摘要:

J. Chem. SOC., Faraday Trans. I, 1987,83, 1621-1629 Self-consistency of the Percolation Model as applied to a Macrofluid-like Water-in-Oil Microemulsion Rolf Hilfiker and Hans-Friedrich Eicke" Institut fur Physikalische Chemie der Universitat Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland A self-consistent picture of the structure and structural changes in a water-in-oil microemulsion (as a model of supermolecular fluids) close to the percolation threshold has been established. The self-consistency is evidenced by the demonstration that light-scattering and viscosity data for these systems can be quantitatively predicted by a model based on fractal cluster distribution. The model parameters were determined from electro-optic Kerr-effect measurements. Complex fluids (i.e.supermolecular liquids or one-component macrofluids) have recently given rise to unexpected interest from chemists and physicists as an easily accessible model for testing current liquid-state theories. An example of such a fluid evolved from the study of water-in-oil (W/O) ' microemulsions ' using three-component [water, Aerosol OT (AOT) and oil (e.g. iso-octane)] liquid-liquid dispersions. We have already claimed'? that this system represents a reasonably ideal case of a one-component ma~rofluid.~ All the above-mentioned terminology expresses the fact that the colloidal particles (i.e. the aqueous nanophases) are the only 'macroscopic' objects in the suspension; in other words, the solvent and other dissolved species are taken to possess no structure and no pairwise correlation either with each other or with the nanophases.A large variety of different phenomena have been observed with these fluids. A particularly striking observation is the occurrence of percolation well below the phase separation of the system, where the nanophases segregate from the supernatant oil. The electric conductivity, electric birefringence, light scattering, dynamic viscosity and other physical parameters display in a more or less pronounced fashion the percolation of the nanodroplets. Recently, we succeeded in describing the time-dependent relaxation of the electric birefringence close to the percolation threshold using a scaling m0de1.~ The fractal nature of the clusters in the percolation regime will prove to be of considerable importance. This property has been confirmed experimentally, e.g.by permittivity measurements5 which are particularly sensitive to a change of dimensionality and is also predicted theoretically.6T Here we report and interpret Kerr-effect, light-scattering and viscosity measurements on the water-AOT-iso-octane system. In view of the simplicity and high precision of Kerr-effect measurements, very reliable data about these complex fluids can be obtained. We maintain that, from a physical model of the system which correctly interprets the Kerr-effect results, a self-consistent picture of these fluids should evolve. One should expect accordingly that the basic assumptions which have to be introduced to interpret the birefringence data should also predict results of quite different experimental techniques which are applied to the study of these supermolecular fluids. In particular, we show that a description of the system using parameters derived from Kerr-effect measurements quantitatively predicts the results of light-scattering and viscosity measurements.16211622 Structure of a Water-in-Oil Microemulsion Experiment a1 Preparation of Complex Fluids (Supermolecular Liquids) Ternary mixtures of water, Aerosol OT (sodium di-2-ethylhexyl sulphosuccinate, AOT), and iso-octane (2,2,4-trimethylpentane) were prepared. Water was deionized and doubly distilled ; iso-octane was of the highest grade commercially available from Fluka, Switzerland, and used as received. Aerosol OT from Fluka was purified as described in ref. (1).A constant mole ratio of H20 and AOT, i.e. [H,O]/[AOT] = w, = 60 was used throughout this work. From previous work1 it is known that under these conditions a colloidal liquid-liquid dispersion with aqueous nanodroplets of a remarkable mono- dispersity in the size range of typically 10 nm radius is formed. The droplets are covered by a monomolecular layer of AOT. The system is thermodynamically stable and responds reversibly to variations of temperature and/or changes in composition. Instrumental The principal arrangement to conduct electro-optic Kerr-effect measurements is de- scribed in ref. (4). Light-scattering experiments were performed with a small-angle laser-light-scattering photometer. The light source was a polarized 5 mW He-Ne laser (A = 632.8 nm). The scattering angle and volume in air were 5.8" and 8.6 x cm3, respectively.The scattering cell was equipped with two 5 cm thick quartz windows, to avoid contributions to the scattering due to reflexions from the cell and the air/quartz interfaces. Thermostatting of the scattering cell was possible within f 0.05 K. Refractive- index increments were determined with a Brice-Phoenix differential refractometer (BP-2000 V), also employing a He-Ne laser. Theoretical Considerations and Results According to our introductory remarks we consider electro-optic Kerr-effect experiments a particularly suitable starting point for discussing the self-consistency of the complex fluid model. The most substantial information at present about the H20-AOT-iso-octane system stems from the time-dependent rise (or decay) of the electric birefringence.Since we interpret the Kerr effect as resulting from electric- field-induced orientation of nanodroplet cluster^,^ the distribution of cluster sizes must be determined. This distribution is predicted by a theoretical ansatz of Stauffer.'j, According to this author, the number of s-clusters per unit volume, N,, at a given concentration is related to the cluster size, s, by the relation N, = c, s-@ exp ( - c, d) (1) with 0 = 1.5 and = 1.0 for nanodroplet concentrations below the critical percolation concentration (ccrit). c, and c, are constants. The nanodroplet clusters are characterized by the fractial dimension p - l (p = 0.5 for three-dimensional systems and for which c < Ccrit). Accordingly, we express the cluster radius, R,, as R, = C ~ S P (2) where ck is a constant.The Kerr constant (B) is proportional to the square of the difference between the refractive indices of the solvent (nsolv) and the droplet clusters of mass s(n,), as well as to the square of the cluster volume (K):8R. HilJiker and H-F. Eicke 1623 The fractal dimension of the fully solvated nanodroplet clusters implies that their indices of refraction (n,) become dependent upon the cluster size. While the refractive index of a ‘cluster’ with s = 1 is equal to that of a single nanodroplet (rind), ns tends towards the refractive index of the dispersion medium with increasing cluster size, owing to increasing inclusion of oil. If we base our considerations on the ‘one-component macrofluid’ concept, the refractive index of the nanodroplet cluster can be expressed by n s = A d rind + 4solv nsolv (4) where &,d and 4solv are the volume fractions of the nanodroplets and the solvent within the cluster, respectively.The cluster volume can be related to the cluster size and the volume of one nanodroplet, & = CRS3P5. ( 5 ) vo’ by cR is directly available from ck in eqn (2). From eqn (4) and the relation for the volume fraction 4nd = s&/ VS, the difference in refractive index is finally derived: n, - nsolv = cgl slW3p(nnd - nsolv). (6) Inspection of eqn (3), (5) and (6) indicates that the Kerr constant scales with the square of the cluster size. We are now in a position to describe the time-dependent rise of the electric birefringence after application of an electric field.For this we assume that: (i) the birefringence is due to an orientation of nanodroplet clusters via a Debye mechanism, (ii) the cluster distribution function given in eqn (1) is appropriate, and take into account that (iii) An (cc B) is proportional to s2. The expression that results is*: (7) with z, = znds3’, where Tnd is the Debye relaxation time of a single nanodroplet. The latter can be evaluated using the hard-sphere radius of the droplet and the viscosity of the solvent. c1 is a constant. By measuring Anrise(?) and assuming z, as above we are able to determine the concentration- and temperature-dependent quantity c2. This parameter is related to the typical cluster size, styp, which is in turn connected with the relative concentration of nanodroplets with respect to their critical percolation concentration (cCrit).Stauffers has shown the relationship between typical cluster size and the weighed-in concentration of co Anrise(?) = [ 1 - exp (- t / z s ) ] c, s2-@ exp ( - c, d) s-1 droplets (c) to be where (r = 0.46 for three dimensions. It is now most relevant4 to the following that Styp cc c;W (9) The steep rise of sty* with concentration close to the percolation transition, which can also be induced by a temperature increase, is displayed in the inset of fig. 1. Thus, combination of eqn (8) and (9) allows one to establish a relation between c, and c: with C/o = 2.17. The experimentally determined c, values can be fitted perfectly with an equation of (1 1) the form c - Ccrit( T) 2.17 I - c,(c, T ) = a+bl Cwit(T) * A closer examination showed that x in ref.(4) is 1.1624 Structure of a Water-in-Oil Microemulsion I I I 0.1 0.2 0.3 clg ~ r n - ~ Fig. 1. Concentration dependence of styp (inset) and c, [see eqn (ll)] for the system H , G AOT-iso-octane ([H,O]/[AOT] = 60); T = 296 (m), 298 (A), 300 (0) and 301 K (0). \ 0 \ 0 \ T 296 290 30 0 302 TIK Fig. 2. Temperature dependence of the critical percolation concentration, cCrit, as obtained from fig. 1 and eqn (1 1). The constant a was found to be 0.094, independent of temperature. Apart from the fact that this constant does not affect any of the conclusions to be drawn, the finite value of a is probably an artifact produced by the time domain of the experiments: the larger the cluster size, the greater the probability for their disintegration during the field-induced orientation process.The validity of the explanation was proved experimentally, as the value of a decreased when the observation time was shortened: a larger number of clusters remain intact and contribute fully to the birefringence amplitude. Fig. 1 shows the c, dependence on the weighed-in concentration of nanodroplets. The various plots correspond to different temperatures. The fitted parameters determine sensitively the critical percolation concentration cCrit, for a particular nanodroplet size (in the present case w, = 60), presented in fig. 2. The value of cCrit depends on the interaction between the particle^;^ an attractive interaction decreases cCrit. The observedR.Hiljiker and H-F. Eicke 1625 decrease of cCrit with increasing temperature reflects an increasingly attractive interaction between the nanodroplets. At high temperature this results eventually in the observed phase separation of the macrofluid. Discussion Predictions Concerning the Concentration-dependent Light-scattering and Dynamic-viscosity Data of Complex Fluids Using the experimentally determined and characterized parameter c2(c, T), we now extend our observation to predict typical patterns of the concentration-dependent light-scattering and viscosity data of one-component macrofluids. The results will prove to be of more general validity than might have been expected from the particular example studied in these experiments. Small-angle Light-scattering as a Function of Concentration A combination of the cluster distribution function below the percolation threshold with the hard-sphere approximation of Percus-Yevick, including a suitable polydispersity correction, allows one to calculate the concentration dependence of the scattering intensity for small scattering angles.The reduced scattering intensity Ri (= Re, solution - Ri, solvEnt) of monodisperse hard spheres can be calculated in the Percus-Yevick approximation, as R;I = KCM[(I -4)41/[(1 +2w1 (12) where K = 4n2n2((an/ac)2/(L; N) = K'(an/ac)2, c is the weighed-in concentration of solute (H20 and surfactant) in mass per unit volume, A4 is the molecular weight and 4 the volume fraction of the hard spheres, A, the vacuum wavelength of the scattered light, n the refractive index of the solution, anlac the refractive index increment and N Avogadro's number.A polydispersity of the scattering entities can be considered by modifying egn (12), giving (13) O0 an ~ b , polydisp = [ zl ( z)s ~s cs] [(I - 4app)*/(l+ ~ p p ) 2 1 PPD where the quantity in the large brackets refers to clusters of size s, and PPD is a polydispersity correction which can be calculated if the size distribution of clusters is known. Note that, instead of the volume fraction of hard spheres, 4, which appears in eqn (12), eqn (13) contains $app, the apparent volume fraction of fractal clusters. In this assumption the cluster volume is no longer proportional to s, but to s3p; thus 4app is larger than 4. The total volume fraction of clusters is equal to the sum of the volume fractions of the individual cluster sizes, and substituting eqn (1) yields The appearance of & in eqn (14) can be eliminated using the mass balance: 00 x N,s&=4 8-11626 Structure of a Water-in-Oil Microemulsion where 5 is the volume of a single nanodroplet and 4 the volume fraction of nanodroplets.4 is known from the weighed-in amounts of water and surfactant and the corresponding partial molar volumes of the components. Combining eqn (14) and (1 5) yields a2 c s3p-@ exp ( - c, si) s-1 Eqn (1 6) allows one to calculate fractal cluster volume fractions. The ratio of the apparent and weighed-in volume fraction of nanodroplets is called a, where which is straightforwardly determinable from experiments. The refractive-index incre- ment is determined in the following way. Making the same assumption as for eqn (4) and substituting c = 4d (where d is the density), we obtain The refractive index increment is thus For the case s = 1, (an/&), = (1 /dnd) (nnd -nsolv), where (an/&), is the experimen- tally determined refractive-index increment.Substituting eqn (6) into eqn (19) yields We are now in a position to evaluate all quantities occurring in eqn (1 3) with the help of the cluster size distribution function below the critical percolation concentration [eqn (1 l)]. The molecular weight of an s-cluster follows from eqn (5): where M , is the molecular weight of a single nanodroplet. The concentration of an s-cluster in mass per unit volume is simply N, and M, are obtainable from eqn (1) and (21).Eqn (l), together with the condition that the weighed-in concentration (in mass per unit volume) of water and surfactant (co) C, = N,M,/N. (22) can be described by a2 c0 = C sNsMl/N (23) s=1 makes it possible to express the constant c, [from eqn (l)] as coN/Ml 2 sl-@ exp ( - c, sc) a2 c, = S-1 From eqn (l), (22) and (24) we find for the concentration of s-clusters CR (ds/dnd) C0S3P-@ exp ( -C2 S') a2 c, = C sl-@ exp ( - c, si) s-1R. HilJiker and H-F. Eicke 1627 A 40 - - I - mz 30 - I E 2 20- 1 is' s N 0 M 1 - 0 10 - 8 8\ \ 8. \ I I 1 0.1 0.2 0.3 clg ~ r n - ~ Fig. 3. Light-scattering results for a H,O-AOT-iso-octane system, with [H,O]/[AOT] = 60 and temperature as a parameter. T = 296 (e), 298 (o), 300 (A) and 301 K (0). The curves are theoretical predictions using parameters obtained from Kerr-effect measurements (see fig.1). Finally, we have to consider the polydispersity correction which was introduced by Pusey et aI.l0 into the Percus-Yevick approximation of the hard-sphere model. The factor PPD in eqn (1 3) can be calculated from where the A,, are the vth moments of the radial distribution functions.1° With the distribution function of cluster sizes from eqn (1) and the relation between cluster size and radius from eqn (2), the A,, can be calculated using Having compiled all quantities necessary to calculate the relative scattering intensity of a complex fluid, we arrive at a generalization of eqn (1 2) : 03 s-1 Fig. 3 represents the experimental light-scattering data with plots of eqn (28). The agreement between the theoretically predicted curves and the data is indeed remarkable.There is only one variable parameter involved [cR, see eqn (16)]; this parameter is, however, restricted to be temperature-independent and limited to a range of values set by viscosity measurements. Hence the degree of self-consistency is certainly striking. Moreover, particular features are correctly predicted, such as the shift of the scattering1628 Structure of a Water-in-Oil Microemulsion 2.0 a 1.5 1 / /Ir / A / A 0 / A a 1.0 I I I I 0.1 0.2 0.3 4 Fig. 4. Values for a = (q5app/4) obtained from viscosity. (0, 296 and A, 298 K) and Kerr-effect measurements (0, 296 and a, 298 K) of a H,O-AOT-iso-octane system with [H,O]/[AOT] = 60. intensity maxima with increasing temperature to lower weighed-in concentrations of water and surfactant. It should be emphasized that the increase of the scattering power with temperature by a factor 1.5 is surprisingly low, in view of the strong increase of styp by a factor of ca.3. This is a direct consequence of the fractal nature of the clusters: the apparent volume fraction of clusters close to the percolation threshold [see eqn (1 6)] is much larger than the weighed-in volume fraction, which leads according to the Percus-Yevick equation (12) to a decrease in scattering power of the solution. Dynamic Viscosity as a Function of Concentration The self-consistency covers phenomena that are in principal even more involved, such as the dynamic viscosity of the complex fluids. The classical Einstein model of the dynamic viscosity loses its validity if the volume fraction (q5app) is no longer small (compared to 1).Under the assumption of no mutual hindrance of the particles, Guth'l and Guth and Simha12 extended the Einstein model to larger volume fractions: where qsp = (q/qsoIv)- 1. From a fit of eqn (29) to experimental measurements of viscosity, q5app values were obtained from which a, a measure for the degree of fractal cluster formation could be evaluated [see eqn (1 7)]. A comparison of a-values derived from viscosity measurements and those theoretically calculated according to eqn (16) with c,(c, T ) data derived from Kerr-effect measurements is shown in fig. 4. The agreement between calculated and experimental points for two temperatures is satisfac- tory, as long as the temperatures are not too close to the phase separation (T, - T > 5 K).Furthermore, fig. 4 displays the interesting detail that the qisc data do not extrapolate to a = 1 for 4 + 0 but to 1.3. Thus (4app)4+0 = 1.34. This indicates that the hydrodynamic radius exceeds the geometrically estimated one by ca. 10% owing to the solvation of the nanodroplets by the dispersion medium. Again, the anomalous behaviour of the viscosity is a direct consequence of the fractalR. HilJiker and H-F. Eicke 1629 nature of the clusters. As the percolation transition is approached, the clusters grow in size, which leads to a concomitant increase of the apparent volume fraction of solute [eqn (1 6)] and, hence, to viscosities larger than the classically expected values. Conclusion We have shown that from the assumptions necessary to interpret electro-optic Kerr-effect measurements of complex fluids with fractal cluster structure, a self-consistent descrip- tion of these fluids emerges.This is verified by the demonstration that a minimum set of assumptions suffices to predict quantitatively the results of experiments of a rather different nature. The temperature- and concentration-induced changes in the structure of these systems are correctly reproduced; in particular, owing to the difference between the fractal (Hausdorm and Euclidean dimensions, the large viscosity increase and the surprisingly low scattered intensity rise are seen to be consistent, which is apparent from an inspection of eqn (28). We are convinced that the discussed model is a basis for other systems with similar structure and should predict results obtained by other experimental techniques. We are grateful to Dr V. Kim for light scattering and to Mr H. Hammerich for viscosity measurements. We also acknowledge financial support from the Swiss National Science Foundation. References 1 H-F. Eicke, R. Hilfiker and M. Holz, Helv. Chim. Acta, 1984, 67, 361. 2 R. Hilfiker, H-F. Eicke, S. Geiger and G. Furler, J. Colloid Interface Sci., 1985, 105, 278. 3 J. B. Hayter, Faraday Discuss. Chem. SOC., 1983, 76, 7. 4 H-F. Eicke, R. Hilfiker and H. Thomas, Chem. Phys. Lett., 1985, 120, 272. 5 H-F. Eicke, S. Geiger, F. A. Sauer and H. Thomas, Ber. Bunsenges. Phys. Chem., 1986,90, 872. 6 D. Stauffer, Phys. Rep., 1979, 54, 1 . 7 D. Stauffer, A. Coniglio and M. Adam, Advan. Polym. Sci., 1982, 44, 105. 8 C. T. O’Konski, in Molecular Electro-optics (Plenum Press, New York, 1981). 9 S. A. Safran, 1. Webman and S. Grest, Phys. Rev. A, 1985, 32, 506. 10 P. N. Pusey, H. M. Fijnaut, and A. J. Vrij, Chem. Phys., 1982,77, 4270. 1 1 E. Guth, KolloidZ., 1936, 74, 147. 12 E. Guth and R. Simha, Kolloid Z . , 1936, 74, 266. Paper 6/ 1599 ; Received 4th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301621

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 1631-1639 Photocatalytic Dehydrogenation of Liquid Alcohols by Platinized Anatase Falah H. Hussein and Robert Rudham" Department of Chemistry, University of Nottingham, Nottingham NG7 2RD The photocatalytic dehydrogenation of liquid methanol, ethanol, propan- 1-01 and propan-2-01 has been investigated using platinized anatase and 366 nm U.V. radiation over the range 278-303 K. Activities and activation energies for carbonyl-compound formation were effectively identical for the four alcohols on Pt(0.5)/Ti02 prepared by photodeposition. The activation energy of 20 kJ mol-' is associated with photoelectron transport through the anatase to the Pt particles. With Pt(O.S)/TiO, prepared by impregnation and H, reduction, identical activities for the four alcohols were achieved after 0, treatment.From the effect of u.v. intensity on the rate of propanone formation at 293 K, the limiting quantum yield was found to be 0.45 for photodeposited Pt(0.5)/Ti02 under N, and 0.82 for support anatase under 0,. Arrhenius plots for reaction at reduced U.V. intensities showed that the activation energy fell to zero when these quantum yields were achieved. Different mechanisms follow from the manner in which photoholes are surface trapped; two mechanisms for dehydrogenation with a limiting quantum yield of 0.5 are discussed. Suspensions of platinized anatase photocatalyse the dehydrogenation of alcohols in the liquid and in a q ~ e o u s ~ ~ ~ ~ ~ - ~ and benzene8 solution. Only radiation of energy equal to, or greater than, that of the anatase band gap is effective, unless the anatase is sensitized by dyes or surface d ~ p i n g .~ Anatase is a more efficient support than r ~ t i l e , ~ whilst other supported transition metals are less active than p l a t i n ~ m . ~ , ~ - ~ The primary reaction product other than H, is the corresponding aldehyde or ketone, which was shown to be produced in almost equivalent yield for ethanol and propan-1-oll and for aqueous propan-2-01. However, when methanal and ethanal are produced in significant concen- tration further catalytic or photocatalytic steps may occur, with methanal yielding HCO,H, CO, and H,5 and ethanal yielding CH,CO,H, CH,, CO, and H,.699 Further reaction beyond the primary product might account, at least in part, for the variation in photocatalytic reactivity of different alcohols when expressed in terms of H, produ~tion.~?~ Pichat et al.,l using platinized anatase (4.85 wt % Pt) prepared by impregnation and H, reduction, obtained the activity sequence : methanol > ethanol > propan- l-ol z propan-2-01 M butan- l-ol > 2-methylpropan-2-01 0.A similar se- quence was obtained3 with platinized anatase (2.0 wt % Pt) prepared by photodeposition : methanol > ethylene glycol > ethanol > propan-1-01 > propan-2-01 > butan-1-01 > butan-2-01 > 2-methylpropan-2-01 z 0. Presently, rates of carbonyl-compound for- mation are used to compare the reactivities of methanol, ethanol, propan-1-01 and propan-2-01 as a function of temperature on platinized anatase (0.5 wt% Pt) prepared by impregnation and H, reduction and by photodeposition.Pichat et al., measured rates of H, production from propan-1-01 with platinized anatase (0.5 wt% Pt) over the range 233-328 K; they suggested that reaction of photoholes with surface alkoxide ions was rate determining at T > 313 K, whilst desorption of H, from Pt was rate determining at T < 263 K. Photocatalytic production of propanone from propan-2-01 yielded linear Arrhenius plots over the range 278-303 K 16311632 Photocatalytic Dehydrogenation of Alcohols for 0.5 wt% Pt, Pd, Rh photodeposited on a n a t a ~ e . ~ A single activation energy of 20 kJ mol-1 was ascribed to trap-hindered photoelectron transport through the anatase to the metal particles. Photocatalytic dehydrogenation of alcohols on platinized anatase is associated with relatively high quantum yields; values between 0.1 and 0.45 have been reported for different alcoh01s.l~~ Since upper limiting values of the quantum yield should assist in the mechanistic interpretation of the reaction, we presently report on the effects of radiation intensity on quantum yield and activation energy for propanone production from propan-2-01 on both anatase and platinized anatase.Experimental The apparatus used to determine photocatalytic activities at precise temperatures in the range 278-303 K has been described.1° In all experiments 150 mg of catalyst was suspended in 20 cm3 of analytical reagent-grade alcohol and irradiated with filtered 366 nm radiation from a single Thorn ME/D 250 W mercury lamp.Uranyl oxalate actinometry at 298 K (a = 0.49) showed that the unattenuated radiation entering the reaction vessel from the four such lamps used was 2.7 x (lamp 2), 3.2 x lo-' (lamp 3) and 2.4 x lop7 (lamp 4) ein s-l.? With platinized anatase catalysts, an inert atmosphere was achieved by flushing the vessel containing the reaction mixture for 30 min with a 20 cm3 min-l flow of pure N, at atmospheric pressure; the flow was maintained during subsequent photoreaction. For experiments made with the anatase support, pure 0, was used in the same procedure. After centrifuging, 0.2 cm3 samples extracted from the reaction mixture were analysed for propanone by gas-liquid chromatography,ll or for methanal, ethanal or propanal by spectrophotometric measurements of the hydrazone formed from their interaction with 2,4-dinitrophenylhydrazine.12 These analytical methods gave excellent linear Cali- bration plots for each carbonyl compound, whilst comparative analyses of propanone in propan-2-01 showed that the two methods gave identical results.Full details of the preparation of the platinized anatase catalysts from Degussa P25 anatase (50 m2g-l) are given in our earlier paper.4 The same nomenclature is retained: method A indicates H,PtCl, impregnation followed by H, reduction at 753 K,l whilst method C indicates photodeposition from H,PtCl, methanol-methanal Pt(x)/TiO, indicates a catalyst with a platinum content of x wt% . (lamp l), 3.0 x Results Different Alcohols with Hydrogen-reduced Catalyst A dark reaction was observed when Pt(O.S)/TiO, prepared by method A was contacted with methanol, ethanol, propan-1-01 or propan-2-01 at 293 K under a N, atmosphere.With ethanol, propan- 1-01 and propan-2-01 the reaction obeyed first-order kinetics with respect to departure from a common equilibrium yield of the appropriate carbonyl compound of 0.45 x mol drnp3. With methanol the situation is less certain, since the yield of methanal fell short of this value and tended to decrease after a reaction time of 5 h. Plots of logl,(C, - C,)/C, against t, where Ct and C, are the carbonyl compound concentrations at time t and co, are shown in fig. 1 ; the slopes yield the following activity sequence: propan-2-01 > propan-1-01 > ethanol > methanol. As in previous work,* we attribute the dark reaction to interaction between alcohol and adsorbed OF, which must be allowed to reach completion before meaningful photocatalytic measurements can be made.Having achieved this condition the photocatalytic dehydrogenation of methanol, ethanol and propan-2-01 was followed at 293 K using lamp 1. Fig. 2 shows that linear reaction progress plots were obtained for products from ethanol and propan-2-01, I' 1 ein = 1 mole of photons.F. H . Hussein and R. Rudham 1633 0.0 - b8 \ n 2; I -1.0 4 s Y c( -2.0 Fig. 1. First-order plots for product formation by the dark reaction at 293 K on H, reduced Pt(O.S)/TiO,: 0, methanal; 0, ethanal; A, propanal; A, propanone. 2 Y 0.5 3.0 2.0 1 .o 50 100 tlmin 50 100 Fig. 2. Product formation from different alcohols on H, reduced (method A) and photodeposited (method C) Pt(OS)/TiO,: 0, methanol; 0, ethanal; A, propanal; A, propanone.(i) Method A as prepared, 293 K; (ii) method A after 0, treatment at 753 and 293 K; (iii) method C, 293 K; (iv) method C, 278 K. whereas that of methanal from methanol was curved. It is possible that the anomalous behaviour of methanol arises from the further catalytic or photocatalytic reaction of the methana15v1* to yield products undetected by the present spectrophotometric method. The dark reaction associated with H, reduced Pt(0. 5)/Ti0, was completely eliminated by contact with 0, at 1 atmf for 16 h at 753 K. Following such treatment, methanol, ethanol, propan- 1-01 and propan-2-01 yielded a single linear reaction progress plot at t 1 atm = 101 325 Pa.54 FAR 11634 Photocatalytic Dehydrogenation of Alcohols 293 K with lamp 2. However, fig. 2 shows that the overall level of photocatalytic activity was considerably reduced by the 0, treatment, although lamp 2 was slightly more intense than lamp 1. Different Alcohols with Photodeposited Catalyst No dark reaction was detected with any of the four alcohols with photodeposited Pt(O.S)/TiO,. Fig. 2 shows that a common linear reaction progress plot was obtained for ethanol, propan-1-01 and propan-2-01 both at 293 and 278 K using radiation from lamp 2. For methanol there was a pronounced deviation from linearity at 293 K, but the initial rate was the same as that for the other alcohols. The reaction progress plots in fig. 2 suggest a single activation energy for alcohol dehydrogenation.To verify this, the activation energy for each alcohol was determined from the slopes of a number of reaction progress plot^^^^^ measured at temperatures between 278 and 303 K with lamps 2 or 3. The activation energies were : methanol, 19 f 1 kJ mol-1 ; ethanol, 20 f 1 kJ mol-1 ; propan-1-01, 20 f 1 kJ mol-l; propan-2-01, 20 f 1 kJ mol-l. When Pt(O.S)/TiO,, prepared by method C, was subjected to the same H, treatment used in method A, H, at 1 atm for 16 h at 753 K followed by cooling to 293 K in N, before exposure to air, a dark reaction capacity was generated. With propan-2-01 an equilibrium yield of 0.49 x lo-, mol dm-3 of propanone was observed. Following completion of the dark reaction, linear reaction progress plots were obtained for photocatalytic reaction; five plots at different temperatures gave an activation energy of 9_+ 1 kJ mol-l.This value is close to that of 8 f 1 kJ mol-1 previously found for Pt(O.S)/TiO, prepared by method A.4 The final product concentrations in the runs presented in fig. 2 can be used to calculate the total turnover numbers achieved per surface site during reaction. These range between 2 and 11 per surface Ti ion or between 30 and 170 per supported Pt atom. These are lower limiting values, since it is highly unlikely that all ca. 5 x 10l8 m-2 surface Ti ions contribute or that Pt is totally dispersed as accessible atoms. Nevertheless, the numbers serve to show that reaction is photocatalytic rather than a photoassisted surface reaction restricted to a single monolayer. Similar considerations apply to all runs used to determine the activation energies presented above, where the minimum total turnover number achieved was 6 per surface Ti ion or 90 per supported Pt atom.Effect of Radiation Intensity on Quantum Yield Rates of propanone formation from propan-2-01, calculated from the slopes of linear reaction progress plots, were determined as a function of radiation intensity for the support anatase under 0, at 293 K using lamps 3 and 4. Similar rate measurements were made with photodeposited Pt(0.25)/Ti02 and Pt(O.S)/TiO, under N, at 293 K using lamp 3. Variations in radiation intensity, I, were achieved by neutral density filters for which the optical density at 366 nm had been experimentally determined. Experiments at low radiation intensities were made over prolonged reaction times up to 8 h; this ensured that accurate rates were determined and that total turnover numbers exceeded 3 per surface Ti ion for all three catalysts or 90 and 45 per Pt atom with Pt(0.25)/Ti02 and Pt(O.S)/TiO,, respectively.For the support anatase plots of reaction rate us. 1°.5 were straight lines passing through the origin, although points at I < 0.5 x lo-' ein s-l lay below the lines. More acceptable plots were previously obtained for the photo-oxidation of propan-2-01 on low-area anatase and rutile at 293 K.l0 However, with Pt(0.25)/Ti02 and Pt(O.S)/TiO, plots of reaction rate us. I or 1°a5 exhibited pronounced curvature. In view of this, the effect of radiation intensity on photocatalytic activity is presented in terms of quantum yields in fig.3. Quantum yields, given by the reaction rate expressed in mol s-l divided by the radiation entering the reaction vessel expressed in ein s-l, assume that all 366 nmF. H. Hussein and R. Rudham 1.0 n 7 0.5- € € m 71 0 - v) ; 0.0- c, c! v 1635 - 0.8 0.6 s ..I h +I 3 m cf 0.4 0.2 1 .o 2 .o 3.0 I/ lo-' ein s-l Fig. 3. Quantum yields for propanone formation at 293 K as a function of radiation intensity: 0, support anatase with lamp 3; A, support anatase with lamp 4; 0, Pt(O.S)/TiO, with lamp 3; A, Pt(0.25)/TiO2 with lamp 3. I I I 3.3 3.4 3.5 3.6 1 0 3 ~ 1 ~ Fig. 4. Arrhenius plots for propanone formation on support anatase, 0, and Pt(O.S)/TiO,, 0, at different radiation intensities: (i) 3.2 x loe7, (ii) 0.65 x and (iii) 0.14 x ein s-l.54-21636 Photocatalytic Dehydrogenation of Alcohols radiation is absorbed by the catalyst. Such an assumption is justified by the observation that variations in catalyst mass between 100 and 300 mg had no effect on the reaction rate at 293 K with either support anatase or Pt(0.5)/Ti0,.4 Effect of Radiation Intensity on Temperature Dependence Rates of propanone formation from propan-2-01 were determined as a function of temperature with support anatase under 0, and photodeposited Pt(O.S)/TiO, under N, at three different radiation intensities from lamp 3. With experiments at low radiation intensity, prolonged reaction times ensured accurate rates and total turnover numbers that were indicative of photocatalysis. The results are presented as Arrhenius plots in fig.4. With I = 3 . 2 ~ ein s-l both catalysts yield linear plots with a common activation energy of 20+ 1 kJ mol-l. With I = 0.65 x ein s-l and I = 0.14 x ein s-l, the catalysts exhibit both temperature-dependent and tempera- ture-independent regions ; for the temperature dependent region the activation energy is ca. 20 kJ mo1-l. Note that the temperature at which the reaction ceases to be temperature dependent increases with increasing I and is lower for Pt(O.S)/TiO, than for support anatase. Discussion Different Alcohols Unlike published studies of the photocatalytic dehydrogenation of liquid-phase alcohols by platinized P25 anata~e,l9~ the present work shows the reaction rate to be independent of the reactant alcohol with Pt(O.S)/TiO, prepared by photodeposition. However, with Pt(O.S)/TiO, prepared by impregnation and H, reduction this situation was only achieved at diminished activity following 0, treatment under conditions which restored stoichiometry to the anatase support.Although previous authors1v3 followed production of H, rather than carbonyl compound, further dehydrogenation of the primary product was unlikely to have augmented the rate of H, production to any measurable extent other than for methanol. This follows from the observations1 that H, was produced in almost equivalent yield to ethanal and propanal from ethanol and propan- 1-01, respectively, and that propanal was ca. 100 times less active than propan-1-01. However, further dehydrogenation of methanal5$l4 would account for the lower rates of formaldehyde production found in the present work (fig.2) and the higher rates of H, evolution from methanol reported in the published work. The most probable source of the differences between the three studies lies in the methods of catalyst preparation. In this connection, it has recently been shown15 that the choice of sacrificial organic reagent used in the photodeposition of platinum on anatase affects both the degree of metal dispersion and the photocatalytic activity for ethanol dehydrogenation. Reaction rates at ambient temperature varied from 0.03 pmol s-l using ethanol as the sacrificial reagent to 0.01 pmol s-l using methanoic acid. If differences in incident U.V. intensity are neglected, this variation in activity is comparable with that for ethanol dehydrogenation at 293 K in the three studies: 0.04 pmol s-l, Pichat et al.;lO.l pmol s-l, Borgarello and Pelizzetti;3 0.025 pmol s-l, method A following 0, treatment; 0.13 pmol s-l, method C .We have previously shown4 that the activation energy for the dehydrogenation of propan-2-01 was consistently 20 kJ mol-1 using photocatalysts prepared by the photo- deposition of Pt, Pd or Rh on P25 anatase. We now find identical activities and a common activation energy of 20 kJ mol-1 for photocatalytic dehydrogenation of four alcohols on Pt(O.S)/TiO, prepared by photodeposition on the same batch of anatase. These new observations give considerable support to the suggestion4 that the activation energy is associated with the trap-hindered transport of photoelectrons through the anatase to the metal particles rather than with any surface chemical process.The low activation energyF. H. Hussein and R. Rudham 1637 of 8-9 kJ mol-l given by Pt(O.S)TiO, subjected to H, treatment at 753 K, either during or after catalyst preparation, is associated with a reduction in photoelectron trap depth which accompanies the generation of non-stoichiometry in the anatase support. Changes in activation energy arising from modifications to the solid-state properties of TiO, are also observed for the photocatalytic oxidation of propan-2-01 on pure and cation doped rutile. lo Effects of U.V. Intensity and Temperature on Quantum Yield The effects of varying the incident U.V. intensity on the photocatalytic oxidation of propan-2-01 by TiO, have been considered by Egerton and King.16 It was shown that a direct dependence of reaction rate upon intensity is to be expected at low intensities where photoelectron-photohole recombination is negligible, whereas a square-root dependence is expected at high intensities where recombination is dominant.In the absence of competing reactions, the first condition gives the maximum quantum yield set by the reaction mechanism, whilst the second gives diminished values due to the recombinative loss of photoelectrons and photoholes. Failure to obtain satisfactorily linear plots for reaction rate against either I or I o s 5 reflects an intermediate situation in the present experiments. Since the total turnover numbers associated with linear reaction progress plots indicate a single catalytic process, extrapolation of the plots in fig.3 to zero radiation intensity is justified. This gives limiting quantum yields of 0.45 f 0.01 with Pt(0.25)/Ti02 and Pt(O.S)/TiO, under N, and 0.82 f 0.02 with support anatase under 0,. This disparity in limiting quantum yield indicates mechanistic differences between the two types of photocatalyst system. Any increase in reaction rate at constant radiation intensity necessarily means an increase in quantum yield, so that Arrhenius plots would become horizontal if rates corresponding to the maximum quantum yield were achieved at high temperatures. In this region of ‘zero activation energy’, all photoelectrons would be free from traps or, more importantly, recombination centres. For the Arrhenius plots in fig.4 the horizontal portions correspond to quantum yields of 0.45 f 0.02 for Pt(O.S)/TiO, and 0.68 f 0.02 for support anatase at a radiation intensity of 0.65 x ein s-l, and 0.48 0.02 for Pt(O.S)/TiO, and 0.76f0.02 for support anatase at a radiation intensity of 0.14 x ein s-l. For Pt(O.S)/TiO, these quantum yields are in excellent agreement with the maximum value of 0.45 f 0.0 1 found by extrapolation, but with support anatase the agreement is less satisfactory. The absence of a horizontal portion in the Arrhenius plots for reaction at the full radiation intensity of 3.2 x lo-’ ein s-l, in previous work4 and presently with other alcohols, arises from the failure to achieve the necessary quantum yield within the experimental temperature range. It follows that the form of the Arrhenius plots is determined by electronic processes within the anatase support rather than by different surface processes being rate controlling at high and low temperatures.Mechanistic Considerations There is general agreement concerning the initial steps in photocatalytic reactions on platinized anatase. The absorption of a photon by the anatase forms an exciton which dissociates to a photoelectron and a photohole: (1) Pt/TiO, + hv -+ e--h+ -+ e-+ h+. To effect reaction it is necessary that photoelectrons are trapped at platinum crystallites (Pt,), whereas photoholes are trapped at negatively charged surface species : Pt, +e- -+ Pt; S-+h++ S.1638 Photocatalytic Dehydrogenation of Alcohols Presumably the extent of metal dispersion affects the pathlength between photoelectron formation and trapping at Pt,, and consequently the extent to which e- and h+ are lost in non-radiative recombination.With support anatase under 0, the photoelectrons are surface trapped to yield adsorbed 0;. This is subsequently consumed according to a mechanism p r o p o ~ e d ~ ~ 1 ~ ~ for the photocatalytic oxidation of alcohols on rutile, where the limiting quantum yield of unity is in reasonable agreement with the present value of 0.82 & 0.02. Infrared studies18-20 have shown that alcohols react readily with titanium dioxide surfaces to form surface alkoxide ions, and this process can be expected to occur with the fully hydroxylated anatase surface of the present catalysts before commencing irradiation.Pichat et ale1 proposed a mechanism for alcohol dehydrogenation in which surface alkoxide ions are the photohole traps; it is given below for propan-2-01, but is equally applicable to other alcohols. Interaction of a photohole with the propan-2-oxide ion releases a propan-2-oxyl radical and exposes an incompletely coordinated Ti4+ ion : The radical can yield propanone either by the loss of a hydrogen atom: or by hydrogen transfer between two such radicals Dissociative chemisorption of a further alcohol molecule regenerates the propan-2-oxide ion and releases a proton: The role of Pt; is to convert protons into hydrogen atoms, catalyse their combination and release H, in a process that is essentially the reverse of hydrogen spillover. If reaction proceeds by reactions (4), (5) and (7), then Me,CHO-.- - Ti4+ + h+ 4 Me,CHO' + Ti4+. Me,CHO' -+ Me,CO + H, 2Me,CHO' -+ Me,CO + Me,CHOH. Me,CHOH + Ti4+ -+ Me,CHO- - - - Ti4+ + H+. (4) ( 5 ) (6) (7) Pt;+H++H + Pt,/H -+ Pt,+H, (8) 'H but if it proceeds by reactions (4), (6) and (7) then Overall reaction was consideredl to proceed by reactions (4), (5), (7) and (8), giving a limiting quantum yield of unity. However, were the reaction to proceed by reaction (4), (6), (7) and (9), the limiting quantum yield would be 0.5, which is more in keeping with the present experimental value. There is no direct evidence for photohole trapping at surface alkoxide ions, so that alternative trapping at surface hydroxyl ions to give hydroxyl radicals : must also be considered.Hydrogen abstraction by OH is favoured as a step in photocatalytic reactions on platinized Ti02.6v21*22 It also forms the basis of a mechanism for the photocatalytic oxidation of alcohols on rutile,1°*17 which is considered to be applicable to the present experiments with support anatase under 0, and can be modified for dehydrogenation on platinized anatase under N,. Following interaction of OH with propan-2-01 : propanone formation with a limiting quantum yield of 0.5 could occur in two ways: Pt, + H+ -+ Pt,-H + Pt, +4H2. (9) OH-+h+ -+OH (10) Me,CHOH + OH -+ Me,cOH + H,O 2Me,cOH -+ Me,CO + Me,CHOH Me,cOH +OH -+ Me,CO + H,O. (1 1) (12) (13)I;. H. Hussein and R. Rudham 1639 Interaction of the H,O from reactions (11) and (13) with Pt;: Pt, + H,O + Pt,-H + -OH + Pt, + 4H2 + OH- releases H, and regenerates OH- which adsorbs on the anatase surface.It follows that there are two possible mechanisms for the photocatalytic dehydro- genation of alcohols on platinized anatase, each following from the manner in which the photohole is surface trapped. However, a decision between these mechanisms is precluded when the rate-determining process is normally that of photoelectron transport within the anatase support. We thank the University of Salah Al-deen, Iraq, for a scholarship to F.H.H. References 1 P. Pichat, J-M. Herrmann, J. Disdier, H. Courbon and M-N. Mozzanega, Nouu. J. Chim., 1981, 5, 2 P. Pichat, M-N. Mozzanega, J. Disdier and J-M. Herrmann, Nouu. J. Chim., 1982, 6, 559. 3 E. Borgarello and E. Pelizzetti, Chim. Ind. (Milan), 1983, 65, 474. 4 F. H. Hussein and R. Rudham, J. Chem. SOC., Faraday Trans. I , 1984,80, 2817. 5 T. Kawai and T. Sakata, J. Chem. SOC., Chem. Commun., 1980, 694. 6 T. Sakata and T. Kawai, Chem. Phys. Lett., 1981, 80, 341. 7 S. Teratani, J. Nakamichi, K. Taya and K. Tanaka, Bull. Chem. SOC. Jpn, 1982, 55, 1688. 8 F. H. Hussein, G. Pattenden, R. Rudham and J. J. Russell, Tetrahedron Lett., 1984, 25, 3363. 9 J. F. Houlihan, R. J. Pollock and D. P. Madacsi, Electrochim. Acta, 1983, 28, 585. 627. 10 P. R. Harvey, R. Rudham and S. Ward, J. Chem. SOC., Faraday Trans. 1, 1983,79, 1381. 1 1 R. B. Cundall, R. Rudham and M. S. Salim, J. Chem. Soc., Faraday Trans. I , 1976,72, 1642. 12 D. E. Jordan and F. C. Veatch, Anal. Chem., 1964,36, 120. 13 D. Duonghong, E. Borgarello and M. Gratzel, J. Am. Chem. SOC., 1981, 103,4685. 14 S. Naito, J. Chem. SOC., Chem. Commun., 1985, 1211. 15 H. Nakamatsu, T. Kawai, A. Koreeda and S. Kawai, J. Chem. SOC., Faraday Trans I , 1986,82, 527. 16 T. A. Egerton and C. J. King, J. Oil Color. Chem. ASSOC., 1979, 62, 45. 17 P. R. Harvey, R. Rudham and S. Ward, J. Chem. SOC., Faraday Trans. I 1983, 79, 2975. 18 Y. M. Shchekochikhin, V. N. Filimonov, N. P. Keier and A. M. Terenin, Kinet. Catal., 1964,5, 94. 19 P. Jackson and G. D. Parfitt, J. Chem. Soc., Faraday Trans. I , 1972, 68, 1443. 20 C. H. Rochester, J. Graham and R. Rudham, J. Chem. SOC., Faraday Trans. I , 1984,80, 2459. 21 I. Izumi, F-R. F. Fan and A. J. Bard, J. Phys. Chem., 1981,85,218. 22 W. W. Dunn, Y. Aikawa and A. J. Bard, J. Am. Chem. Soc., 1981, 103,6893. Paper 6/1618; Received 7th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301631

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem Soc., Faraday Trans. 1, 1987,83, 1641-1649 Basicity of Water in Dipolar Aprotic Solvents using the 1,4,8,11 -Tetramethyl-l,4,8,ll-tetra-azacyclotetradecane Nickel@) Cation as a Probe of Electron-pair Acceptance Shoji Yamasaki, Yoshimasa Yanai, Etsuro Iwamoto* and Takahiro Kumamaru Department of Chemistry, Faculty of Science, Hiroshima University, Hiroshima 730, Japan Yuroku Yamamoto Environmental and Safety Engineering Department, Fukui Institute of Technology, Gakuen, Fukui 910, Japan. The basicity of water in 1,2-dichloroethane (1 ,ZDCE), nitrobenzene (NB), nitromethane (NM), N,N-dimethylacetamide (DMA), and 1,2-DCE-DMA mixtures has been examined using spectral data of 1,4,8,11 -tetramethyl- 1,4,8,11 -tetra-azacyclotetradecane nickel@) ([Ni(TMC)I2+}. The water molecules in 1,2-DCE, NB and NM, in which the monomeric water dominates, are barely coordinated to the cation, but in 1,ZDCE-DMA mixtures, in which the water molecules hydrogen bond with DMA molecules, they are easily coordinated. The equilibrium constant KNiw = [Ni(TMC)H,02+] /[Ni(TMC)2+][H20] was 1.9 in pure water, and 6.7 in a mixed solvent of 1.0 DMA fraction, 11.9 in 0.75, 23.2 in 0.50, 56.0 in 0.25 and 0 in 0.The results emphasize that hydrogen bonding of the water molecules with DMA increases the basicity of water molecules by polarization of 0-H bonds and leads to water coordination. The monomeric water is not coordinated to the square-planar nickel(r1) cation. 1.r. spectra are presented to support these findings. The chemical interactions of water in organic solvents have been extensively studied from the viewpoint of the electron-pair- and proton-donating ability of water.The proton- donating ability was discussed by association of water with solvent molecules and other solutes identified by means of i.r.l-l0 and n.m.r.ll spectroscopies and a vapour pressure method.l29 l3 The electron-pair-donating ability or basicity of water has been discussed in connection with its coordination to metal ions14* l5 and proton accepting in hydrogen-bonding reactions.ls Gutmannl' proposed donor number (DN) as a measure of the basicity of solvents which is ascribed to the solvation enthalpy of antimony pentachloride by the donor solvent in 1,2-DCE and gave a donor number of 18 to water. POPOV~~ measured the cheacal shifts of the 23Na nucleus in various solvents and found a correlation of the shift with donor number for each solvent, leading to a donor number of 33 for water.Furthermore, different values of 28,19 30,20 and 40.321 for the donor number of water were evaluated by investigations of the solvent exchange rate for nickel@), of e.s.r. spectra of copper(I1) complexes and of visible spectrum data for bis(acety1aceto- nato)oxovanadium(Iv), respectively. Gutmann22 introduced the term ' bulk donicity ' in order to characterize the basic properties of water: when water molecules are self- associated the donicity of the oxygen atoms increases by polarization of the 0-H bonds. Quite recently, Marcus21 also discussed whether or not bulk solvents and isolated solvent molecules in an inert solvent can be described by the same parameters.Despite an enormous amount of publications concerned with water, there are few studies concerning 16411642 Basicity of Water the donor reactivity of the water molecules, depending on their complexity in hydrogen- bonding interactions, to a given electron-pair acceptor. In view of the general importance of the water reactivities in non-aqueous solvents, it is of interest to investigate further the nature of water basicity. In this work, in order to study donor reactivities or basicity of water in 1,2-DCE7 NB, NM and DMA we select [Ni(TMC)I2+ as a probe of electron-pair acceptance. This metal chelate cation exists in solutions as an equilibrium mixture of a red, square-planar, low-spin species and a green, square-pyramidal, high-spin species according to the (1) reaction where S is an electron-pair-donor ~ o l v e n t .~ ~ - ~ ' The position of the equilibrium can be estimated spectrometrically using the absorption band at 520 nm due to the square- planar chromophore and may be related to the basicity of the donor molecule. In the four organic solvents mentioned above the metal chelate solution is red, showing that the basicity of those solvents is very low. When a trace amount of water is added to the red DMA solution, the solution becomes greenish. The basicity of water in those solvents was related to the equilibrium constant of reaction (1). [Ni(TMC)I2+ + S g [Ni(TMC)SI2+ Experiment a1 The metal chelate eletrolyte [Ni(TMC)](ClO,), was prepared according to the method in the l i t e r a t ~ r e .~ ~ ~ ~ * Infrared spectra for this salt showed that no NH absorptions are p-esent. Analysis was as follows: found C, 32.62; H, 6.40; N, 10.84; calculated: C, 32.71 ; H, 6.27; N, 10.90. Nitrobenzene (reagent grade, Wako Pure Chemical Industries Ltd) was purified by fractional distillation under reduced pressure through a 1.2 m column after washing with sulphuric acid, sodium carbonate and distilled water, successively. Nitromethane, 172-dichloroethane and dimethylacetamide (reagent grade, Wako Pure Chemical Industries Ltd) were purified by fractional distillation without washing. All solvents were used after standing over molecular sieves (type 4A) for one week. The water content in solvents was determined coulometrically by using a Hiranuma model AQ-3 aquacounter. The sample solution for measurement of water contents was taken directly from the spectrum cell just after measuring the absorbance in a dry box under a dry nitrogen atmosphere.Visible spectra were recorded on a Hitachi model 288A double-beam spectrophotometer equipped with thermostatted 5 cm quartz cells at 25 f 0.01 "C. The high-resolution proton chemical shift was measured using a Hitachi model R-600 lH Fourier-transform n.m.r. spectrometer at 60 MHz at ambient probe temperature 35 f 1 "C. Tetramethylsilane was used as the inner standard. Infrared spectra were recorded on a Nihon Bunko model A-102 instrument with 0.05 mm CaF, cells. All solutions were prepared by weight. Densities were measured by using an Anton Paar model DMA 02D digital density meter at 25 & 0.002 "C.Results and Discussion The metal chelate [Ni(TMC)I2+ cation with R,S,R,S-nitrogen configuration forms a square-pyramidal coordination geometry about nickel(r1) by axial ligation with a donor X-Ray studies on the green crystal [Ni(TMC)S](ClO,), or where S is the azide dimethylf~rmarnide~~ and acetonitrile31 showed that the S donor is bonded to the nickel ion on the same side as the four nitrogen substituents and the nickel ion is 29 to 34 pm out of the tetra-aza plane toward the S ligand. Such a movement prevents a sixth ligand from bonding in the other axial position owing to steric hindrance between the sixth ligand and the methylene groups in the skewed -NH[CH2I3NH- moieties. This holds even in solutions and is supported by the conductance results ofS. Yamasaki et al.1643 [Ni(TMC)](ClO,), in isodielectric nitrobenzene-acetonitrile mixtures : the first associa- tion constant decreases markedly with increasing fraction of acetonitrile because of the axial coordination of basic acetonitrile as a fifth ligand, whereas the second association constant shows little dependence on the fraction.32 The ionic association with the perchlorate ion may occur in mixtures containing 1,2-DCE with a low dielectric constant, although the metal chelate salt was found to be completely dissociated in DMA.32 However, the basicity of the perchlorate ion is very weak and it is reasonably assumed that there is no ionic association effect on reaction (1). Interconversion of R,S,R,S- into R,S,S,R-configuration subject to six-coordination is negligible.33 Thus, the donor solvent coordination is expressed by reaction (1). Typical visible spectra of [Ni(TMC)](ClO,), in DMA solutions containing water are shown in fig.1. The absorption peak at 520 nm is referred to the 4-coordinate planar complex, and the peaks at 392 and 654nm are referred to the 5-coordinate square- pyramidal complex. It is shown that the 5-coordinate species increases as the water content increases and the existence of ca. 3 mol dm-3 water completely displaces equilibrium (1) to the right. Fig. 2 shows the dependence of molar absorptivity at 520 nm on the water concentration in various solvent systems. Water does not significantly influence the molar absorptivity in 1,2-DCE7 NB and NM, whereas in the 172-DCE-DMA systems water decreases it, showing that the water molecules present in DMA can easily be bonded to the central nickel.In table 1 the molar absorptivities are summarized together with the corresponding water concentration. In order to obtain the molar absorptivity of the square-planar cation in water, concentrated sodium perchlorate media have been ~ ~ e d . ~ ~ y ~ ~ In a 6.22 mol dm-3 solution of NaClO,, the peaks at 392 and 654 nm disappeared completely and a molar absorptivity of 217 dm3 mol-1 cm-l at 520 nm was obtained. Table 1 also gives the molar absorptivity in the 172-DCE-DMA systems which were extrapolated to zero concentration of water. The equilibrium constant (KNiw) for water according to reaction (1) was evaluated using the equation36 l/(&o-&’) = l/(~o-&s) + ~/(E~-E~)KTN~w[H,OI (2) where E’ is the apparent molar absorptivity at 520 nm and E, is the molar absorptivity of the water-coordinated chelate cation.In this calculation the concentration of water bonded to the cation is negligible compared with that of the unbonded water since the metal chelate concentration used for measurements is in the range of 6 x lo-, to 1.5 x mol dm-3. It is assumed that water molecules are not associated with each other. The equilibrium constant was calculated using the data of which the saturation fraction lay between 0.2 and 0.8, except for the neat DMA with the highest E, value.37 A typical plot is shown in fig. 3. However, the E, value for the 0.75 DMA mole fraction obtained from the intercept was negative.Thus the intercept obtained from this equation is very sensitive to the data uncertainty and ass~mptions.~~ Therefore, with a constant value of E , = 10, which was obtained by extrapolation of the data in fig. 1, KNiw values were also evaluated at each concentration of water using the equation KNiw = (E, - E ’ ) / ( E ’ - E,)[H,O]. (3) The KNiw values obtained in two ways in two runs are shown in table 2. DCE, NM and NB, for which the DN values are 0,2.7 and 4.4, respectively, are hardly coordinated to the nickel(rr), as expected (table 1). In view of the fact that N,N- dimethylformamide (donor number 26.6) can be bonded to the nickel,30 it is interesting that DMA is also hardly bonded to it in spite of its large donor number (27.8).An investigation for this subject is now undertaken. In water, a molar absorbance of 75 dm3 mol-1 cm-l indicates that 70% of the nickel chelate is square-pyramidal. It should be noted that, although water molecules in DCE, NM and NB are hardly bonded to the nickel, the water coordination ability is1 644 0.7 0.6 0.5 3 -E 0.4- s: 2 0.3- 0.2 0.1 0) Basicity of Water - - - - - /i I 500 600 700 wavelength/nm 400 Fig. 1. Visible spectrum variation with the water concentration of a 3.68 x mol dm-3 solution of [Ni(TMC)](ClO,), in DMA at 25 "C. Water concentration (mol dm-9: 1, 4.54 x lo+; 2, 5.51 x 3, 1.25 x 10-l; 4, 2.47 x 10-l; 5, 4.73 x lob1; 7, 2.21. I 200 ," I E I - I E l E m '0 x c, .C( .2 Y I3 100 2 s h - 0 E 1 E *l 0 4 0 12 [ H,O]/ mol dm-3 Fig. 2. Variation of molar absorptivity at 520 nm with the water concentration in various solvents at 25 "C I, I,2-DCE; 2, NM; 3, NB; 1,2-DCE-DMA mixtures (DMA mole fraction): 4, 1.0; 5, 0.75; 6, 0.50; 7, 0.25.S.Yamasaki et al. 1645 Table 1. Molar absorptivities at 520 nm ( E ~ ~ ~ ) of [Ni(TMC)](ClO,), in various solvents at 25 "C solvent dissolved water '520 [H,O]/mol dm-3 /dm3 mol-1 cm-' 1,2-dichloroethane nitrobenzene (NB) (1,2-DCE) nitromethane (NM) water aqueous 6.22 mol dm-3 sodium perchlorate solution 9.25 x 10-5 2.16 x 10-3 1.86 x 10-3 1.22 x lo-' 1.75 x 10-l 1.37 55.6 38.1 228 f 12 223 f 20 218 & 7 217 & 5 181 f 1 177f3 1 90a 75b 7OC, 71d 217b 1,2-DCE-DMA mixtures '250 (DMA mole fraction) [H,O]/mol dm-3 /dm3 mol-1 cm-' 1.0 0.75 0.50 0.25 20 1 175 156 151 a Ref.(27). * At 513 nm. Ref. (26). Ref. (23). 0.0 6 m I 0.04 0 I n u: I 0 u, w 0.02 0 20 40 60 [ H2 0 ]-'/mol dm-j Fig. 3. The Benesi-Hildebrand plot for 1,2-DCE-DMA mixtures at 25 "C. DMA mole fraction: 1, 1.0; 2, 0.75; 3, 0.50; 4, 0.25.1646 Basicity of Water Table 2. Equilibrium constant (KNIw) for the water coordination to [Ni(TMC)I2+ in various solvents at 25 "C solvent KNiWa KNiWb 1 ,ZDCE-DMA mixtures (DMA mole fraction) 1 .o 4.6 f 0.2 6.6 f 0.2 5.5 k 0.5 6.8k0.1 nitrobenzene nitromethane water 0.75 4.6f 1 . 1 11.9f 1.0 5.0- 1.1 11.8f0.7 0.50 21.0 f0.8 23.3 t 0 . 6 20.6 f 0.9 23.1 f 1.5 0.25 52.1 4.4 55.8 & 2.4 62.4+ 1 . 1 56.0 _+ 1.9 0.0 0 0 0 1.9 1 . 0 5 ~ a The Benesi-Hildebrand equation [eqn (2)]. With an E, value of 10 [eqn (3)]. Ref.(26). strengthened in the 1,2-DCE-DMA system. Now we must consider the degree of water self-association and water-solvent interactions. Water self-association in polar organic solvents has been investigated by vapour pressure,l19 3 9 9 40 i.r.27 41 and lH n.m.r.4 data. Much evidence indicated that in 1,2-DCE trimeric species were favoured over dimer with a trimerization constant of 4.60 dm6 m01-~.~O This means that only 10% of the dissolved water is associated at 0.1 mol dm-3 at 25 "C. In nitrobenzene the monomer-dimer equilibrium with an equilibrium constant of 1.054 dm3 mol-1 predominate^,^^ showing that 15 % of the dissolved water is associated at 0.1 mol dm-3. Below 0.1 mol dm-3 water, the fraction of monomeric water increases more. Thus, it seems that the monomer of water is not bonded to the nickel@) and the water trimer is less effectively bonded if it has the cyclic It has been suggested that when a hydroxy group is hydrogen bonded to a base there occurs a polarization of the 0-H bond in the direction of the oxygen atom.43p44 The cation association constants of water in acetonitrile are larger than those of any alchoh01.l~ This was explained in terms of the difference in charge shift accompanying hydrogen bonding: water is doubly hydrogen bonded to acetonitrile while alcohols are singly bonded and therefore the basicity of the water oxygen is more enhanced.Hankins and M o s k o ~ i t z ~ ~ carried out accurate SCF calculations of the interaction potentials for pairs of water molecules and showed that charge transfer takes place from the proton acceptor to the proton donor.Gutmanng6 claimed the importance of the electron changes within the molecules induced by coordination and formulated the outer-sphere effects for electron-pair-donor ligands. This was applied to the explanation for the increased donicity (DN = 33) in water as solvent compared with that (DN = 18) of water molecules in 1,2-DCE, and the term 'bulk-donicity' has been introduced. In fact 70% of the chelate cation in pure water at 25 "C ligates water molecules, although it does not ligate the isolated water in the organic solvents used here (table I), indicating that the coordination ability of bulk water corresponds to only ca. 0.24 mol dm-3 water in DMA (fig. 1). In 1,2-DCE-DMA mixtures, KNiw increases with decreasing the fraction of DMA.It is likely that the variation of the DMA basicity plays an important role. Fig. 4 shows internally hydrogen bonded.S. Yamasaki et al. 160- 1647 3 t cz I I I I 1 I mole fraction of water Fig. 4. lH n.m.r. spectra of water and DMA referenced to TMS. 1, water; 2, DMA, trans-N-methyl; 3, DMA, cis-N-methyl. 0, Neat DMA; 0, 0.50 mole fraction DMA. 0.2 0.4 0.G 0.8 1 0 the dependence of lH chemical shift of water and N-methyl groups of DMA on the water fraction in the 1.0 and 0.50 DMA fraction systems. The change in the chemical shift of N-methyl groups was ascribed to that in the degree of dimerization of DMA molec~les.~~~ 48 Since the chemical shifts of N-methyl groups are independent of the water fraction in both, it may be said that the dimerization equilibrium remains constant in each system.However, in the mixed solvent of 0.5 mole fraction DMA, the difference in the chemical shift between cis and trans methyl groups is slightly smaller compared with that in the 1 .O mole fraction system, suggesting that the degree of dimerization of DMA decreases. This leads to the results in table 1 that the molar absorptivity at zero water concentration in mixtures decreases with decreasing the DMA fraction, showing that DMA molecules become more bonded. Water protons in the 0.50 mole fraction solvent have a higher field shift compared with those in neat DMA. This may be attributed to the existence of free 0-H protons in the former as stated below. 1.r. spectra presented in fig. 5 clearly show the DMA-H,O association.Spectrum 1 in the 0.25 mole fraction mixture shows the characteristics expected for 1 : 1 water-DMA species: the sharp, high-frequency band at 3653 cm-l is assigned to stretching of the free 0-H bond and the broad low-frequency band at 3441 cm-l is associated with the hydrogen-bonded 0-H group. A narrow band at 3275 cm-l which is independent of the solvent composition can be attributed to the overtone of the bending vibration of water. Increasing the DMA fraction decreases the sharp band and produces a broad band at 3525cm-l with a shoulder at ca. 3480 cm-l, which is ascribed to the two bonded 0-H groups for a 1 : 2 water-DMA species. As in neat DMA, the free 0-H stretching band almost disappears and in the 0.25 mole fraction mixture the shoulder at 3525 cm-l is seen; it can be said that in neat DMA most water molecules form a 1 : 2 complex involving one water molecule and two DMA molecules and the 1 : 1 complex becomes predominant with decreasing DMA mole1648 .- I m *g e E Y Basicity of Water 1 V I 3800 3600 3400 3200 wavenum ber1cm-l Fig.5.1.r. spectra of water in 1,2-DCE-DMA mixtures. DMA mole fraction (water concentration in mol dm-3): 1, 0.25 (0.0477); 2, 0.50 (0.188); 3, 0.75 (0.0793); 4, 1.0 (0.0435). fraction. It was reported that the equilibrium constant for the 1 : 1 H,O-DMA association in 1,2-DCE was 2.2 at 25 0C.49 This value indicates that water molecules in a diluted concentration of water below 0.1 mol dmb3 is associated with DMA molecules in 0.87 mole fraction for the DMA 0.25 mole fraction solvent.KNiw in the 0.25 mole fraction mixture is larger than that in neat DMA. This is either because the basicity of the oxygen atoms of water for the 1 : 1 species is stronger than that for the 1 : 2 species or because coordination of the double hydrogen-bonded water molecules is sterically hindered by hydrogen-bonding DMA molecules. However, the fact that the DMA molecules are more coordinated in lower DMA mole fraction mixtures seems to support the former because the oxygen atoms of DMA with the stronger coordination ability produce a stronger basicity of the oxygen atoms of water by a polarizing effect through hydrogen bonding. It can be said that dimerization of the DMA molecules due to dipole-dipole interactions decreases the electron density on oxygen atoms of the DMA molecules.The lower frequency (3441 cm-l) of the bonded 0-H stretching also supports the strong hydrogen-bonding interactions. Mohr et al.' investi- gated the frequencies of the 0-H stretching absorption bands of 1: 1 complexes of water and various bases in carbon tetrachloride and showed that the bonded 0-H stretching band for NM and NB was at 3610 and 3585 cm-l, respectively, and that for dimethylformamide was at 3490 cm-l. Thus, it is said that the oxygen atoms of water in NB and NM are much less polarized compared to those in the amides. It can be concluded that the monomeric water molecules in 1,2-DCE, NM and NB of which donor number corresponds to 18 cannot be coordinated to the square-planar metal chelate cation. The complex formation of water with both water itself and the basicS.Yamasaki et al. 1649 DMA increases the coordination ability of water, depending on the degree of charge transfer to the oxygen atoms through hydrogen bonding. This work was supported by a Grant-in Aid for Scientific Research no. 61 134043 for the Ministry of Education, Science and Culture of Japan. References 1 S. C. Mohr, W. D. Wilk and G. M. Barrow, J . Am. Chem. SOC., 1965,87, 3048. 2 L. B. Magnusson, J. Phys. Chem., 1970, 74, 4221. 3 M. Saneganik, C. Klofutar, S. Poljk and L. Smerkar, J. Znorg. Nucl. Chem., 1970, 32, 1659. 4 D. N. Glew and N. S. Rath, Can. J. Chem., 1971,49, 837. 5 D. R. Cogley, M. Falk, J. N. Butler and E. Grunwald, J. Phys. Chem., 1972, 76, 855. 6 D. B. Henson and C. A. Swenson, J.Phys. Chem., 1973, 77, 2401. 7 P. McTigue and P. V. Renowden, J. Chem. SOC., Faraday Trans. I , 1975, 71, 1784. 8 A. Le Narvor, E. Gentric, J. Laurason and P. Saumagne, J. Chem. SOC., Faraday Trans. 1, 1976, 72, 9 M. C. R. Symons, Chem. SOC. Rev., 1983, 12, 1. 1329. 10 K. B. Patee, G. Eaton and M. C. R. Symons, J. Chem. SOC., Faraday Trans. I , 1985,81, 2775. 1 1 F. Takahashi and N. C. Li, J. Am. Chem. SOC., 1966,88, 11 17. 12 S. D. Christian, A. A. Taha and B. W. Gash, Q. Rev. Chem. SOC., 1970, 24, 20, and references therein. 13 S. D. Christian, E. E. Tucker and A. Mitra, J. Chem. SOC., Faraday Trans. I , 1976, 73, 537. 14 D. R. Cogley, J. N. Butler and E. Grunwald, J. Phys. Chem., 1971, 75, 1477. 15 G. W. Stockton and J. S. Martin, J. Am. Chem. SOC., 1972,94, 6921.16 M. C. R. Symons, Acc. Chem. Res., 1981, 14, 179. 17 V. Gutmann and E. Wychera, Znorg. Nucl. Chem. Lett., 1966, 12, 257. 18 R. H. Erlich, E. Roach and A. I. Popov, J. Am. Chem. SOC., 1970,92,4989. 19 S. Funahashi and R. B. Jordan, Znorg. Chem., 1977, 16, 1301. 20 T. Ogata, T. Fujisawa, N. Tanaka and H. Yokoi, Bull. Chem. SOC. Jpn, 1976, 49, 2759. 21 Y. Marcus, J. Solution Chem., 1984, 13, 599. 22 U. Mayer and V. Gutmann, Struct. Bonding (Berlin), 1972, 12, 113. 23 E. K. Barefield and F. Wagner, Znorg. Chem., 1973, 12, 2435. 24 K. D. Hodges, R. G. Wallmann, E. K. Barefield and D. N. Hendrickson, Inorg. Chem., 1977,16,2746. 25 A. Anichini, L. Fabbrizzi, P. Paoletti and R. M. Clay, Znorg. Chim. Acta, 1977, 24, L21. 26 N. Herron and P. Moore, Znorg.Chim. Acta, 1979, 36, 89. 27 P. Moore, J. Sacchinidis and G. R. Willey, J. Chem. SOC., Dalton Trans., 1984, 1323. 28 E. K. Barefield, F. Wagner, A. W. Heringer and A. R. Dahl, Znorg. Synth., 1975, 16, 220. 29 M. J. D’Aniello Jr, M. T. Mocella, F. Wagner, E. K. Barefield and I. C. Paul, J. Am. Chem. SOC., 1975, 30 S. F. Lincoln, T. W. Hambley, D. L. Pisaniello and J. H. Coates, Aust. J. Chem., 1984, 37, 713. 31 I. S. Crick, B. F. Hoskins and P. A. Tregloan, Inorg. Chim. Acta, 1986, 114, L33. 32 E. Iwamoto, Y. Yanai, S. Yamasaki, T. Yokoyama and T. Kumamaru, unpublished results. 33 P. Moore, J. Sachinidis and G. R. Willey, J. Chem. SOC., Chem. Commun., 1983, 522. 34 C. K. Jorgensen, Acta Chem. Scand., 1957, 11, 399; 919. 35 K. Sone and M. Kato, 2. Anorg. Allg. Chem., 1959, 301, 277. 36 (a) H. A. Benesi and J. H. Hildebrand, J. Am. Chem. SOC., 1949,71,2703; (b) R. L. Scott, Recl. Trav. 37 D. A. Deranleau, J. Am. Chem. SOC., 1969, 91, 4044. 38 M. H. Lift and J. Wellinghoff, J. Phys. Chem., 1977, 81, 2644. 39 J. R. Johnson, S. D. Christian and H. E. Affsprung, J. Chem. SOC. A, 1967, 1924. 40 J. R. Johnson, S. D. Christian and H. E. Affsprung, J. Chem. SOC., 1966, 77. 41 C. Jolicoeur and A. Cabana, Can. J. Chem., 1968,46, 567. 42 J. R. Holmes, D. Kivelson and W. C. Drinkard, J. Am. Chem. SOC., 1962,84,4677. 43 R. U. Lemieux and A. A. Pavia, Can. J. Chem., 1969,47,4441. 44 P. A. Kollman and L. C. Allen, J. Am. Chem. SOC., 1971,93,4991. 45 D. Hankins and J. W. Moskowitz, J. Chem. Phys., 1970,53,4544. 46 V. Gutmann and R. Schmid, Coord. Chem. Rev., 1974, 12, 263. 47 M. Rabinowitz and A. Pines, J. Chem. SOC. B, 1968, 11 10. 48 R. C. Neuman Jr, W. Snider and V. Jonas, J. Phys. Chem., 1968, 72, 2469. 49 D. Mueller, Ph. D. Dissertation (University of Oklahoma, 1966). 97, 192. Chim. Pays-Bas, 1956, 75, 787. Paper 6/ 1639; Received 1 lth August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301641

出版商:RSC    

年代:1987

数据来源: RSC