C,H,, (% 1 0 0 8.6 10.7 10.8 11.3 20.0 30.8 43.1 56.0 74.0 100 290 3.09 x lop3 286 3.31 x 10-4 299 6.61 x 10-4 286 3.52 x 10-4 304 8.31 x 10-4 285.5 1.38 x lop4 284.5 1.64 x lop4 309 6.17 x 292 3.12 x 10-4 310 6.10 x 10-4 297.5 3 . 4 0 ~ lop4 298.5 3.1 x lo-, 327.5 6 .2 4 ~ lop4 300 inactive 333.5 2.51 x 10-4 350 3.41 x 10-4 300 inactive 3 50 8 . 6 ~ 366 1.38 x 10-4 inactive up to 400 "C inactive up to 420 "C 10.4 11.7 15.2 16.4 5.5 7.0 1.5 4.7 14.1 1 .o 2.3 2.3 2.4 4.7 7.3 6. I 7.4 - - 2.6 6.2 2.1 3.1 8.3 12.5 traces 1.5 - 0.3 1.7 2.4 - 0.6 - 1.2 - 0.3 - 1.9 - 0.5 - traces traces 0.6 - 0.7 1.5 1.6 - 4.7 - 2.7 - - - - 25.1 29.2 25.0 49.2 11.3 15.6 4.7 7.3 29.2 3.9 5.3 5.4 9.9 11.6 - 7.6 15.1 - - 55.6 53.9 39.0 32.9 82.9 73.3 93.3 86.8 56.4 97.1 93.3 92.4 91.6 84.7 78.0 81.6 74.8 - - 100 100 not determined 100 100 100 100 99 82 f.c.c. (ca. 75 at % CU) phase + b.c.c. (PdCu) phase not determined 100 a Assuming that the roughness factor of all films is equal to 1; geometrical surface area 400 cm2. * Expressed as the percentage of a reactant consumed in the formation of a given product divided by total consumption.Defined in the text. catalytically inactive, the presence of the intermetallic compound will not be discussed further. We could not measure the surface composition of our Pd-Cu and Pd-Ag alloy films. Therefore for further interpretation of kinetic data it is necessary to use literature data on surface segregation in both alloy s y ~ t e r n s . ~ ~ - ~ ~ REACTION OF NEOPENTANE AND n-BUTANE WITH HYDROGEN OVER Pd-Cu( 1 1 1) AND Pd-Ag( 1 1 1) ALLOY FILMS These reactions were carried out in the presence of a ten-fold excess of hydrogen. The alkane partial pressure was ca. 1.85 Torr. Tables 1 4 report the results; the reproducibility of the results is estimated to be 20-30%.The calculation of turnover frequencies and initial product distributions selectivities was performed in the way described in ref. (2). Usually it was intended to work under very low conversion (1-3 % ) and so most of the experiments were carried out at only two temperatures (conversion at the third temperature was too high). In order to compare the catalytic behaviour of different alloys at one temperature it was necessary to make interpolations or even, in exceptional cases, extrapolations. 48 FAR 11450 ISOMERIZATION OF ALKANES ON Pd-Cu, Ag FILMS Table 2. Reaction of n-butane on ( 1 1 1) oriented Pd-Cu films turnover film reaction frequencya initial product distribution (% )h degree of (1 1 1 ) composition, temperature /molecule orientationC c u (at% ) /"C atom-ls-' CH4 C,H6 C,H, i-C4H,, (% 1 0 9.6 15.5 17.8 35.3 54.4 73.5 100 272 2.28 x 10-4 294.5 6.54 x 10-4 289 1.23 x 10-4 312 7.38 x 10-4 302.5 2.35 x 10-4 316 2.39 x 10-4 331.5 5.34 x 10-4 305 9.4 x 10-5 324 2.50 x 10-4 354 6.9 x 10-5 368 9.1 x 10-5 425 7.9 x 10-5 388 5 x 10-6 inactive up to 420 "C 32.6 13.6 43.0 26.6 24.6 40.5 19.6 4.6 34.2 18.4 8.4 34.6 27.1 6.6 39.9 21.0 6.3 42.6 20.6 9.1 41.4 26.5 5.5 29.3 23.1 6.6 29.0 5.8 13.5 18.4 10.4 12.6 24.8 25.4 25.4 49.3 17.8 36.4 38.5 10.9 8.3 41.7 38.6 26.4 30.1 28.9 38.7 40.9 62.3 52.2 7.3 - 100 100 100 95 100 mainly PdCu (b.c.c.phase)d 85e 100 a-c As in table I . Slightly marked superstructure lines. Most probably the f.c.c. alloy (not PdCu,), no superstructure lines. DISCUSSION Pd-Cu ALLOYS Table 1 shows the results for neopentane conversion over Pd-Cu(l11) films.As in the case of Pd-Au(ll1) alloys,2 selectivity towards isomerization is high but it now increases with copper content. These results do not confirm the existence of the 'hydrogenolytic' effect of copper in mixed Pd-Cu ensembles, as has been reported for copper-rich Pt-Cu alloy^.^-^ Unfortunately, the activity of the 74.0 at% Cu alloy was far below the detection limit. The fact that the degree of (1 1 1) orientation was worse with copper-rich alloys (table 1) does not seem to affect the catalytic properties. The isomerization activity for alloys having > 30 at% Cu is very low. Fig. 1 shows the rate of isomerization of neopentane on Pd-Cu(ll1) alloys at 300 "C. It is clear that the rate is similar to the rate of neopentane isomerization over Pd-Au( 1 1 1).2 Again, palladium alloys with ca.10 at % of a diluent metal (now Cu) are more active than pure Pd. It was argued previously2 that there is no serious surface segregation of the Pd-Au system and the same situation seems to exist for the Pd-Cu system. Van Langeveld et aZ.14 have studied the surface composition of Pd-Cu alloy films by A.e.s. and photoelectric work-function measurements. Their experimental results were in good agreement with the calculated values (assuming a broken-bond model) and showed only slight surface enrichment in copper. Assuming in the present case the absence of surface segregation (as in the case of Pd-Au alloys2) we conclude that there is no change in the catalytic activity of the palladium sites surrounded by copper atoms as compared with palladium adjacent to Au atoms.For n-butane isomerization, fig. 2 shows a similar relationship, suggesting the same mechanism for the isomerization of neopentane and n-butane. Again, alloys with ca. 10 at% Cu are more active than pure Pd and the relationship resembles the catalyticz. KARPINSKI, w. JUSZCZYK AND J. STACHURSKI 1451 Table 3. Reaction of neopentane on ( I 11) oriented Pd-Ag films turnover film reaction frequencya initial product distribution (% )' degree of (1 1 1) composition, temperature /molecule orientationC &(at%) "C atom-'s-' CH, C,H, C,H, I-C,H,, i-C5Hl, PA 1 0.8 317.5 335 1.9 316.5 337.5 6.1 338 359.5 9.2 360 392 12.8 333.5 365 2 . 9 0 ~ lo-, 3.0 4.42 x lo-" 9.0 8 .2 ~ 1.7 3 . 4 0 ~ 1O-O 1.7 1.75 x lop4 3.1 9 . 3 ~ 10-5 1.9 1.1 x 10-5 5.1 6.2x 10-5 7.3 2.5x 10-5 5.9 1.18 x 10-4 7.4 - 0.2 3.6 5.7 0.6 traces - traces - - traces 0.6 6.2 5.9 0.4 - - traces - - 8.6 88.2 100 18.2 62.6 6.1 91.6 99.5 8.6 89.7 8.7 89.4 I00 17.7 78.6 14.5 74.3 99.5 25.1 67.6 17.1 70.9 not determined 22.2 70.0 a- As in table 1. Table 4. Reaction of n-butane on (1 1 1) oriented Pd-Ag films turnover film reaction frequencya initial product distribution (%)b degree of (1 11) composition, temperature /molecule orientationC Ag (at % ) /"C atorn-ls-' CH, C,H, C,H, I-C,H,, (% 1 2.5 319 339 7.2 347 369 8.6 36 1 386 15.0 362 386 5.2 x 10-5 2.61 x 10-4 5.2 x 10-5 7.9 x 10-5 5.50 x 10-4 5.4 x 10-5 4.26 x 10-4 2.36 x 17.0 2.8 6.0 6.0 7.7 0.9 4.7 2.0 10.4 4.3 2.0 5.9 7.9 2.4 2.1 3.8 32.1 48.4 16.1 28.0 44.9 62.6 48.2 56.4 48.0 100 39.4 75.3 100 65.3 40.3 not determined 29.4 41.2 99.5 37.8 a-c As in table 1 behaviour of Pd-Au(ll1) films in n-butane reaction [fig.2 in ref. (2)]. Similarly the selectivity for isomerization increases with the content of inactive metal (table 2). Therefore replacing one ligand (Au) by another (Cu) in the alloy does not alter the catalytic behaviour of palladium very much. Whether this is proof of the absence of the ligand effect is still uncertain, but it is possible that by changing the ligand only minor changes in the electronic structure of the palladium alloy have been introduced. Uncertainty in the assessment of the surface composition and the lack of control of defect density do not allow small quantitative changes to be considered.Note that for one Cu-rich alloy (75.5 at% Cu) it was possible to observe predominant hydrogenolysis of n-butane (table 2). The result is in good agreement with the results published by Ponec and coworkers for P ~ - C U ~ - ~ and Pd-CulO alloy catalysts, where they argued against the ligand effect. It is possible that their arguments may be strengthened by the results obtained in this work, i.e. the same catalytic behaviour of Pd-Au and Pd-Cu alloys. However, the presence of the maximum in the isomerization activity against alloy composition curve still has to be explained. It is 48-21452 ISOMERIZATION OF ALKANES ON Pd-Cu, Ag FILMS 0 20 40 60 80 100 c u (at %) Fig. 1. Catalytic activity of Pd-Cu(ll1) alloys for the isomerization of neopentane at 300 "C: 0, this work; (b, from ref.(2); A, rough estimates from the activity at other temperatures (see table 1). I 0 0 20 40 60 80 100 Cu (at %) Fig. 2. Catalytic activity of Pd-Cu( 1 1 1) alloys for the isomerization of n-butane at 3 17 "C: 0, this work; A, rough estimates from the activity at other temperatures (see table 2). doubtful as to whether progress will be made without careful surface control in the sense of the surface composition, surface defects and the extent of surface carburization. Pd-Ag ALLOYS The situation with regard to the results for the Pd-Ag( 11 1) system is not as good, as we were limited to Pd-rich alloys (up to 10-1 5 at % Ag in the bulk) because larger amounts of silver deactivated palladium completely (tables 3 and 4 and fig.3 and 4). No maximum for the isomerization rate was found, but a rather dramatic decrease of the activity with the Ag content was noticed. The situation is more complicated than with Pd-Cu alloys because surface enrichment in silver is predicted by theory15 and has been confirmed by the majority of experimental studies.16-19 From recent work by Kuijers and Ponec19 it follows that the alloy of ca. 10 at% Ag in the bulk shouldz. KARPINSKI, w. JUSZCZYK AND J. STACHURSKI 1453 6 7 6 - I 0 2 5 3 5 \ $ 4 22 5 $ 3 E Y 2 1 0 0 5 10 15 Ag (at %) Fig. 3. Catalytic activity of Pd-Ag(ll1) alloys for the isomerization on neopentane at 317 "C. 0 0 5 10 15 Fig. 4. Catalytic activity of Pd-Ag( 1 1 1) alloys for the isomerization of n-butane at 344 "C.Ag (at %) have 40-80 at % Ag in the surface layer, depending on whether the alloy is equilibrated in vacuo (ca. 80%) or after interaction with CO (ca. 40%). Therefore one should not be surprised by the steep decrease in the isomerization activity with increasing silver content. The absence of a maximum in fig. 3 or 4 (cf. Pd-Cu and Pd-Au alloys) does not necessarily confirm the existence of the ligand effect. Such a maximum, if it exists,1454 ISOMERIZATION OF ALKANES ON Pd-Cu, Ag FILMS would be difficult to find (probably somewhere between 0 and 1 % Ag in the bulk). If we do suppose that a maximum exists for ca. 1 at% Ag, then, remembering that the lateral homogeneity of our films was not very good, we must consider the activity of the film as the average value of the ‘integrated’ activity, i.e.measured over the film with a reasonable concentration gradient (say between 0 and 2 at% Ag). Thus the activity maximum could be easily overlooked for the Pd-Ag system. Furthermore, no changes were found in the selectivity for isomerization for all three of the palladium systems: Pd-Au2 and Pd-Cu and Pd-Ag. This suggests that the ligand effect does not occur for alkane isomerization over palladium alloys. However, more careful study (with a good surface control, especially for Pd-Ag alloys) is needed in order to establish the origin of the synergistic effect in the isomerization of alkanes on alloys. We thank Prof. W. Palczewska for helpful discussion during the course of this research. This work was carried out as part of Research Project 03.10.Z . Karpinski, Nouu. J . Chim., 1980, 4, 561. Z . Karpinski, J . Catal., 1982, 77, 118. N. A. McKervey, J. J. Rooney and N. G. Samman, J . Catal., 1973, 30, 330; J. K. A. Clarke and J. J. Rooney, Adu. Card., 1976, 25, 125. V . Ponec, Adu. Catal., 1983, 32, 149. J. W. A. Sachtler and G. A. Somorjai, J . Cutal., 1983, 81, 77. H. C. De Jongste, F. J. Kuijers and V. Ponec, Proc. 6th Znt. Congr. Catal. (The Chemical Society, London, 1976), vol. 2. p. 915. H. C. De Jongste, V. Ponec and F. G. Gault, J . Catal., 1980, 63, 395. H. C. De Jongste and V. Ponec, Proc. 7th Int. Congr. Catal., Tokyo, 1980 (Elsevier, Amsterdam, 1981), vol. A, p. 186. M. J. P. Botman, H. C. De Jongste and V. Ponec, J . Catal., 1981, 68, 9. lo H. C. De Jongste and V . Ponec, J . Catal., 1980, 63, 389. N. Msrtensson, R. Nyholm, H. Calen, J. Hedman and B. Johansson, Phys. Rev. B, 1981, 24, 1725. I 2 J. A. Nicholson, J . D. Riley, R. C. G. Leckey, J. G. Jenkins and J. Liesegang, J . Electron Spectrosc. Relat. Phenom., 1979, 15, 95. I:’ S. Hiifner, G. K. Wertheim and J. H. Wernick, Phys. Reu. B, 1973, 8, 451 1. l4 A. D. Langeveld, H. A. C. M. Hendrickx and B. E. Nieuwenhuys, Thin Solid Films, 1983, 109, 179. 15 F. L. Williams and D. Nason, Surf. Sci., 1974, 45, 377. l 6 R. Bouwman, G. J. M. Lippets and W. M. H. Sachtler, J. Catal., 1972, 25, 350. l 7 B. J. Wood and H. Wise, Surf: Sci., 1975, 52, 151. I H F. J. Kuijers and V. Ponec, Ned. Tijdschr. Vacuumtech., 1978, 16, 318. l 9 F. J. Kuijers and V. Ponec, J . Catal.. 1979, 60, 100. (PAPER 4/ 1388)

ISSN:0300-9599    

DOI:10.1039/F19858101447

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 81, 1455-1457 Photon Correlation Spectroscopy of a Coagulating Suspension of Illite Platelets BY BRUCE E. NOVICH AND TERRY A. RING* Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 021 39, U.S.A. Received 4th September, 1984 The self-beat intensity correlation spectra scattered from coagulating suspensions of illite platelets have been analysed to determine colloidal stability ratios. The stability ratio results as a function of KCl concentration were used to evaluate the effective Hamaker constant for the illite platelets. The effective Hamaker constant obtained by this method compares favourably with those obtained by direct measurement on crystallographically identical mica plates, either in KNO, solution or, after applying the geometric mixing rule, in a vacuum.Photon correlation spectroscopy (P.c.s.) has been used routinely to measure the z-averaged translational diffusion coefficient of colloidal suspensions. In considering the initial stages of coagulation of a narrow size fraction of platelets it is found that singlets and doublets are present in the suspension. Thompson,l Bargeron2+ and Chu and coworkers4~ studied bimodal distributions of polystyrene latex spheres in water and observed that P.C.S. was sensitive to small numbers of larger particles in the presence of large numbers of small particles. With sufficiently precise data they were able to determine the two particle diameters and the ratio of scattering amplitudes. This approach is used to study the initial stage of coagulation.For a bimodal distribution the field autocorrelation function, g’(z), is given by g’(z) a exp (- rl T) +fexp (- Tz Z) (1) where T i is the decay constant with i = 1 for the singlet and i = 2 for the doublet. The relative strength of r2 with respect to rI6 is given by where Ni is the number density of particles and Ii(m, K ) is the scattering intensity per particle of relative refractive index m and at scattering vector K = (47c/i2) sin (8/2) for which ;I is the wavelength of light in the medium. The decay constant, Ti, is given Ti = . 9 i K 2 (3) where gi is the z-averaged diffusion coefficient. The diffusion coefficient can be related to the hydrodynamic diameter, di, by the Stokes-Einstein equation where k is Boltzmann’s constant, Tis the absolute temperature and y is the viscosity of the medium.The hydrodynamic diameter of the doublet can be approximated by the surface diameter of the doublet for low Reynolds numbers.’ Thus d2 = 4 2 4 . 14551456 P.C.S. OF ILLITE PLATELETS Before coagulation begins g'(z) will be given by the first term of eqn (1). A logarithmic plot of g'(z) gives I', as the slope and N , I,(m, K ) as the intercept.8 As doublet formation proceeds, the second term of eqn (1) increases from zero, since N, increases and N , decreases. If rl and r2 are not too different, g'(z) can be approximated by a single-exponential curve : ( 5 ) g'(z) cc exp ( - Tz). For a particular coagulation experiment we know I', and TZ. Noting eqn (1) we find that any change in the measured value of r is due to changes i n 5 assuming that the concentration of triplets and higher-order species is negligible.From eqn (2) we see that f is proportional to N,. The coagulation of singlets to form doublets obeys the following second-order rate law : = k,Nf dN, dt where k , is the rate constant. In the initial stage of coagulation we assume that N , is known and is essentially constant. Thus the initial slope offwith time is proportional to dN2/dt, allowing the rate constant k, to be determined to within a proportionality constant. When the proportionality constant [I,(m, K ) / N , Il(m, K)] is not known, as is the case for platelets, the experimental stability ratio, W, can be determined from the ratio of rate constants for rapid-, k,, and slow-, k,, coagulation conditions, since the two proportionality constants cancel : W = k,/k,.(7) Coagulation experiments were performed on a -0.5 pm fraction of natural illite from the Source Clay Repository (Imt-1). Using cumulant analysis the mean size was 0.39 pm and the standard deviation 0.14 pm. The particle concentration of the stock solution was determined by dry-weight analysis, using a density of 2.81 g ~ m - ~ . ~ The Malvern K7025 correlator used in this study was similar to those described in the literature.lO9 l1 A 632.8 nm line source was used with scattering at 90". The 120-channel correlator was controlled by a Commodore 3032 computer, which also analysed the correlation functions to yield at various coagulation times.Optimal operating parameters for the correlator were a 200 ps sample time and 2 x lo5 samples, which allowed data to be taken and analysed at a rate of one per min using a least-squares method. Selection of these parameters required a compromise between accuracy and speed of sampling and analysis. Speed was critical under unstable conditions; data for a measurement of were obtained sufficiently rapidly that the change in r was always < 10% of the previous value. Specific coagulation experiments were performed by diluting a stock sol into a solution having the desired salt concentration and pH, contained in a 3 cm3 glass cuvette. The cuvette was quickly capped, mixed and transferred to the sample chamber of the correlator for analysis. The correlator gave as a function of time.From the preceding analysis f (and therefore N,) was determined as a function of time, since rl and TZ were known. In no case were the data used withf > 0.5, assuring that the fraction of triplets and higher orders was negligible. The coagulation rate constant, k,, was evaluated from the initial slope, df/dt It-,, and N , using eqn (6). The stability ratio, W, was determined by eqn (7) for rapid- and slow-coagulation conditions. The KCl stability curve for illite at pH 10 is shown in fig. 1. The critical coagulation concentration Ccrit is the concentration corresponding to the intersection of the slow- and rapid-coagulation sections of the curve. Ccrit at pH 10 was 0.202 mol dm-3 KCl.B. E. NOVICH AND T. A. RING 1457 -2 I 0 log,, ([KCL] /mol dm-3) Fig.1. Stability ratio W( = k,/k,) plotted against KC1 concentration for illite at pH 10. The effective Hamaker constant was calculated for illite using an augmented Reerink and Overbeek12 analysis,13 which employed the parallel-plate form of the interaction energies. The derivation uses values of Ccrit, the specific surface area (89 m2 g-l)14 and the cation-exchange capacity (27 mequiv. per 100 g).14 The effective Hamaker constant for illite was 2.50 x J, which was within the experimental error of that measured by Israelachvilli and Adams15 for crystallographically identical mica plates in KNO, solution and equal to the value calculated using the geometric mixing rule16 and A,, = 1.3 x J, determined by Tabor and Winterton17 for mica plates in vacuum, and A,, = 4.38 x J, determined by Krupp et al.18 for water.We thank Dr R. Torrence Martin for helpful discussions and criticism during the preparation of this manuscript. D. S. Thompson, J. Chem. Phys., 1971,54, 141 1. C. B. Bargeron, Appl. Phys. Lett., 1973, 23, 379. C. B. Bargeron, J. Chem. Phys., 1974,60, 2516. S . P. Lee and B. Chu, Appl. Phys. Lett., 1974, 24, 201. F. C. Chen, W. Tscharnoth, D. Schmidt and B. Chu, J. Chem. Phys., 1974, 60, 1675. B. Chu, Laser Light Scattering (Academic Press, New York, 1974), p. 230. T. Allen, Particfe Size Measurement (Chapman and Hall, London, 3rd edn, 1981), p. 104. a D. E. Koppel, J. Chem. Phys., 1972, 57, 4814. * T. W. Lambe and R. V. Whitman, Soil Mechanics (Wiley, New York, 1968), pp. 29-40. lo H. Z. Cummings and P. W. Pusey, in Photon Correlation Spectroscopy and Velocimetry, ed. H. Z . l 1 G. R. Weise and T. W. Healy, J. Colloid Interface Sci., 1975, 51, 427. l2 H. Reerink and J. Th. G. Overbeek, Discuss. Faraday SOC., 1954, 18, 74. l 3 B. E. Novich and T. A. Ring, Cfays Cfay Miner., in press. l4 H. Van Olphen and J. J. Fripiat, Data Handbook for Clay Materials and other Non-Metallic Minerals l6 J. N. Israelachvilli and G. E. Adams, J. Chem. Soc., Faraday Trans. I , 1978, 74, 975. l6 J. Visser, Adv. Colloid Interface Sci., 1972, 3, 331. l7 D. Tabor and R. H. S. Winterton, Proc. R. SOC. London, Ser. A, 1969, 312, 435. l a H. Krupp, W. Schnabel and G. Walter, J. Colloid Interface Sci., 1972, 39, 421. Cummings and E. R. Pike (Plenum Press, New York, 1977), pp. 164-199. (Pergamon Press, New York, 1979), p. 346. (PAPER 4/ 153 1)

ISSN:0300-9599    

DOI:10.1039/F19858101455

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1985, 81, 1459-1465 Thermodynamics of Ethanol at Low Concentrations in Mixtures of Cyclohexane and 1,4-Dimethylbenzene BY HAN T. FRENCH AND ROBIN H. STOKES* Department of Chemistry, University of New England, Armidale, New South Wales 2351, Australia Receiued 2 1 st September, 1984 For the two-component system cyclohexane + 1,4-dimethylbenzcne we report (u) vapour- pressure measurements at 318.15 K and (b) enthalpy-of-mixing measurements at 288.15,298.15 and 318.15 K. From these data we obtain activity coefficients at 318.15 and 298.15 K. Several compositions of this system are then used as mixed solvents, in which the enthalpy of dilution of ethanol is measured from cu. 0.1 mole fraction of ethanol down to below 0.001 mole fraction. The limiting partial excess enthalpy of ethanol is thus obtained as a function of the mole fractions of cyclohexane and 1,4-dimethylbenzene in the mixed solvent. This function is far from linear, and the actual values are compared with those calculable from the single-solvent data previously published.Previous papers from this laboratory'. have presented and interpreted measure- ments of activity coefficients and excess enthalpies for solutions of ethanol (1) in cyclohexane (2) and in 1,4-dimethylbenzene (3). Comparisons of the limiting values in various solvents of the partial excess enthalpy of ethanol and of the activity coefficients of ethanol at vanishing ethanol concentration are of particular interest because we are in this case dealing with the interactions of the free ethanol molecule with the solvent, uncomplicated by hydrogen bonding between alcohol molecules.However, because this hydrogen bonding is relatively strong, it is necessary to carry the measurements to very low alcohol concentrations ( < 0.001 mole fraction) in order to obtain valid extrapolations to infinite dilution. (The extent of the non-ideality of these solutions is similar to that of 2: 2 electrolytes in water.) The limiting partial excess enthalpies show a sharp contrast between aliphatic and aromatic solvents. In h e ~ a n e , ~ hexadecane4 and C5-, C,- and C,- c:ycloparaffins3 at 25 "C the range is only from 23 to 24 kJ mol-l. In benzene3 and 1,4-dimethylben~ene~ the values are, respectively, 15.8 and 15.6 kJ mol--l. It was shown in an earlier paper2 that the properties of ethanol in 1,4- dimethylbenzene at moderate concentrations (up to ca.0.1 mole fraction of ethanol) could be calculated rather well from those of the cyclohexane solutions as interpreted by the association model using the same association constants, combined with an enthalpy and equilibrium constant for the solvation of free hydroxyl group by the n-electron system. A further test of this model is now made by examining the limiting partial excess enthalpies of ethanol in mixtures of the two solvents. The theoretical treatment of these requires knowledge of the activity coefficients of the components of the mixed solvent, which were also measured. 14591460 THERMODYNAMICS OF ETHANOL Table 1. Liquid-phase mole fraction of 1,4-dimethylbenzene, total pressure, activity coefficients and excess Gibbs free energy for the binary system cyclohexane(2) + 1,4- dimethylbenzene(3) at 3 18.15 K 0.030 I 1 0.050 23 0.075 99 0.100 34 0.150 9 0.201 6 0.251 2 0.301 8 0.351 2 0.401 1 0.450 7 0.451 9 0.491 0 0.533 5 0.571 3 0.609 0 0.651 3 0.698 4 0.750 6 0.801 1 0.850 6 0.899 3 0.926 0 0.949 4 0.964 84 0.982 29 29.222 28.716 28.096 27.532 26.352 25.198 24.074 22.9 10 21.774 20.605 19.432 19.392 18.436 17.392 16.429 15.448 14.319 13.016 11.514 10.007 8.463 6.881 5.986 5.200 4.664 4.041 0.3481 0.3433 0.3213 0.2942 0.2538 0.2 157 0.1824 0.1545 0.1294 0.1077 0.0868 0.0868 0.0736 0.0598 0.0495 0.0404 0.03 14 0.0229 0.01 55 0.0096 0.0055 0.0027 0.0018 0.0006 0.0002 0 0.0008 0.0010 0.0026 0.0052 0.01 10 0.0191 0.0288 0.0395 0.05 16 0.0647 0.0797 0.0797 0.09 14 0.1060 0.1 186 0.1318 0.1471 0.1646 0.1842 0.2045 0.2239 0.2437 0.2534 0.2705 0.2804 0.2859 29.8 48.2 70.8 90.4 125.9 155.3 178.2 196.2 208.8 216.7 219.2 219.3 218.7 215.1 209.3 201.3 189.7 173.7 152.3 128.0 100.9 71.2 53.9 37.8 26.5 13.4 EXPERIMENTAL Materials were purified as described in our previous papers.Vapour pressures were measured in a continuous-dilution apparatus and converted into activity coefficients by the methods of Barker5 and Marsh.6 These measurements (table 1) were made only for the binary (2 + 3) system, not for its mixtures with ethanol, as the analysis of three-component systems on the basis of total vapour pressure only is of doubtful reliability. The enthalpies of mixing of the (A + B) system were measured at three temperatures in isothermal displacement calorimeters.6 The experimental results are given in table 2.The activity coefficients of I .4-dimethylbenzene were converted to 298.15 K as follows. First the HE data at each temperature were fitted to polynomials of the form H E / x , x , = Aix;. (1) i=O,4 The partial excess enthalpies of 1,4-dimethylbenzene were calculated for the round mole fractions of interest at each temperature (table 3). The comparatively small temperature variation of HF means that an average value over the range 298-3 18 K is sufficiently accurate to convert the lnf3 values of table 1 into values at 298.15 K. The values at round concentrations are given in table 4 as lnfi. To obtain excess enthalpies of ethanol in mixtures of 1,6dimethylbenzene + cyclohexane the mixed solvents of round mole fraction composition were prepared in the isothermal displacement calorimeter by adding the calculated volume of one component to known volumes of the other.H.T. FRENCH AND R. H. STOKES 1461 Table 2. Enthalpies of mixing of cyclohexane(2) and 1,4-dimethylbenzene(3) 0.018 41 0.037 06 0.070 78 0.102 1 0.1589 0.2094 0.006 99 0.020 24 0.033 45 0.044 24 0.055 9 0.077 2 0.112 5 0.145 5 0.171 5 0.197 7 0.221 5 0.243 5 0.018 64 0.037 41 0.071 64 0.103 2 0.161 0 0.212 0 3.018 2.967 2.9 14 2.857 2.763 2.683 2.900 2.854 2.849 2.836 2.814 2.782 2.720 2.660 2.619 2.571 2.538 2.507 2.755 2.732 2.68 1 2.629 2.540 2.468 0.2541 0.3 124 0.3622 0.4 184 0.4655 0.4884 0.2639 0.2830 0.3026 0.3209 0.3380 0.3530 0.3681 0.3934 0.4148 0.4357 0.4545 0.4709 0.2566 0.3151 0.365 1 0.42 16 0.4687 0.4938 T = 288.15 K 2.618 0.4461 2.541 0.4683 2.479 0.5160 2.414 0.5748 2.362 0.6282 2.340 0.6926 T = 298.15 K 2.481 0.4897 2.456 0.5010 2.431 0.5002 2.408 0.5171 2.388 0.5362 2.370 0.5569 2.352 0.5771 2.323 0.5923 2.299 0.6120 2.277 0.6310 2.258 0.6513 2.242 - T = 318.15 K 2.406 0.4466 2.317 0.4687 2.264 0.5164 2.207 0.5696 2.157 0.6284 2.135 0.6928 2.376 2.352 2.303 2.247 2.200 2.148 2.224 2.21 1 2.2 10 2.194 2.176 2.156 2.138 2.121 2.107 2.092 2.074 - 2.169 2.146 2.102 2.056 2.009 1.964 0.7428 0.8024 0.8712 0.9105 0.9520 0.9761 0.6738 0.6959 0.7194 0.7451 0.7734 0.8052 0.8461 0.8887 0.9304 0.95273 0.98389 - 0.7435 0.8024 0.8709 0.9104 0.9532 0.9761 2.109 2.066 2.020 1.999 1.977 1.974 2.057 2.040 2.021 2.003 1.981 1.959 1.931 1.905 1.883 1.860 1.826 - 1.926 1.891 1.853 1.839 1.832 1.852 Table 3.Partial excess enthalpies of 1,4-dimethylbenzene(3) in mixtures with cyclohexane(2) T/K x3 = 0.0 0.1 0.25 0.50 0.75 288.15 3070 2186 1280 447 98 298.15 2920 2079 1213 430 91 318.15 2780 2003 1166 408 93 The quantities were so chosen that a few cm3 of mercury remained in the mixing vessel. After temperature equilibration of the solvent, ethanol was added in a normal enthalpy-measurement run. A 5 cm3 piston burette was used with an electronic revolution counter on the motor drive giving 2240 counts per cm3. This run was stopped at ca. 0.1 mole fraction of ethanol and a quantity of the final solution was drawn back into another burette. A fresh batch of solvent of the same composition as the first was then prepared, and the stock solution from the first stage was added to it in a second-stage dilution run.1462 THERMODYNAMICS OF ETHANOL Table 4.Limiting partial molar excess enthalpies of ethanol( 1) in cyclohexane(2) and 1,4-dimethylbenzene(3) and their mixtures at 25 “C AKransfer /kJ mol-’ lim HF /kJ mol-l 1nX obs. calc.a 0.00 24.0 0.420 (0) (0) 0.10 21.6 0.344 - 2.4 - 2.3 0.25 20.2 0.21 1 -3.8 -4.2 0.50 18.5 0.08 1 - 5.5 - 6.0 0.75 16.5 0.017 - 7.5 - 7.4 1 .oo 15.6 0.000 - 8.4 - 8.4 a Calculatedbyeqn(8)withQ = 1.49; h, = - 12.6 kJ mol-1;(D13-D12) = -0.9 kJ mol-l; a3 = x&O. Calculation of the results was done in the usual way, with the weighted mean molar mass of the solvent used in the same way as the molar mass of a pure solvent.Thus the excess enthalpies given in table 5 refer only to the addition of ethanol to the solvents already mixed. If the total enthalpy of mixing of all three components is required it can be evaluated by combining the data of table 2 with those of table 5. The excess enthalpies in the second dilution step were calculated as usual from the relations (2) H”/x, = H”/x,(stock solution) + AH/n, where A H is the enthalpy increase on adding to the mixed solvent an amount of stock solution containing n , moles of ethanol. RESULTS AND DISCUSSION Fig. 1 shows the limiting partial excess enthalpy of ethanol at infinite dilution in the pure and mixed solvents as a function of solvent composition. This graph is more conveniently regarded as giving the enthalpy of transfer of a single non-hydrogen- bonded ethanol molecule from cyclohexane to the mixed solvents or to pure xylene, as shown by the right-hand set of ordinates.We now examine the extent to which its shape can be accounted for in terms of a ‘solvation’ of the free hydroxyl group by the aromatic 71 system. According to that model, the monomeric ethanol and 1’4-dimethylbenzene are subject to a ‘ solvation ’ equilibrium with equilibrium constant Q and enthalpy of solvation h,. It follows that at infinite dilution of ethanol in a pure or mixed solvent in which the 1’4-dimethylbenzene has activity a,, the fraction of solvated ethanol molecules is Qa,/( 1 + Qa,) and the limiting enthalpy of transfer is h, Qa,/( 1 + Qa,). Using the values of h, and Q arrived at in the previous work along with the 1,4-dimethylbenzene activities of table 4 gives the broken curve shown in fig.1. This indicates a more extreme curvature than the experimental data, and we conclude that the model is oversimplified. There are two natural ways of improving it: by taking account of entropy effects due to the three molecules involved and by allowing for differences in the van der Waals interaction energy between ethanol and the respective solvent components. [In the earlier treatment the same interaction parameter (6 J ~ m - ~ ) was used for both solvents.] According to the Scatchard-Hildebrand model, using volume-fraction statistics for the entropy contribution, the limiting activity coefficient of a species i (molar volume 6 ) at infinite dilution in a solvent of molar volume V, is lim lnfi = In (F/ V,) + D Y,/RT (3)H.T. FRENCH AND R. H. STOKES 1463 Table 5. Enthalpy changes for the addition of ethanol( 1) to mixed solvents comprising 1,4-dimethylbenzene(3) and cyclohexane(2) (see footnotes for symbols) 0.000 42 0.000 927 0.001 90 0.002 90 0.003 79 0.004 53 0.005 05 0.000 423 0.000 863 0.001 284 0.001 654 0.002 08 0.002 43 0.002 87 0.003 21 0.004 00 0.001 18 0.002 22 0.003 08 0.004 10 0.005 1 1 0.005 97 0.006 91 0.007 86 0.001 63 0.002 20 0.002 65 0.003 37 0.003 93 0.004 65 0.005 78 21.5, 21.60 21.5, 21.3, 20.9, 20.4, 20.0, 20.1, 20.19 20.11 19.8, 19.7, 19.7, 19.6, 19.5, 19.3, 18.3, 18.1, 17.8, 17.6, 17.4, 17.2, 16.9, 16.4, 16.3, 16.3, 16.2, 16.1, 16.0, 15.9, 17.9, x; = 0.1 0.006 64 18.91 0.007 93" 17.97 0.008 46 17.57 0.010 82 15.98 0.013 02 14.67 0.015 10 13.66 0.0 16 67 13.13 0.004 71 19.10 0.005 47 18.82 0.006 16 18.53 0.006 85 18.22 0.007 58 17.87 0.008 27 17.54 0.009 56 16.90 0.010 81 16.28 0.01 1 26 16.08 0.008 72 16.75 0.009 08" 16.53 0.010 77" 16.14 0.012 24" 15.72 0.012 48" 15.63 0.013 97" 15.18 0.01 5 46" 14.78 0.016 18" 14.52 0.006 75 15.85 0.007 65 15.70 0.008 63 15.55 0.009 09 15.49 0.011 14 15.17 0.011 64a 15.04 0.013 01 14.83 X; = 0.25 xi = 0.5 xi = 0.75 0.017 28 0.031 70a 0.046 37" 0.060 95" 0.076 86" 0.094 O l a 0.012 05 0.013 22a 0.022 76a 0.034 78" 0.044 43a 0.056 59" 0.068 36" 0.075 44a 0.0 17 94 0.031 82" 0.046 30a 0.058 29" 0.065 69a 0.073 12b 0.084 82b 0.098 72b 0.023 50a 0.034 66a 0.046 48a 0.057 37a 0.078 1" 0.087 9a 0.097 0" 12.76 9.27 7.49 6.43 5.66 5.08 16.08 15.18 11.89 9.51 8.30 7.24 6.51 6.16 14.03 11.10 9.18 8.1 1 7.59 7.07 6.51 5.98 12.98 11.08 9.74 8.79 7.50 7.04 6.68 xi is the mole fraction of 1,4-dimethylbenzene in the mixed solvent before the addition of ethanol and x1 is the mole fraction of ethanol in the three-component mixture, so that the solvent components are then at mole fractions x, = xi( 1 -x,) and x, = (1 -xi) (1 -xl).Superscript letters (a) or (b) denote runs in which pure ethanol was added to solvent; where no superscript is present the point is from a run in which a stock solution of ethanol was added to the solvent.1464 THERMODYNAMICS OF ETHANOL 0.0 0.2 0.4 0.6 0.8 1.0 x; Fig. 1. Limiting partial excess enthalpies of ethanol (1) at infinite dilution in cyclohexane(2) and 1,4-dimethylbenzene(3) and their mixtures at 25 "C.0, measured values; (-) and (---) calculated from equilibrium solvation model with and without van der Waals term, respectively. x: is the mole fraction of 1,6dimethylbenzene in solvent. Left-hand ordinate, lirn (x, + 0) HF; right-hand ordinate, AG (transfer) = lim HF(3) - lim HF(2). where D is an interaction energy density (derivable in principle, but not accurately enough in practice, from the cohesive energy densities of the pure liquids). Using this expression for the known limiting activity coefficients of ethanol in cyclohexane and in 1,4-dimethylbenzene at 25 "C, and allowing for the solvation in the latter solvent, we have for the free energy of transfer of ethanol(1) from cyclohexane(2) to 1,4-dimethylbenzene( 3) AG,,,,,f,,/RT = lirn 1nf1(3) - lirn 1nf1(2) = 2.939 -4.23 1 = - 1.342 (4) so that from eqn (3) (D13 - Dlz) V,/RT- In (1 + Q ) = 1.277.Trial suggests a value of -900 J mol-1 for (D3-D2)K; this is of reasonable magnitude and leads to Q = 1.49 at 25 "C. For the limiting excess enthalpies of ethanol in the two solvents lim H33) = AHdiss +D13 V, + h, Q / ( 1 + Q) = 15.56 kJ mol-l ( 5 ) 21-+Q lim H32) = AHdiss+Dl2 V, = 24.02 kJ mo1-l x1-a where AHdiss represents the enthalpy change on converting 1 mol of pure alcohol to a pure-liquid monomeric form, Subtracting and using the assumed value of -900 J mol-l for (D13 -D12) V,, we find for the enthalpy of solvation h, = - 12.6 kJ mol-l. (7)H. T. FRENCH AND R.H. STOKES 1465 This is reasonably consistent with - 13.0 kJ mol-l obtained from the temperature dependence of Q, using the data from 13.3 to 45 "C for the two solvents. We can now calculate the limiting enthalpies of transfer from cyclohexane to the mixed solvents, using for the solvation contribution and assuming the van der Waals contribution to be linear in the mixed solvent composition : (8) AHtransfer = hs Qa3/(1+ Qa3) + (Di3 - DiJ V, xi- Table 5 and fig. 1 show that the calculated results agree quite well with the experimental ones. CONCLUSION The enthalpies of transfer of ethanol at infinite dilution from cyclohexane to mixtures of cyclohexane and 1,4-dimethylbenzene can be accounted for in terms of an equilibrium solvation of the hydroxyl group by the aromatic molecule provided that a reasonable allowance is made for the changing van der Waals interaction. This work was made possible by a grant under the Australian Research Grants Scheme. We thank Sean O'Shea for assistance with the enthalpy measurements. R. H. Stokes and M. Adamson, J. Chem. SOC., Faraday Trans. 1, 1977, 73, 1232; R. H. Stokes, J . Chem. SOC., Faraday Trans. 1, 1977, 73, 1140. R. H. Stokes and H. T. French, J . Chem. SOC., Faraday Trans. 1, 1980, 76, 537. R. H. Stokes and C. Burfitt, J . Chem. Thermodyn., 1973, 5, 623. H. T. French, A. Richards and R. H. Stokes, J. Chem. Thermodyn., 1979, 11,671. J. A. Barker, Austr. J. Chem., 1953, 6, 207. K. N. Marsh, Trans. Faraday SOC., 1968, 64, 883. (PAPER 4/ 1634)

ISSN:0300-9599    

DOI:10.1039/F19858101459

出版商:RSC    

年代:1985

数据来源: RSC