摘要:

J. Chem. Soc., Faraday Trans. 1, 1986,82, 3357-3365 The Electrical Conductance of Molten Lead@) 9,lO-Dihydroxyoctadecanoate and some Binary Mixtures with Lead@) Octadecanoate M. Sola Akanni" and P. Chuckwulozie Mbaneme Department of Chemistry, University of r e , Ile-I$e , Nigeria Data are presented for electrical conductances of molten lead@) 9,lO- dihydroxyoctadecanoate and the system lead@) octadecanoate-lead@) 9,lO- dihydroxyoctadecanoate. The lead@) 9,lO-dihydroxyoctadecanoate is prepared from 9, lO-dihydroxyoctadecanoic acid which in turn is obtained from the oxidation of cis-9-octadecenoic acid by hydrogen peroxide in methanoic acid. For the mole fraction of lead@) 9,lO-dihydroxyocta- decanoate < 0.02, the Arrhenius plots for conductance of the mixtures show curvature similar to the behaviour of pure, unsubstituted lead@) carboxylates, while above this mole fraction the plots display a maximum resembling that of pure lead@) 9,lO-dihydroxyoctadecanoate.Conductances are reproducible with temperature cycling up to the maximum, but show permanent decrease when cycled beyond the maximum for the mole fractions of 9,lOdihydroxyoctadecanoate 2 0.03. The maximum is inter- preted in terms of the current carriers (Pb2+ ions) reacting with the dihydroxy groups to form a bridged cyclo-acid. The observed low conductance of pure lead@) 9,lO-dihydroxyoctadecanoate and the decrease in conductance of the mixtures compared with lead@) octadecanoate is suggested to be due to the relatively small dissociation of the dihydroxy soap. Activation energies for conductance in the low-temperature region show a steady decrease with increasing mole fraction of lead@) 9,lO-dihydroxyoctadecanoate up to a certain composition and then increase.This behaviour is attributed to a change in the microscopic structure of the melt owing to the increasingly dominant role of the substituted hydroxy groups. The activation energy for the pure dihydroxy soap is close to those of other lead@) soaps, suggesting that the major charge carrier is probably the same, i.e. the Pb2+ ion. Previous studies from our laboratory have provided information on the temperature dependence of electrical conductivity, density and molar volume of some molten, straight-chain, divalent metal carb~xylates.l-~ Data are also available in the literature on the temperature dependence of electrical conductivity, density and molar volume of some binary mixtures in which the intermolecular forces in the pure molten soaps were systematically perturbed by the addition of polar and non-polar material^.^-^ Maxima have been observed in the plots of log (conductivity) against inverse temperature for several molten inorganic salts.8 Gregory and Tartar,9 working on a metal carboxylate similar to that employed in the present study, examined the effect of substitution of hydroxy groups in the middle of the carbon chain of a fatty acid soap by measuring the electrical conductance of various concentrations of potassium 9,lO-dihydroxyocta- decanoate in warm (60 "C), dilute solutions of potassium hydroxide.The potassium 9,lO-dihydroxyoctadecanoate was found to be more soluble at this temperature than potassium octadecanoate. The former salt solutions were also found to have higher densities and higher equivalent conductances than the solutions of the latter salt of corresponding (equal) concentrations.More valuable information can be obtained on the influence of substitution of 33573358 Electrical Conductance of Molten Lead(@ Compounds hydroxy groups in the middle of the carbon chain of metal carboxylates by studying the electrical conductance of the soaps in the molten state. Such a study is particularly desirable, as data on the conductance of molten metal 9,lO-dihydroxyoctadecanoates are not available in the literature. Hence the requirement for the present work, which reports the conductance behaviour of molten lead@) 9,lO-dihydroxyoctadecanoate and some lead@) octadecanoate/lead(II) 9,lO-dihydroxyoctadecanoate mixtures.Experimental Materials All the fatty acids employed were from B.D.H. and were stated to have a minimum 99% purity by g.1.c. assay. They were used without further purification. Lead(@ nitrate was AnalaR grade from B.D.H. Preparation of 9,lO-Dihydroxyoctadecanoic Acid 9,lO-Dihydroxyoctadecanoic acid was prepared by the oxidation of cis-9-octadecenoic (oleic) acid with hydrogen peroxide in methanoic (formic) acid, as reported by Swern et al.l0 The crude product obtained melted at 356 K. This is considerably lower than the melting point of 363-365 K reported for the crude product obtained by Swern et al. However, recrystallisation from hot ethanol three times raised the melting point to 366 K.Further recrystallisation no longer changed this value and the acid was considered pure enough for use, as the reported melting point of the recrystallised product of this acid (prepared from oleic acid) was 367 K.l0 Note that 9,10-dihydroxyoctadecanoic acid has also been prepared by Gregory and TartarB by the oxidation of oleic acid with alkaline permanganate in dilute solution. The pure crystals they obtained melted at 404.6 K. This value agrees with the melting point of 404-405 K obtained by us and 403-403.5 K obtained by Swern et al. for pure 9,10-dihydroxyoctadecanoic acid prepared by the oxidation of trans-9-octadecenoic acid with H,O,-HC0,H. These melting points can be rationalised if it is recalled that addition by H,O,-HC0,H to a double bond often occurs in the trans-position, while addition by alkaline permanganate is usually in the cis-position.ll Thus (a) the cis-alkene (oleic acid) + H,O,-HC0,H leads to a racemic (DL) mixture (previously reported melting point 367 K; present melting point 366 K); (b) the cis-alkene (oleic acid) + alkaline permanganate leads to the meso product (reported melting point = 404-405 K); (c) the trans-alkene (elaidic acid) + H,O,-HC0,H leads to the meso product (previously reported melting point 403-403.5 K; present melting point 404-405 K).It can therefore be seen that the stereochemistry of the acid has a significant effect on its melting point. Preparation of the Soaps The lead(@ octadecanoate (PbA,) used for this work had been prepared and characterised in our laboratory.l, The lead@) 9,1O-dihydroxyoctadecanoate, Pb(OH),A,, was pre- pared by metathesis in alcohol solution.l* 12$ l3 The 9,10-dihydroxyoctadecanoic acid (prepared as described above) was dissolved in hot absolute ethanol and a stoichiometric amount of solid potassium hydroxide added.The stoichiometric quantity of lead nitrate was dissolved in the minimum amount of water and this was added slowly, with stirring, to the alcohol solution. The resulting mixture was left for at least 2 h for precipitation to be completed. The precipitate was then filtered off, washed with copious quantities of ethanol followed by distilled water and acetone. The crystals were dried under vacuum at 362 K and finally recrystallised twice in hot propan-1-01.The 9,10-dihydroxyoctadecanoic acid obtained from both the cis- and trans-9- octadecenoic (racemic and meso) acids was used in turn in preparing the soaps, and theM. S. Akanni and P . C. Mbaneme 3359 Table 1. Melting points and C, H and Pb analyses of Pb(OH),A, soaps prepared in different waysa melting point/K theory found theory found theory found (A) (B) 411-413 51.59 52.20 8.42 8.58 24.72 25.21 437-439 51.59 50.86 8.42 8.42 24.72 25.32 a (A) Prepared from the dihydroxy acid obtained from oleic acid, racemic mixture. (B) Prepared from dihydroxy acid obtained from elaidic acid, meso product. products were all fine crystalline powders. The i.r. spectra showed them to be free of an excess of organic acid. Results of elemental analyses for carbon, hydrogen and lead, together with the melting points, are presented in table 1.The melting-point values in the table show that the effect of the stereochemical configuration observed for 9,lO-dihydroxyoctadecanoic acid is retained in the soaps. The Pb(OH),A, prepared from 9,lO-dihydroxyoctadecanoic acid (racemic mixture obtained from oleic acid) was used for all our conductance measurements because a considerable amount of the soap is required to cover the electrodes of the cell, and only oleic acid was available in large quantities. Measurements The procedures for the measurement of molar volume and conductance have been described elsewhere.l* The method of preparing the mixtures was the same as previously reported. l4 Results The Arrhenius plots for the pure soaps and the mixtures are shown in fig.1. At low mole fractions (XPb(OH)2A2 < 0.02) curvature was observed at high temperatures. This behaviour is similar to that observed for pure lead(n)l and cadmium(~r)~ carboxylates and has been interpreted in terms of a simple theory in which the soap dissociates according to the reaction (1) If the assumption that the M2+ ion is the major charge carrier in the melt is made and that it moves by a simple activated process, then MA, e M2+ + 2A-. AHg + AH/3 2.303RT logk = log Q - where AHf and AH are the enthalpies of activation for the movement of M2+ and for dissociation, respectively. NeA AS2 + AS/3 log ' = log (%)+ 2.303R (3) where ASf and AS are the entropies of activation for the movement of M2+ and for dissociation, respectively, N is Avagadro's number, e is the charge on an electron, A is3360 Electrical Conductance of Molten Lead@) Compounds --.- DD o -2.60 1 -2 -80 * * a * * * * * -3.00 * t * * * -3.201 I I I I r * 2.00 240 2.20 2.30 2.40 2-50 2.55 103 KIT Fig.1. Semilogarithmic plot of conductivity against inverse temperature for lead@) octadecanoate- lead(@ 9,lO-dihydroxyoctadecoate (PbA,-Pb(OH),A,) systems : 0, pure PbA, ; x , XPb(OH)sA2 = 0.01 ; 0, XPb(()H)BAe = 0.02; +, XPb(OH)&e = 0.03; 0, XPb(OH)~A~ = 0.05; ., XPb(OH)2Ae = O-lo; a, XPb(OH)*Ae = Oe50; *, pure Pb(0H)2A,m Table 2. Values of intercept and low-temperature limiting slopes for conductance (AH+ AH/3) correlation XPb(OH)nAa log Q /kJ mol-l coefficient 0.00 0.01 0.02 0.03 0.05 0.10 0.50 1 .oo 3.399 3.239 3.004 2.489 2.03 1 1.432 1.926 2.51 1 43.6 Ifr 2.0 0.9968 42.7 f 1.5 0.9964 40.9 f 1.6 0.991 1 37.1 f 1.8 0.993 1 33.7 f 2.2 0.9961 29.3 f 1.2 0.9958 3 1.5 f 2.4 0.9988 38.3 f 2.1 0.9957 a pre-exponential factor in the relationship between the electrical mobility of M2+ and the absolute temperature and Vm is the molar volume of the pure soaps or the mixtures. At low temperatures, where the degree of dissociation is small, plots of logk against 1/T should be linear, with slopes equal to (AHg +AH/3)/2.303R. In table 2, values of AH* = (AHt + AH/3) are presented as calculated by a least-squares method from these plots, together with the correlation coefficients. The molar volumes of selected mixtures were studied.Fig. 2 shows plots of Vm against temperature for 0.10,0.30 and 0.60 mole fractions of Pb(OH),A,.M. S.Akanni and P . C. Mbaneme 3361 t x x X X X X O 0 O0 O ooo 0 O 0 600 400 420 4 40 4 60 4 80 T/K Fig, 2. Plot of molar volume against temperature for the systems lead(@ octadecanoate-lead(@ 9,lOdihydroxyoctadecanoate: x , 0.10; 0, 0.30 and 0, 0.60 mole fractions of lead@) 9,l Odihydroxyoctadecanoate. Discussion The electrical conductance of lead(@ 9,lO-dihydroxyoctadecanoate is relatively lower than that of lead(@ octadecanoate and, unlike the curvature observed in the plot of log (conductivity) against inverse temperature for the latter, a maximum is obtained for the former. The low conductance of Pb(OH),A, may arise as a result of (i) low dissociation of the soap or (ii) high viscosity of the melt.In order to examine which of these factors is responsible for the low conductance observed, the electrical conductance of a binary mixture of Pb(OH),A, and octadecanoic acid (0.20 mole fraction) was measured. Octadecanoic acid melts to give a relatively more mobile liquid than Pb(OH),A,. The soapacid mixture showed much lower conductance values than pure Pb(OH),A,, suggesting that the low conductance of the mixture is probably a result of either the low dissociation of the soap component or a decrease in the concentration of Pb2+ ions. To verify this behaviour further and to confirm the maximum observed in the plot of logk against 1/T for Pb(OH),A,, the electrical conductance of various binary mixtures of lead(I1) octadecanoate and lead@) 9,1O-dihydroxyoctadecanoate, PbA,- Pb(OH),A,, were measured.As can be seen in fig. 1, the actual conductance decreases with an increase in Pb(OH),A, composition. The dissociation equilibria producing the charge carriers which lead to curvatures at low concentrations of Pb(OH),A, (&b(OH)2A2 0.02) are (CH,[CH,],,COO),Pb e Pb2+ + 2CH,[CH,],,CO, ( 5 ) (CH,[CH,],-CH-CH-[CH,],COO),Pb Pb2+ + 2CH,[CH,],-CH-CH-[CH,], COT. I I I I OH OH OH OH (6) Conductances were reproducible with temperature cycling from ca. 10 K above the melting point to ca. 480 K for these low concentrations. The decrease in conductance on adding more Pb(OH),A, may be attributed to a decrease in the entropy terms3362 Electrical Conductance of Molten Lead(xI) Compoundr Table 3. Low-temperature limiting slopes for conductance for lead soaps compound (AH$ + AH/3)/kJ mol-1 lead(@ octadecanoate 43.6 & 1.7 (this work) lead(@ octadecanoate 43.0f 1.5 [ref.(l)] lead@) cis-9-octadecenoate 45.3 & 2.0 [ref. (411 lead(@ trans-9-octadecenoate 42.9 & 1.5 [ref. (4)] lead@) 9,lO-dihydroxyoctadeanoate 38.3 & 1.8 (this work) (AS# + AS/3). This will necessarily mean a decrease in Pb2+ concentration arising from the lower dissociation of Pb(OH),A, than PbA,. A similar effect, but in an opposite trend, has been observed in conductance behaviour of lead@) dodecanoate-lead(@ ethanoate mixture. At high mole fractions (XPb(OH)rA2 2 0.03) the behaviour of the mixtures is similar to that of pure Pb(OH),A,. Conductances in these cases were reversible with increasing temperature up to the maximum, and beyond showed a permanent decrease.A possible explanation for the conductance to increase initially and then decrease with continuous increase in temperature might be a reduction in the concentration of the current carriers in the melt at high temperature. Such a reduction in concentration of Pb2+ ions can occur if the reactant Pb2+ ion attacks the hydroxy groups [reaction (611, resulting in the formation of a lead cyclo-anion : OH OH /”\ 0 0 I I I 1 Pb2+ + CH3[CH2],CH-CH-[CH2],CO~ + CH,[CH2],CH-CH-(CH2),C0, + 2H+. (7) The higher the temperature, the more reaction (7) is favoured and the lower the conductance becomes, as more of the charge carriers (Pb2+ ions) become involved in bond formation. Note that the protons do not remain as charged ions in the melt. If they were present as charged ions, the conductance would increase (rather than decrease) significantly to manifest their presence as hydrogen ions are expected to be highly mobile.Instead, the protons react with the cyclo-anion to form the lead cyclo-acid: Thus the gradual formation of the lead cyclo-acid (which is a non-conducting species or which is at best of very low conductance) should be responsible for the fall in conductance, after the initial rise, with increase in temperature. The formation of the lead cyclo-acid is made possible because the attack by the Pb2+ ion occurs at carbons 9 and 10 of the carboxylate anion. It may be possible as an alternative for attack to occurM. S. Akanni and P. C. Mbaneme 3363 at carbons 9 and 9 or 10 and 10 of the anions; then a bridged structure, instead of a cyclo-acid, will be obtained as shown below: Pb2+ + 2CH,[CH,],-CH-CH[CH2]],CO~ I I 1 OH OH CH,-[CH,],-CH-CH-[CH,],CO, I I 0 OH I Pb + 2H+ I 0 OH I I 1 CH,-[CH,],-CH-CH-[CH,],CO, R-CH-CH-R’-CO,H R-CH-CH-R’-CO,H I I I I 0 OH 0 OH I 1 Pb I I 0 OH OH 0 I I I I s-cis-form s-trans-form. (9) Pb L 7 R-CH-CH-R’-CO,H H0,C-R’-CH-CH-R For attack on carbons 9 and 9, the products are similar to those obtained above for attack on carbons 10 and 10, as can be seen below : R-CH-CH-R’-CO,H I I R-CH-CH-R’-CO,H I I OH 0 HO 0 I I Pb 7 I I Pb L OH 0 I I 0 OH I I R-CH-CH-R’-CO,H H0,C-R’-CH-CH-R s-cis-form s-trans-form.For the formation of these bridged acids, the furnished protons react with the bridged anions in one stage, and the symmetry of the bridged acid could be the driving force for its formation.To confirm the existence of an acid in the PbA,-Pb(OH),A, melts at a high composition of the second component or in pure Pb(OH),A, following conductance measurements, and following the cooling that usually leads to solidification of the melts, a small quantity of the solid was dropped into a solution of saturated sodium hydrogen carbonate (NaHCO,) in a test tube. Copious effervescence occurred and the liberated3364 Electrical Conductance of Molten Lead(@ Compounds gas was observed to turn limewater milky. Further evidence in support of the presence of acid in the systems is found in the i.r. spectra of KBr pellets of samples of the solids, which showed the appearance of a C=O absorption band at ca.1690 cm-l. This band was absent in the spectra of mixtures (XPb(OH)2A2 0.03) or pure Pb(OH),A, that was not subjected to conductance measurement. Furthermore, if the reactions leading to the formation of the lead bridged cyclo-acid discussed above are correct, then the system PbA,-Pb(OH),A, or pure Pb(OH),A, would give a non-ideal liquid at high temperatures. Deviation from the ideal behaviour of an isotropic liquid has been tested by plotting the molar volume against temperature.15 The results showed that when a liquid (melt) behaved ideally the plots were linear, and when deviation from ideality occurred, curvature was displayed within the range of temperatures employed. Fig. 2 shows the results obtained from some of the mixtures.Curves, rather than straight lines, are obtained for the mixtures, and the behaviour is suggestive of deviation from ideality, showing, perhaps conclusively, that some sort of reaction is taking place in the molten PbA,-Pb(OH),A, or pure Pb(OH),A, systems. This interpretation is in agreement with the work of Boston et all6 on the conductivity maximum observed for Al,Cl,. The explanation was based on a chemical reaction in which traces of water reacted with Al,Cl, to produce HCl. However, it has also been suggested that the maximum observed in the plot of log (conductivity) against inverse temperature for some molten inorganic salts might be due to the possibility of the cation-anion complexes becoming more abundant at higher temperature^.^' Such a tendency, in which the coulombic force plays a significant role, would be considerably reduced in molten organic .ionic salts such as metal carboxylates (soaps), as they generally melt to give more mobile liquids than the fused inorganic salts.Thus an explanation based on the formation of cation-anion complexes in PbA,-Pb(OH),A2 or pure Pb(OH),A, would not be sufficient to account for the observed maxima in fig. 1. The activation enthalpy for conductance (AH* = AHg +AH/3) in table 2 decreases up to a certain composition of Pb(OH),A, and then increases. The breaking and making of bonds occurring in the melt should lead to a change in the microscopic structure of the melt, especially when the effect of the substituted dihydroxy groups increases and starts to play a dominant role.Such effect would certainly affect the activation energy for conductance. A case in support of this view is the interpretation of thermal data on lead@) dodecanoate-lead@) oxide mixtures in which the addition of PbO is suggested to change the structure of the liquid phase from small, essentially spherical micelles into long, cylindrical rni~e1les.l~ The AH* value of 43.6 kJ mol-1 obtained for pure PbA, is in agreement with the reported value, as can be seen in table 3. The value is also close to the reported AH* values for some unsaturated, molten, pure lead(I1) carboxylates (table 3). The closeness of the activation energy values in this table confirms that the Pb2+ ions are the major charge carriers in molten lead@) carboxylates. References 1 M. E. Ekwunife, M. U. Nwachukwu, F. P. Rinehart and S. J. Sime, J. Chem. Soc., Faraday Trans. I , 2 U. J. Ekpe and S. J. Sime, J. Chem. Soc., Faraday Trans. I , 1976, 72, 1144. 3 S. 0. Adeosun, W. J. Sime and S. J. Sime, J. Chem. SOC., Faraday Trans. I , 1976,72,2470. 4 S. 0. Adeosun, A. 0. Kehinde and G. A. Odesola, Thermochim. Acta, 1979,28, 133. 5 S . 0. Adeosun, Thermochim. Acta, 1979,32, 1 19. 6 S. 0. Adeosun and M. S. Akanni, Thermochim. Acta, 1980,39, 35. 7 S. 0. Adeosun, M. S. Akanni and H. D. Burrows, Thermochim. Acta, 1980,42, 233. 8 L. F. Grantham and S. J. Yosim, J. Chem. Phys., 1966,45, 1192. 9 N. W. Gregory and H. V. Tartar, J. Am. Chem. Soc., 1948,70, 1992. 1975,71, 1432. 10 D. Swern, G. N. Billen, T. W. Findley and T. T. Scanlan, J. Am. Chem. SOC., 1945, 67, 1786. 11 I. L. Finar, Organic Chemistry, vol. 2, Stereochemistry and the Chemistry of Natural Products (William 12 R. D. Vold and G. S. Hattisngdi, Ind. Eng. Chem., 1949, 41, 23 1 1. Clowes, London, 5th edn, 1975), p. 170.M. S. Akanni and P . C. Mbaneme 3365 13 S. M. Nelson and R. C. Fink, J. Chem. Soc., 1954,4412. 14 S. 0. Adeosun, W. J. Sime and S. J. Sime, Thermochim. Acta, 1977, 19, 275. 15 M. S. Akanni, H. D. Burrows, F. Oshodi and H. C. Sime, Thermochim. Acta, 1982,54, 289. 16 C. R. Boston, S. J. Yosin and L. F. Grantham, J. Chem. Phys., 1969,51, 1669. 17 W. Fisher, K. Heinzinger, W. Hezzog and A. Klemm, 2. Naturforsch., Teil A , 1962, 17, 799. Paper 512160; Received 9th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203357

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82, 3367-3380 Optical Anisotropies of Alkylcyanobicyclohexyls and Related Compounds Patrick Navard*T and The Late Paul J. FloryS IBM Research Laboratory, Sun Jose, California 95120-6099, U.S.A. Mean-square optical anisotropies, y2, of cyanocyclohexane, bicyclohexyl, and the trans-4-cyano-trans-4‘-n-alkylbicyclohexyls, NC-C,H,,-C,H,, --C,Han+?, with rn = 2, 5 and 7, have been determined from depolarized Rayleigh light-scattering measurements conducted on dilute solutions in carbon tetrachloride. The results are treated on the basis of ad- ditive contributions of group anisotropies: r,, = Aa,, - 2Aa,, and TccN. = Aa,,, - AqH, where the Aa denote the anisotropies of the polariz- abilities of the indicated bonds. Diverse conformations of the bicyclohexyl system are taken into account.The value = 0.68 A3 required for bonds comprising this group appreciably exceeds the value of r,, = 0.53 8, for the n-alkanes and presumably for the alkyl chains. Similarly, rccN = 3.3 A3 is required for the cyano group attached to bicyclohexyl, compared with 2.15 A3 for the same group in acyclic nitriles and in cyanocyclohexane. These differences appear to be due to inductive effects that are known to impair a simple additive scheme. Contributions of members of the alkyl chain beyond the first to the tensor & representing the anisotropy of the polariz- ability virtually vanish when averaged over the configuration of the chain. Hence, properties such as the electric birefringence should be independent of the length of the alkyl chain under conditions that allow the latter to be in a random configuration. The quadratic tensor invariant, y2 = iTr(&&), should increase gradually with chain length.The relatively large alternating effects in the values of y2 for odd and even values of m are explained by the persistence of correlations in the directions of successive bonds of the alkyl chain. The effect of a departure of the principal axis of the polarizability tensor from the major geometric axis of the molecule is illustrated and its relevance to the anisotropic dispersion forces operative in liquid crystals is pointed out. The orientation-dependent energy attributable to the anisotropy of the polarizability is much smaller than would be required by the observed nematic-isotropic transition temperatures.This suggests that the large dipole moment of the molecules may be responsible for the nematic stability of the cyanobicyclohexyls. The cyano group occurs prominently among compounds that, under suitable conditions, manifest liquid crystallinity. Representative examples include trans-4-trans-4’-disubsti- tuted bicyclohexyl compounds (I) and the p,p’-biphenyl derivatives (11) where R is an alkyl group, -(CH,),-H, with rn = 3-10. t Present address: Ecole Nationale Superieure des Mines des Paris, Centre de Mise en Forme des $ Deceased 8th September, 1985. Mattriaux Sophia-Antipolis, 06565 Valbonne, France. 33673368 Optical Anisotropies of Alkylcyanobicyclohexyls The anisotropy of the optical polarizability plays an important role in most liquid- crystalline compounds.2~ Hence, it is of interest to study the optical anisotropy of the nitrile group.The work of Le Fkvre et aL4 on the static electric birefringence of lower alkyl nitriles indicates differences, Aa, between the polarizabilities parallel and perpendicular to the CN bond of ca. 2 A3. This value is substantially smaller than the anisotropy found for the phenyl group in most compounds; it is only slightly greater than the anisotropy associated with the C-C1 and C-Br bond^.^ The electric moment of the cyano group, ca. 3.6-4.0 D,* is quite large. The small size of this group enhances its anisotropic interactions, both dipolar and dispersive, with neighbouring molecules. A puzzling aspect of the manifestations of liquid crystallinity among compounds of these types is the fact that the cycloaliphatic nitriles exhibit nematic-isotropic transition temperatures comparable with those of the corresponding aromatic nitriles (II).s This is surprising in view of the greater anisotropy conferred by the phenylene groups in the latter compounds compared with the cyclohexylidene groups.This should promote molecular alignment in the cyanobiphenyls and hence raise their transition temperatures. In order to explore these issues more thoroughly we have determined the optical anisotropies of various cycloaliphatic compounds, including nematogenic substances of type (I), by performing depolarized Rayleigh scattering (DRS) measurements on their solutions in carbon tetrachloride. Evaluation of the anisotropy of the polarizability conferred by the cyano group in the molecules in question may thus be determined. Similar measurements on corresponding aromatic compounds are presented in the following paper.' Experiment a1 Materials Carbon tetrachloride, benzene and cyclohexane were of analytical grade.Cyanocyclo- hexane and bicyclohexyl were purchased from Aldrich Chemicals. The trans-4-cyano- trans-4'-alkylbicyclohexyls were obtained from Merck and were used without purification. Depolarized Rayleigh Scattering Intensities of the depolarized Rayleigh scattering (DRS) by solutions of the various com- pounds in carbon tetrachloride were measured at an angle of 90" using polarized radiation from an He-Ne laser (A = 632.8 nm) according to methods previously des~ribed.~? 8-11 All measurements were conducted at 22 2 "C.Measuring the attenuation of the intensity by narrow bandpass filters having bandwidths of 1.2 and 3.5 nm, respectively, provided a basis for subtracting the collision-induced components of the scattering from that due to the intrinsic molecular ani~otropy.~~ The intensity, Zmol, of the DRS attributed to the solute in solution was obtained from these corrected intensities, and Is?ggent for the solution and solvent, respectively according to the relation (1) zmol = polution- mol (1 - 4) P?gZent where 4 is the solute volume fraction. Intensities Zmol were converted into absolute Rayleigh ratios RHv; mol by comparison with Icc14(3.5), the intensity of scattering by carbon tetrachloride observed using the 3.5 nm filter.Thus * 1 D x 3.3356 x C m.P . Navard and P . J . Flory 3369 / -0- 5 - I I I I I I I 0.00 0.10 0.20 0.30 0.40 volume fraction of solute, q!~ Fig. 1. Apparent optical anisotropies of bicyclohexyl (0) and cyanocyclohexane (0) plotted against the volume fraction 4 of the solute in carbon tetrachloride. where n" is the refractive index of the medium. The (reference) Rayleigh ratio RS+(3.5) for carbon tetrachloride under these conditions as previously determined by Patterson8 is (7.4 f 0.1) x The calibration was confirmed by measurements on benzene. Pike et a1.12 found R,v+vv = 11.84 x cm-l for the absolute Rayleigh ratio of liquid benzene and pu = 0.432 for its depolarization ratio, both being determined with unfiltered radiation at II = 633 nm. It follows that REV = 2.56 x cm-l; our measurements on the neat liquid, made without filters, yielded RHV = 2.50 x cm-l.The calibration is also confirmed by measurements on benzene in a dilute solution of CCl, which, as reported in the following paper,' yields results in close agreement with previous meas~rements.~~ l3 cm-1. Optical Anisotropies Apparent optical anisotropies, yipp, were calculated according to where p is the solute number density and A is the wavelength in vacuo. Extrapolation of yipp to infinite dilution yields the intrinsic molecular anisotropy y2 for the solute, i.e. Y 2 = lim (Yipp)g+o. (4) An alternative internal field correction to eqn (3) can be found in the 1iterat~re.l~ Results Values of yipp, determined as described, are plotted in fig. 1 and 2 against the volume fractions for several of the compounds investigated.In fig. 2 and in the text, the various compounds are identified by notation as follows: ch stands for cyclohexyl or trans- cyclohexylidene, and NCch2C, for trans-4-cyano-trans-4'-alkylbicyclohexyl, where m is the number of carbon atoms in the n-alkyl group. The straight lines in fig. 1 and 2 were determined by the method of last squares. Values of the intrinsic optical anisotropy, y2, or its configurational average, ( y2), were obtained by extrapolation to 4 = 0. The results deduced from these measurements and from others not included in the figures are collected in table 1. Values of y2, or ( y 2 ) , deduced by calculations from tensors of the constituent groups as described below, are given in the final column of table 1.3370 60 50 W s -3 01 r- 40 30 Optical Aniso trop ies of A lky lc yano bic yclo hexy Is I I I I I I I 0.00 0.10 0.20 0.30 0.40 volume fraction of solute, 4 Fig.2. Apparent optical anisotropies of NCch,C, (O), NCch,C, (e), and NCch,C, (A) plotted against the volume fraction 4 in carbon tetrachloride. Table 1. Summary of principal results for cycloaliphatic compounds compound ~~ observed calculat ed" cyclo hexane/ch cyanocyclo hexane/NCch bicyclohexyl/ch, NCch,C, NCch,C, NCch,C, NCch,C, NCch,C, NCch,C, NCch,C, 0.9 1.12 4.7 - 33.4 33.8 32.5 35.6 34.2 34.4 36.1 35.0 37.9 36.5 5.6 (5.6) - - - - "The values listed were calculated using the following parameters, expressed in A3: TCc = 0.53, r'& = 0.68 and rCCN = 3.30. The Anisotropy Tensor The optical anisotropy y2, evaluated from experiments according to eqn (2) and (4), is related to the polarizability tensor a according to y2 = ijTr(&&) ( 5 ) 6 = a-iETra (6) where & is the anisotropic part of a, i.e.P.Navard and P. J. Flory 337 1 E being the unit tensor. If the principal axes of cii (and of a) coincide with those of the reference frame of choice, cii may conveniently be resolved into its cylindrical (Aa) and acylindrical (Aat ) parts according to 6 = AaA+AatAt (7) where A = diag (#, -$, -+) (8) At = diag (O,& -+) (9) and Aa = a, - (a, + %)/2 (10) Aat = %-a, (1 1) y2 = (At~)~++(Aat)~. (12) a,, a,, and 05 being the principal components of ii. It follows that The foregoing relationships express the connections between the molecular anisotropy and the experimentally measurable quantity y2.The molecular tensor ii is formulable, in principle, as a sum of contributions from constituent groups. It provides the link between chemical constitution and y2 and other manifestations of the optical anisotropy. If, as in the case of the compounds considered here, the individual groups (consisting of C-C and N-C bonds) are cylindrically symmetric, then their respective contributions Aa suffice for formulation of 8. The phenyl and phenylene groups of the aromatic compounds treated in the following paper' are acylindric ; hence contributions Aat for these groups must be included. The additional experimental measurement required for evaluation of A d may be furnished by the molar Kerr constant deduced from the electric birefringence.Analysis of Results Two Cartesian coordinate systems referred to the cyclohexyl group are shown in fig. 3. In one of these, Xo &Zo, the Xo axis is taken parallel to the equatorial terminal bonds, and the Xo y0 plane bisects the cyclohexyl group; the direction of the Zo axis perpendicular to this plane is chosen to complete a right-handed coordinate system. The axes of the second reference frame, XYZ, obtained from the first by a rotation z about the common axis 2, are symmetry axes of cyclohexane (but not of the cyclohexyl or cyclohexylidene groups). The X axis is drawn through the mid-points of the pendant equatorial bonds. The transformation from Xo yOZ0 to XYZ is expressed by COST -sin7 R,(z)= [s;z c y z g 1.(13) For tetrahedral bonding z = sin-,(+). If the nematogenic molecule is approximated by a rod, the coordinate system XYZ offers the advantage that its X axis is close to the rod axis (geometric long axis). Hence it may be preferred for the analysis and interpretation of the mesomorphic state in compounds of the kinds to be considered. The Xo & Zo system is easier to employ in the formulation of the polarizability tensor from its constituents. Hence, the following analysis is carried out in this reference frame. The cyan0 group, when present (in the equatorial position), is taken to occur on the left-hand side of the cyclohexyl (or bicyclohexyl) core, its symmetry axis being parallel to X,,; the alkyl group is then attached on the right-hand side. On the assumption that all the bond angles are tetrahedral, each of the carbon-carbon bonds comprising the cyclohexane molecule and the molecules of its cycloaliphatic3372 Optical Anisotropies of Alkylcyunobicyclohexyls \ Fig.3. Cyclohexylidene group in the chair form. Bonds are labelled according to direction: e for equatorial, a for axial, and + and - for guuche+ and gauche-, respectively. Also indicated are alternative positions for bonds of a pendent n-alkyl group assigned on the tetrahedral lattice. Directions of the axes of the right-handed Cartesian coordinate systems X,, GZ,, and XYZ are shown. homologues is parallel to one of the four axes identified in fig. 3 by symbols e (for equatorial), a (for axial), + (for gauche+) and - (for gauche-), respectively.The molecular anisotropy tensor can be assembled from tensors Be etc., one for each C-C bond and the associated deficit of two C-H bonds eliminated in forming the C-C bond8$ 1 5 9 l6 The cycloaliphatic molecules may be considered to be formed in this fashion from isotropic methane molecules, for example. For a C-C bond parallel to the Xo axis, i.e. for an equatorial bond, it follows that f i e = Tcc A = TCc diag (3, -4, -4) (14) where TCc = Aacc - 2AaeH (15) with AaCH and Aacc [see eqn (lo)] formally identified with the anisotropies of the C-C and C-H bonds, re~pectively.~* 8* l5 The corresponding tensor fi, for an axial bond (see fig. 3) is obtained from Be by rotation about Zo through the tetrahedral angle, the transformation given by eqn (1 3) being used for this purpose.Thus, with z = cos-l(- k), 6, = R,(z) Be RZ(z) r-2 - 2 4 2 o 1 = (rcc’g) I - 3 O I * Rotations of Ba by +2n/3 about X,, yield - 1 T243 * (17) 1 -2 d2 +d6 3 &+ = ( r c c l g ) (Redundant off-diagonal elements are omitted from these symmetric tensors and others DRS measurements by Patterson8 on n-alkane homologues yield 2 _ff???i3 when a small correction,5 - 0.01 A3, is applied to compensate for the error arising from taking the angles to be exactly tetrahedral. Using the tensors given by eqn (14), (16) and (17) one may formulate tensors for the various cyclic aliphatic molecules, subject to the limitations of additivity. Contributions of any other groups that are present can be incorporated similarly.P. Navard and P. J. Flory 3373 C y clo hexane Application of the procedure to cyclohexane in its stable chair confirmation, with all bond angles tetrahedral (see fig. 3), yields & = 2(8,+8+.+ a _ ) 4 4 4 2 0 It follows from eqn (5) that YEh = 4r& = 1.12 A'. The experimental result obtained from DRS measurements on solutions in CCl,, as reported in table 1, is lower than this calculated value. The difference, however, does not exceed the experimental error in the determination of the small optical anisotropy of cyclohexane. The apparent molecular anisotropy calculated from the DRS of the neat liquid, but not recorded in table 1, is 0.8 A', in approximate agreement with the value in dilute solution.? C y anocyclohexane This molecule, denoted by NCch, may assume either of two conformations. In one the cyano group is in the equatorial position and in the other it is axial to the ring.Determinations of the equilibrium ratio of the cis to trans isomers in 1,4- alkylcyanocyclohexanes in which the steric requirements of the chosen alkyl group preclude its occurrence in the axial position provide a basis for evaluating the proportions of the two conformers in NCch. According to the results of Allinger and Szkrybalo,18 the equilibrated mixture contains 59% of the trans isomer in which CN is in the equatorial position; Rickborn and JensenlB found 56%. On the plausible premise that these percentages reflect the preference for the equatorial conformer in NCch,18-20 we adopt the mean of these results, namely 0.575, as the fractional Occurrence of this conformer. The anisotropy tensor for NCch may be derived, formally at least, from that of cyclohexane by taking account of the replacement of a hydrogen atom by CN, either in the equatorial or axial position.The effect of such substitution is represented in a reference frame with the X axis parallel to the C-CN bond by rCCN A where AaccN being the anisotropy associated with the CN group and the C-CN bond. If the CN group occupies the equatorial position &Cch, e = &h+ rCCN A = ~ [ 4rCC + 6rCCN 4 4 2 rcc 0 - iorcc-3rCCN 0 6rCC - 3rCCN Rotation of rCCN A about the Zo axis through the tetrahedral angle, the transformation being effected through use of eqn (13), gives 4rCC-2rCCN 4d2 rCC-2d2rCCN 0 - i o r C C + 5rCCN 0 ] (22) [ 6rcc - 3rCCN aNCch, a = t t The DRS measurements reported by Foulani and Clkmentl' on dilute carbon tetrachloride solutions of n-alkanes and cyclohexane and on pure cyclohexane follow a pattern similar to that found here.Their values of y2 are much larger, however, for reasons pointed out previously.*3374 Optical Anisotropies of Alkylcyanobicyclohexyls for the axial conformer [cf. eqn (1 6)]. From these expressions one obtains [according to (23) (24) Combining eqn (23) and (24) according to the roportions of the equatorial and axial forms indicated above, substituting TCc = 0.53 13, equating the result to the observed value (Y&Cch) = 4.7 A6 (see table 1) and solving the resulting quadratic expression, one obtains rCCN = 2.15 A3. For purposes of comparison, we note that molar Kerr constants of acetonitrile, isobutyronitrile and pivalonitrile determined from electric birefringence measurements by Le Fkvre et aL4 together with dipole moments tabulated by McClellan21 yield r C C N = 1.75 A3, 2.12 A3 and 2.28 A3 for these respective compounds.Our result is in close agreement with the value for the second of these compounds, which is the nearest analogue of cyanocyclohexane. eqn (511: YkCch, e = 4r&2+trCC rCCN+r&N ~Rcch, a = 4Qc -4rcc ~ C C N + QCN. Bic yclohex y 1 Raman spectra22 reveal three principal conformers of this compound: (i) the symmetric equatorial-equatorial conformer, ee-sym, in which the coordinate systems Xo y0 Zo and X& 2; for the two cyclohexyl rings are parallel, (ii) the equivalent but chirally related pair, ee &, obtained from ee-sym by rotations of & 2n/3 about the central C-C bond joining the two rings, and (iii) the symmetrical equatorial-axial conformer, ea-sym, in which the Xo y0 and Yo Yo planes for the two rings are coincident.Other conformers generated by rotations about the central C-C bond in ea are precluded by steric repulsions. Integrated intensities of the Raman lines identified with these three conformers indicate mole fractions of 0.356, 0.546 and 0.098, respectively.22 The similarity in the abundances of each of the three ee conformers, namely ee-sym, ee+ and ee-, is consistent with the conclusion reached by Scott et al.23 from their investigations on 2,34imethylbutane, according to which the energies of the two conformers of this molecule, analogous to ee-sym and ee of bicyclohexyl, respectively, differ by < 0.1 kcal mol-l.The anisotropy tensors for conformers of bicyclohexyl are most readily derived by combining the group tensors given by eqn (14), (1 6) and (1 7) for the several C-C bond directions on the tetrahedral lattice. We thus obtain see, sym = 5&,+4&, +4&- 14 8 4 2 0 see+ = 58,+2riia+2&+ - +4dT 4 2 4 2 &246 =(rhc/9) -11 f 4 4 3 [ -3 and, in the reference frame Xo YoZo embedded in the equatorial cyclohexyl group, &a, sym = 38,+2aa+4&+ +4&- -2 4 4 2 0 = G c / 9 ) [ -7 ;]*P . Navard and P . J. Flory 3375 A prime is appended to I?& in these expressions to permit the choice of a value of rl,, that departs from that for n-alkanes; see below. They yield The experimental result, (&,) = 5.6 As, in conjunction with the fractions of the several conformers given above, yields = 0.68 A3.The difference compared with TCc = 0.53 A3 obtained from measurements on n-alkanes may be due to the presence of two trisubstituted carbon atoms in bicyclohexyl, or to the enhancement of inductive effects arising from the greater size and compactness of this molecule. Whatever the cause may be, rhc is assigned below to CH, groups that are part of the bicyclohexyl system, in interests of consistency in the correlation of results. Alkyl Cy anobicyclohexyls Alkyl substitution in the trans-4 position should prevent the occurrence of conformations in which both the alkyl group and the central C-C bond are joined axially to the ring thus substituted.According to evidence cited above, on the other hand, the CN group imparts only a small preference for the equatorial form. The ring bearing the CN group might therefore be expected to adopt the axial, axial (chair) form to a limited extent. Inasmuch as only one of the rings of the trans-4-cyano-trun~-4'-alkylbicyclohexyl may assume the axial conformation, a maximum of 5% of this conformer is to be expected. Two frequencies are observed in the Raman spectra2, of compounds of this series. These correspond to the ee-sym and eef forms found in bicyclohexyl. In the ethyl homologue, which we denote by NCch,C,, the intensities indicate the occurrence of these conformers in the proportion 0.47/0.53. For both NCch,C, and NCch,C, the intensities are in the ratio 0.48/0.52.A frequency corresponding to the ea form is not observed in any of the alkyl cyanobicyclohexyls, from which we infer that the fractional amount of this conformer may be substantially smaller than the upper limit estimated above. The anisotropy tensors for the alkyl cyanobicyclohexyls are readily formulated by adding the contributions of the cyano and alkyl groups to that of the bicyclohexyl residue. Thus (29) &NCchzR = achz+rCCN The first term on the right-hand side is given by eqn (25) or (26), depending on the conformation of the bicyclohexyl residue ; the ea-sym conformation whose tensor is given by eqn (27) may be ignored according to the Raman spectra. The last term in eqn (29) represents the contribution of the alkyl group, which depends on its conforma- tion.The alkyl group of the ethyl homologue, NCch,C,, may occur in either of two chirally related conformations depending on the torsion about the (equatorial) bond joining the ethyl group to the cyclohexylidene ring. In one of them, which we denote by R,, the C-C bond of the ethyl group takes the position labelled + in fig. 3; in the other conformation, R-, it takes the - position in fig. 3. In the R, conformation the terminal C-C bond is periplanar to the + bond in the adjoining cyclohexylidene ring and syn to the - bond. In conformation R- these relations are reversed. The third (staggered) conformation, in which the terminal C-C bond is syn to both the + and - bonds of the cyclohexylidene ring, may be ignored because of its appreciably higher energy.,* Inasmuch as the two chirally related conformations are locally equivalent, it suffices to consider only one of them.We choose the conformation R,. Having made a choice at this juncture, one is obliged to distinguish the conformations ee+ and ee- of the bic y clo hex y 1 sy s tem . In formulating the optical anisotropy tensors for the relevant conformers of NCch,C,, and also its higher homologues, we take rkc = 0.68 A3 for C-C bonds in the3376 Optical Anisotropies of Alkylcyanobicyclohexyls bicyclohexyl ring system in order to maintain consistency with the analysis of ( y 2 ) for ch,. The C-C bonds of the alkyl group are assigned Tcc = 0.53 A3. The anisotropy tensor ii for each of the conformers ee-sym, ee+ , and ee - , with the R, conformation assigned to the ethyl group in each instance, may then be formulated and the corresponding y2 may be obtained as a function of rCCN.Combining these expressions in the proportions 0.47,0.265 and 0.265 and setting the result equal to the observed value of ( y 2 ) , namely 33.8 A6 (see table l), one obtains r C C N = 3.43 A3, which is much greater than the value deduced above from y 2 for NCch. Agreement between calculated and experimental values of ( y 2 ) for the three homologous alkylcyanobicyclohexyls with m = 2, 5 and 7 (see below) is optimized by taking TCCN = 3.3 A3. Optical anisotropies y2 calculated for each of the three conformers of NCch,C,, ee-sym, ee+ and ee-, respectively, are 35.6, 28.7 and 30.6A6. Averaging these values as above, we obtain ( y 2 ) = 32.4 A6, the calculated value recorded in table 1.Computations for the homologues with longer n-alkyl groups were carried out by generating the various conformations with bonds arrayed on the tetrahedral lattice in which the cyclohexyl groups are represented. Each conformation was weighted in approximate accordance with the scheme and analysis successfully applied to n-alkanes and related chain m01ecules.~~~ 26 Specifically, for a bond following one assigned to the trans state, conditional probabilities of 0.50, 0.25 and 0.25 were adopted for its assignment to trans (t), gauche+ (g+), and gauche- (g-), respectively. For a bond following a g * placement, the conditional probabilities were approximated by 0.70,0.30 and 0 for t, g* and g”, respectively, successive gauche assignments of opposite sign being disallowed with an error that is insignificant for the present purpose.All eligible conformations of the alkyl group were generated and each was weighted in this manner. As stated above, the conformation for the first bond was so assigned as to place the second bond in the R, orientation identified in fig. 3.7 The second bond of the chain is subject to the conditional probabilities applicable to a bond whose predecessor is g+ (see fig. 3). Adoption of Tcc = 0.53 A3 for each bond of the alkyl group fulfills requirements for specification of the anisotropy tensor for every chain confirmation. Combination with the tensor for the cyanobicyclohexyl residue in each of its three conformations, ee-sym, ee + and ee - , according to eqn (29) yields the corresponding molecular tensors ii and the associated values of y2.Weighting the bicyclohexyl conformations in the proportions 0.48, 0.26 and 0.26, respectively, as indicated by the Raman intensities, and weighting the conformations of the alkyl group in the manner described, one obtains the average ( y 2 ) over all conformations of the molecule. Results of these calculations are given in the last column of table 1. Satisfactory agreement with the experimental measurements is achieved through adoption of rCCN = 3.3 A3; see above. Comparison of the calculations for alkyl groups of successive lengths reveals a pronounced alternation in the values of (7,) for successive members of the series, those with an odd number m of carbon atoms being greater by 1.5-2.2 As than the mean for m - 1 and m + 1.This oscillation finds explanation in the preferential incidence of bonds with the equatorial orientation among those separated from the ring by an even number of intervening bonds. Electric Birefringence and the Averaged Anisotropy Tensor Anisotropies of the electric polarizabilities of nematogenic molecules are frequently evaluated from measurements of electric birefringence. For dipolar molecules, the t The alternative assignment R- of the torsion about the first bond yields the equivalent set of chirally related conformations, for each of which y2 is identical to that for the corresponding member of the set adopted.P . Navard and P . J. Flory 3377 molar Kerr constant ,K thus determined usually depends principally on the quantity /?, defined by were p is the dipole moment and pT is its transpose. The additional contribution due to the induced polarization amounts to < 2% of the electric birefringence of the cyanobicyclohexyls.The term in /? makes the major contribution owing to the large dipole moment. Inasmuch as the induced polarization is proportional to ( y 2 ) , an appropriate allowance for its contribution is straightf~rward.~-ll It is important to observe that the quantity p depends on the average ( a ) of the tensor 6 over all configurations of the molecule. The averaging is conveniently performed in the Xo Y , Zo reference frame in which the dipole moment remains parallel to one of the axes. The DRS, on the other hand, depends on the tensor invariant y2, a quadratic quantity that is required to be averaged over all configurations of the molecule, including those of the alkyl group.Whereas ( y 2 ) must increase, without limit, with the length of the n-alkyl chain, /? must converge with chain length, the absence of long-range correlations between bonds of the chain being assumed. In fact, contributions of bonds beyond the first to ( 6 ) are found to be negligible for a tetrahedral chain subject to the configurational statistics adopted above. These bonds are distributed nearly uniformly among the four directions of the tetrahedral lattice. Hence, the molecular tensor 6NCch2R represented in Xo &Zo and averaged over all configurations of the alkyl chain may be approximated B = ( P T W (30) by (&NCchpR) (aNCch2C~) (31) where the angle brackets denote configurational averaging that includes the alternative conformations of the bicyclohexyl residue.This relation holds for an alkyl chain of any length m, provided that it is in a random configuration in (approximate) conformity with the established configurational statistics of such chains.26 The error does not exceed - +0.05 h for m > 2. It follows from eqn (29) that were 6chz is given by eqn (25) or (26), depending on the conformation of the bicyclohexyl residue, and Be is the contribution of the (equatorial) pendant bond according to eqn (14). Since the dipole moment is directed along the Xo axis, the quantity p, defined by eqn (30), depends on the 1,l element of 6 ; i.e. A1 ternatively B = f p 2 Z (34) where Kii = q(6NCCh2R)11 is the cylindric part of the anisotropy referred to the Xo axis.Inasmuch as the 1,l element of acn2 is the same for the different conformations of the bicyclohexyl residue according to eqn (25) and (26), it follows from eqn (14), (31) and (32) that (3 5 ) regardless of the apportioning of the conformations of the bicyclohexyl system between ee-sym and ee f . Eqn (34) and (35) should apply to all homologues, irrespective of the length of the alkyl chain, provided that the chain configuration is random; see above. According to the values of the parameters adopted above, Z = 5.42 A3. Assi ning p the value 3.6, D, typically found for aliphatic nitriles,2l we obtain j? = 48.1 D2 i3. The molar Kerr constant reported by Dunmur and Tomes27 for NCch2C, yields B = 98 D2 A3.The reason for the large discrepancy in this instance is not apparent. Results of electric birefringence and DRS measurements on analogous aromatic compounds are in better agreement, as shown in the following paperm7 - ~a = irkc + rCCN + rcc3378 Optical Anisotropies of Alkylcyanobicyclohexyls Discussion The X-axis in fig. 3 more nearly approximates the geometric long axis of the molecule. With reference to the nematic state, the anisotropy tensor should therefore be presented in the XYZ system rather than Xo y0 Zo. Transformation of (aNCchlR) given by eqn (3 1) and (32) to this reference frame may be effected by use of R,(z), expressed by eqn (13) with z = sin-l (i). In the nematic fluid, uniform rotation about the long axis of the molecule, approximated by X, may be expected.The cylindrical (traceless) tensor, obtained by averaging (aNCchlR) over rotations about the X axis, is determined by its 1,l element in XYZ or by the corresponding anisotropy m, which is three-halves of this element. For the significant conformers of bicyclohexyl we thus obtain and = 8(rccN+rcc)+yr& = 4.66 A3. (37) These values are appreciably lower than Z = 5.42 A3, as is their weighted mean, 4.88 A3. The comparison illustrates the effect of a departure (19.5') of the principal axis of the tensor from the geometric axis of the molecule. The anisotropy rather than Z would appear to be quantity relevant to the stabilization of the nematic state by anisotropic London dispersion forces.2q The axial ratios for the alkyl cyanobicyclohexyls are much smaller than would be required for anisotropy of molecular shape alone to induce liquid crystallinity. The axial ratio x of the n-amyl (m = 5) derivative estimated as described 29 is ca.4.0. According to theory3 an axial ratio of 6.4 is required to induce nematic liquid crystallinity in a pure fluid without assistance from orientation-dependent interactions between molecules. The fact that x for the alkyl cyanobicyclohexyls is much smaller than this figure suggests that the intervention of strong orientation-dependent interactions should be required to sustain liquid crystallinity. A characteristic temperature P may be defined as the ratio of the orientation- dependent energy E: associated with contact between two mutually aligned segments divided by the Boltzmann constant, k.A segment is taken to be the portion of a rod-like molecule having a length equal to the mean breadth of the molecule. The number of segments in the molecule is thus identified with x. According to theoryl33 28y 29 P E E,*/k = 2ap* k-1(AdB)2 (38) where p* is the characteristic pressure, estimated to be ca. 500 J ~ m - ~ for the compounds considered, is the isotropic polarizability of the molecule and c is a constant approximately equal to 0.05-0.1 .13* 29 The molar polarization calculated for NCch,C, is 5 = 32.7 A3. Thus (AdD)2 = 0.023, a very small value. Substitution in eqn (38) yields P = 12-24 K, which is quite negligible. Results of d.s.c. measurements on alkyl cyanobicyclohexyls are given in table 2. According to theory,13 the observed nematic-isotropic transition temperatures TNI would require characteristic temperatures T* z 300 K, i.e.values very much larger than the anisotropy of polarizability would indicate. If the observed enthalpy transition ANN1 (table 2) is identified with the orientation- dependent energy -Eorient per mole, then P may be evaluated by use of the relation~hipl~, 28, 29 where s is the order parameter. Assuming s = 0.7, we thus obtain T* = 160 K for NCch,C,. Obviously, the orientation-dependent energy attributable to the anisotropy of the polarizability is much smaller than the observed heat of transition. As shown above, it is also much less than would be required to account for the stability of a nematic phase comprising molecules of axial ratio 4. is the hard-core volume of a segment, ca.70 cm3, Eorient = -+xRPs2 (39)P . Navard and P . J. Flory 3379 Table 2. Results of differential scanning calo- rimetry measurements of alkyl cyanobicyclo- hexyls, CNch,C, AHNI AsNI rn TNI/OC /kJ mol-1 /J mo1-l K-l 3 80 1.05 3.0 4 79 0.80 2.3 5 85 1.30 3.6 7 86 0.88 2.4 The large dipole moment may be responsible for the nematogenic properties of the cyanobicyclohexyls, as Brownsey and Leadbetter30 have suggested. The small size of the CN group allows two of them to approach within distances where the dipole-dipole energy may fall in the range of 10-20 kJ mol-l. The magnitudes of these interactions are manifested in the large heats of vaporization of the aliphatic nitriles. In acetonitrile, for example, dipolar interactions may account for more than half the energy of vap~rization;~~ the polar contribution is estimated as 17.5 kJ m01-l.~~ The local ordering resulting from these interactions in the isotropic liquid may reduce the configurational entropy below that for non-interacting rod-like particles of axial ratio x = 4. Pairing of cyano end groups in the nematic state, with the alkyl bicyclohexyl groups oppositely directed, would yield a dimeric species having an axial ratio > 6.It should be necessary, however, to allow for interactions between the CN groups in the isotropic state as well, for the reasons mentioned. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 S. Chandrasekhar, Liquid Crystals (Cambridge University Press, Cambridge, 1977). W.Maier and A. Saupe, Z. Naturforsch., Teil A, 1959, 14, 882; 1960, 15, 287. P. J. Flory and G. Ronca, Mol. Cryst. Liq. Cryst., 1979, 54, 289. R. J. W. Le Fevre, B. J. Orr and G. L. D. Ritchie, J. Chem. SOC., 1965, 2499. C. W. Carlson and P. J. Flory, J. Chem. SOC., Faraday Trans. 2, 1977,73, 1505. G. W. Gray, Philos. Trans. R. SOC. London, Ser. A, 1983,309, 77. P. J. Flory and P. Navard, J. Chem. SOC., Faraday Trans. 1, 1986,82, 3381. G. D. Patterson and P. J. Flory, J. Chem. SOC., Faraday Trans. 2, 1972,68, 1098. U. W. Suter and P. J. Flory, J. Chem. SOC., Faraday Trans. 2, 1977,73, 1521. B. Erman, D. C. Marvin, P. A. Irvine and P. J. Flory, Macromolecules, 1982, 15, 664. P. A. Irvine, B. Erman and P. J. Flory, J. Phys. Chem., 1983,87,2929. E. R. Pike, W. R. M. Pomeroy and J. M. Vaughan, J. Chem. Phys., 1975,62, 318. P. A. Irvine and P. J. Flory, J. Chem. SOC., Faraday Trans. 1, 1984, 80, 1821. T. Keyes and €3. M. Ladanyi, Adv. Chem. Phys., 1984,56, 411. R. P. Smith and E. M. Mortenson, J. Chem. Phys., 1960,32, 502. R. A. Sack, J. Chem. Phys., 1956,25, 1087. P. Foulani and C. Clement, Bull. SOC. Chim. Fr., 1969, 10, 3462. N. L. Allinger and W. Szkrybalo, J. Org. Chem., 1962, 27,4601. B. Rickborn and F. R. Jensen, J. Org. Chem., 1962, 27,4606. M. Manack, Conformation Theory in Organic Chemistry (Academic Press, New York, 1965), vol. 3. A. L. McClellan, Tables of Experimental Dipole Moments (Rahara Enterprises, El Cerrito, California, U.S.A., 1974), vol. 11. J. F. Rabolt and P. Navard, in preparation. D. W. Scott, J. P. McCullough, K. D. Williamson and G. Waddington, J. Am. Chem. SOC., 1951, 73, 1707. P. J. Flory, Statistical Mechanics of Chain Molecules (Wiley-Interscience, New York, 1969), A. Abe, R. L. Jernigan and P. J. Flory, J. Am. Chem. SOC., 1966,88,631. See ref. (24), pp. 133-152. D. A. Dunmur and A. E. Tomes, Mol. Cryst. Liq. Cryst., 1983,97,241. pp. 208-210.3380 Optical Anisotropies of Alkylcyanobicyclohexyls 28 P. J. Flory and G. Ronca, Mol. Cryst. Liq. Cryst., 1979, 54, 31 1 . 29 M. Ballauff and P. J. Flory, Ber Bunsensges. Phys. Chem., 1984, 88, 524, 530. 30 G. J. Brownsey and A. J. Leadbetter, J. Phys. Lett., 1981, 42, L135. 3 1 R. Kumar and J. M. Prausnitz, Solutions and Solubilities, vol. I11 of Techniques of Chemistry series, ed. A. Weissberger (J. Wiley and Sons, New York, 1971), chap. V. Paper 5/223 1 ; Received 18th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203367

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82, 3381-3390 Optical Anisotropies of Alkylcyanobiphenyls, Alkoxycyanobiphenyls and Related Compounds The Late Paul J. Floryt and Patrick Navard*$ IBM Research Laboratory, Sun Jose, California 95 120-6099, U.S.A . Mean-square optical anisotropies, y2, of cyanobenzene, anisole, cyano- biphenyl, p,p’-cyano-n-alkylbiphenyls, NC-CgH4-C,H4-C,H2,+1, with m = 6, 8 and 9, and p,p’-cyano-n-alkyoxybiphenyls, NC-C,H4-C,H4-O-C,H2,+,, with m = 6 and 9, have been determined from intensities of the depolarized Rayleigh scattering of their dilute solutions in carbon tetrachloride. Results are compared with previously reported determinations of optical anisotropies by other means. Treatment of the data according to a constitutive scheme corresponding to that applied in the preceding paper demonstrates consistency of the main results and enables calculations to be performed for homologues not included in the experiments.The difference between the polarizabilities parallel and per- pendicular to the long axis of the cyanobiphenyl group is Aa = 20.6 A3 in the alkyl series and 23.3, A3 in the alkoxy series; the difference is attributable to para substitution of the phenyl group by oxygen. Bonds of the randomly configured alkyl groups make a negligible contribution to the configuration- ally averaged anisotropy tensor, (a). Their contribution to ( y a ) is small but not negligible. The relevance of the anisotropies, Aa, to the manifestation of liquid crystallinity in the alkyl and alkoxy homologues is pointed out.The alkyl- and alkoxy-cyanobiphenyls are prominent among the more extensively investigated nematogenic substances. The anisotropies of the electric polarizabilities of members of the former series have been evaluated by various methods, and the results are generally discordant. The anisotropies of the alkoxy derivatives have not been determined. Electric birefringence measurements on benzene solutions of p,p’-cyano-n-amylbi- phenyl, NCph,C,, ‘reported by Dunmur et ally2 yielded = 17.5 A3. Coles and Jennings3 found = = 27 & 5 A3 from optical and d.c. Kerr measurements on the same molecule dissolved in carbon tetrachloride. Refractive indices and electric permittivities measured in the vicinity of the clearing temperatures of nematic fluids, 1-5 besides providing relevant information on the nematic state and the nematic-isotropic transition, may be put to the ancillary purpose of evaluating the apparent molecular anisotropies. From the refractive indices and per- mittivities parallel and per endicular to the director of nematic NCph,C,, Horn4 treat the results.Similar measurements by Dunmur et aZ.l yield = 30-35 A3. The n-heptyl homologue has been investigated also, but less extensively. Permittivity measure- ments of Lippens et aL5 on NCph,C, in the nematic state and the determinations of the order parameter by Karat and Madhusudanas yield= = 13.5-1 5 A3 for this compound. Results for various homologues of the series should not be expected to differ perceptibly; see the Discussion. Parameters such as AZ that ostensibly refer to the molecule are obtained Aa = 18 and 22.5 8: for this compound, depending on the procedure used to t Deceased 8th September, 1985.$ Present address : Ecole Nationale Supkrieure des Mines de Paris, Centre de Mise en Forme des Materiaux, Sophia-Antipolis, 06565 Valbonne, France. 338 13382 Optical Aniso tropies of A lky lcyanobiphen y Is 70 + P “r- 60 50 400 W 5 a 4 “* 350 300 0.00 0.10 0.20 volume fraction of solute, $ Fig. 1. Apparent optical anisotropies of cyanobenzene (a) and cyanobiphenyl(0) plotted us. volume fraction in carbon tetrachloride. subject, of course, to intermolecular induction and other effects when determined from measurements on the neat fluid. In this paper we present results of depolarized Rayleigh scattering (DRS) measurements on dilute solutions in carbon tetrachloride leading to the evaluation of the mean-squared optical anisotropies, y2, of representative mesogenic alkyl- and alkoxy-cyanobiphenyls and related compounds.Anisotropies of the various compounds are correlated and compared with data obtained by other methods as reported in the literature. Experiment a1 Carbon tetrachloride and benzene used in the measurements reported below were of analytical grade. Benzonitrile (cyanobenzene, NCph) was obtained from Pfalz and Bauer ; anisole and cyanobiphenyl (NCph,) from Aldrich Chemicals; and p,p’-cyano- n-octylbiphenyl (NCph,C,) from B.D.H. p,p’-Cyano-n-hexylbiphenyl (NCph,C,) and n-nonyl (NCph,C,), n-hexoxyl (NCph,OC,) and n-nonoxyl (NCph,OC,) analogues were kindly made available by Dr R.J. Cox of the IBM Research Laboratory. The procedures described in the preceding paper’ and in the references cited therein were followed. Results Measured values of y2 for several of the compounds investigated are plotted in fig. 1 and 2 against volume fractions in carbon tetrachloride. The compounds are identified by adaptation of the notation used in ref. (7). Intrinsic optical anisotropies y2, or their configurational averages ( y 2 ) , obtained by linear least-squares extrapolations of these plots and those for other compounds not shown are listed in the second column of table 1 .P. J. Flory and P. Navard 3383 0.10 0.20 volume fraction of solute, 4 Fig. 2. Apparent optical anisoiropies plotted against volume fraction in carbon tetrachloride for p,p'-cyano-n-hexylbiphenyl (O), p',p'-cyano-n-nonylbiphenyl (0) and p,p'-cyano-n- nonoxybiphenyl (A).Analysis of Results The reference frame chosen for representation of the aromatic compounds is shown in fig. 3. X , Y and 2 are the principal axes of a right-handed Cartesian coordinate system embedded in a 'core' consisting of phenyl or biphenyl together with the cyano group, if present. The tensor for the core may be formulated from cylindrical and acylindrical contributions, Aa and Aat, respectively, in accordance with eqn (7 j ( l 1 ) of ref. (7) the axes 1, 2 and 3 being identified with X , Y and 2. The choice of the X axis identified in fig. 3 as the quasicylindrical axis is appropriate for higher members of the series of compounds to be considered.Although ill-suited for benzene, with its cylindrical axis along 2, the anisotropy tensor for this molecule may be so representeda$ by taking A g h = iA& = 1.84 A3 as is required by the mean of the experimental values of y2 quoted in table 1. C y anobenzene According to electric birefringence measurements of LeFevre et aZ.,1° at 1 = 589 nm the molar Kerr constant of cyanobenzene (benzonitrile) is ,K = 1.06, x lo-* cm5 statvolt-2 mol-lt with &defined as in ref. (8). This result yields [see eqn (30) of ref. (7)] /I E pT&p = 6.99 D2 A'$ after deducting the small contribution due to induced polarization that is proportional to y2. Inasmuch as the dipole moment p is directed along the X axis = $p2Aa. (1) t statvolt = 300 V. $, 1 D z 3.3356 x lo-"" C m.112 FAR 13384 Optical Anisotropies of Alkylcyanobiphenyls Table 1. Summary of principal results for aromatic compounds Y2 or <r’>lA6 compound observeda calculatedb / A 3 benzene benzonitrile biphenyl (ph,) (NCPh) NCPh, NCPhZC, NCph2C5 NCph2C6 NCph2C7 NCPhZC8 NCPhZC9 anisole (phOC,) NCph,OC, NCph,0C7 NCph,0C8 NCph,OC, NCph,0C6 13.6 13.29 1 3.813 54.9 14013 326 _. 450 458 46 1 - 29.3 - 593 - - 596 - 453.0 459.2 454.9 458.8 455.9 458 590.4 596.8 592 597 593 - 1.84 - - 6.5, 11.2, 17.7 20.6 20.6 20.6 20.6 20.6 20.6 4.1, 23.3, 23.3, 23.3, 23.3, 23.3, a Values from the literature are identified by superscripts denoting the reference. Parameters used in the calculations are: r,, = 0.53 A3, r,, = 0.37 A3, the value of AaNCphz given in the fourth column and AaL,,,, = 4.3 A3.For benzene, biphenyl and cyanobiphenyl the core comprises the entire molecule. For benzonitrile and anisole it is phenyl. For all others it is the cyanobiphenyl (NCph,-) species. LeFevre and LeFevrell found p = 4.02 D for cyanobenzene. McClellan’sf2 tabulations indicate p = 3.99 D as the best value among closely a reeing results from several Substitution of this value in eqn (12) of ref. (7), together with the observed value of y k c p h given in table 1, yields Aa!kcph = 4.0 A3, with an uncertainty of ca. kO.5 A3. The result is larger than the values, 3.3-3.9 A3, found for monohalogenated derivatives of benzene,13 but the difference lies within the ranges of experimental error. Previous results** 9 9 13-15 indicate that the acylindrical contribution A d to the anisotropy of the phenyl, or phenylene, group is little affected by substitution.Subtraction of rCCN = 2.15 A3, as follows from the o tical anisotropy of anisotropy contributed by the phenyl group in benzonitrile. By comparison with for benzene the exaltation attributable to substitution by CN is ca. 2.6A3. A large inductive effect of CN is consistent with its promotion of electron delocalization. sources. Taking p = 4.00 D, one obtains Aa!Ncph = 6.5, R from the equation above. cyanocy~lohexane,~ from Aa,,,, = 6.5, A3 yields Aaph = 4.4 8: for the cylindrical Biphenyl and Cyanobiphenyl According to DRS measurements published re~ently,~ y2 x 140 Ae for biphenyl (ph,) dissolved in carbon tetrachloride. The Cotton-Mouton constant deduced from magneticP. J.Flory and P. Navard Y A I 3385 NC Fig. 3. n-Alkyl-cyanobiphenyl molecule with the alkyl chain in the all-trans planar conformation. The indicated torsion about the central bond of biphenyl renders the actual conformation non-coplanar. Torsional rotations about the bonds of the alkyl chain also may occur as indicated. The alkoxy chain is similarly represented but with bonds 1 and 2 (Cph-O and 0-C, respectively) fixed in planar (trans) conformations. birefringence measurements of LeFkvre and Murthy16 is ,C = 6.99 x cm3 G-, mol-l. The relevant invariant analogous to y 2 is iTr(&t) = (45kT14nN) ,C (2) where 2 is the traceless tensor representing the magnetic anisotropy, k is the Boltzmann constant and N is the Avogadro number. We thus obtain :Tr(@) = AaAx +- $ A d Ax? = 1.68 x J cm3 G-, Results of Lasheenl' on the magnetic susceptibility of biphenyl in the crystalline state yield Ax = 9.14 x in the same units. Substitution of these values in the equation above and simultaneous solution with eqn (12) of ref.(7) using the quoted value of y2 lead to A%,, = 1 1.2~813 and Ag,, = 4.3 813. Magnetic anisotropy parameters obtained from the tabulations of Schmalz et al.18 yield closely similar results. If allowance is made for the twisting of the phenyl groups about the long axis,9 then J Gb2 molecule-l and Ax? = 20.3 x where y/ is the dihedral angle between the planes of the benzene rings. According to an analysis of vibrational spectral9 and n.m.r.,2O w x 40". Substitution of this value together with A4,, in the equation above yields A g , x 2.8 A3, in tolerable agreement with values of 3.0-3.9 A3 for other substituted benzene derivatives.8* 13-15 The value A%, = 5.6 A3 for the cylindrical part of the anisotropy for each phenyl group in biphenyl reflects a large exaltation by its partner.Inasmuch as Aat is usually little affected by para substitution of a cylindrical substituent, we may assume that the value, 4.3 A3, of A 4 , , for biphenyl holds also for p-cyanobiphenyl (NCph,). From &--,ha given in table 1 we thus obtain AaNCphn = 17.7 813 according to eqn (12) of ref. (7). The contribution of A g h 2 accounts for only ca. 4% of y2 in this instance. The foregoing deduction of AaNCphz is supported by the electric birefringence measurements of Lalanne et aL21 which, in conjunction with the dipole moment ( p = 4.00 D) conferred by CN, yields 17.3 A3 for this quantity. The same authors21 determined y2 for NCph,, and for the homologues NCph,C,-NCph2Cl, as well, from DRS measurements carried out by a procedure that appears to resemble ours.Their values, however, are about twice those we have obtained. The disparity suggests an error in calibration. Comparison of h N C p h Z with A%h2 gives 6.4 A3 for the combined contributions to be 1 12-23386 Optical An is0 trop ies of A lky lcyano b ip hen y 1s attributed jointly to CCN and to its increase of A%,,,. If 2.15 A3 is allocated to rCCN according to results for cyanocyclohexane and the aliphatic nitriles,? the balance to be attributed to the increase is ca. 4.2 A3, an effect substantially greater than that associated with substitution of CN in benzene.That the increase in biphenyl should be larger follows from the greater delocalization provided by the second phenyl group. Alk ylc y anobiphenyls The conformations of the n-alkyl chains in the p,p’-cyanoalkylbiphenyls were treated similarly to the trans-4-cyano-trans-4’-alkylbicyclohexyls as described in ref. (7). The rotation b1 about bond 1 (see fig. 3) joining the n-alkyl chain to the phenylene ring is impeded by a moderate steric repulsion between the second methylene group of the chain and the hydrogen atom at the ortho position on the ring.22 In the coplanar conformation with 41 = 0, this repulsion is estimated at 200-400 cal mol-1 relative to the ‘perpen- dicular’ conformation where q51 = 7r/2.t In satisfactory approximation we have represented the continuum of torsional angles by two discrete states, one at dl = 0 and the other at 4 = z/2.For each conformation of the alkyl chain y2 was calculated separately for the respective torsional states of the first bond. The former state was given a weight of one-half compared to unity for the latter. The calculated averages ( y 2 ) over all conformations, discussed below, are little affected by plausible variations in the relative weighting of the two conformations. Even a two-fold change alters ( y 2 ) by only ca. I As. Inspection of models shows that rotation about the second skeletal bond is little influenced by the assignment given to the first bond. Hence, weightings of 0.5,0.25 and 0.25 were assigned to the t, g+ and g- states, respectively, for bond 2.?7 237 24 Conformations of succeeding bonds were weighted exactly as for the alkyl groups of the alkylcyano- bicyclohexyls treated in ref.(7). With the bonds arrayed on a tetrahedral lattice the tensors for the four lattice directions are Be = rcc diag (8, -8, -$) -2 2 4 2 0 (3) (4) where, adapting the notation used in ref. (7), we identify e with the direction of bond 1 in fig. 3, a with the direction of bond 2 in the planar conformation with dl = 0 and & with the directions of bond 2 for 41 = +27r/3. The tensor for the alkyl group in a specified conformation was taken to be the appropriate sum of the tensors given by eqn (3x5). Alkyl substitution may be expected to increase Aa for the cyanobiphenyl core.According to results for toluene, isopropylbenzene and t-butylbenzene* and their halogenated derivatives,13 the exaltation of Aa by alkyl substitution on benzene or on the p-phenyl group is 1-2 &. The effect should be greater for the biphenyl group. Optimization of agreement between calculations and experimental results for the alkylcyanobiphenyls requires ActNCph2- = 20.6 A3. (The dash appended to the subscript in this instance distinguishes the cyanobiphenyl radical from the corresponding molecule.) This corresponds to an increase of 2.9 A3 compared with Aa for the cyanobiphenyl t 1 cal = 4.184 J.P . J. Flory and P . Navard 3387 molecule. With Aahcph,- = 4.3 A3, the anisotropy tensor attributed to NCph2- in the alkyl cyanobiphenyls becomes dNCph2- = diag (1 3.73, - 4.72, - 9.02) in A3.The molecular tensor is obtained as the sum of eqn (6) and the tensor for the alkyl group. The tensors for q51 = n/2 are most easily evaluated by retaining the alkyl group tensor in the coordinate system in fig. 3 with q5 = 0 and rotating the core through n/2, whereupon its tensor is given by eqn (6) with the second and third elements interchanged. The values of ( y 2 ) given in the third column of table 1 have been obtained by averaging the y2 thus calculated over all conformations of the alkyl group, the orientations q51 = 0 and n/2 of each conformation being weighted as specified above. Oscillation of ( y 2 ) between even and odd homologues is large compared with the trend with chain length. Reasons therefore are the same as those cited for the similar behaviour in the alkylcyanobicyclohexyls.7 Anisole That the skeletal atoms of this molecule are coplanar is confirmed by crystallographic investigation^.^^^ 26 As should be expected, delocalization of electrons through the phenyl-oxygen bond is sufficient to override the comparatively weak repulsion between the methyl group and the ortho CH of phenyl.The contribution of the oxymethyl group to the anisotropy of the polarizability depends on rco, defined by were Aaco is the anisotropy ascribed to the C-0 bond. According to DRS measurements on diethyl ether and on the dimethyl ethers of ethylene glycol and of oligomers of the polyoxyethylene series,27 Tco = 0.37 A3. Taking the bond angle at oxygen to be tetrahedral, we thus obtain by addition of the tensors for phenyl and for each of the two C-0 bonds rco = A C ~ ~ ~ - A % ~ (7) for the molecular anisotropy of the polarizability of the anisole molecule.Again, $h is resolved into its cylindrical (A%,) and acylindrical (A4h) components; see eqn (7)-(11) of ref. (7). The observed value of y2 for anisole, given in table 1, together with AGh = 3.3 A3 as indicated by results8* l3? l4 for analogous derivatives of benzene, yields A%h = 4.l5ii3 for the phenyl group in anisole. This result is consistent with A%h = 4.0k0.15 A3 found for the phenyl group in diphenyl ~arb0nate.l~ Electric birefringence measurements of Aroney et aZ.28 on anisole in dilute solution afford an alternative route to evaluation of the anisotropy of its polarizability. They found ,K = 2.6 x 10-lo cm5 statvot2 mol-1 for the molar Kerr constant defined as above.8 The contribution of the dipolar term to the electric birefringence is comparatively small in this instance owing to the modest magnitude of the dipole moment, p = 1.24, D,28 and its unfavourable direction z = 105-1 19' from the X axis,29 p being measured in the positive sense.Consequently, the induced polarization accounts for approximately half of the birefringence. The difference gives /3 = 0.86 D2 A3. This result may be compared with the value obtained according to eqn (30) of ref. (7) and eqn (8) from the anisotropy of the polarizability deduced from DRS measurements. Substitution of the parameters given above in eqn (8) and execution of the operations prescribed by eqn (30) of ref.(7) with z = 105-1 10" yieldp = 0.72-0.89 D2 A3. The value3388 Optical Anisotropies of Alkylcyanobiphenyls deduced from the molar Kerr constant falls within this range. Hence the alternative methods appear to be in good agreement. Alkox ycy anobiphen yls Steric repulsion between the ortho hydrogen and the second methylene of the alkoxy group is much greater in gauche conformations of the 0-CH, bond (see fig. 3) than in the corresponding conformations of the alkyl group.24 The difference is a consequence of the shorter length of the C-0 bond compared to C-C. The g' conformations for bond 2 are virtually excluded on this account and hence may be disregarded. Thus, the first three bonds of the alkoxy group may be taken to be coplanar with the adjoined phenylene group.Owing to the smaller radius of the oxygen atom compared with methylene, discrimination between g and t conformations of bond 3 is negligible for our purposes; see the analysis of conformational interactions in oxyalkane chains by Abe and Mark30 in this connection. Accordingly, we weight them equally. Conformations of succeeding bonds of the alkoxy chain are weighted as for the alkyl chains in the a1 k ylc yano bic yclo hex yls7 and the a1 kylcy anobip hen yls . The molecular tensor ii may be expressed by fi=d NCph2+ii, + 82 + a 3 + iiR' (9) where contributions of individual bonds are identified by subscripts as numbered in fig. 3, and iiR' is the sum of contributions of bonds i = 4 to (m+2), inclusive, m being the number of carbons in the alkoxy group. Inasmuch as the first three bonds of the alkoxy group are presumed to be fixed relative to the cyanobiphenyl residue, the first four terms in eqn (6) may be combined to yield + 6R'.(10) 1 8 = 6 N C p h z + r 0 &o - Srcc 0 9 CO+&C v r c o 0 0 0 -&o-SLT [The second term follows from the corresponding term in eqn (8) with the contribution ii3 = r c c A of bond 3 added thereto.] The last term depends on the conformation specified by the torsional angles 43, . . ., dm+l, or by the corresponding rotational isomeric states. Upon taking TCc = 0.53 A3, Tco = 0.37 A3 and Aa~,,,,- = 4.3 A3 as above, we find agreement with the experimental determinations of ( y 2 ) to be optimized by letting AaNcphz- = 23.35 A3; see table 1. Comparison with the values of this parameter for cyanobiphenyl and for the alkylcyanobiphenyls indicates the very large increase, amounting to 5.6, A3, for the alkoxy group compared with 2.9 A3 for the alkyl group.Persistent oscillations in ( y 2 ) with the length of the chain are again revealed by the calculations. The reasons cited above apply here as well. Discussion The anisotropies of the polarizabilities of the cyanobiphenyls are much greater than those of the cyanobicyclohexyls investigated in ref. (7). The difference in the values of the mean-square anisotropies ( y 2 ) is most striking. These quantities for the cyano- biphenyls are dominated by the core; the contribution of the alkyl or alkoxy tail accounts for only ca. 5 % of ( y 2 ) . Moreover, the first bond of the alkyl group, or the first three bonds of alkoxy group, account for most of this latter contribution for chains of the length considered here.As is apparent from the calculated values in table 1, the trend with chain length is small. It is sustained without limit, however, as we have pointed out.'P . J. Flory and P . Navard 3389 The molecular anisotropy tensor ( 6 ) for the alkyl- and alkoxycyanobiphenyls averaged over all conformations of the pendant group and over rotations of the molecule as a whole about the long molecular axis (X) is well approximated by the sum of the cylindrical part of 6 for the core and for those bonds that are in a fixed array, i.e. bonds up to and including the first one about which torsional rotation is permitted. As found for the alkylcyanobicyclohexyls,7 the averaged contribution of remaining bonds of the alkyl group may be neglected, regardless of the latter’s length, provided that its configuration is random.For the p,p’-cyano-n-alkylbiphenyls we thus obtain with Aacore = AaNCph2- (see table 1). This value lies within the fairly broad range of previous estimates 1-5 quoted in the Introduction. For the n-alkoxycyanobiphenyls the first three bonds of the alkoxy group are included. It follows from eqn (10) that - AaNCph20R = Aacore+$rCO+rCC (12) = 24.1 A3 (see table 1). These values of r a should be applicable to estimation of the role of orientation- dependent interactions in promoting liquid cry~tallinity.~~ 22 The molecular axis X approximates the geometric long axis of the mesogen. Hence may be used directly for this purpose, in contrast to the circumstances encountered in ref.(7) (see the Discussion therein). The larger value of da for the alkoxy-cyanobiphenyls compared with the value for the alkyl-cyanobiphenyls may be presumed to contribute to the higher clearing temperatures observed for the former. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 D. A. Dunmur, M. R. Manterfield, W. H. Miller and J. K. Dunleavy, Mol. Cryst. Liq. Cryst., 1978,45, 127. D. A. Dunmur and A. E. Tomes, Mol. Cryst. Liq. Cryst., 1983, 97, 241. H. J. Coles and B. R. Jennings, Mol. Phys., 1978, 36, 1661; P. V. Kolinsky and B. R. Jennings, Mol. Phys., 1980,40,979. R. G. Horn, J. Phys., Paris, 1971, 39, 105. D. Lippens, J. P. Parneix and A.Chapoton, J. Phys., Paris, 1977, 38, 1465. P. P. Karat and N. V. Madhusudana, Mol. Cryst. Liq. Cryst., 1976, 36, 51. P. Navard and P. J. Flory, J. Chem. SOC., Faraday Trans. I , 1986,442, 3367. U. W. Suter and P. J. Flory, J. Chem. SOC., Faraday Trans. 2, 1977,73, 1521. P. A. Irvine and P. J. Flory, J. Chem. SOC., Faraday Trans. I , 1984,80, 1795. R. J. W. LeFevre, B. J. Orr and G. L. D. Ritchie, J. Chem. SOC., 1965, 2499. C. G. LeFkvre and R. J. W. LeFtvre, J. Chem. SOC., 1954, 1577. A. L. McClellan, Tables of Experimental Dipole Moments (Rahara Enterprises, El Cerrito, California, 1974), vol. 11. E. Saiz, U. W. Suter and P. J. Flory, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 1538. B, Erman, D. C. Marvin, P. A. Irvine and P. J. Flory, Macromolecules, 1982, 15, 664. P. A. Irvine, B. Erman and P. J. Flory, J. Phys. Chem., 1983,87, 2929. R. J. W. LeFkvre and D. S. N. Murthy, Aust. J. Chem., 1968,21, 1903. M. A. Lasheen, Philos. Trans. R. SOC. London, Ser. A, 1964, 256, 357. T. G. Schmalz, C. L. Norris and W. H. Flygare, J. Am. Chem. SOC., 1973,95, 7961. R. M. Barratt and D. Steele, J. Mol. Struct., 1972, 11, 105. L. D. Field, S. Sternhell and A. S. Tracey, J. Am. Chem. SOC., 1977,99, 5249. J. R. Lalanne, B. Lemaire, J. Rouch, C. Vancamps and A. Proutiere, J. Chem. Phys., 1980,73, 1987. P. J. Flory and G. Ronca, Mol. Cryst. Liq. Cryst., 1979, 54, 31 1. A. Abe, R. L. Jernigan and P. J. Flory, J. Am. Chem. SOC., 1966, 88,631. P. J. Flory, Statistical Mechanics of Chain Molecules (Wiley-Interscience, New York, 1969), pp. T. H. Goodwin, M. Przybylska and J. M. Robertson, Acra Crystallogr., 1950,3, 279. 133-1 52.3390 Optical Anisotropies of Alkylcyanobiphenyls 26 H. G. Norment and I. L. Karle, Actu Crystullogr., 1962, 15, 873. 27 G. D. Patterson and P. J. Flory, J. Chem. SOC., Furuduy Trans. 2, 1972,68, 1 1 1 1. 28 M. J. Aroney, R. J. W. LeFkvre and Shu-Sing Chang, J. Chem. SOC., 1960,3173. 29 M. H. Lombroso and G. Dumas, Bull. Soc. Chim. Fr., 1955, 651. 30 A. Abe and J. E. Mark, J. Am. Chem. Soc., 1976,98,6468. Paper 512232; Received 18th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203381

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82,3391-3399 Thermal Desorption and Infrared Studies of Butylamine adsorbed on SiO,, A1,0, and CaO Rolf Sokoll,* Hartmut Hob& and Irmtraut Schmuck Sektion Chemk, Friedrich-Schiller-Uniuersitat, 6900 Jena, DDR The adsorption of n-butylamine on SiO,, y-Al,O, and CaO at the solid/ vapour interface has been studied by infrared spectroscopy and temperature- programmed desorption (t.p.d.). The infrared spectroscopic results show that adsorption of n-butylamine on SiO, leads to the formation of hydrogen bonds between surface hydroxy groups and amine molecules. Adsorption on y-Al,O, and CaO involves coordinative interactions between amine molecules and Lewis-acidic surface sites ( A P and Ca2+ cations), but hydrogen bonding also occurs. In the t.p.d.spectra one (SiO,) or two (Al,O,, CaO) strong maxima appear caused by the desorption of n-butylamine and butyronitrile, respectively. Desorption of unchanged butylamine at low temperatures can be attributed to amine molecules hydrogen-bonded to surface hydroxy groups of the oxides. However, there are some phenomena which indicate the possibility that molecules coordinatively bonded to weak Lewis-acidic surface centres contribute to this maximum. On the other hand, only butylamine molecules coordinatively bonded to strong Lewis-acidic centres are dehydrogenated and desorb as butyronitrile at higher temperatures. In our previous work concerning the adsorption of n-octadecylamine on SiO,,l Y-A~,O~,~ Mg03 and a-Fe,OS4 at solid/liquid interfaces it was assumed that adsorbed amines undergo dehydrogenation on the oxide surfaces (with the exception of SO,) at elevated temperatures, leading to the desorption of the corresponding nitriles.To obtain further information about this reaction, which can be important in the practical use of NH-containing substances (e.g. as corrosion inhibitors) at elevated temperatures, we investigated the adsorption and desorption behaviour of n-butylamine by infrared spectroscopy and temperature-programmed desorption (t.p.d.) using oxides with different surface properties (different types of acidic surfaces centres) as adsorbents. Following the results obtained by the investigation of butylamine - a-Fe,O, adsorbate^,^ the interactions of butylamine with highly dispersed SiO,, y-Al,03 and CaO are now described.Experiment a1 Materials Samples of CaO were obtained by vacuum decomposition (973 K, 1 x Pa) of pressed discs of CaCO, in an infrared cell. The B.E.T. surface area of the resulting CaO sample was ca. 50 m2 g-l. X-ray powder diffraction analysis confirmed that the samples were polycrystalline CaO. The other oxides were Aerosil300 (Degussa) and Aluminium- oxid C (Degussa/B.E.T. surface area ca. 80 m2g-l). n-Butylamine was purified by distillation and degassed immediately before use in the adsorption experiments. Apparatus and Procedures Infrared spectra were recorded by a Specord 75 IR spectrometer (VEB Carl Zeiss, Jena) coupled with a KRS 4200 computer (VEB Robotron) in the range 4000-1200 cm-l. For 33913392 Adsorption of n-Butylamine on SiO,, Al,O, and CaO Table 1.Wavenumbers (cm-l) of NH-bands of butylamine after adsorption on various oxides at beam temperature, and of gaseous butylamine SiO, 3360 3296 3155 1586 Y-Al,O, 3312 3222 3120 1578 CaO 3338 3266 3157 1589 3330 3254 3156 1588 3315 3226 3126 1580 M a a a-Fe,O,b buty lamine gas-phase 3403 3340 3180 161 5 a Unpublished results, From ref. (4). monitoring the desorption spectra and for identification of the desorption products a CH8 mass spectrometer (Varian Mat) connected to the t.p.d. apparature was used. To obtain consistent surface states the oxide samples were initially activated under vacuum (p = Pa) at 1023 K for 1 h. All other experimental conditions and the procedures of measuring the infrared and t.p.d. spectra were the same as in ref.(4). Results Infrared Absorption Spectra Table 1 summarizes the wavenumbers of the NH-bands of butylamine after adsorption on the various oxides and subsequent evacuation at beam temperature. Therefore the only phenomena now described are those which were in the infrared spectra after stepwise heating of the butylamine/oxide adsorbates. SiO, Fig. 1. shows the infrared spectra of freshly activated SiO, and of SiO, evacuated at various temperatures after the adsorption of butylamine at beam temperature. The intensities of all the infrared bands of adsorbed butylamine decrease continuously with increasing temperature. On reaching the final heating at 573 K the major part of the adsorbed species has desorbed. In the same way, the broad absorption band of surface hydroxy groups interacting with adsorbed butylamine (between 3200 and 2500 cm-l) disappears, and the sharp band of isolated hydroxy groups nearly falls to its initial intensity after heating at 573 K.The phenomena observed by infrared spectroscopy during thermal treatment of butylamine/Al,O, adsorbates are illustrated in fig. 2. Three characteristic temperature ranges can be distinguished. (a) Between beam temperature and ca. 473 K the intensities of all the bands of adsorbed butylamine decrease. No change in the band positions is observed, with the exception of that of the NH,-bending vibration, which is slightly shifted to lower wavenumbers (1 578 + 1574 cm-l). (b) At ca. 523 K a new infrared band appears at 1643 cm-1 and the band of the NH,-bending vibration is again shifted to lower wavenumbers [1569 cm-l; see fig.2(B), curve (b)]. Heating the adsorbate at still3393 R. Sokoll, H. Hobert and I. Schmuck I 1 I I I 1 1 1 I , I 1 I l q 3800 3500 3 000 2500 1800 1500 1200 wavenumberlcm-' Fig. 1. (A) Infrared spectra of SiO,: (a) in vacuum; (b)-(d) after adsorption of butylamine from the vapour phase and subsequent evacuation at (b) beam temperature, (c) 373 K and (d) 573 K. (B) Difference spectra: (1) @)-(a); (2) (c)-(a); (3) (d)-(a). I- I w wavenum ber/cm-' Fig. 2. (A) Infrared spectra of Al,O, : (a) in vacuum; (b) - (d) after adsorption of butylamine from the vapour phase and subsequent evacuation at (b) beam temperature, (c) 553 K and (d) 773 K. (B) Difference spectra: (1) (b) - (a); (2) (c) - (a); (3) ( d ) - (a).higher temperatures leads to a loss of the intensity of these bands. (c) Finally, at temperatures > 723 K the infrared spectra show only two small bands at ca. 1549 and 1473 cm-l [fig. 2(B) curve (c)]. CaO Fig. 3 shows the infrared spectra of freshly activated CaO and of CaO evacuated at various temperatures after adsorption of butylamine at beam temperature. With increasing temperature all the bands of adsorbed butylamine are reduced in intensity3394 r n .x 1 d 3 E .s W * P) c E Q Adsorption of n-Butylarnine on SiO,, Al,O, and CaO 3800 3500 3000 0 t teoo 1500 1 w avenumber/crn-' 1 00 Fig. 3. (A) Infrared spectra of CaO: (a) in vacuum; (b)-(d) after adsorption of butylamine from the vapour phase and subsequent evacuation at (b) beam temperature, (c) 473 K and (d) 623 K.(B) Difference spectra: (1) @)-(a); (2) (c)-(a); (3) (d)-(a). T/K Fig. 4. Thermal desorption spectrum of butylamine adsorbed on SiO, at room temperature. with no change in frequency. Only the band of the NH,-bending vibration shows a small shift to lower wavenumbers at elevated temperatures (1 589 + 1583 cm-l). Thermal Desorption Spectra All desorption experiments were carried out with oxide samples saturated by butylamine vapour at room temperature. When the t.p.d. spectra exhibited more than one strong maximum, additional experiments were made with smaller initial coverage of the oxide surfaces. The results can be summarized as follows.R. Sokoll, H. Hobert and I. Schmuck 3395 T/K Fig. 5. Thermal desorption spectra of butylamine adsorbed on y - A1,0, at room temperature.(A) : (a) Complete saturation of the surface; (b) partially covered surface. (B): (a) total amount; (b) and (c) desorption products represented by the most intense peak in their mass spectrum: (b) butylamine, (c) butyronitrile. (Q 1 400 500 600 700 T/K Fig. 6. Thermal desorption spectra of butylamine adsorbed on CaO at room temperature. (A): (a) Complete saturation of the surface; (b) partially covered surface.( B): (a) Total amount; (b) and (c) desorption products represented by the most intense peak in their mass spectrum: (b) butylamine, (c) butyronitrile. Fig. 4 shows that in the case of butylamine/SiO, adsorbates only one strong desorption peak appears at ca. 423 K (S,). Two additional maxima at ca. 673 K (SII) and 843 K (SIII) are very small.Mass-spectrometric studies showed that butylamine is the main desorption product (S, and SII), whereas maximum SIII is formed by the desorption of ethylene and propylene. The t.p.d. spectra of butylamine/Al,O, adsorbates consist of two strong peaks at ca. 423 K (A,) and 623 K (AII) [fig. 5(A)]. With decreasing initial coverage of the oxide3396 Adsorption of n-Butylamine on SO,, Al,O, and CaO surface maximum A, is reduced in intensity, whereas maximum A,, is little affected. This result clearly shows the existence of different adsorption centres on the surface of A1,0,. Fig. 5 (B) illustrates the product distribution : butylamine causes the low-temperature maximum A,, whereas maximum A,, is formed by the desorption of butyronitrile. When the initial surface coverage is drastically diminished a third desorption peak can be detected at ca.693 K (Arrr) which usually is overlapped by the strong maximum A,,. Small amounts of propylene and hydrogen cyanide desorb in this temperature range. The t.p.d. spectra of butylamine/CaO adsorbates show two maxima at ca. 423 (C,) and 523 K (C,,) [fig. 6(A)], which are caused by the desorption of butylamine (C,) and butyronitrile (C,,), respectively [fig. 6 (B)]. Discussion Many workers have investigated the adsorption of various organic molecules on oxide surfaces by infrared spectroscopy. From these results it can be concluded that mainly coordinatively unsaturated metal cations and surface hydroxy groups act as adsorption centres. Our previous results for the adsorption of n-octadecylamine on Mg03 suggest that it is unlikely that surface oxygen ions are involved in the interactions between n-butylamine and CaO. Therefore, our results will now be discussed with regard to the following two forms of adsorbate complexes: type A, M-OH.* .NH,R, and type B, M”+ * -. NH,R. All infrared NH bands of butylamine adsorbed on the surfaces of the three oxides investigated are shifted to lower wavenumbers in comparison with those of the pure gaseous amine (table 1). This means that butylamine always acts as an electron-pair donor bonded to the oxide surfaces via its nitrogen atom.lV5 However, these frequency shifts (Av) are different in the various cases, thus referring to the existence of different types of amine-adsorbent interactions. SiO, Butylamine adsorbed on SiO, shows the smallest frequency shifts of the NH-stretching vibration bands (Av w 20 cm-l).From this, together with the observed simultaneity of butylamine desorption and restoration of the unperturbed silanol band, it can be concluded that the formation of hydrogen bonds between butylamine molecules and surface hydroxy groups of SiO, is the main interaction (type A). This interpretation is confirmed by the t.p.d. spectrum of butylamine/SiO, adsorbates. One strong maximum indicates that butylamine desorbs from only one type of weakly acidic surface centre. This result agrees with other infrared spectroscopic studies of the adsorption of aliphatic amines on SiO, at the solid/vapour [e.g. ref. (6) and (7)] and solid/liquid [e.g.ref. (1) and (8)] interfaces. A1203 Adsorption of butylamine on Al,O, causes greater frequency shifts of NH-stretching and NH,-bending vibration bands than in the case of SiO,. This means that butylamine interacts with active sites on the A1,0, surface which are more acidic than the surface hydroxy groups of SO,. Therefore the feature determining interaction between butyla- mine and Al,O, is that of the nitrogen atom with coordinatively unsaturated A13+ cations on the solid surface (type B), which is in agreement with other literature data [e.g. ref. (2) and (9)) However, there also occurs the formation of hydrogen bonds between adsorbed butylamine molecules and surface hydroxy groups of Al,O,, which is indicated by the loss of intensity of the unperturbed OH-bands (3785, 3722 and 3690 cm-l) after adsorption at beam temperature.These two adsorption states are also indicated by theR. Sokoll, H. Hobert and I. Schmuck 3397 two strong maxima in the t.p.d. spectrum of butylamine/Al,O, adsorbates. Moreover, maximum A, shows a tailing-off at the high-temperature side. This phenomenon can be connected with the observed frequency change of the NH,-bending vibration between beam temperature and 523 K, and is possibly due to slightly different adsorption states from which desorption of unchanged butylamine occurs (butylamine hydrogen-bonded to hydroxy groups and coordinated to weak Lewis-acidic centres). Butylamine molecules adsorbed on strong Lewis-acidic centres (desorption maximum AII) are dehydrogenated according to the following scheme : CH,CH,CH,CH,NH, + CH,CH,CH,CH=NH + CH,CH,CH,CN.(1) -Hz -H2 This proposal is strengthened by the appearance of an infrared band at 1643 cm-1 which can be interpreted as a (C=N)-stretching vibration of the imine species.11- l2 On the other hand, adsorbed butyronitrile was not detected by infrared spectroscopy, so that nitrile must be formed from the imine species immediately at the moment of desorption. The infrared bands remaining after heating the butylamine/Al,O, adsorbates at temperatures higher than 723 K (1549 and 1473 cm-l) can be discussed in different ways. However, the most probable explanation seems to be that these bands are caused by carboxylate species formed from strongly adsorbed amine molecules. This assumption is supported by our results for the adsorption of acetic acid on Al,O,.We found two bands at 1555 and 1472cm-l which can be assigned to the stretching vibrations of the (CO;) group.ll* l3 Moreover, from other investigations the oxidation of organic compounds on the Al,O, surface is well known [e.g. ref. (14)]. A comparison of our results with other literature data shows that the proposed existence of different types of Lewis-coordinated amine species was also detected by the adsorption of pyridine on v-A1203.15 However, the product distributions which we found in our t.p.d. experiments disagree with those reported by Koubek et all6 These workers investigated the adsorption of amines on Al,O, by t.p.d. under a flow of helium with a chromatographic column. They found that the high-temperature maxima in their desorption chromatograms were caused by the products of deammination (olefins and ammonia) and disproportionation (di- and tri- alkylamines and ammonia) reactions of amines.The formation of nitriles was not mentioned. Therefore we also analysed the main desorption products by infrared spectroscopy in the following way. After adsorption of butylamine on Al,O, and evacuation at 553 K (most of the unchanged amine desorbed), the adsorbate was heated to 753 K in the closed infrared cell and then cooled to room temperature. By this process the products which form the high-temperature maximum in the t.p.d. spectra were removed from the Al,O, surface and then readsorbed. Fig. 7 shows the result. Two bands are visible at 2290 and 2241 cm-l which can be ascribed to coordinatively and hydrogen-bonded nitrile, re~pective1y.l~ No bands of unsaturated products appear. Additional experiments showed that after adsorption of di- and tri-alkylamines on Al,O, olefins desorb, in addition to nitriles, at the high-temperature maximum.Both products (identified by mass spectrometry and infrared spectroscopy) are formed by a reaction which is analogous to scheme (1): e.g. /CH,CH,R’ RCH,N --+ RCH=NCH,CH,R’+ RCN + CH,=CHR’. (2) ‘H -H2 -Hz In no case could measurable amounts of disproportionation products be detected. We are thus unable to explain this disagreement; but we assume that the differences mentioned are due to different experimental conditions. This is also indicated by the results of Brey and Cobbledick,18 who investigated catalytic reactions of butylamines over alumina in a flow reactor.As main reaction products they found varying amounts3398 Adsorption of n-Butylamine on SO,, Al,03 and CaO 3200 3 000 2400 w avenumberl cm - ' 200 Fig. 7. Infrared spectra of Al,O,: (a) in vacuum; (b) after adsorption of butylamine from the vapour phase and subsequent evacuation at 553 K; (c) after (b) and subsequent heating at 753 K in a closed infrared cell. of olefins and nitriles which depended on, for example, the pretreatment conditions of the catalyst. CaO No literature data are available for the infrared spectroscopic investigation of amine adsorption on the surface of CaO. However, it was reported that the features determining the interactions of pyrrolelO and n-~ctadecylamine~ with MgO are those of the nitrogen atom with Mg2+ cations on the solid surface.From table 1 it can be seen that the NH-stretching vibration bands of butylamine adsorbed on CaO are shifted to lower wavenumbers to a greater extent than after adsorption on SO,. On the other hand, this shift is smaller than in the case of butylamine/MgO adsorbates. Therefore we assume that, analogous to the above mentioned literature results, butylamine molecules interact with Ca2+ cations on the oxide surface (type B). However, hydrogen bonds are also formed between adsorbed butylamine and surface hydroxy groups, which is indicated by the decreasing band intensity of isolated Ca-OH species (3695 cm-l) and by the appearance of a broad absorption maximum at ca.3570 cm-l after adsorption at beam temperature. As in the case of Al,03, two maxima in the t.p.d. spectrum of butylamine /CaO adsorbates confirm the assumed existence of these different adsorption states. Again, it is possible that amine molecules contribute to the maximum CI by interacting with weak Lewis-acidic surface centres. The desorption temperature of butyronitrile is lower than in the case of butylarnine/Al,O, adsorbates because of the higher Lewis acidity of A13+ cations in comparison with those of Ca2+ cations on the surface of the corresponding oxides. References 1 R. Sokoll and H. Hobert, 2. Phys. Chem. (Leipzig), 1986,267, 241. 2 R. Sokoll and H. Hobert, 2. Phys. Chem. (Leipzig), 1986,267,321. 3 R. Sokoll and H. Hobert, J. Chem. Soc., Faraday Trans. I , 1986,82, 1527. 4 U. Marx, R. Sokoll and H. Hobert, J. Chem. SOC., Faraday Trans. I , 1986,82, 2505.R. Sokoll, H. Hobert and I. Schmuck 3399 5 E. L. Zhukova and I. I. Shmanko, Opt. Spektrosk., 1972,32, 514. 6 F. H. Van Cauwelaert, F. Vermortele and J. B. Uytterhoeven, Discuss. Furuduy Soc., 1971, 52, 66. 7 K. Hirota, K. Fueki and T. Sakai, Bull. Chem. SOC. Jpn, 1962, 35, 1545. 8 C. H. Rochester and G. H. Yong, J. Chem. Soc., Furuduy Truns. I , 1980,76, 11 58. 9 T. Morimoto, J. Imai and M. Nagao, J. Phys. Chem., 1974,78, 704. 10 J. A. Lercher, Ch. Colombier and H. Noller, 2. Phys. Chem. N.F., 1982, 131, 11 1. 11 L. J. Bellamy, The Infrured Spectra of Complex Molecules (Methuen, London, 1958). 12 H. Gunzler and H. Biick, IR-Spektroskopie, Eine Einfuhrung (Verlag Chemie, Weinheim, 1975). 13 L. H. Little, Infrured Spectra of Ahorbed Species (Academic Press, New York, 1966). 14 M. Deflin, I. M., Eltantawy and M. Baverez, J. Cutul., 1978, 54, 345. 15 C. Morterra, A. Chiorino, G. Ghiotti and E. Garrone, J. Chem. Soc., Furuduy Trans. I , 1979,75,271. 16 J. Koubek, J. Volf and J. Pasek, J. Cutul., 1975, 38, 385. 17 H. Knoezinger and H. Krietenbrink, J. Chem. Soc., Furuduy Truns. I , 1975, 71, 2421. 18 W. S. Brey and D. S. Cobbledick, Ind. Eng. Chem., 1959, 51, 1031. Paper 512244; Received 20th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203391

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1986,82, 3401-3406 Normal Coordinate Analysis of Molecules adsorbed on Zeolite Surfaces Part I. - Cyclopropane adsorbed on Sodium Faujasites and Mordenites Olga Zakharieva-Pencheva* Physical Faculty, University of Sofia, A . Ivanov Boulevard 5, 1126 Sofia, Bulgaria Horst Forster and Jurgen Seebode Institute of Physical Chemistry, University of Hamburg, Bundestrasse 45, 0-2000 Hamburg 13, Federal Republic of Germany Vibrational analysis of adsorbed polyatoms provides valuable information on the degree of participation of different bonds in the interaction with the adsorbent. As an example, normal coordinate analysis has been applied to cyclopropane in the free and the adsorbed state in sodium faujasites and mordenites of different Si/Al ratios using a valence force field.The alterations of the force constants can be interpreted in terms of changes of the molecular geometry and electronic structure as well as by the hindrance of the bending motion due to the interaction with the internal surface. The lack of influence of the lattice module supports the location of the cyclo- propane molecule near the cations. Cyclopropane has been utilized in catalytic research as a common test molecule for the study of hydrocarbon conversion, e.g. on zeolites.l-1° Despite the disagreement in details of the geometry of the sorption complex8J1-13 there is an overall conformity that cyclopropane is adsorbed near the cations, which could also be supported by electron ~pectroscopy.~~~ l5 Since 1936 there has been growing interest in the vibrational spectroscopy of cycl~propane.l~-~~ Also the force field of cyclopropane is of particular interest.lg9 2 2 v 25-31 Most of the force constants are derived in symmetry c~ordinates,~~~ 2 2 ~ 25-277 32 which cannot easily be transformed into force constants based on other coordinates.The valence force field is physically well founded because internal coordinates have a simple physical meaning and the force-constant matrix using these coordinates has its simplest form. Recently, Spiekermann et aL31 developed the force field for cyclopropane, using a valence force model for the interatomic interactions. The bending coordinates for the displacement of the internal ring angles were ignored. The neglect of the interactions involved in these coordinates seems to be questionable.We carried out a normal coordinate analysis using a harmonic valence force field for both the free cyclopropane molecule and cyclopropane adsorbed on the internal surface of sodium faujasite and mordenite. Procedure A computer program for the complete and fragmentary solution of the vibrational problem was applied, which is described in detail in ref. (34). A total of 27 internal coordinates was introduced (see fig. 1): 3 CC bond stretches, 6 CH bond stretches, 3 CCC bendings, 3 HCH bendings and 12 CCH bendings. Data for the geometry of cyclopropane are given in the 33? 35-41 For calculations we used 15 1.4 pm for 34013402 Cyclopropane Adsorbed on Zeolites Fig. 1. Internal vibrational coordinates of the cyclopropane molecule.the CC and 108.2 pm for the CH bond lengths, taken from Jones and St~icheff~~ and a CCH angle of 117.91”, obtained by an ab initio procedure with a 4-21G basis The HCH angle chosen represents a state of carbon ranging between sp2 and sp3 hybridization. The force-constant matrix elements were generated by the inverse problem calculation. For the gas-phase frequencies, data from Duncan and Burns were used.26 A modified least-squares fit was applied.42 In order to take into account the non-linearity of the force constant calculation, constraints were imposed on the values of the parameters changed at each iteration step. Both the limits of variation of each parameter and the maximum step for every iteration were established. Different combinations of the force constants under consideration and the corresponding frequencies were used to reveal the stability of the solution.In spite of all these efforts, the inverse problem calculation was very difficult and the convergence was poor. This forced us to include also the deuterated molecule into parts of the calculation. In each case three samples of sodium mordenite with Si0,/A1203 ratios of 1 1.7, 15.9 and 23.0 and sodium faujasite with Si02/A1203 ratios of 3.8, 4.6 and 7.0 were made available to us by the Institute of Organic Chemistry of the Bulgarian Academy of Sciences, Sofia. The spectra were recorded on a Digilab FTS 14 E Fourier transform spectrometer. Details of sample characterization, pretreatment, gas dosing as well as the experimental set-up have been described el~ewhere.~~ Results and Discussion The force constants of the free cyclopropane molecule obtained in this way are listed in table 1.The somewhat higher value of force constant F4, 4(CH) compared to alkanes as well as to non-strained cycloalkanes is due to the higher s character of this bond.34 The vibrational frequencies calculated with this force field are given in table 2. Comparison with the experimental values shows a good agreement. The analysis of the vibrational modes revealed some of the fundamentals being of a mixed type. The v,, (E’) vibration, for example, involves CCC bending, CC stretching as well as HCC bending coordinates. The same coordinates, but with different values of the corresponding participation coefficients, are included in the vlo(E’) fundamental.This explains the different assignments of these vibrations in the literature, vl0 to a ring deformation and v,, to a CH, wagging 26* 27 or vice versa.32 The v3 vibration (A;) is essentially a ring-breathing mode in which, however, CCH and HCH angle deformation coordinates0. Zakharieva-Pencheva et al. 3403 Table 1. Calculated force constants F (in units lo2 N m-l) of the free and sorbed cyclopropane molecule F free sorbed F free sorbed 4,1 4,2 4 , 8 4 , 1 7 4 , 4 4 , 5 4 , 6 4 , 1 s 4,16 &,14 4,15 4.2756 0.0641 -0.1282 - 0.2643 - 0.400 5.5923 0.0256 0.0192 0.2895 + 0.3916 - 0.0797 4.2564 0.0641 -0.1282 - 0.2573 - 0.400 5.5916 0.0192 0.0192 0.2895 0.3916 - 0.0797 40,10 0.4580 43,13 0.7862 &16 0.1496 & 2 1 0.0145 e3, 22 - 0.3862 e3, 23 - 0.0649 43, 26 - 0.0382 43,14 0.0100 F;3,27 0.0519 4 4 , 1 4 o'4481 0.4580 0.7855 0.0092 0.1488 0.0137 -0.3855 - 0.064 1 -0.0374 0.0443 0.4389 Table 2.Observed and calculated frequencies/cm-l of cyclopropane gas and cyclopropane adsorbed in sodium mordenites (Si02/A1203 ratios 1: 11.7; 2: 15.9; 3:23.0) and faujasites (SiO,/Al,O, ratios 1 : 3.8; 2: 4.6; 3 : 7.0) cyclopropane cyclopropane on cyclopropane on frequency 1 2 3 1 2 3 state symmetry number obsd calc. obsd obsd obsd obsd obsd obsd calc. free mordenite NaY c yclopr opane molecule in sorbed E" E' E' A ; + E V l v2 vs v4 v5 V6 v7 V8 Vll v10 v11 v12 '13 '14 '9 + 2v14 '2+'ll 3038 3044 3040 3040 1479 1481 1463 1464 1188 1189 - - 1126 1125 - - 1070 1068 - - 3102 3108 3102 3100 854 854 844 844 3025 3024 3025 3025 1438 1437 1433 1433 868 869 864 864 3082 3082 3085 3082 1188 1186 - - 739 742 - - 2921 - 2935 2939 2875 - 2882 2882 2862 2860 2083 - 2106 2104 1881 - 1890 1892 1029 1032 - - - - 2631 - - - 3039 1460 1462 1462 1460 - 3101 840 3023 1433 863 3080 - - 2939 2882 2862 2102 1887 - - - 3101 3101 3101 847 846 844 3021 3021 3025 1434 1434 1435 863 863 866 3092 3090 3090 T - - 2104 2104 2102 1890 1890 1897 2615 2615 2617 3042 1463 1185 1129 1070 3109 843 3023 1432 1033 865 3084 1186 750 also participate.The analysis of the v, (A;) eigenvectors shows participation coefficients of the internal CCH and HCH bending coordinates. On the other hand, the v, (A;), v, (A$), v, (E') and v,, (E") fundamentals represent pure CH stretching vibrations. The calculated values were highest for the displacement of the H atoms.Also, the v4(A3, v5(Ai), v,(A:) and vI4(E") frequencies contain exclusively CCH bending coordinates. The experimental frequencies of cyclopropane adsorbed on faujasites and mordenites3404 Cyclopropane Adsorbed on Zeolites with different SiO,/Al,O, ratios are also assigned in table 2. A relatively weak interaction, i.e. weak chemisorption with the zeolite surface, is observed as demonstrated by the small shifts of some of the frequencies. The ring-breathing vibration is expected to give the largest change, but unfortunately this band cannot be seen owing to the strong absorption of the zeolite skeleton in this region. Only four of the bands in the accessible range are downscale shifts observed: v, from 1479 to 1462 cm-l, v7 from 854 to 844 cm-l, v,(E’) from 1438 to 1433 cm-l and vll from 868 to 863 cm-l.Only the v,, fundamental of cyclopropane adsorbed in faujasites shows a displacement of ca. 10 cm-l to higher frequencies. As can be seen from table 2, the combination bands were also shifted an appreciable amount by this interaction which, however, should not be discussed here, as they were not included in the vibrational analysis. Normal coordinate analysis was also extended to the adsorbed cyclopropane molecule. The force field was obtained by the inverse problem calculation, fitting the calculated to the experimental frequencies. The zeolite surface was not directly taken into account. The force field of the perturbed molecule is compared with that of the free molecule in table 1.The main change involves the Fl, ,(CC) force constant. This is caused by the withdrawal of IT electron density towards the sodium cation, in front of which the cyclopropane molecule is located, giving rise to a weakening of the CC bond. On the other hand, the F14, 14(HCH) force constant diminishes, which can be explained by an increase of the HCH angle forced by adsorption. The enlargement of this angle shifts the C atoms closer to the sp2 hybrid state, resulting in a strengthening of the exocyclic CH bonds and a weakening of the endocyclic CC bonds. Both effects are responsible for the decrease of the CC bond strength. As an experimental result the ring deformation mode vll, in which the CC stretching coordinates are also contained, shifts to lower frequencies.The vll fundamental represents mainly the ring deformation, but CCH bending and HCH angle deformation coordinates also participate with smaller co- efficients. The Jacobian matrix elements for this frequency are not very large. This is reflected in the constancy of F,,,, lo(CCC) and the slight changes of F13, ,,(CCH,CCH) and the angle interaction force constants. The stronger effects on the ring-breathing vibrations cannot be checked by experiment owing to the complete obscuration of this region by the zeolite absorption as mentioned above. The two effects cited above counteract on the CH bonds: while these bonds are weakened by the withdrawal of electron density, their bond strength increases by enlargement of the HCH angle. As a result, for both the free and the adsorbed cyclopropane, the F4, ,(CH,CH) force constants are nearly the same and also the positions of the symmetric and asymmetric CH stretching bands (vl, v6, v,) remain unaltered; only v,, shows a 10 cm-I upscale shift on faujasites as mentioned above.The decrease of the CH, deformation frequencies v, and v9 is essentially due also to the increase of the HCH angle during adsorption. The largest change is observed with the v, mode, which shifts downscale by ca. 18 cm-l. The calculated fundamentals show a considerable participation of the HCH and CCH bending coordinates. Also all CCH and HCH angle deformation coordinates contribute to this vibration. The Jacobian matrix elements indicate that this frequency is sensitive to q3, ,,(CCH,CCH), q4, ,,(HCH,HCH) as well as the interaction constants C3, 14, 4,, 15, q3, ,,, q3, ,,, e3, 23, In case of the v, vibrational mode, not all of the HCH angle deformation coordinates contribute, which explains the small shift of ca.6 cm-l. The Jacobian matrix has very high values for the CCH interaction force constants and the diagonal force elements e4, 14 and small ones for the aA/aFl, 13, aA/aF4, 13 and aA/aF4, 15 elements (A = 4z2c2v2). Concerning the v, fundamental, in which the HCH and CCH angle deformation coordinates participate, a frequency shift of ca. 10 cm-l is observed. From the Jacobian matrix a high sensitivity of this frequency from F13,13 and all the CCH-CCH interaction constants can be derived, while the aA/W elements, where F stands for P 14, 14 and F13,14, &3,26 and 4 3 .2 7 .0. Zakharieva-Pencheva e t al. 3405 respectively, are zero. The calculated atomic displacements for this vibration show that chiefly the H atoms are moving. The frequencies of all four modes (v2, v,, v,, vll) depend strongly on e3,13 and the different CCH-CCH interaction force constants. The slight decrease of &, 13 in comparison with that of the free molecule may be explained by the presence of several effects. For the decrease of this force constant the reduced CC electron density may be responsible, as mentioned before, but may largely be compensated by enlargement of the HCH angle and by the hindrance of deformation motion due to the interaction with the surface. Also the four frequencies depend very heavily on the off-diagonal force elements, which express the interaction of the CCH-CCH angle deformation.As can be seen from table 1, all these interaction force constants slightly decrease in the adsorbed state. As cyclopropane interacts only weakly with the adsorbent, all frequency changes are small. Comparing the experimental frequencies obtained for cyclopropane in faujasites and mordenites with different Si02/A120, ratios, no definite influence of this parameter can be detected, which fits the finding that cyclopropane is adsorbed in front of the cations, representing the centres of sorptional interaction in the zeolite pore system. Conclusion Normal coordinate analysis turns out to be a valuable tool also for the study of adsorbed molecules and gives more insight into details of the changes of molecular properties due to the interaction with the solid surface.The results obtained with cyclopropane in faujasites and mordenites are promising, although this adsorbate seems to be less suited to this kind of investigation and only a part of the cyclopropane fundamental and combination bands can be observed and used for this analysis owing to the strong absorption of the alumosilicate skeleton in the low-frequency range. We thank Professor Gribov and Professor Dementiev for the opportunity of using their vibrational programs. The financial support of the Universities of Hamburg and Sofia, of the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie is gratefully acknowledged. References 1 Z. M. George and H. W. Habgood, J.Phys. Chem., 1970,74, 1502. 2 N. T. Tam, R. P. Cooney and G. Curthoys, J. Catal., 1976,44, 81. 3 P. Fejes, I. Hannas, I. Kiricsi and K. Varga, Acta Phys. Chem. Szeged, 1978, 24, 119. 4 I. Kiricsi, I. Hannus, K. Varga and P. Fejes, J. Catal., 1980, 63, 501. 5 S. H. Abbas, T. K. Al-Dawood, J. Dwyer, F. R. Fitch, A. Georgopoulos, F. J. Machdo and 6 H. F. Leach and C. E. Marsden, Stud. Surf. Sci. Catal., 1980,5, 141. 7 H. Forster, S. Franke and J. Seebode, J. Chem. Soc., Faraday Trans. I , 1983,79, 373. 8 H. Forster and J. Seebode, Zeolites, 1983, 3, 63. 9 P. Fejes, I. Kiricsi, H. Forster and J. Seebode, Zeolites, 1984, 4, 259. 10 P. Fejes, H. Forster, I. Kiricsi and J. Seebode, Stud. Surf. Sci. Catal., 1984, 18, 91. 11 N. T. Tam, P. Tsai and R. P. Cooney, Aust.J. Chem., 1978,31, 255. 12 J. Howard and I. J. Braid, Zeolites, 1985, 5, 101. 13 V. W. Cruz, P. C. W. hung and K. Seff, J . Am. Chem. Soc., 1978,100,6997. 14 K. Klier, Adv. Chem. Ser., 1971, 101, 480. 15 K. Klier and M. Ralek, J. Phys. Chem. Solids, 1968, 29, 951. 16 G. Herzberg, Molecular Spectra and Molecular Structure, Vol. II. Infrared and Raman Spectra of 17 A. W. Baker and R. C. Lord, J. Chem. Phys., 1955,23, 1636. 18 P. M. Mathai, G. G. Shepherd and H. L. Welsh, Can. J. Phys., 1956,34, 1448. 19 L. M. Sverdlov and E. P. Krainov, Opt. Spectrosc., 1957, 3, 54. 20 C. Brecher, E. Krikorian, J. Blanc and R. S. Halford, J . Chem. Phys., 1961,35, 1097. 21 D. L. Duncan and D. C. McKean, J . Mol. Spectrosc., 1968, 27, 1 17. 22 Hs. H. Gunthard, R. C. Lord and T. K.McCubbin Jr, J. Chem. Phys., 1956,25, 768. S. M. Smyth, Stud. Surf. Sci. Catal., 1980, 5, 127. Polyatomic Molecules (Van Nostrand Reinhold, New York, 1945).3406 Cyclopropane Adsorbed on Zeolites 23 J. B. Bates, J. Chem. Phys., 1973,58,4236. 24 S . J. Daunt and H. F. Shurvell, J. Raman Spectrosc., 1974, 2,463. 25 S. J. Cyvin, Spectrochim. Acta, 1960, 16, 1022. 26 J. L. Duncan and G. R. Burns, J. Mol. Spectrosc., 1969,30,253. 27 I. W. Levin and R. A. R. Pearce, J. Chem. Phys., 1978,69,2196. 28 T . Hirokawa, M. Hayashi and H. Murata, J. Sci. Hiroshima Univ., Ser. A. 1973,37, 282. 29 B. Galabov and H. Moms, J. Mol. Struct., 1973, 17, 421. 30 J. R. Robins, S. J. Daunt and H. L. Shurvell, J. Raman Spectrosc., 1976, 5, 41 1. 31 M. Spiekermann, D. Bougeard and B. Schrader, J. Mol. Struct., 1980, 60, 55. 32 A. Komornicki and F. Pauzat, J. Phys. Chem., 1983,87, 3847. 33 W. J. Jones and B. P. Stoicheff, Can. J. Phys., 1964,42, 2259. 34 L. A. Gribov and V. A. Dementiev, Methods and Algorithms of Calculations in the Theory of the Vibrational Spectra of Molecules (Science, Moscow, 198 1). 35 C. van Alsenoy, J. N. Scarsdale and L. Schafer, J. Chem. Phys., 1981,74,6278. 36 K. B. Wiberg and J. J. Wendoloski, J. Am. Chem. SOC., 1982,104, 5679. 37 L. M. Sverdlov, M. A. Kovner and E. P. Krainov, Vibrational Spectra of Polyatomic Molecules (Science, 38 R. J. Elliott and W. G. Richards, J. Mol. Struct., 1982, 87, 211. 39 M. D. Harmony, V. W. Laurie, R. L. Kuczkowski, R. H. Schwendeman, D. A. Ramsey, F. J. Lovas, 40 0. Bastiansen, F. N. Fritsch and K. Hedberg, Acta Crystallogr., 1964, 17, 538. 41 J. L. Duncan, J. Mol. Spectrosc., 1968, 25, 451. 42 V. A. Dementiev and L. A. Gribov, Zh. Prikl. Spektrosk., 1971, 14, 889. 43 H. Forster and R. Seelemann, J. Chem. SOC., Faraday Trans. I , 1978, 74, 1435. Moscow, 1970). W. J. Lafferty and A. G. Maki, J. Phys. Chem. Ref. Data, 1979,8, 619. Paper 512280; Received 30th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203401

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82, 3407-3413 An llB Nuclear Magnetic Resonance Study of the Reaction of the Tetrahydroxyborate Ion with Polyhydroxy Compounds J. Graham Dawber* and Stuart I. E. Green Department of Chemistry and Biology, North Staflordshire Polytechnic, Stoke-on-Trent ST4 2DE An llB n.m.r. study of the formation of complexes between the tetrahy- droxyborate ion, B(OH);, and various polyols has been carried out using ethane diol, propane-l,2-diol, propane- 1,3-dioi, glycerol, mannitol and sorbitol. The results show quite clearly that the polyols can form 1 : 1 and 1 : 2 boratepolyol complexes, and that the complexation is not limited to hydroxy groups on adjacent carbon atoms but can also involve those on alternate carbon atoms, although in the latter case the reaction is less favourable.Attempts have been made to calculate the equilibrium constants for the various equilibria in equimolar mixtures of polyol and borate ion. The reaction of polyhydroxy compounds (polyols) with boric acid has been known for many years, particularly as a means of increasing the acid strength of boric acid for its volumetric ana1ysis.l The relationship between the structural characteristics of the polyol and its influence upon the acidity of boric acid was reviewed by Boiseken,2* who also suggested that the formation of a 1 : 1 monochelated borate complex was followed by the formation of a bichelated complex containing two molecules of polyol. However, it was suggested later by Nickersong that only the 1 : 1 complex was formed in the case of mannitol, although a re-examination of the data by Campbell5 supported the interpret- ation that a 1 : 2 complex also exists.Nickersoqs however, still maintained that the complex was solely of the 1 : 1 stoichiometry. It has also been suggested7-lo that it is the borate ion, rather than boric acid, which is complexed by the polyol, and this is supported by recent polarimetric measurements.ll~ l2 Because the stoichiometry of the complexes is not ~ e r t a i n , ~ ~ - l ~ discrepancies occur among values derived for the association constants, since assumptions made about the stoichiometry of the complex lead to varying methods of calculation.89 16-18 It is also commonly assumed that only adjacent hydroxy groups in the polyol are involved in the formation of the borate complexes, although in the case of the inisitols it is thought that the three hydroxy groups on alternate carbon atoms are involved in a tridentate coordination with the anion.l7* l9 It was the purpose of the present work to study the formation of complexes between various polyols and the tetrahydroxyborate ion, B(OH);, making use of llB n.m.r.measurements, in attempt to remove some of the uncertainty concerning the stoichiometry of the complexes. Although borax has commonly been used in the characterisation of carbohydrates, the tetrahydroxyborate ion offers some advantages for the study of polyol-borate complexes because of its simplicity relative to the borax anion, and also because NaB(OH), has a much higher solubility. At comparable concentrations of borate, the association constants for the reaction of borax and B(0H); with mannitol and sorbitol have been shown to be similar, while that for boric acid is considerably smaller.12 Although the llB n.m.r.signal produces a broad line in the case of boric acid (I = 3/2 nucleus), for B(0H); the line is considerably narrowed, thereby making it more 34073408 llB N.M.R. of B(OH), - Polyol Complexes amenable to study in terms of changes in chemical shift. The polyols studied included ethane diol, propane- 1,2-diol, propane- 1,3-diol, 2-methylpentane-2,4-diol, glycerol, mannitol and sorbitol. Experimental Solutions of NaB(OH), in water (ca. 0.6 mol dm-3) were made by neutralising approp- riate amounts of H3BO3 (A.R.) and NaOH (A.R.). Incremental amounts of the polyols (G.P.R. grade) were added in increasing amounts to the NaB(OH), solution, and the llB n.m.r.spectrum was measured after each addition of polyol. The llB n.m.r. proton-decoupled spectra were measured with a Jeol FX-90Q Fourier- transform spectrometer (proton resonance at 89.55 MHz, llB resonance at 28.75 MHz) using a tip-angle of 45" and a pulse repetition time of 2 s. The spectral responses over 4500 Hz were acquired into 8K and zero-filled to 16K data points, and an experimental broadening of 0.3 Hz applied prior to Fourier transformation. The llB experiments were carried out in H,O with the n.m.r. spectrometer locked-on to D,O in a capillary in the n.m.r. tube. External referencing of the llB spectra was made by taking the signal for H3BO3 in water as 6 = 0.0 ppm.On this scale the signal for boron trifluoride etherate complex came at 6 = -20.0 ppm. Results and Discussion The simplest of the polyols is ethane diol and its effects upon the B(0H); spectrum are shown in fig. 1. The peak due to free B(OH), at 6 = - 17.3 ppm is gradually replaced by a peak at S = - 13.4 ppm as the diol is added. The latter peak is likely to correspond to formation of a 1 : 1 borate-polyol complex, and at higher polyol concentration is gradually reduced in intensity and replaced by a further peak at 6 = -9.7 ppm corresponding to a 1 : 2 borate-polyol complex. The peak at 6 = - 9.7 ppm is considerably broadened and in spectral width resembles that of H,BO, and we ascribe this to relaxation phenomena rather than any effects due to symmetry alone.Thus the stoichiometry of the B(OH);-ethane diol system does not correspond to a single complex, but a two-stage equilibrium as shown below: CHZOH CH2-0 OH - I x CH2-0 OH I \ HO OH I CH2OH CHzOH A'2 iH2-0,d0-$2 / \ CH2-0 O-CH2 CHZOH I + HoJO-r2 HO / \ 0-CHI - The peaks of the various spectra in fig. 1 were integrated in order to give an estimate of the relative proportions of the various species. This can only be approximate, since the relaxation times of the llB nucleii in its various species will be different. The results of the integrations are shown in fig. 2, corresponding to the reaction scheme above. The llB spectra for the propane-1,2-diol-borate system were very similar to those for the ethane diol system, with corresponding peaks for the 1: 1 and 1 : 2 complexes at approximately the same resonance positions.The changes in relative proportions of the various species were obtained from the n.m.r. integrations and it appears that the propane- 1,2-diol-borate complex is formed more easily than the ethane did-borate complex (see equilibrium constant in table 1). In both of the above cases the hydroxy groups of the polyol involved in forming theJ. G. Dawber and S. I. Green 3409 -I””’”I -1 0 -20 - - 6 (PPm) 8 (PPm) -1 0 -20 -1 0 -20 Fig. 1. IlB n.m.r. spectra of the ethanediol (m,)-NaB(OH), (ml) system. m,/m, = (a), 0, (b) 0.6, (c) 2.6, ( d ) 9.7 and (e) 38.1. mzlm 1 Fig. 2. The ethanediol (mz)-NaB(OH), (m,) system; relative proportions of various species : 0, NaB(OH), (6 = - 17.3); x , 1 : 1 complex (6 = - 13.4); 0, 2: 1 complex (6 = -9.7). complex were on adjacent carbon atoms.The question arises whether or not the hydroxy groups on alternate carbon atoms of a polyol can participate in complex formation, e.g. with propane-1,3-diol, the results for which are shown in fig. 3. Here it can be seen that the B(0H); peak at - 17.3 ppm is gradually replaced by a peak at higher field3410 L llB N.M.R. of B(0H); - Polyol Complexes Table 1. Association constants for NaB(OH),- polyol systems (calculated at rnz/ml = 1.0) Kl KZ polyol /mol-l dm3 /rnol-l dm3 ~~ ethanediol 0.74 0.29 propane- 1,2-diol 2.2 0.55 propane-l,3-diol 1.1 0.08 glycerol 6.7 1.3 mannitol 137 21 sorbitol 278 24 L v -‘“7 -15 6 (ppm) -20 -1 5 - 20 Fig. 3. llB n.m.r. spectra of the propane-1,3-diol (rnz)-NaB(OH), (m2) system.mz/rnl = (a) 0, (b) 0.3, (c) 0.8, ( d ) 1.9 and (e) 13.0. (- 18.3 ppm) when the polyol is added, and at higher polyol concentration a further peak appears at even higher field (- 18.6 ppm). This shows unequivocally that complex- ation can occur across the alternate carbon atoms and that a 1:l borate-polyol complex is followed by formation of a 1 : 2 borate-polyol complex. Thus the complexation reaction is not solely restricted to hydroxy groups on adjacent carbon atoms. This finding was confirmed with an experiment involving B(0H); and 2-methylpentane- 2,4-diol although only low polyol concentrations were possible due to limited solubility in water. The relative proportions of the various species for the propane-1,3-diol system were obtained from the integration of the peaks.The interesting feature concerning the chemical shifts is that the llB signals for the complexes involving alternate carbon atoms of the polyols are at higher field relative to B(OH);, whereas for the complexes involving adjacent carbon atoms are at lower field. It should be possible, therefore, to distinguishJ . G. Dawber and S . I. Green 341 1 - ~ . , . . ” , . . ” , ’ ” . , . . ” , ~ -10 -20 -10 -20 6 ( P P d 6 ( P P d Fig. 4. llB n.m.r. spectra of glycerol (m2)-NaB(OH), (ml) system. rn2/m1 = (a) 0, (b) 1.3, (c) 3.3, ( d ) 7.1 and (e) 21.0. both types of complex where this might occur. Such a possibility arises in the case of glycerol, CH,OHCHOHCH,OH. The llB n.m.r. spectra of its complexes with B(0H); are shown in fig.4 where it can be seen that both types of signal occur, those at low field corresponding to complexation across adjacent carbon atoms, and those at high field corresponding to complexation across alternate carbon atoms. For both types, the evidence shows that 1: 1 and 1 : 2 complexes are formed. Again the spectra were integrated to give the relative proportions of the various species. The isomeric polyols mannitol and sorbitol when complexed with B(0H); were also studied and the relative proportions of the various borate species are shown in fig. 5 and 6. In both cases the results show the readiness of the polyols to complex and also indicate the presence of both 1 : 1 and 1 : 2 complexes, which had been previously suggested by the polarimetric studies.12 In the case of mannitol there was some suggestion of a slight amount of complexation across alternate carbon atoms but less evidence in the case of sorbitol.For both mannitol and sorbitol the rapidity of the disappearance of the llB signal due to B(0H); at low concentrations of oolyol may suggest that each polyol molecule could become involved with several bcrate anions in this region of low polyol concentration (m,/ml < 0.25). In principle it is possible to calculate the equilibrium constants for the various equilibria from the integration of the llB n.m.r. peaks. We can represent the equilibria Kl in the form: A- -k P f AP-3412 llB N.M.R. of B(OH), - Polyol Complexes m z h 1 Fig. 5. The mannitol (m,)-NaB(OH), (ml) system; relative proportions of various species: 0, NaB(OH), (6 = - 17.3); x , 1 : 1 complex (6 = - 13.4); 0, 2: 1 complex (6 = -9.5); a, reaction across alternate carbon atoms (6 = - 18.2). Fig.6. The sorbitol (m,)-NaB(OH), (ml) system; relative proportions of various species: 0, NaB(OH), (6 = - 17.3); x , 1: 1 complex (6 = - 13.6); 0, 2: 1 complex (6 = -9.5). where A- is the borate anion and P is the polyol. If m, and m, are the moles of borate and polyol, respectively, in all their forms, and a, and y are the fractions of the total llB n.m.r. signal for A-, AP- and AP,, respectively, then [A-] = am,, [AP-] = Bm, and [AP,] = ym,, and [PI = mz-m,(j7+y). It can then be shown that andJ. G. Dawber and S. I. Green 341 3 The values of Kl and K, changed by factors of ca. 1.5 over the wide concentration range of m,/ml.This will be due to several factors such as non-ideal solutiop behaviour, inaccuracies in obtaining integration values when there was overlapping of peaks, and the relationship of llB concentration, in its various forms, with n.m.r. signal changing slightly. In order to provide a reasonable comparison between the various polyols, the values of Kl and K, were calculated for equimolar mixtures of polyol and borate, i.e. m,/m1 = 1 .O (with m, at 0.6 mol dmP3). The values are given in table 1, where it can be seen that increasing the number of OH groups in the polyol markedly increases both Kl and K,, and in the cases of mannitol and sorbitol the association constants are considerable. It appears from this data that sorbitol complexes with B(0H); more readily than mannitol with the Kl values being in the ratio 2.0.Interestingly, although the actual numerical values of the association constants for these polyols were different from the polarimetric studies,12 their ratio for sorbitol/mannitol was 1.9, and from the acid-base studiesll was 1.5. References 1 M. G. Mellon and V. N. Morris, Znd. Eng. Chem., 1924, 16, 123. 2 J. Boeseken, Advan. Carbohydr. Chem., 1949, 4, 189; 1949, 12, 81; J. P. Sickels and H. P. Schultz, J. Chem. Educ., 1964,41, 343. 3 J. Boeseken and N. Vermaes, J. Phys. Chem., 1931,35, 1477. 4 R. F. Nickerson, J. Znorg. Nucl. Chem., 1968, 30, 1447. 5 G. W. Campbell, J. Znorg. Nucl. Chem., 1969, 31, 2626. 6 R. F. Nickerson, J. Znorg. Nucl. Chem., 1970, 32, 1400. 7 A. Deutsch and S. Asoling, J. Am. Chem. Soc., 1949, 71, 1637. 8 G. L. Roy, A. L. Lafferiere and J. 0. Edwards, J. Znorg. Nucl. Chem., 1957,4, 106. 9 S. D. Ross and A. J. Cattoti, J. Am. Chem. Sac., 1949, 71, 3563. 10 T. P. Onak, H. L. Landesman, R. E. Williams and I. Shapiro, J. Phys. Chem., 1959, 63, 1533. 11 J. G. Dawber and D. H. Matusin, J. Chem. Soc., Faraday Trans. 1 , 1982,78, 2521. 12 J. G. Dawber and G. E. Hardy, J. Chem. Soc., Faraday Trans. 1, "1984,80, 2467. 13 H. B. Davis and C. J. B. Mott, J. Chem. Soc., Faraday Trans. 1 , 1980,76, 1991. 14 R. Larsson and G. Nunziata, Acta Chem. Scand., 1970, 24, 2145. 15 M. Mazurek and A. S. Perlin, Can. J. Chem., 1963, 41, 2403. 16 P. Antikainen, Ann. Acad. Sci. Fenn., Sect. AZZ, 1954, 56, 3; Suom. Kemistil. B, 1958, 31, 255. 17 S. J. Angyal and D. J. McHugh, J. Chem. Soc., 1957, 1423. 18 J. J. Kankare, Anal. Chem., 1973, 45, 2050. 19 R. M. Williams and R. H. Attala, in Solution Properties of Polysaccharides, ed. D. A. Brant (American Chemical Society, Washington, D.C., 1981). Paper 6/093; Received 13th January, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203407

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. Sac., Faraday Trans. I , 1986,82, 3415-3430 Diffusion Phenomena and Metal Complex Formation Equilibria Part 1 .-CdII-Thiourea Systems in Aqueous and Mixed-solvent Media David R. Crow School of Applied Sciences, The Polytechnic, Woluerhampton WVl 1 SB Changes in polarographic diffusion current of the cadmium(1r) aquo complex when consecutively complexed with thiourea and substituted thioureas are directly related to commonly used polynomial functions of free-ligand concentration involving overall formation constants. Pseudo- formation curves are defined and their analysis exploited as a means of obtaining formation-constant data; these are essentially in agreement with those generated by more conventional methods when these are applicable. It is shown that the average diffusion coefficient of a system of metal complexes in equilibrium depends upon the free-ligand concentration in a manner which is a function of both the nature of the ligand and the solvent medium.Introduction and Theory The average value of the diffusion coefficient of the various simultaneously diffusing mononuclear complexed species in a system of metal-ligand step equilibria may be expressed as follows : N N N Eqn (1) assumes that each species may be allocated a characteristic ' polarographic' diffusion coefficient which is independent of the presence of other species. More (2) succinctly - D = a.,,D,+a,D,+ ... aNDN where a,,, a,.. .a, are the degrees of formation and Do, D,. . .DN are the individual diffusion coefficients of aquo and complexed species.Eqn ( I ) is of general form, applicable to a variety of mean parameters1 and to mean diffusion behaviour of various types of associated specie^.^-^ Dj and pj are the diffusion coefficient and overall formation constant of the species of coordination numberj, while [XI is the free-ligand concentration. The polynomial function &[XI has been much exploited in the determination of formation constants; its value may be obtained by a variety of methods including d.c. polarography, in which case it is a function of the change in half-wave potential of an aquo ion when complexed, provided that the electrode reactions are strictly re~ersible.~ Recently &[XI data have been shown to be functions of changes in diffusion current induced by complexation.6 with [XI must be some function of the complexing process; the size (and shape) of the diffusing species has changed.The almost invariable reduction in D suggests, as is intuitively reasonable, that the effective size of the diffusing It is clear that the variation of 113 3415 FAR 13416 Diffusion Phenomena particles has, on average, increased. Stepwise replacement of coordinated water by a larger new ligand species progressively increases the average volume of depolarizer particles reacting at a dropping mercury electrode and produces a corresponding reduction in observed diffusion current. It has been shown that, for an essentially constant complex-ion geometry, the change in diffusion current (Aid) of a metal aquo ion when complexed progressively with another ligand species is directly proportional to the ligand number (A) over the major part of the latter's range as realised by experimental conditions.7 y The full expression relating Aid and A has been derived as7 A = k(Aid--) (Aid)2 X kAid 2idM (3) where k is a constant and idM is the diffusion current of the aquo ion. The term is frequently small in comparison with Aid and allows the simpler approximate form of eqn (3) to be used for many systems. Since F, [XI may, in principle, be expressed in the alternative forms N F, [XI = z 83 Klj = &[XIE it follows that or - n - - A i -- d In F,' [XI dln[X] d - k where F,' [XI = (4 [X])llk and In F,' [XI = j:' Aid d In [XI. Graphs of Aid us. In [x] thus have a similar sigmoid shape to that shown by the formation curve; such pseudo- formation curves have successfully formed the basis for the estimation of consecutive overall formation constants provided k may be determined.2 The legitimate approxi- mation introduced into eqn (3) allows its expression in the forms: - In eqn (6) the constant k" is the product of k and k (the Ilkovic constant), while D , and D, are, respectively, the diffusion coefficient of the aquo ion and the mean diffusion coefficient shown by a system of species characterised by R .k" is characteristic of the metal-ligand system studied and of conditions such as solvent and temperature, but is independent of specific parameters of individual indicator electrodes. Eqn (6) implies the direct proportionality of the diffusion coefficients DMxj, of individual complexes to the squares of their coordination numbers.This has been verified for neutral ligands by the direct application of eqn (1) using experimentally determined D and F, [XI data combined with reliable values of formation constants obtained by independent means? Individual diffusion coefficients were estimated from graphical analysis based upon that of Leden.9 Expression of eqn (1) in the formD. R. Crow 3417 leads to - D(l +PMX+BMX2 [XI2+ ... ) = DM+DMXBMX [Xl+DMX2BMX~ ..- (7) It has been shown that eqn (7) may be used to obtain values of BMX...PMxN when the functions (D- DMx). . .(D- DMx,) are available.1° A graphical analysis, not unlike that of Leden, is employed, but there are some unusual features; initially the function G, is calculated (at experimental ligand concentrations) defined by and to (DM-D) = ( D - D ~ ~ ) ~ ~ ~ [ X l + ~ ~ ~ ( D ~ D ~ ~ ~ ) ~ ~ x It has been found to be most appropriate to obtain BMx by plotting values of G, against (D-DMx)and to estimate PMx in terms of the value of G, in the limiting condition [XI = 0; (D - DMx)[xl-o.Development of new functions of general form Gj-1- (D - Dj-1) Pj-1 [XI Gj = = ( D - D , ~ , ) P , ~ , + - - . +(D-DnnxN)PMxNIXIN-j (9) allows estimation of constants Dux, from plots of Gj us. @-DMxj) and extrapolation to the limiting value (Gj),,,,,; (D-&X,)[X],o. The final function, G,, is unique in that for each value of ligand concentration It was considered desirable for the present work, as for that published earlier and based upon pseudo-formation curves, to apply at least some independently determined diffusion current data reported in the literature.Here a serious problem was encountered; although many authors have been at pains to obtain reliable half-wave potentials, similar care does m t usually seem to have extended to measurements of diffusion currents as functions of free-ligand concentration. That both parameters are significant is apparent from the defining equation5 for F, [XI in terms of AE+ at 298 K, uiz. (w1 F, [XI = antiloglo 16.9 16n AE+ + log in which ZM and I , represent the diffusion current constants of aquo and complexed ions, respectively, n is the number of electrons exchanged in the electrode process, while the numerical factor includes universal constants and absolute temperature at a value of 298 K.For the case of incoming ligands which are considerably larger than water, the ratio IM/Zc becomes significantly different to unity at ligand concentrations corresponding to maximum coordination. In this respect two problems arise. First, as the present work will show, in the later stages of coordination Zc appears not always to reflect the primary complexation process. Secondly, there are clearly differing attitudes on the part of authors as to the relative importance of the current and potential terms in eqn (10). In many cases the log(ZM/Zc) term is dismissed as being of negligible size relative to the larger term in AEi (and, in fairness, this may sometimes be so). In other instances the current term has been included in calculations of &[XI, but one is left with the impression that this is but a token gesture to the demands of the equation.Variations in values of diffusion currents which are reported and used often suggest, by their unexpected irreguarity, that slight variations in metal-ion concentration have occurred 113-23418 Diflusion Phenomena from solution to solution. In cases where the most careful control of experimental conditions has been exercised, a wide range of concentrations covered and a large number of data points determined, a much more regular variation of diffusion current with ligand concentration is observed. It was suggested by Bond1' that after the generation of successive formation constants from &[XI data a recalculation of such data via the defining polynomial should be performed in terms of the values of the constants proposed. There are often large discrepancies observed between the values of 6 [XI deriving from primary experimental AEi data and those reported on the basis of derived equilibrium data.While such discrepancies may depend to some extent upon the quality of the potential values and very significantly upon the number of data points (which are frequently substantially too few), it will be shown in the present paper that incorrectly assessed current corrections can be a considerable source of error. In fact, it has become clear that an observed diffusion current, arising from the mass transfer of a variety of complex species determined by a prevailing value of free-ligand concentration, has significance of two kinds.First, it is an integral part of the expression for &[XI; secondly, and quite independently, it is a function of A. The latter property implies application, in principle, to systems showing electrochemically irreversible behaviour in which AEi is without direct thermodynamic significance. A major problem with the determination of AEi values over wide ranges of ligand concentration is the very large number of working solutions required. Inherent cumulative errors, which can so easily enter the graphical extrapolation procedures usually adopted, can most readily be minimised by the availability of a large number of data points. Preparation of the large number of solutions desirable often proves costly in terms of both time and money, but the extended timescale thus imposed on the measurements themselves can cause further problems. Reference potentials may change (AEi values are really required to f 0.1 mV) and, unless particular care is taken to ensure the constancy of the concentration of metal ion from one solution to the next (normally ca.lo-, mol drn-,), variation of diffusion current over the range of solutions will show erratic rather than systematic variation. It is preferable to generate individual working solutions in situ, via dilution wherever possible, and data used for the systems reported upon here were obtained by such means. These systems, all involving the cadmium ion, were chosen because of the proven reversibility of the electrode processes. In contrast to its behaviour in a 50% dioxan-water solvent, the CdII/Cd(Hg) system remains reversible in the 50% ethanol-water medium, even in the presence of the various thioureas considered.12 Formation-constant data are thus known by independent means, allowing reliable assessment of the significance of current as a reflection of the complexation process.Experiment a1 AnalaR grade cadmium sulphate and potassium nitrate were used as the sources of depolariser ions and supporting electrolyte, respectively; AnalaR thiourea was used as complexing agent. Generation of experimental working solutions was effected by the ' titration ' procedure described e1~ewhere.l~ To an initial carefully measured volume of a working solution containing mol dm-3 Cd2+ and 0.1 mol dm-3 KNO, were added suitable volume increments of a solution containing 0.1 mol dm-, KNO, and 2.0 mol dm-, thiourea.Between each of the latter additions the mean diffusion current at a constant applied potential of - 1.00 V us. SCE was measured using a manual polarographic circuit. This incorporated a Tinsley potentiometer type 3387B and a calibrated Pye ' Scalamp' d.c. microammeter. No maximum suppressor was used and with a natural drop-time of 2.3 s, distortion of the polarographic waves by second-order maxima was negligible. This was confirmed by means of a PAR 174A polarographD. R. Crow 3419 operated in d.c. mode at a 2 s drop-time. In any case the applied potential at which currents were measured in the manual technique was well removed from the half-wave potential. Alternatively, full polarograms might just as well have been obtained instrumentally between additions of ‘ titrant’ as has been described;13 indeed the latter procedure is more appropriate for systems which show a large and less systematic movement in half-wave potential, such as is frequently found in the case of the Ni2+ ion.l3 However, the visual monitoring of currents in the manual technique afforded rather better precision and allowed assessment to kO.1 PA.Clearly, while the concentration of supporting electrolyte remained unchanged during addition of ‘titrant’, those of the depolariser ion and ligand altered. The directly observed reduced diffusion current was derived from both the complexation of the Cd2+ ion and its dilution by the added solution. It was necessary to carry out a ‘blank’ run, using a working solution containing supporting electrolyte only, so that any effect upo~t the capacitance current of the added ligand could be measured and allowed for.Assuming the maintenance of near-ideal volumes, various ligand concentrations were estimated via a simple volume correction, v i z . C , = mu/( V + v ) , where C , = analytical concentration of ligand in a working solution, V = initial volume of solution taken into the cell, v = incremental volume of added solution and m = concentration of the ligand/supporting electrolyte ‘ titrant ’. Changes in current (id)eff, induced by the complexing process alone were calculated by applying the ‘ amperometric ’ correction factor to the observed diffusion currents (id)obs, viz. (id)eff = (id)obs (V+ v ) / V.In order for calculated currents and ligand concentrations to be determined accurately it was necessary to use initially dry polarographic cells and components and to measure initial volumes of working solutions particularly carefully. The maximum dilution attempted was 50%. In the case of the cadmium-thiourea system, no significant departure from a final volume of 100cm3 was observed following an overall dilution of 50cm3 of working solution with 50 cm3 of ‘ titrant’. Great care was taken with the deoxygenation process, effected with ‘white spot’ nitrogen; this was passed through two Dreschel bottles containing working solution thermostatted to the same temperature (298 K) as that in the polarographic cell. The rate of passage of nitrogen through the solution was constant, slow and never such as to allow violent agitation; between additions of ‘ titrant’ slow passage of nitrogen was maintained over the surface of the solution.Such precautions reduced to a minimum any tendency of the solvent to evaporate. Even with care taken to presaturate the gas with solvent, too vigorous deoxygenation was found to cause very slight but, in terms of the magnitudes of current differences required, significant increases in diffusion current due to evaporation. This procedure differs somewhat from that used by Lane et aZ.,12 the data used here for the systems involving cadmium and substituted thioureas in 50% v/v ethanol-water deriving from these workers. Lane’s initial working solution contained the maximum analytical concentration of ligand together with depolariser and supporting electrolyte.After recording the polarogram, the solution was diluted progressively with another containing the same concentration of cadmium ion and supporting electrolyte, but no ligand. Various ligand concentrations produced were calculated by application of the appropriate dilution factor. Data reported by Lane et aZ. for cadmium complexes with thiourea and substituted thioureas are almost unique in that a large number of effective working solutions, slightly differing in ligand concentration, were produced with not only accurate half-wave potentials but also diffusion currents measured in each case. It is particularly noticeable that in the case of both Lane’s data and the experimental results obtained in the present study, variation of diffusion current with ligand concentration is clearly defined and largely uncomplicated by irregularities often encountered with individually prepared solutions.In manipulating data obtained experimentally and reported by Lane, smoothed curves were usually drawn through points corresponding to the somewhat inconvenient values3420 Digusion Phenomena [ X]/mol dm-3 Fig. 1. Variation of diffusion current for mol dm-3 Cd2+ in the presence of increasing concentrations of various complexing agents : , thiourea; B, methylthiourea; a, ethylthiourea; 0, diethylthiourea (data derived from Lane et reported for 50% v/v ethanol-water solvent, left-hand scale applies); 0, thiourea (experimental points determined by dilution in aqueous solution, right-hand scale applies).of ligand concentration produced by the dilution sequence. Interpolated values of diffusion current were obtained and used at more convenient thiourea concentrations. Values of mean diffusion coefficient were estimated via the Ilkovic equation in terms of that for the cadmium ion, which was taken as 7.20 x cm2 s-l in aqueous solution and as 2.50 x loAs cm2 s-l in the mixed solvent. Results and Discussion Use of Changes in Diffusion Current Relative magnitudes of the variation of diffusion current of loA3 mol dm-3 Cd2+ with total ligand concentration are shown in fig. 1 (since [XItotal 4 [Cdltotal; [XItotal x D(]free). Here the data for thiourea and substituted thioureas in the mixed solvent are taken from Lane et aZ.;12 in some instances directly reported data are plotted, while in others interpolated values at rounded values of total ligand concentration are given.Experi- mental data obtained by the present author for the thiourea system in aqueous solution are shown on an expanded current scale for clarity. The 32 data points obtained for this latter case are necessary to clarify the shape of the curve, which is based upon relatively smaller current changes than those obtained for most of the other systems. There is clear evidence that the curve for the system involving the unsubstituted ligand in aqueous solution is largely free from a progressive distortion which is observed to varying extents, at higher values of concentration of substituted ligand, in the mixed solvent. The nature of this distortion is made clearer by inspection of the relationship between Aid and R for systems affected.For the case of cadmium-methylthiourea in the mixedD. R. Crow 342 1 OS6I 0.14 ( a ) 52 / id ?" 0.5 Y ii \ 4.0- 1.0 2.0 3.0 4.0 d.2 ' d.3 6.4 d.5 6.6 d.7 ' i.8 d.9 1:O o'ooK.l [MTUl/mol dm-3 a *C1 3 1 d - 0.15 1.4 1.2 3 1.0 -0.10 ;.= d, ;3 -3 0.8 0.6 OX 0.2 - 0.05 0.1 0.2 0.3 OX 0.5 0.6 0.7 0.8 0.9 1.0 [MTU]/mol dm-3 Fig. 2. (a) Diffusion current based parameters for the system cadmium-methylthiourea (MTU). a, Plot of Aid vs. calculatedn; 0, observed variation of diffusion current with MTU concentration; W, corrected variation in terms of extrapolated Aid for higher ligand concentrations; 0, variation of log(IM/Ic) based on observed Ic; @, corrected variation of log(IM/Ic) based on curve with filled squares.(b) a, Variation of observed Aid with ligand concentration; 0, variation of corrected Aid with concentration of MTU; @, variation of the term Aid-(Aid)2/2!dM with [MTU]; 0, variation of (Aid)2/2idM (expanded scale) with ligand concentration. solvent, Leden's analysis of functions based on Lane's half-wave potential data was repeated ; values of formation constants previously reportedf2 were confirmed and used to estimate values of A. It is clear from fig. 2(a) that Aid is a linear function of A until values of the latter reach the region of 3.6. Serious departure from linearity sets in after this point, which corresponds to a ligand concentration of 0.3 mol dm-3. The implication3422 Diffusion Phenomena is that the primary complexation process is essentially complete by this concentration and thereafter observed currents begin to reflect the onset of secondary processes (or conceivably geometrical changes in the predominating complex species). Expected values of Aid and id may be obtained by extrapolation to H = 4, as may also those of the function log(ZM/Ic) of eqn (10); these are shown for comparison in fig.2(a). The magnitude and significance of the term (Aid)2/2idl in eqn (3) is shown in fig. 2(b), which endorses the previous observation that beyond 0.3 mol dm-3 of ligand, functions of current no longer directly reflect the main complexation process. Integration of the pseudo-formation curves, derived from eqn (9, was carried out in order to obtain values of Fi [XI over a range of ligand concentrations for six systems; an exercise which, in the case of those showing marked distortions of the id vs.[XI curves, was carried out only for a range of concentration for which, so far as could be judged, this interference had not yet set in. Corresponding values of F, [XI were calculated from Lane’s half-wave potential data and these were checked against values predicted using his published formation constants. In some cases of serious discrepancy between values deriving from these two sources, the reported experimental data were reworked to produce rather different values of constants. Such discrepancy was most marked for the data reported for the parent thiourea complexes in aqueous solution and has been commented upon previously.6 The rather unlikely sequence14 p1 = 24, p2 = 51, p3 = 40, p4 = 3590, derived from the graphical interpretation of experimental half-wave potential shifts was less compatible with the original &[XI data than the sequence p1 = 20, pz = 100, p3 = 400, p4 = 3000 arising from an alternative interpretation of the data.An attempt to integrate the pseudo-formation curves drawn on the basis of Lane’s rather scattered and sparse published current data yielded p1 = 20, p2 = 180, p3 = 800 and p4 = 2500. Finally, the pseudo-formation curve based upon the large number of data points shown in fig. 1 (open circles) produced p1 = 23 & 2, p2 = 140 f 30; p3 = 300+ 150 and p4 = 1450+400. The family of plots shown in fig. 3 demonstrates the clear relationship between log (Fl [XI) data deriving from the respective pseudo-formation curves and log (F, [XI) based upon independent parameters.If, however, pseudo-formation curves are to be used independently to estimate formation constants, independent means are required for calculation of the ratio k = log (F, [X])/log (Fi [XI). From eqn (4) and (5) it follows that at ligand concentrations corresponding to the development of maximum coordination Thus, values of k may in principle be obtained from measurement of the maximum slope of the In (Fi [XI) vs. In [XI graph and assumption of a reasonable value for nmax. The latter is not always a clear-cut matter,6 in which case the unique value of k must be established by essentially trial-and-error means. For all the systems investigated here the maximum coordination number with respect to added ligand was expected to be 4.Graphs of log (Fi [XI) vs. log [XI are shown in fig. 4 for three systems and values of k determined from the limiting slopes of the curves for all the systems are compared with the values of the reciprocal slopes, B/Ai,, of fig. 3 in table 1. The agreement is remarkably good and allows some confidence to be placed in F , [ X ] data based upon assessment of an appropriate value of k by means of eqn (5’). This is the case even for the system involving ethylthiourea, for which from fig. 1 it is clear that whatever the limiting value of the current corresponding to maximum coordination, it is obscured by a large distortion. Graphical integration of the pseudo-formation curve for this system was carried out at appropriate intervals of ligand concentration up to 0.15 mol dm-3 only; generation of corresponding F, [XI data by use of k = 4.18 was followed by Leden analysis, which worked very satisfactorily to yield the data p1 = 50, bz = 1600, /I3 = 6000 and p4 = 4.6 x lo5.Formation-constant data reported by Lane et aE. @Il = 198, b2 = 400,D. R. Crow 3423 log (Fo) [ x 1 Fig. 3. Correlation between the function F,' [x] (deriving from integration of the pseudo-formation curve) and F, [x] (calculated from half-wave potentials). 0, Thiourea (aqueous solution). Non-aqueous: 0, diethylthiourea; a, dimethylthiourea; 0, methylthiourea; 0, ethylthiourea; ., thiourea. I - 2.0 -1.0 -0.5 log [XI Fig. 4. Examples of graphs of log (Fi [x]) us.log w] for the determination of k from limiting slopes. Complex systems involving Cd2+ ion and (a) thiourea (aqueous medium) ; (b) methylthiourea (non-aqueous) ; (c) dimethylthiourea (non-aqueous).3424 Diffusion Phenomena Table 1. Comparison of values of k in the expression Ti = kAid derived from two sources for systems involving the cadmium ion and thioureas ligand thioureaa thioureab meth ylthioureab dimethylthiourea* ethyl thioureab dimethyl thioureab 2.31 9.26 3.13 2.72 4.00 2.53 2.33 9.25 3.07 2.82 4.18 2.50 a Aqueous solution. * Ethanol-water solvent. p3 = 6.2 x lo3, #I4 = 4.8 x lo5) are seen to be at some variance with respect to the values of PI and #I2 found by the present analysis. Matters were improved, however, by careful replotting of Lane’s half-wave potential data, estimation of interpolated shifts from rounded values of lower ligand concentrations and recalculation of the factor log (ZM/Zc).The latter was carried out in terms of a limiting value for the diffusion current of 4.10 PA, as suggested by extrapolation of the linear Aid vs. E plot (using A values based on constants deriving from the pseudo-formation curve). Subsequent Leden analysis of the resultant F,[x] data gave rise to values of the constants agreeing closely with those deriving from analysis of current variations, viz. #I1 = 51, #Iz = 1600, /I3 = 6000 and = 4.7 x lo6. The theoretical limiting current is quite different from experimental diffusion currents for the ligand concentration range 0.15-0.7 mol dm-3. In fact, at 0.7 mol dm-3, the experimental diffusion current (3.30 PA) gives a value for Aid which is almost twice the theoretical value.Use of Mean Diffusion Coefficient, D, and Species Diffusion Coefficients, DMxj Plots of @ as a function of R (derived from half-wave potentials) for several systems are given in fig. 5 ; as a reflection of properties of the complementary pseudo-formation curves, these are linear for most of the range of values adopted, the characteristic constant k” following from the slope. Values of this constant for systems involving cadmium and a range of thioureas in both aqueous and mixed ethanol-water media are collected in table 2. Confirmation of the validity of eqn (3) and (6) permits the acquisition of two pieces of information; (i) corrected values of IM/Zc and (ii) values of diffusion coefficients of individual complexes.So far as the first is concerned, data presented in table 3 show the comparative effects on the values of F, [x] and derived functions of (a) ignoring the term log (ZM/Zc) in eqn (lo), (b) including experimental values which become increasingly inappropriate as the ligand concentration increases and (c) using corrected values. Once it is possible to assess the values of individual diffusion coefficients it is possible to use eqn (1) and its manipulation according to expressions (7H9) to determine refined values of formation constants. When provisional values of such constants have been estimated from a pseudo-formation curve, the ligand number may be calculated as a good approximation for selected values of free-ligand concentration.This then allows an estimate to be made of the limiting value of Aid, which corresponds to full development of the maximum coordination number. Calculation of the diffusion coefficient, DMxN, is then possible and interpolation for other values between DM and DMxy is straightforward. For example, at [thiourea] = 0.5 mol dm-3 (in aqueous solution), Aid = idM-idc = 14.15- 12.79 = 1.36 PA. From analysis of the pseudo-D. R. Crow 3425 22 - 20 - l.b/ 0- 0 I I I I 1 1 2 - 3 4 Fig. 5. Verification of eqn (6); plots of @ us. R for systems of cadmium complexes: 0, thiourea; a, methylthiourea (aqueous medium). Non-aqueous : (>, thiourea; m, methylthiourea; 8, dime thy1 thiourea. Table 2. Values of the constant k” [eqn (5)J for systems involving cadmium and a range of thioureas in both aqueous and mixed ethanol-water media ~~ ~~~ k”/cm-’si k” (non-aq) ~ ligand water ethanol-water k (as) thiourea 13.19 3 1.85 2.41 methylthiourea 5.25 12.25 2.33 dimethylthiourea 5.42 9.98 1.84 ethylthiourea 4.78 13.81 2.89 trimethylthiourea - 11.85 - diethylthiourea - 8.26 - formation curve, R x 3.06 at this concentration of ligand, so that k = 7i/Aid = 3.06/1.36 = 2.250.Thus, when A 4 4, (Aid)lim = 4/2.25 = 1.778 PA. The limiting value of current, (id)lim, corresponding to full coordination, is therefore 1t.15- 1.78= 12.37 PA. At zero ligand concentration with DM = 7.20 x cm2 s-l, D b = 2.683 x cm s*, so that the Ilkovic constant, given by id =Fa, becomes 14.15/2.683 x = 5.274 x lo3 PA cm-l sk Therefore, &m A+4 = Dhx4 = (id)lim X 10-3/5.274 = 2.345 x cm2 s-l.From eqn (6), when R = 1, 2 and 3, DMX, DMxz and DMx3 become, respectively, 6.750, 6.315 and 5.895 x cm2 s-l. Complete analysis of half of the experimental data, obtained for the cadmium-thiourea system in aqueous solution, conducted according to the principles outlined is presented in table 4. Corresponding graphs of functions G5 us. both [XI and (D- D,,S are shown in fig. 6 (these include all the experimental data). These graphs show not only data cm s-i and DMX = 5.499 x3426 Diflusion Phenomena Table 3. Comparative effects of differing IM/Ic ratios on the variation of F, [XI and 4 [XI with ligand concentration 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 .o 0.3 0.4 0.5 0.6 0.7 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8 877 2.510 6 147 12 584 23 467 39 854 63 597 101 484 233 694 1 572 3 169 5 909 9 727 643 3 052 9 083 23 131 50 409 101 622 162 173 1108 3 205 7 896 16 287 30 786 53 048 85 219 135 987 31 1 956 2 231 4 633 8 923 15 416 887 4 079 12 509 32 976 75 250 158 485 261 201 MTU 1105 1086 3 193 3 277 7 838 7 804 16071 15931 30002 29202 50953 49433 81 308 78 720 129 747 119 747 DMTU 302 307 91 1 892 2 077 2 085 4 206 4 212 7 861 7 668 12943 12923 869 859 3 833 4 048 11 435 12437 29 121 29926 63461 61 519 127 935 113 324 204 162 192 553 ETU 89 257 85 050 88 488 89 130 91 114 91 614 91 951 97 053 20 410 21 750 21 288 21 453 22 171 21 700 336 000 343 143 333 050 354 280 376 633 413 201 387 470 117 443 112 158 116 544 117 745 121 630 123 768 124 878 131 524 30 000 31 988 31 832 32 750 34 727 35 598 477 375 463 777 462 800 508 792 565 945 648 090 627 595 117 523 111 725 115 556 116 078 118 333 118 711 118 946 125 316 28 890 30 220 29 368 29 453 29 551 29 550 525 250 433 410 420 845 447 112 474 980 520 851 488 341 a Systems involve cadmium and the ligand species methylthiourea (MTU), dimethylthiourea (DMTU) and ethylthiourea (ETU).Values calculated from interpolated data of Lane et al. in 50% ethanol-water. Values calculated from AE: without IM/& correction. Values calculated from AE1 with IM/Zc correction estimated from experimental diffusion current data reported by Lane. Values calculated from polynominal in [XI using formation constants derived from the pseudo-formation curve. values calculated from AE: with modified IM/Ic correction.derived from individual experimental points (which develop inevitable scatter) but also data derived from interpolated values of mean diffusion coefficient at more convenient values of free thiourea concentration. Values of formation constants agree closely with those derived from the pseudo-formation curve and with those reported earlier on the basis of similar analysis of many fewer and independently published data points. In view of the very small overall current changes which occur for this case, the resulting analysis is remarkably satisfactory: matters are greatly improved for systems which, based upon bulkier ligands such as the substituted thioureas, show more dramatic changes. In fig. 7 are shown values of D calculated from experimental currents together with values calculated from different combinations of D,,, and bMx, data.Conclusions Variation of Aid with ligand concentration is a reliable reflection of complexation processes taking place and is related to the characteristic formation curve for a given system.Table 4. Derived Gj values and associated (B-O,,) functions required for the calculation of formation constants in the system cadmium-thiourea D, = 6.750 x lo-' D, = 6.315 x lo-' D, = 5.895 x D, = 5.499 x [thiourea] /mol dm-s 106D 106G, 106(D-D,) p,(D-D,) 106G2 106(D-D2) B2(D-D2) 106G3 106(D-D3) /3,(;fi-D3) 106G4 106(D-D4) G,(D-D4) O.ooO0 0.0198 0.0392 0.0769 0.1481 0.2456 0.305 1 0.3871 0.4375 0.461 5 0.5075 0.5507 0.6207 0.7500 0.7879 7.200 7.043 6.920 6.739 6.487 6.240 6.1 11 6.0 12 5.910 5.895 5.852 5.818 5.750 5.731 5.717 - 7.929 7.140 5.995 4.813 3.909 3.569 3.069 2.949 2.828 2.656 2.510 2.336 1.959 1.882 0.450 0.293 0.170 -0.01 1 -0.263 -0.510 -0.639 - 0.738 - 0.840 -0.855 -0.898 -0.932 - 1.000 - 1.019 - 1.033 - 5.86 3.40 - 0.22 - 5.26 - 10.20 - 12.78 - 14.76 - 16.80 - 17.10 - 17.96 - 18.64 - 20.00 - 20.38 - 20.66 - 104.49 95.41 80.82 68.01 57.45 53.59 46.06 45.14 43.18 40.62 38.41 35.99 29.79 28.61 0.885 0.728 0.605 0.424 0.72 - 0.075 -0.204 - 0.303 - 0.405 - 0.420 - 0.463 - 0.497 - 0.565 -0.584 -0.598 - - - 59.36 24.08 - 10.50 -28.56 - 42.42 - 56.70 - 58.80 - 64.82 -69.58 -79.10 -81.76 - 83.72 - - - 279.1 296.7 276.7 269.3 228.6 232.8 221.0 207.8 196.1 185.4 148.7 142.6 1.305 - - 0.844 0.592 0.345 0.2 16 0.1 17 0.015 0.000 -0.043 - 0.077 -0.145 -0.164 -0.178 - - - - - - 56 30 4 0 -11 - 20 - 38 - 43 - 46 - - - - - - 699 513 518 479 43 1 392 360 256 239 1.701 - - - - - 0.612 0.513 0.41 1 0.3% 0.353 0.319 0.251 0.232 0.218 - - - - - - b 1412 # 1000 1261 2 1210 1221 1229 1434 1103 1096 81 B 2 83 8 4 from [G3/(~-D3)l[xl-o 20f 1 140f20 260f 100 1200f300 from similar analysis of Lane's data 20f2 160f30 460f100 1164f200 from pseudo-formation curve 23f2 140f30 3005 150 14505400 w P h, 4Diflusion Phenomena 3428 9 - 8- 7- 6 - G1 5 - 4- 3 - 2 - 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 -1.0 - 0.5 0.0 9.5 @--D1) r I [TUIlmol dmV3 I 0.1 0.2 0.3 Ok 0.5 0.6 0.7 I 0.8 0.5 1D I @ - 0 3 ) Fig.6. Graphs of Gj us. corresponding (D-DMxj) functions and thiourea concentration for cadmium complexes. (a) Plot of G , us. (0- 0,) yields /3, = 9.0/0.450 = 20 (intercept of G, vs.[TU] curve is difficult to identify from that plot alone). (b) Plot of G, us. (D-D,) yields /3, = 124/0.885 = 140 (intercept of G, us. [Tv] curve is difficult to identify from that plot alone). (c) Plot of G, us. (0 -4) yields P3 = 340,’ 1.305 = 260 (intercept of G, vs. [Tu] is easily identifiable as 340). 0, Experimental points; 0, interpolated points; a, interpolated points for Gj us. [Tu] plots.D. R. Crow 3429 [ thiourea]/mol dm-3 Fig. 7. Values of D, derived from different sources, plotted as a function of thiourea concentration. 0, Experimental points. Other points are calculated on the basis of the individual diffusion coefficients proposed in table 4, but using various sets of equilibrium data obtained from independent analysis : symbol 8, 8 2 P 3 8 4 20 159 460 1164a 0 21 140 362 1254* 0 20 140 260 12OOc Ref.(10). Unpublished analysis. Present analysis. The ' pseudo-formation ' curve frequently shows quite clearly the values of ligand concentration at which maximum coordination is approached and beyond which new interactions become significant. - Many reported data, based upon measurements of half-wave potentials, derive from ligand concentrations, which favour formation of the higher complexes in a series. E.g. in the case of cadmium-ethylthiourea in the non-aqueous medium, although the maximum coordination is almost reached at a ligand concentration of 0.15 mol dm-3, reported potential and current data reach well towards 0.9 mol dm-3. Between these concentrations increasing errors in the log (ZM/Zc) function are observed.This introduces error into the DeFord-Hume treatment on two counts : (a) inconsistencies occur in the3430 Difusion Phenomena values of 4 [XI functions for higher ligand concentrations and introduce distortions, particularly in the linear graphs to be expected for ultimate and penultimate functions; (b) insufficient experimental data corresponding to the earlier stages of complex formation can provide misleading values of formation constants for lower complexes, e.g. for the case cited, H increases from zero to 0.98 within the ligand concentration range 0.0-0.02 mol dm-3, but this is represented by a single experimental datum point. Whatever experimental and computational methods used, it is required not only to have access to sufficient data; these must adequately represent, and as evenly as possible, the progressive state of complexation. Reliable values of diffusion currents, being capable of fairly rapid accumulation by the method described, can prove to be of considerable significance in this respect. The principles outlined allow the variation in diffusion coefficients for systems showing irreversible (but diffusion-controlled) behaviour at an indicator electrode to be used for the evaluation of equilibrium data. This has received preliminary confirmation in the case of nickel c~mplexes.~~q l5 The author thanks the referees for their helpful comments and their suggestions for improvement of the manuscript. References 1 H. Irving, Advances in Polarography (Pergamon Press, New York, 1960), p. 42. 2 J. M. Corkill and T. Walker, J. Colloid Interface Sci., 1972, 39, 621. 3 E. L. Cussler, AZChE J., 1980, 26, 43. 4 D. G. Hall, J. Chem. SOC., Faraday Trans. 2, 1985,81, 1599. 5 D. D. DeFord and D. N. Hume, J. Am. Chem. SOC., 1951,73, 5321. 6 D. R. Crow, Talanta, 1982,29, 733; 739. 7 D. R. Crow, J. Electroanal. Chem. Znterfacial Electrochem., 1968, 16, 137. 8 D. R. Crow, Polarography of Metal Complexes (Academic Press, London, 1969), p. 95. 9 I. Leden, Z . Phys. Chem. (Leipzig), 1941, 188, 160. 10 D. R. Crow, Talanta, 1983,30, 659. 11 A. M. Bond, Coord. Chem. Rev., 1971,6, 377. 12 T. J. Lane, J. W. Thompson and J. A. Ryan, J. Am. Chem. SOC., 1959,81, 3569. 13 D. R. Crow, Talanta, 1984,31,421. 14 T. J. Lane, J. A. Ryan and E. F. Britten, J. Am. Chem. SOC., 1958,80, 315. 15 D. R. Crow, Electrochim. Acta, 1983,244, 1799. Paper 61146; Received 20th January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203415

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82, 3431-3438 Kinetics of the /?-Hydroxy Elimination Reactions from the Pro t oporphyrin Iron(III)-CHRCH,OH Complexes in Aqueous Solutions A Pulse-radiolytic Study Yacov Sorek Nuclear Research Centre Negev, Beer-Sheva, Israel Haim Cohen" Nuclear Research Centre Negev and Coal Research Centre, Ben Gurion University of the Negev, Beer-Sheva, Israel Dan Meyentein* Chemistry Department, Ben Gurion University of the Negev, Beer-Sheva, Israel The free radicals 'CH,CH,OH and 'CH(CH,)CH,OH react with iron(n)protoporphyrin, FeIIPP, to form the transient complexes (PP)Fe"I-CHRCH,OH. The spectra of these complexes are reported. The transient complexes decompose in a first-order process via the reaction (PP)Fe1I1-CHRCH2OH --* FelIIPP + OH- + CH,=CHR. The specific rates of the latter reactions are 80 and 40 s-l for the two systems, respectively.The results are compared with those for analogous processes. Redox processes of iron-porphyrin complexes are of importance owing to their role in many enzymatic The high reactivity of iron-porphyrin complexes towards different redox agents suggests that they might react with aliphatic free radical^.^? lo Such radicals are formed in living organisms mainly by the absorption of ionizing radiation, but also in the presence of different chemicals, e.g. halothanes.1° The interaction of free radicals with iron-porphyrin complexes is therefore of interest. Recent studies9 have shown that several aliphatic free radicals, 'R, react with iron(I1)-porphyrin complexes, FeIIP, via FeIIP + 'R + PFeIII-R.(1) The products of reaction (1) are complexes which have an iron-carbon o-bond. Some of these complexes are relatively stable,s whereas others decompose via the heterolytic process : H2O PFeIII-R --+ FeIIIP + RH. Reaction (2) was observed even for free radicals which are powerful reducing agents, e.g. 'CH,OH, *CH(CH,)OH and *C(CH,),OH.ll Recent studies have shown that the decomposition of a series of complexes of the type L,Mn-CR1R,CR,R40H occurs via the @-elimination of OH- ions :12 (3) L,Mn-CR,R,CR,R,OH + MnL,+, + RlR,C=CR,R4 + OH-. The rate of reaction (3) depends strongly on the nature of the central cation M and the ligands L. 343 13432 /?-Hydroxy Elimination Kinetics It seemed of interest to study the mechanism of reaction of aliphatic free radicals with /?-hydroxy substituents and with iron(I1)porphyrins.If reactions analogous to reaction (3) occur, they might be of importance in different radiobiological proces~es.~~ Further- more, it was hoped that the measurement of the rates of /?-elimination processes in complexes with iron as the central cation might shed light on the factors affecting the rate of this process. In the present study the reactions of iron@) protoporphyrin, FeIIPP, with the free radicals 'CH,CH,OH, 'CH,CH(CH,)OH, 'CH(CH,)CH(CH,)OH and 'CH,C(CH,),OH are reported. Experimental Materials Solutions of FeIIPP were pKepared by reducing chlorohaemin obtained from Sigma Chemical Co. Stock solutions of FeIIIPP at pH 11 .O were reduced by the introduction of Adams catalyst, PtO,, and bubbling dihydrogen for ca.6 h. The stock FeIIPP solution thus obtained was transferred into an all-glass syringe through a glass sinter. The completion of the reduction was checked spectrophotometrically . The spectrum of FeIIPP thus obtained is identical to that reported in the literature.1° Thus one can reduce FeIIIPP by this technique, which is preferable as it does not introduce any additives into the solution. Solutions with the required composition were prepared by mixing this stock solution with the appropriate alkene-saturated solutions containing all other chemicals required. Finally, water saturated with N,O was injected into the syringe so that the partial pressure of alkene and N,O were 9: 1. All other chemicals were of analytical grade and were used without further treatment.All water solutions were prepared with heat-distilled water, which was further purified using a Millipore filtering system. Irradiation The pulse-radiolytic experiments were carried out using the electron linear accelerator of the Hebrew University of Jerusalem. 5 MeV, 200 mA 0.1-0.5 ps pulses with a dose of 100-500 rad per pulse were applied. A detailed description of the experimental set-up and the techniques used for data analysis has been published previously.12a Steady-state irradiations were carried out in a 6oCo y-source with a dose rate of 2.5 x lo5 rad h-l. Analysis The yield of dihydrogen as well as the yield of alkenes was determined by gas chromatography using a thermal-conductivity detector. The H, used for the reduction of FeIIIPP was eliminated before the irradiation by bubbling Ar through the solution.The yield of H, was calibrated by irradiating with the same dose N,O-saturated blank solutions containing 1 x mol dm-, NaBr for which G(H,) = 0.40.14 In any case, blanks in the absence of FeIIPP were done. Qualitative identification of alkenes was done by f.t.i.r. spectra using a Nicolet MXS infrared spectrometer, in a 14 cm gas-sampling cell and mass-spectral analysis was used to check isotopic enrichment where needed. Production of Free Radicals The radiolysis of water may be summed up by the equation Y , e- H,O + eLq, 'OH, 'H, H,, H,O,, H,O+, OH- (4) the yield of the products being Geq = GOH = 2.65, G, = 0.60, G,, = 0.45 andY. Sorek, H. Cohen and D. Meyerstein 3433 GHIO8, = 0.75.14 In solutions containing high concentrations of solutes the free-radical yield is somewhat larger and the yield of H, and H,O, is somewhat ~ma1ler.l~ The free radicals formed are homogeneously distributed in the solution within c 100 ns of the radiation being absorbed.The hydrated electron reacts with N,O according to e,-g+N,O + N,+0- k = 8.7 x log dm3 mol-1 s-l l5 ( 5 ) followed by 0-+ H,O + OH+OH- k = 1.7 x log dm3 mol-1 s-l. (6) Thus in neutral or alkaline N,O-saturated solutions, [N,O] = 2.2 x lo-, mol dm-,, more than 90% of the primary free radicals are transformed into 'OH radicals, the rest being hydrogen atoms. When (CH,),COH is added to the solution, the following reactions occur: OH/H + (CH,),COH + H,O/H, + 'CH,C(CH,),OH (7) ko, = 4.2 x lo8 dm3 mo1-1 s-l l6 and k , = 8 x lo4 dm3 mol-1 s-l l7 thus forming the 'CH,C(CH,),OH free radical.In solutions containing ethanol or propan-2-01 the following reaction occurs : OH/H + RCH(CH,)OH + 'CR(CH,)OH + H,O/H, (8) where R = H or CH,. The specific rates of the reactions with 'OH are 1.7 x log and 2 x lo9 dm3 mol-l s-l la and with H atoms 2.5 x lo7 and 1.8 x lo8 dm3 mol-l s-l l9 for CH,CH,OH and (CH,),CHOH, respectively. In these solutions ca. 13% of the free radicals are formed by /?-hydrogen abstraction, i.e. the radicals formed are *CH,CH,OH and 'CH,C(CH,)HOH.20* 21 In alkene-saturated solutions the 'OH is added to the double bond forming the appropriate /I-hydroxyalkyl radical : (9) R,R,C=CR,R, + 'OH + R,R,C'--CR,R,OH (in this study R, = R, = H; R,, R, = H or CH,).The specific rates are 4.8 x log and 6.5 x log dm3 mol-l s-l for CH,=CH, and CH(CH,)=CH,, respectively.22 E.s.r. data indicate that the product of the reaction 'OH + CH,=CHCH, is the secondary free radical 'CH(CH,)CH,OH and not the primary free radical 'CH,CH(OH)CH3.23 Results and Discussion When CH,=CHR (R = H,CH,) saturated solutions containing 10% v/v of N20 and (1-3) x mol dm-, FeIIPP at 10 < pH < 13 are irradiated two consecutive reactions are observed.? The first obeys a pseudo-first-order rate law, the rate of reaction being proportional to [FeIIPP]. The specific rates of reaction between 'CHRCH,OH and FeIIPP calculated from these results are summed up in table 1. The spectra of the short-lived transients obtained in these reactions (fig.1) resemble each other, but differ slightly from the spectrum which is attributed to FelPP/PPFelI1-Hll and more considerably from the spectra attributed to PPFe111CR,R20H.11 The difference in the spectra of the transients observed from that of FeIPP and/or PPFeIII-H as well as the difference in the final products (see below) suggests that the transients observed are PPFeIII-CHRCH,OH in analogy to PPFelll-CR,R,OH.ll The difference in the spectra of the latter two types of complexes is attributed to the effect of a-hydroxy substituents on the spectra of such complexes.11*12a The spectra of t In a previous study'l we have shown that the reactions of Fe**PP with 'CR,R,OH free radicals seem to be independent of the degree of aggregation of Fe'IPP. There is, therefore, no reason to expect an effect of aggregation on the reactions reported here.Table 1.Yields, spectra of products and specific rates of reaction of hydroxyalkyl radicals with FeIIPP solute:a 0.1 mol dm-3 0.1 mol dm-3 0.1 mol dm-3 9: 1 9: 1 N,O-satd N,O-satd He-satd satd satd CH,CH,OH (CH,),CHOH (CH313COH C,H*/N,O C3H6/N20 reacting 87 % 'CH(CH,)OH 87 % 'C(CH,),OH free radical : 13 % 'CH,CH,OH 13 % 'CH,CH(OH)CH, easb 'CH,CH,OH 'CH(CH,)CH,OH LE/nm 4/dm3 mol-I cm-I LE/nm 4 d m 3 mol-l cm-I k;t+Fe~~pp/dm3 mol-1 s-l k( transien t -+ products)/ s- l GH.2 GCH*=CH~ 560" 570" 550" 560e 550e 13 000" 11 OOO" 10 000" 7 0ooe 8 OOOe 4 200" 5 300" g - 680" 660" g g 4 x loEd 6 x loEd 6 x losd 6 x loEe 5 x l o g e 250" 300" 2.5 x 103" 80f loe 40f loe 0.88" 0.88" 2.95" 0.7 f 0.3e9 la 0.7 f 0.3e9 - - 0.6f0.1e a All solutions also contained 2 x towards FeIIPP and Fe'PP." of the short-lived transient.2: 1 mixture of alkene:N,O. Total dose 1.2 x lo5 rad. mol dm-3 Fe'IPP. Maximum of absorption of the short-lived transient. Only one absorption band observed. The 'CH,C(CH,),OH radicals formed in these solutions were shown to be unreactive This work. f Molar extinction coefficient Solution composition: 1 x loA3 mol dm-3 FeIIPP at pH 10.0 saturated with a Data from ref. (1 1).Y. Sorek, H. Cohen and D. Meyerstein 3435 10000 4 I - I m 0 5 0 0 0 E a \ 10oc 45 0 510 650 h/nm Fig. 1. Spectra of products formed by a pulse producing 3.3 pmol dm-3 of free radicals in a solution containing 2 x mol dm+ FeJIPP at pH 10.0 saturated by a 9: 1 mixture of C,H, and N,O.0, 500 ps after the pulse; A, 50 ms after the pulse. PPFeIII-CHRCH,OH are indeed similar to those of deuterporphyrin Felll-CH,.Ba Thus we attribute the first process observed to the reaction The short-lived intermediates PPFeIIICHRCH,OH decompose in processes obeying first-order rate laws. The rates [independent of the solute concentration, pH (in the range studied), pulse intensity and wavelength at which kinetics were followed] are listed in table 1. The spectrum of the final product in each of the two systems is identical to that of FeIIIPP (fig. I), as in the case of the reactions of FeIIPP with the a-hydroxyalkyl free radicals.ll There are three plausible mechanisms which are expected to obey a first-order rate law for the decomposition of the PPFeIII-CHRCH,OH complexes : FeIIPP + 'CHRCH,OH + PPFeIII-CHRCH,OH. (10) HZO PPFeIII-CHRCH,OH + Fe'PP + HOCHRCH,OH HZO PPFeIII-CHRCH,OH -+ FeIIIPP + CH,RCH,OH H2O PPFeIII-CHRCH,OH + FeIIIPP + H,C=CHR + OH-.Reaction (1 1) would be followed byll H2O FeIPP + FeII'PP + H,.3436 B-Hydroxy Elimination Kinetics Table 2. Specific rates of B-hydroxy elimination from L,M"-CH,CH,OH L,MnCH,CH,OH rate of reaction/s-l (dmg-H),Colll-CH,CH,OHa 0.7[H,O+Ib (H,0),Cr11*-CH,CH,0H2+ PPFelll-CH,CH,OH CU'I-CH,CH,OH~~ 2.0 + 1.43 x 104[H,0+]C 3.2 x 10, + 3.8 x 1O7[H30+Id 80(10 d pH < 13) a dmg-H = dimethylglyoxomato. Ref. (24). Ref. (12a). Ref. (124. The rate of reaction (14) is faster than that of reaction (1 l), thus one does not expect to observe FeIPP as a second intermediate. In order to check whether reactions (1 1) and ( 14) indeed describe the decomposition mechanism of PPFe111-CHRCH20H we measured the yield of dihydrogen in these systems; the results are summarised in table 1.The results clearly indicate that dihydrogen is not formed via reactions (1 1) and (14) as G(PPFeIII-CHRCH,OH) x 6 (i.e. Ges9+GOH), whereas the observed yield of dihydrogen is G(H,) = 0.7 +0.3. Most of the latter is due to the dihydrogen formed as a primary product of the radiolysis of water GH, = 0.45, the rest is probably due to the competition between N,O and FeIIPP for the reaction with e&. The product of the latter reaction is Fe1PP,l1 which yields dihydrogen via reaction (14). In order to determine which of the other mechanisms of decomposition of PPFelI1-CHRCH,OH [reactions (12) and (1 311 is correct we decided to measure the yield of olefin in these reactions.For this purpose N,O-saturated solutions containing 0.1 mol dm-, CH,CH,OH, 4 x lo-, mol dm-, FeIIPP at pH 10.0 were irradiated in the 6oCo y-source. The yield of C2H, was determined by g.c. Under these conditions one expects that 13% of the hydroxy radicals will react to yield 'CH,CH,OH radicals (see above). The observed yield of ethylene under these conditions [GC2H4 = 0.6k0.1, table 11 is in full accord with the expected yield of 'CH,CH,OH radicals and therefore proves that reaction (1 3) and not reaction (12) describes the mechanism of decomposition of PPFe111-CH,CH20H. When solutions containing propan-2-01 or propan- 1-01 instead of ethanol were irradiated, propene in a similar yield was identified as a product both by g.c.and f.t.i.r. No effort was made to determine the accurate yield of propene. In order to check whether the B-hydroxy elimination runs directly or via a short-lived intermediate of the type PPFeIII-CRCH,, as has been earlier suggested for analogous reactions,12a the experiments with the ethanol and propan-2-01 solutions were repeated in D,O solutions. If PPFeIII-CR=CH, is formed as an intermediate its decomposition in D20 would yield CDR=CH,. Mass spectrometry indicated that no deuterium, above natural abundance, is incorporated into the ethylene and propene formed. Thus reaction (13) proceeds without the formation of an intermediate of the type suggested,l2&9 in analogy to the mechanism of decomposition of (H20),-Cr-CH,C(CH,)20H2+.12f The radicals 'CH,C(CH,),OH and 'CH(CH,)CH(CH,)OH do not react with FeIIPP under our experimental conditions, probably owing to steric hindrance.Finally it is of interest to compare the specific rate of B-hydroxy elimination from PPFelll-CH,CH,OH with analogous reactions reported in the literature, table 2. The results indicate that the acid-independent term in the rate expression changes by over six orders of magnitude for the systems studied. [For (~~~-H),CO~~~-CH,CH,OH theY. Sorek, H. Cohen and D. Meyerstein 3437 acid-independent term is < 1 x s - ~ . , ~ ] The results thus indicate that the rate of fl-elimination depends strongly on the nature of the central cation and/or its ligands.The exact values of the equilibrium constants for the homolytic dissociation of the metal-carbon bond in these systems, reaction (1 5), is unknown : (15) Kh L,Mn-CH,CH,OH + L,Mn-l(H,O) + 'CH,CH,OH. However, from measurements for analogous systems one expects that Kh increases along the series ( ~ ~ ~ - H ) , C O I I I - R , ~ ~ (H,0),Cr111-R26 and CUII-R.~~~ It is thus tempting to speculate that the bond strength of the iron-carbon bond in PPFeIII-CH,CH,OH is smaller than in (H,0),-Cr111-CH,CH,0H2+ and larger than in CuII-CH,CH,OH&. We plan to extend our studies on the plausible correlation between Kh and the specific rate of fl-hydroxy elimination processes. We thank Mr S. Melloul for g.c. determinations and Mr D. Carmi for technical assistance.This study was supported by a grant from the Planning and eranting Committee of the Council of Higher Education and the Israel Atomic Energy Commission. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 (a) E. Falk, Porphyrins and Metalloporphyrins (Elsevier, Amsterdam, 1964); (b) B. Chance, R. W. Estabrook and T. Yonetani, Hemes and Hemoproteins (Academic Press, New York, 1966). R. H. Felton and H. Linschitz, J. Am. Chem. Soc., 1966,88, 143. D. W. Clack and N. S. Hush, J. Am. Chem. SOC., 1968,90,4328. D. G. Davis and R. F. Martin, J. Am. Chem. SOC., 1966,88, 1365. T. M. Bednarski and J. Jordan, J. Chem. SOC., 1967,89, 1552. C. Bartocci, F. Scandola, A. Fem and V. Carasiti, J. Am. Chem. SOC., 1980,102, 7068. L. A. Bottomley, L. Olson and K.M. Kadish, in Electrochemical and Spectrochemical Studies of Biological Redox Components, ed. M. Kadish (American Chemical Society, Washington D.C., 1982), p. 279. B. B. Hassinof and I. Pecht, Biochim. Biophys. Acta, 1983, 743, 310. (a) D. Brault and P. Neta, J. Am. Chem. SOC., 1981, 103, 2705; (b) D. Brault and P. Neta, J. Phys. Chem., 1982,86,3405; (c) D. Brault, D. Santus, E. S. Land and A. S. Swallow, J. Phys. Chem., 1984, 88, 5836. (a) R. S. Wade and C. E. Castro, Inorg. Chem., 1985, 24, 2862; (b) D. Mansuy and J. P. Battioni, J. Chem. SOC., Chem. Commun., 1982,638. Y. Sorek, H. Cohen and D. Meyerstein, J. Chem. SOC., Faraday Trans. I , 1985,81, 233. (a) H. Cohen and D. Meyerstein, Inorg. Chem., 1974, 13, 2434; (b) D. A. Ryan and J. H. Espenson, Inorg. Chem., 1982, 21, 527; (c) H.Elroi and D. Meyerstein, J. Am. Chem. SOC., 1978, 100, 5540; ( d ) M. Freiberg and D. Meyerstein, J. Chem. SOC., Faraday Tram. 1, 1980, 75, 1825; 1838; (e) Y . Sorek, H. Cohen, W. A. Mulac, K. H. Schmidt and D. Meyerstein, Inorg. Chem., 1983, 22, 3040; df) H. Cohen, D. Meyerstein, A. J. Shusterman and M. Weiss, J. Am. Chem. SOC., 1984, 106, 1876. See for example M. Dizdaroglu, C. von Sonntag and D. Schulte-Frohlinde, J. Am. Chem. SOC., 1975, 97, 2277; L. Stelter, C. von Sonntag and D. Schulte-Frohlinde, Int. J. Radiat. Biol., 1976, 29, 255; G. Behrens, G. Koltzenburg, A. Ritter and D. Schulte-Frohlinde, Int. J. Radiat. Biol., 1978, 33, 163; and F. Beesk, M. Dizdaroglu, D. Schulte-Frohlinde and C. von Sonntag, Int. J. Radiat. Biol., 1979,36, 565. M. S. Matheson and L. M. Dorfman, Pulse Radiolysis (M.I.T. Press, Cambridge MA, 1969). M. Anbar, M. Bambeneck and A. B. Ross, Nut1 Bur. Stand. Ref. Data Ser., 1973,43. L. M. Dorfman and G. E. Adams, Natl Bur. Stand. Ref. Data Ser., 1973, 46. M. Anbar, Farhataziz and A. B. Ross, Natl Bur. Stand. ReJ Data Ser., 1975, 51. M. Anbar, D. Meyerstein and P. Neta, J. Chem. SOC. By 1966, 742. P. Neta, Chem. Rev., 1972, 73, 533. Farhataziz and A. B. Ross, Natl Bur. Stand. Ref. Data Ser., 1977, 59. K. D. Asmus, H. Mockel and A. Henglein, J. Phys. Chem., 1973,77, 1218. J. K. Thomas, J. Phys. Chem., 1967,71, 1919. 23 W. E. Griffiths, G. F. Longster, J. Myott and P. F. Todd, J. Chem. SOC. By 1967, 530.3438 p- Hy droxy Elimination Kinetics 24 J. H. Espenson and D. M. Wang, Inorg. Chem., 1979,10,2853. 25 J. Halpern, Ace. Chem. Res., 1982,15,238; J. Halpern, Pure Appl. Chem., 1983,55, 1059; J. Halpern, F. T. T. Ng and G. L. Rempel, J. Am. Chem. SOC., 1979, 101, 7124; R. G. Finke, B. L. Smith, B. J. Meyer and A. A. Molinero, Inorg. Chem., 1983, 22, 3677. 26 G. W. Kirker, A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1982,104,1249; J. H. Espenson, Ado. Inorg. Bioinorg. Mech., 1982, 1, 1 . Paper 6/157; Received 21st January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203431

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1986,82, 3439-3445 The Relationship between Immersion Calorimetry and the Parameters of the Water Adsorption Isotherm on Active Carbons Francis Kraehenbuehl, Christian Quellet, Beat Schmitter and H. Fritz Stoeckli" Department of Physical Chemistry of the University, Avenue de Bellevaux 51, CH-2000 Neuchatel, Switzerland It is shown that for active carbons with a single type of hydrophilic adsorption sites, the enthalpy of immersion into water is related to the number of sites, a,,, and to a parameter, c, both of which appear in the Dubinin-Serpinskii equation for the water adsorption isotherm. The correlation was established on the basis of data for 13 different active carbons whose sites were either acidic or of the carbonyl type. It follows that the adsorption branch of the water isotherm can be calculated from data based on the enthalpy of immersion and the micropore volume.It is well known that carbonaceous surfaces are essentially hydrophobic, owing to the relatively low dispersion energies which exist between water molecules and aromatic sheets. Water can, however, be adsorbed preferentially on oxygen-containing centres which usually cover a small fraction of the total surface area. The mechanism of water adsorption has been discussed by Dubinin et a1.l in the fifties, and more recently by Dubinin and Serpinskii,2 who postulated the following relation for the adsorption of (1) water by active carbons : In this equation a represents the amount of water adsorbed at relative pressure pip, and a, is the number of so-called primary centres (usually expressed in mmol g-l of solid); c is the ratio between the rate constants of adsorption and desorption and k is a constant whose magnitude is fixed by the condition that for p l p , = 1 the total amount of water adsorbed in the micropores is a,.Implicitly eqn (1) is valid for a given type of primary sites and in the case of sites with different energies, parameters a, and c become apparent quantities. Note also that eqn (1) applies to the adsorption branch of the water isotherm and there exists no model for the description of the hysteresis loop associated with the desorption of water from the micropores. It has been shown recently by Stoeckli et ~ 1 . ~ that the number of primary sites a, is directly related to the enthalpy of immersion of active carbons into water.For typical industrial carbons, treated in vacuo at 400-600 "C and containing a uniform type of hydrophilic sites (probably of the carbonyl type since their high desorption temperature leads to carbon monoxide*), the relation is (2) where Ahi is in J g-l and a, and a, are in mmol g-l. The quantity Ahi represents the change in enthalpy associated with the transfer of a, mmol of water at room temperature, from the liquid state into 1 g of active carbon. Eqn (2) shows that the excess enthalpy can be divided into contributions from the primary sites and from the bulk of the micropores. The latter, arising from dispersion forces, are relatively small, but since a, 9 a,, the two contributions are of the same order of magnitude. The energies appearing in eqn (1) can be compared with the data from ab initio calculations by PIP0 = a M a , + a) (1 - WI.Ahi = - 25a, - 0.6(aS -a,) 34393440 Water Adorption on Active Carbon 0 0.5 1.0 PIP0 Fig. 1. The influence of the parameter c of eqn (1) on the shape of the water adsorption isotherm. (1) c = 1.0, (2) c = 1.5, (3) c = 2.0, (4) c = 2.5 and (5) c = 3.0, In all cases a,, = 1.00 mmol g-l. Malenkov and Dubinins for the interaction of water with carbon surfaces containing carbonyl sites. As suggested by Barton and Koresh,6 the adsorption of water vapour by active carbons depends on the amount of surface oxides, reflected by a,, and on the size of the micropores. The former parameter plays a major role at relative pressures p / p o < 0.3, whereas the pore dimensions manifest themselves in the region 0.3 < p / p o < 0.5.Mathematical modelling (see fig. 1) and the analysis of experimental data show that parameter c of eqn (1) also plays an important role in the description of water adsorption by active carbons. A better understanding can therefore be gained by relating this parameter to other properties of the system under investigation. Dubinin et aL7 have recently shown that for a given series of carbons with increasing degrees of burn-off, c decreases as the average micropore width increases. These authors had also shown earlierS that in the case of an active carbon subjected to heat treatment in argon at temperatures ranging from 1000 to 1600 "C, there exists a linear relation between constant c and the total surface area S of the micropores, which both decrease with increasing heat treatment.However, in this case a closer examination shows that the relation between c and the pore width does not hold, an indication that c probably depends on different variables. In the present paper we therefore wish to show that there exist a relatively simple relation between c and the enthalpy of immersion into water, as suggested by the analysis of eqn (1) within the framework of thermodynamics.F. Kraehenbuehl, C. Quellet, B. Schmitter and H. F. Stoeckli 344 1 Table 1. Parameters of the water adsorption isotherms and the enthalpies of immersion of 13 different active carbons into watera UO2-1 27.9 UO2-2 40.7 U02-3 70.0 U02-4 51.2 U02-5 56.2 UO3-1 31.2 F02 47.0 N-125 31.6 MSC5-1 23.3 MSC5-2 14.4 SP12 17.1 AU- 1 40.4 AU-2 49.4 26.1 26.1 26.3 28.0 31.0 31.0 32.6 31.5 9.5 9.5 7.0 22.0 35.0 1069 1559 2662 1829 1813 1006 1442 1003 2453 1516 2443 1836 141 1 0.35 0.95 2.01 0.59b O.7Ob 0.22b 1.49 0.49 0.74 0.38 0.58 1.18 0.95 0.50 - 1.03 2.22 - 0.63b 0.52b 0.69b O.7Ob 0.23b - 1.12 0.52 - 0.72 - 0.36 0.53 1.61 - 1.16 - - - - - 1.79 1.85 2.49 1.92 2.04 1.63 1.62 1.66 2.35 2.05 3.10 1.98 1.80 1.15 0.98 0.84 0.9 1 0.97 1.08 0.90 1.10 0.86 1.10 1.14 0.96 1.02 ~~ a ao(l) values obtained from the isotherm and eqn (1); a0(2) values obtained from the enthalpy of immersion and eqn (2) and (12); a0(3) values obtained by back-titration with HCl after neutralization with NaOH in excess. Acidic sites.Experimental Water adsorption isotherms were determined at 293 K, and in some cases at 263 K, on a series of well defined active carbons, characterized as described el~ewhere.~ The equipment was of the McBain type, fitted with transducer pressure gauges.The enthalpies of immersion were measured in a calorimeter of the Calvet type with 180 copper*onstantan thermocouples, corresponding to a sensitivity of 9.5 mV K-l. The equipment, calibrated electrically and with standard systems, has a reproducibility In order to extend the range of hydrophilic centres, carbons with carboxylic groups (UO2-4 and U03-1) were prepared by treating solids U02 and U03, outgassed at 400 "C, with hydrogen peroxide : 2 dm3 of H202 (10 % by volume) were added slowly to 60 g of active carbon, and the mixture was stirred gently for 20 h.The carbons, filtered off, were subsequently washed with water in a Soxhlet for 3 days and finally dried for 6 h at 120 "C in an oven. Sample U02-5 was prepared in a similar fashion starting with solid U-02 but with a 30% solution of H202. Prior to adsorption and immersion experiments, these solids were outgassed for 12 h under vacuum at a temperature of only 120 "C (cf. 400 "C for the other carbons, with sites of the carbonyl type). In order to assess quantitatively the number of acidic sites, samples of carbons UO2-4 and U02-5, previously outgassed at room temperature, were immersed into solutions of NaOH in excess and stirred for 48 h in PVC containers, to avoid base consumption by glass. From the back-titration with HCl the number of primary sites, a,, could be obtained.As shown in table 1, this was found to be in good agreement with the value derived from the adsorption isotherm, eqn (1). The reproducibility of the titration method was & 1 % . The other solids of the U02 and MSCS series were prepared by treating the initial solids at various temperatures between 473 and 873 K. of 1 % .3442 Water Adsorption on Active Carbon Theory Starting from the definition of the isosteric heat of adsorption, qst,lo qst (s)u =j@i (3) and using the water adsorption isotherm, eqn (l), written in the form one obtains Assuming that k does not depend too much on T, an assumption supported by experimental evidence, it follows that 6 lnc 6T -- - - (qst - AHva,)/RT2 where qst is a function of the temperature and of the filling of the micropores, 8 = a/a,. As shown elsewhere,ll in the case of a microporous solid the isosteric heat of adsorption is related to the enthalpy of immersion into the corresponding liquid and all energies being expressed in J mol-l or kJ mol-l of adsorptive (water in the present case).With eqn (7), eqn (6) becomes AHi - Jol 6 In c dB RT2 6T and since c does not depend on 8 for a given system, it follows that 6lnc AHi(T) 6 T RT2 ' -- -- (9) If the enthalpy of immersion does not vary significantly within a given domain of temperature, as suggested by typical experiments between 293 and 307 K, eqn (9) leads finally to (10) The enthalpy of immersion, expressed in J mol-l of water filling the micropores, is obtained by the simple relation where Ahi is in J g-l and a, is in mol g-l.Eqn (1 0) establishes a direct relation between constant c of the water adsorption isotherm and the enthalpy of immersion of the microporous carbon, a quantity which can be measured easily and quickly. Moreover, through eqn (2) and (10) the influence of the primary centres a, on c becomes obvious, provided that these centres are of the same type. c = c, exp (- AHJRT). AH, = Ahi/a, (1 1) Results and Discussion Table 1 summarizes the data obtained from water adsorption and immersion experiments at 293 K with 13 different samples. For sample SP all the data was found in the literature,12 and in the case of samples AU-1 and AU-2 the adsorption isotherms were provided by Dubinin.13 As a typical example, fig. 2 shows the water adsorption isothermF.Kraehenbuehl, C. Quellet, B. Schmitter and H. F. Stoeckli I I M 3a 20 10 0 3443 0 0.5 1 .o PlPo Fig. 2. The adsorption of water vapour by carbon UO3-1 at 293 K (hysteresis loop not shown). 1 .o rl I M - i E 1 0.5 1 .o 0 0 0.5 PlPo Fig. 3. The adsorption of water vapour by graphitized carbon black Vulcan-3G at 293 K (desorption points not shown).3444 Water Adsorption on Active Carbon for sample U03-1 at 293 K. The constants a, and c were obtained from a best fit of the adsorption data to eqn (l), as discussed el~ewhere.~, l4 The multiple linear regression was preferred to the graphical method proposed earlier by Dubinin,2 but in both cases the value of c depends, to some extent, on the number of points and on the pressure range. In the case of samples U02-4 and U02-5, containing only acidic sites, the values of a, obtained from the isotherm are in good agreement with the direct titration. It also appears that the corresponding enthalpies of immersion can be fitted to an equation similar to eqn (l), but with higher specific interactions: Ahi = - 55a, - O.6(aS - a,).(12) The specific interaction is twice as large as observed for the carbonyl groups described by eqn (2), which underlines their different nature. It has also been found that the enthalpy of immersion does not change between 293 and 307 K. As illustrated by table 1, the best fit of c and AHi to eqn (10) leads to the average value c, = 1.00+_0.11. A similar value is obtained for the adsorption of water by the graphitized carbon black Vulcan-3G at 293 K (fig.3), where c = 0.99. In the case of sample U03-1, the analysis of the water adsorption isotherms for 263 and 293 K has lead to identical values of a, (0.22 mmol g-l) and to c = 1.715 and 1.633, respectively. From eqn (9) it follows that Mi = -1048 Jmol-l, in good agreement with the experimental result, - 1006+ 15 J mol-l. Since the individual values of c derived from the isotherms are relatively inaccurate (up to & 10% in some cases), it seems justified to calculate them by using the corresponding enthalpies of immersion in eqn (10) with c, = 1. For active carbons with a known micropore volume and containing only one type of primary centre, parameters a, and c can be derived from the enthalpies of immersion by using eqn (2) or (12) and (10). It follows that the water adsorption isotherm, given by eqn (l), can also be calculated a priori for a range of temperatures.If the two types of sites mentioned above are present simultaneously, the enthalpy of immersion into water is a linear combination of eqn (2) and (12): (13) - Ahi = 25a,, +- Sa,, + 0.6(aS -a,, - aO2) where a,, and aO2 (in mmol H20 g-l) represent the corresponding sites. The back-titration of the acidic sites leads to ao2, and a,, can be calculated from eqn (1 3). By using a,,, aO2 and the corresponding values of c, and c, given by eqn (lo), it is possible to calculate, in principle, an overall adsorption isotherm resulting from the combination of the individual isotherms, eqn (1). Since immersion calorimetry is an accurate and relatively fast technique (1-2 days for the preparation of the solid and the measure of its enthalpy of immersion, as opposed to 2-3 weeks for the determination of a water adsorption isotherm), much time can be saved using the approach outlined in this paper in order to predict the interaction between water vapours and a pure active carbon.We thank M. M. Dubinin for his adsorption data on samples AU-1 and AU-2. References 1 M. M. Dubinin, E. D. Zaverina and V. V. Serpinskii, J. Chern. Soc., 1955, 1760. 2 M. M. Dubinin and V. V. Serpinskii, Carbon, 1981, 19, 402. 3 H. F. Stoeckli, F. Kraehenbuehl and D. Morel, Carbon, 1983, 21, 589. 4 S. S. Barton and B. Harrison, Carbon, 1972,10, 245. 5 G. G. Malenkov and M. M. Dubinin, Izv. Akad. Nauk SSSR (Ser. Khim.), 1984, 1217. 6 S. S. Barton and J. E. Koresh, J. Chern. SOC., Faraday Trans. I , 1983,79, 1147; 1157; 1165. 7 M. M. Dubinin, K. M. Nikolaev, G. A. Petukhova and N. S. Polyakov, Izv. Akad. Nauk SSSR (Ser. Khim.), 1984, 743.I;. Kraehenbuehl, C. Quellet, B. Schrnitter and H. F. Stoeckli 3445 8 M. M. Dubinin, G. A. Andreeva, R. S. Vartapetyan, S. P. Vnukov, K. M. Nikolaev, N. S. Polyakov, 9 F. Stoeckli, F. Kraehenbuehl, A. Lavanchy and U. Huber, J. Chim. Phys., Phys.-Chim. Biof., 1984,81, 10 S . J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity (Academic Press, London, 2nd 1 1 J. H. Clint, J. Chem. Soc., Faraday Trans. I , 1973,69, 1320. 12 B. R. Puri, S. Singh and 0. P. Mahajan, J. Indian Chem. Soc., 1965,42, 427. 13 M. M. Dubinin, personal communication. 14 F. Kraehenbuehl, Ph.D. Thesis (University of Neuchtitel, 1983). N. I. Seregina and D. V. Fedoseev, Izv. Akad. Nauk SSSR (Ser. Khim.), 1982,2425. 785. edn, 1982). Paper 61168; Received 23rd January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203439

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. 1, 1986,82, 3447-3459 Neutron Scattering of Supercooled Water in Silica Gels Christiane Poinsignon and John D. F. Ramsay" Institut Laue-Langevin, 156X, 38042 Grenoble, France and Chemistry Division, AERE, Harwell, Oxfordshire The effects of temperature on the incoherent quasielastic and inelastic neutron scattering of water in two silica gels (56 and 73% w/w SiO,) have been measured by time-of-flight spectroscopy from 300 down to 233 K, where the water is in the frozen state. Quasielastic scattering (QENS) data have been analysed using a model applied by Teixeira et al. in a recent study of the diffusive motions in supercooled water. This model involves a fast and a slow component, corresponding to the rotational and translational motions of the water molecule.The temperature dependence of the rotational dynamics of water in the gels is similar to that in the bulk, as is evident from the activation energy (ca. 10 kJ mol-l). At higher temperatures (2 278 K), translational diffusion is more restricted in the gels, and possibly reflects the effect of long-range surface interactions. However, in the supercooled region where hydrogen bonding is extensive in the bulk, and leads to marked reductions in the rate of diffusion, the translational mobility of water in the gels is similar to that reported for the bulk. Changes in the inelastic scattering, due to intermolecular modes, reflect the increase in hydrogen bonding which occurs on supercooling and also the structural change on freezing. It is also shown that partial freezing can occur at intermediate temperatures (ca.260 K), where the water content of the gel is high and thus possibly less influenced by the silica surface. Understanding of the structure and dynamics of water and aqueous solutions has increased rapidly during the past few years. Progress has been particularly significant in neutron scattering studies, where instrumental developments and advances in data treatment and modelling have been considerable. These studies have demonstrated links between the structure and dynamics of water and highlighted the importance of measurements at low temperatures, especially in the supercooled region where hydrogen bonding is extensive and can lead to anomalous physical properties.2* Of particular note here are the recent quasielastic neutron scattering (QENS) investigations of supercooled water of Teixeira et al.,4 who have shown that anomalous behaviour in this region is reflected by marked reductions in the rate of translational diffusion.There has also been a long-standing interest in the nature of adsorbed water within porous material^.^ Although QENS techniques have more recently been applied, attention here has been directed mainly to studies of water in clay systems6 and, to a much more limited extent, low water coverages on porous silica surfaces' near ambient temperature. In both of these situations a marked influence in the water behaviour is known to occur from interactions with exchangeable cations on the one hand and the hydroxylated surface on the other.Investigations of the structure and dynamics of water in porous solids, at levels close to saturation, have been less numerous. These have included recent neutron diffraction8 and incoherent inelastic scattering st~dies.~ In the latter, where the energy of the intermolecular librational modes have been determined, it has been demonstrated that on cooling these bands show an increasing departure from that of bulk water as the pore size and water uptake are reduced. Indeed, in microporous 114 3447 FAR 13448 C. Poinsignon and J. D. F. Ramsay titania (pore size <2 nm) it has been shown that water remains ‘liquid like’ even at temperatures as low as 10 K.9 In these investigations more significant departures from bulk behaviour have been reported where the water molecules have been close (5 1 nm) to a surface.Over a longer range than this the influence of surfaces is not fully understood, although it is probably of considerable importance. For example, such a knowledge is necessary for an understanding of the interactions between colloidal particles in aqueous media and, in particular, the stabilising role and range of solvation forces. Because the hydrogen-bonded structure of water becomes more extensive in the supercooled state, and leads to marked changes in dynamics, measurements made under these conditions may give a more sensitive reflection of the longer-range perturbation caused by surfaces compared to previous studies made at ambient temperatures. In the present investigation, with these aspects in mind, the quasielastic and low-energy inelastic incoherent neutron scattering of water in two highly concentrated dispersions of colloidal silica (ca.56 and 73% w/w as SiO,, respectively) have been measured in a temperature range from 300 to 233 K, where the water is in the frozen state. The dispersions used here were produced by partially dehydrating a silica sol (particle size ca. 16 nm) which on total dehydration has been shownlO to give a mesoporous silica gel (pore radius ca. 2 nm). From small-angle neutron scattering studies it has also been demonstrated that on progressive concentration of this sol to a hydrated gel there is a corresponding, and almost uniform, reduction in the interparticle separation. Such behaviour is particularly suitable here because the range of influence of the silica surface on the surrounding water, and the effects of the size of the ‘pore space’, can be investigated in a controlled manner.Experiment a1 Materials Two hydrated silica-gel samples, S3a and S3b, containing 56.5 and 72.9% w/w silica, respectively, were prepared by progressive evaporation of a concentrated silica sol (ca. 40% w/w) which has been studied previously.l0V l1 Before evaporation this stock silica sol (S3), which was obtained commercially (Ludox HS, E. I. du Pont de Nemours & Co.), was dialysed repeatedly against water, to give a final pH of ca. 8. A silica-gel sample (ca. 60% w/w) containing deuterium oxide (> 99% D,O) was also prepared from the same sol sample, using dialysis and ultrafiltration procedures already described.ll This was used to determine the background coherent neutron scattering from the silica itself.The mean diameter of the sol particles, as previously determined from transmission electron microscopy, was ca. 16 nm. Surface and porous properties of the dehydrated gel, previously determinedlo from nitrogen adsorption isotherm measurements (at 77 K) after outgassing at 423 K, are given in table 1. The volume fractions of silica in the gels S3a and S3b were 0.37 and 0.55, respectively, assuming a density of 2.2 g ~ m - ~ for the bulk silica. This compares with that of ca. 0.64 for the dehydrated gel (table l), where the particles are in direct contact. Since it has been shown that the packing geometry of the particles is very similar in these concentrated sols and gels,l0 it is possible to calculate approximate average interparticle spacings of 19.4 and 17 nm for S3a and S3b, respectively. On this basis a closest distance of surface separation of 3.4 and 1 .O nm, respectively, can be tentatively estimated in each case, by taking the particle diameter to be 16 nm.Supercooled Water in Silica Gels 3449 Table 1. Surface and porous properties of dehydrated silica gel specific surface pore volume, pore radius, porosity, area, SBET/m2 g-l V,/cm3 g-l r,/nm & 208 0.23 2.1 0.336 Neutron Scattering Measurements were made with the IN6 time-focussing time-of-flight spectrometer12 at the Institut Laue-Langevin, Grenoble.The incident neutron wavelength was 5.1 A and measurements were made at 19 different angles corresponding to wavevectors, Q, in a range from ca.0.2 to 2.1 A-l. The instrument resolution function was determined for each angle with a standard sample of vanadium; at Q = 0.9 A-1 the energy resolution (f.w.h.m.) was 80 peV. Thin gel samples (ca. 0.5 mm) were contained in circular aluminium cells (diameter 50 mm), which were oriented at 135" to the incident beam and contained in a thermostat. Measurements were made at five different temperatures in a range from 300 to 233 K. Theory In investigations of the present type the neutron scattering from the aqueous silica gels will be essentially incoherent because it is dominated by the scattering from the protons of the water. Since the theory of incoherent neutron scattering and its application to the study of the dynamics of aqueous systems has been described extensively elsewherel3v l4 only a general outline of the principles and basis of the dynamic model used here will be given.In the type of experiment performed here the scattered intensity resulting from a beam of monoenergetic neutrons incident on a sample is measured as a function of energy (i.e. time of flight), at different scattering angles. The angle of scattering from the incident beam, 8, defines the momentum transfer in the scattering process, which is given by Q = k-k, (1) where k, and k are the incident and scattered wavevectors, respectively. In the quasielastic region, where contributions from coherent scattering can be neglected, the double-differential scattering cross-section per unit solid angle, SZ, and energy, w, can be written: where k u = E-E, is the energy transfer, ainc is the incoherent cross-section of the proton and S,(Q, w ) is the incoherent scattering law.S,(Q,w) is related by Fourier transformation in space and time to the van Hovel5 self-correlation function, G,(r, t), which describes the ensemble average probability of finding a particle at a position r at a time t, when the same particle was at the origin at t = O : 1 f f Rather than attempt the calculation of G,(r,t) from the measured S,(Q,w), an appropriate dynamic model is generally selected and a comparison made between theoretical fits and experimental data. The model used here4 to describe the motions of the water molecule includes 114-23450 C. Poinsignon and J. D . F. Ramsay contributions from vibrations, rotations and translations which are independent or uncoupled, and can thus be written where x signifies a convolution product of the three motions.These can be described separately as follows. ss(Q, 4 = WQ, 4 x s?<Q, @ x SXQ, 4 (4) Vibrations Contributions from vibrations to the total scattering law are generally reflected by peaks of low intensity in the inelastic region. This contribution of energy reduces the scattering law in the quasielastic region through the Debye-Waller factor exp ( - (u2)Q2), where (u2) is the mean-square amplitude of the vibrations, so that we have Ss<Q, 0) = ~ X P (- (u2>Q2) [s?<Q, W ) x S?(Q, 0)l. ( 5 ) Rotations The isotropic rotational model for this motion of a water molecule is given by SearP as where a is the radius of rotation of the water molecule, taken as 0.98 A (the 0-H distance), jt(Qa) is the elastic incoherent structure factor (EISF), and (22+ 1) j:(Qa) are the quasielastic incoherent structure factors corresponding to the spherical Bessel functions j&).DR is the rotational diffusion constant, which is related to the n.m.r. correlation time z, by4 z1 = 1/60,. (7) With the value of a used here and the range of Q covered, only the first-order terms are applicable. This is demonstrated in fig. 1 , which shows the functions ji(Qa) corresponding to values of 1 of 0, 1,2 and 3 for the conditions appropriate to this study. Translations The translational motion can be described in terms of a random-jump diffusion mode114 where and r(Q), the h.w.h.m., is given by Here z, is the residence time and 0, the translational diffusion constant. DT is related to the mean-square jump length (P)$ by D, = 12/67,.(10) Data Fitting In the fitting of the theoretical made1 to the experimental data a computer program developed at the I.L.L. was used.17 The spectra were thus composed of a sum of Lorentzians, folded with the resolution function, corresponding to translational andSupercooled Water in Silica Gels 345 1 Fig. 1. Elastic incoherent structure factor (a) and quasielastic incoherent structure factors (b) (c) and (d) for the isotropic rotational motion of water molecules, computed with a = 0.98 x in eqn (6), where 1 = 0 (EISF) and 1 = 1, 2 and 3 (QISF) for (a), (b), (c) and (d), respectively. rotational motions as described.The parameters fitted were D,, D,, (u2) and F(Q). The F(Q) factor allows for the purely elastic coherent scattering of the silica, which was determined using a deuterated gel sample. Standard multiple scattering corrections were performed on the original data with the program DISCUSS.'^ At temperatures of 273 K and below the fitting process at low Q became less reliable, especially in the evaluation of DT, since the linewidth r(Q) eventually became con- siderably smaller than the resolution function. Results and Discussion The effects of decreasing temperature on the neutron time-of-flight spectra of water in the silica dispersions are typified in fig. 2 by results obtained at Q x 2.0 A-1 for the gel having a silica concentration of 73 % w/w.These illustrate two important features. First, a marked narrowing of the quasielastic peaks occurs on cooling, which reflects a reduction in the rate of diffusion. Secondly, in the inelastic region the bands due to intermolecular modes of low energy (5 30 meV) become more intense and better defined. A more detailed interpretation of the changes in these two regions can be obtained from a further analysis of such raw data as described below. Quasielastic Scattering The evolution of the quasielastic spectra of water with temperature is shown at a Q value of ca. 2.0 A-1 for both concentrations of silica gel in fig. 3(a) and (b). For the lower concentration (56% w/w SiO,) a gradual narrowing of the quasielastic peak occurs as the temperature is decreased to 273 K but thereafter, at 265 K, an abrupt decrease occurs giving a peak (4) which is only slightly broader than the resolution function ( 5 ) obtained with vanadium.This suggests that all the water remains as liquid at 273 K, but at 265 K partial freezing occurs, leaving a small proportion of water in a supercooled state, which gives rise to the residual broadening, which is more apparent in the wings of the quasielastic peak.3452 C. Poinsignon and J. D . F. Ramsay Fig. 2. Time-of-flight spectra of water in silica gel S3(b) (73% w/w SOz) at different temperatures: (a) 300, (b) 278, (c) 268, (d) 253 and (e) 233 K. Measurements are for Q x 2.0 A. N.B. Shift in ordinates. n 3 1 1 1 1 ( b ) l l l l l 0 8 c energy transfer, ho/eV Fig. 3. Evolution of the quasielastic peak with temperature for (a) silica gel S3(a) and (b) silica gel S3(b).Temperatures in (a) are 300, 278, 273 and 265 K for (l), (2), (3) and (4). In (b), ( 5 ) is the resolution function and temperatures are 300, 278, 268, 253 and 233 K, corresponding to (6), (7), (8), (9) and (lo), respectively.Supercooled Water in Silica Gels 3453 Table 2. Translational and rotational diffusion data for water in silica gel, S2(b) 300 12.2 1.2 0.29 2.1 278 12.1 0.9 0.23 3.1 268 8.3 0.8 0.18 4.7 At the higher silica concentration (73 % w/w SiO,) there are significant differences in behaviour. First, the broadening is less at 300 and 278 K; secondly there is no abrupt decrease in the quasielastic peak width (9) at 253 K which would indicate the onset of freezing. At 233 K, peak (lo), the quasielastic peak is, however, indistinguishable from the resolution function.At this low temperature such behaviour may not nevertheless be unequivocal in demonstrating total freezing of the water, since any liquid water remaining might be expected to give rise to diffusional broadening within the resolution function. It will be shown later that a more reliable indication of the state of water at these lower temperatures can be obtained from the inelastic scattering behaviour. Before proceeding to a description of the inelastic region a more detailed analysis of the quasielastic scattering law followed at the three highest temperatures (300, 278 and 268 K) will be given. Such an analysis is precluded for the lower temperatures owing to the limited quasielastic broadening compared to the instrumental resolution (80 peV, f.w.h.m.) attainable.Typical fits of the scattering law in eqn ( 5 ) to the experimentally determined quasielastic spectra for the dispersion having a concentration of 56% w/w SiO, are shown at three different values of Q for 300 and 278 K in fig. 4(a) and (b), respectively. As previously described, the marked narrowing of the quasielastic peaks on cooling is clearly demonstrated. The effect of water content is seen by comparing the spectra at equivalent Q values (uiz. 1.43 and 1.60 kl) for the dispersion of higher concentration at 300 K (fig. 5 ) with those of fig. 4(a). It is evident that a slight narrowing of the quasielastic peaks occurs. These effects are more clearly demonstrated in fig.6, which shows the dependence of r on Q2. Scattering law parameters for the more concentrated gel, S3(b), are given in table 2. Values of DR and D,, obtained from fitting the other spectra to the scattering law, are shown in fig. 7 for different temperatures, and compared with those obtained by Teixeira et aZ.* for bulk water extending into the supercooled region. At the highest temperature it is apparent that the translational diffusion of water [fig. 7(a)] becomes slower as the silica concentration is increased. This effect becomes less pronounced as the temperature is decreased, and from further extrapolation in the supercooled region it appears that a marked reduction in translational mobility, which occurs in the bulk, may not take place. The influence of the silica seems to be less significant on rotational diffusion [fig. 7 (b)].Thus, despite the limited data available here, it appears that the water in both silica dispersions has a similar Arrhenius behaviour to that shown by the lnDR us. 1/T relationship for bulk water as reported by Teixeira et aL4 From the latter an activation energy for rotational diffusion of 10.5 kJ mol-1 has been deri~ed.~ For bulk water it will also be noted that the effect of supercooling has a less pronounced effect on D , compared to DT. The reason why translational diffusion should be more sensitive to the presence of the silica particles is not clear. This difference, however, may arise from perturbations in the water structure by the silica surface. Thus it is well established from several investigations of the structure and dynamics of water adsorbed on the surface of claysg and porous oxides6t that the first two layers have properties which differ from those in the bulk.lS3454 C.Poinsignon and J. D . F. Ramsay ho/meV r hw/meV Fig. 4. Model fits to quasielastic scattering peaks for silica gel S3(a) at (a) 300 and (l), (2) and (3) correspond to Q/A-l = 1.11, 1.43 and 1.60, respectively. (b) 278 K. It is also significant that translational motion is much more sensitive than molecular rotation to the extent of hydrogen bonding in water, as has already been demonstrated in the studies of supercooled water by Teixeira et aZ.,4 and in the present investigation. Inelastic Scattering (INS) It can be shown that information on all the different types of motion (vibrations, rotations and diffusion) of a scattering system are contained in the velocity autocorrelation (1 1) function, which is given by13 ( N t o ) W o + 0 )Supercooled Water in Silica Gels 3455 holmeV Fig. 5.Model fits to quasielastic scattering peaks for silica gel S3(b) at 300 K. (a), (b), (c) and (d) correspond to Q/A-l = 1.43, 1.60, 1.77 and 1.99, respectively. 300 20c 1 h h 9 L 1 oc C 1 I I I I 2 4 eZ1A-2 Fig. 6. Dependence of the half-width at half-maximum of the quasielastic peak r(Q) us. Qz for water in silica gels at different temperatures. 0 and a, S3(a) at 300 and 278 K; 0 and a, S3(b) at 300 and 278 K; (---) bulk water at 293 K.3456 I I I I 3 - ( a > C. Poinsignon and J. D. F. Ramsay 0.5 \ 0 '\, \ \ \ 0 \ :'a 't \ \ 30t I I I I I I I I 103 KIT 3.4 3.6 3.8 Fig.7. Dependence of (a) translational diffusion constant, D,, and (b) rotational diffusion constant, D,, on temperature. 0 and in (a) and 0 and in (b) correspond to S3(a) and S3(b), respectively. Broken lines show dependence for bulk water from ref. (4). frequencylcm -' I 1 I l I I I I 0 20 40 60 - hw/meV Fig. 8. Effect of temperature on vibrational density of states, Z(w) for water in S3(b). Temperatures are 300, 268, 253 and 233 K for (a), (b), (c) and ( d ) , respectively.Supercooled Water in Silica Gels frequency/cm-' 3457 0 250 5 00 I I I I I I I 0 20 40 60 ho/meV Fig. 9. Effect of temperature on vibrational density of states, Z(o), for Temperatures are 300 and 265 K for (a) and (b), respectively. water in S3(a).where v(to) is the velocity of an atom at time to and v(to + t) the velocity of the same atom at time to + t . The spectral density of expression (1 1) is defined by 1 roo Z(W) = J (u(to)u(to+t)) cosmt dt 311 0 and it can be that Z(o) is obtained experimentally from the relationship Z(m) = o2 lim [SS(Q,w)/Q2] Q+O where Ss(Q, w) is the scattering law as previously defined. Owing to the dominating incoherent cross-section of the hydrogen atoms, the measurements of Z(o) obtained here will reflect the spectral density of states of water and will thus be directly comparable with low-frequency Raman and infrared spectra. The effects of temperature on Z(o) for the gel of highest silica concentration are shown in fig. 8. At 300 K three bands are evident: one of moderate intensity at ca.7.5 meV, a much weaker second one at ca. 19 meV and a third at ca. 57 meV. The first of these, vT1, has been extensively reported20 from Raman and INS measurements on liquid water, but is absent in the infrared spectrum. It has been ascribed21*22 to a hindered translation involving a deformation of a hydrogen bond 0-H. - -0. This band has also been in recent INS studies on aqueous solutions of ZnC1, and has been shown to decrease in intensity with increasing solute concentration, a feature which has been attributed to a break-up of the hydrogen-bond structure. The position of the second band, vTa, is close to that reported from infrared and Raman studies, and has been ascribed to a hydrogen-bond stretching mode2'* 2 3 9 24 or, less specifically, a hindered translational mode.The third broad band, vL, is well documented20 and can be assigned with confidence to a molecular libration. On supercooling to 253 K [fig. 8(c)] two features are evident. First the vT1 band becomes sharper and more intense, and secondly the vL band has shifted to higher energy (ca. 68 meV). Both of these features reflect an increase in the hydrogen bonding in supercooled water; there is, however, no evidence of a phase transition to ice. On further cooling to 233 K marked changes occur [fig. 8(d)] which undoubtedly signal the formation of ice: the band at ca. 7 meV is considerably reduced in intensity, whereas that at ca. 19 meV has increased markedly, both features being typical of the spectrum reported for ice.2o* 25 A new band which appears at ca.34 meV has also been observed with ice26 by Raman spectroscopy, although its assignment is unclear.3458 C. Poinsignon and J. D. F. Ramsay For the gel of lower concentration the characteristics of Z(m) at 300 and 278 K were similar to those previously described. At the lowest temperature studied, 265 K, significant differences were noted which suggest that partial freezing had occurred. This is indicated in fig. 9(b) by the increased intensity of the band at 19 meV without the attendant decrease in that at 7 meV, which was observed at 233 K with the gel of higher silica content. Such an effect indicates that the freezing of water is dependent on its environment. Thus when water is close enough to the silica surface to cause a change of the hydrogen-bonded network which develops in the bulk, retarded freezing may occur.Such surface effects may also explain the less marked changes observed in the translational diffusion constant of supercooled water in the silica dispersions compared with that which occurs in the bulk, as already noted (see fig. 7). Thus the striking reductions of D, reported for bulk supercooled water have been ascribed4 to a considerable increase in hydrogen bonding, which implies that this process may be restricted because of the influence of the silica surface. The present study demonstrates the power of neutron time-of-flight spectroscopy for obtaining information on the dynamics and intermolecular structure of water in porous media such as the silica gels investigated here.On cooling water into the supercooled state a marked reduction in the translational diffusion constant occurs. At high silica/water ratios this effect is, however, less than that recently reported by Teixeira et al. for bulk water. This difference may be ascribed to the effects of surface interactions, which appear to reduce the translational mobility of water at temperatures near ambient on the one hand, and restrict the development of an extensive hydrogen-bonded structure in water in the supercooled state on the other. It has also been shown that the reduction in the rotational diffusion constant is less sensitive to cooling, in accord with the previous findings made with bulk water. The novel method of following supercooling and freezing of water in porous media from changes in the low-frequency inelastic spectra, as discussed here, has considerable potential in future investigations. Here it is shown that partial freezing can be observed, and furthermore there is evidence that the influence of the silica surface delays the onset of freezing. The controversial question as to which is the more important, the size of the pore space or the extended influence of the surface, in depressing the freezing temperature, may indeed be answered in future investigations.These would ideally be made with adsorbents with a range of pore size, containing different multilayer thicknesses of water. We thank Drs M. Bee and A. J. Dianoux for helpful discussions and assistance in the analysis of time-of-fight spectra and Drs J.Teixeira and M. C. Bellisent for personal communications on the Z(o) spectra of supercooled water. The provision of experimental facilities and support at the High Flux Reactor by the I.L.L., Grenoble, is also gratefully acknowledged. References 1 See, e.g., J. Phys. (Paris), Colloq. C7, 1984, 9. 2 C. A. Angell, Annu. Rev. Phys. Chem., 1983,34, 593. 3 F. Franks, in Water, A Comprehensive Treatise, ed. F. Franks (Plenum Press, London, New York, 4 J. Teixeira, M-C. Bellissent-Funel, S. H. Chen and A. J. Dianoux, J. Phys (Paris), 1984, C7, 65. 5 J. Clifford, in Water, A Comprehensive Treatise, ed. F. Franks (Plenum Press, London, New York, 6 J. Conard, H. Estrade-Szwarkopf, C. Poinsignon and A. J. Dianoux, J. Phys. (Paris), 1984,45, 169. 1982), vol. 7, p. 215. 1975), vol. 5, p. 75.Supercooled Water in Silica Gels 3459 7 J. W. Clark, P. G. Hall, A. J. Pidduck and C. J. Wright, J. Chem. Soc., Faraday Trans. I , 1985, 81, 8 D. C. Steytler, J. C. Dore and C. J. Wright, J. Phys. Chem., 1983,87, 2458. 9 J. D. F. Ramsay, H. J. Lauter and J. Tompkinson, J. Phys. (Paris), 1984, C7, 73. 10 J. D. F. Ramsay and B. 0. Booth, J. Chem. Soc., Farahy Trans. I , 1983,79, 173. 11 J. D. F. Ramsay, R. G. Avery and L. Benest, Faraday Discuss. Chem. Soc., 1983,76, 52. 12 B. Maier, Neutron Facilities at the I.L.L. High Flux Reactor (I.L.L., Grenoble, 1983). 13 P. A. Egelstaff, An Introduction to the Liquid State (Academic Press, London, 1967). 14 T. Springer, Quasielastic Neutron Scattering for the Investigation of Diffwive Motions in Solids and 15 L. van Hove, Phys. Rev., 1954,95,249. 16 V. F. Sears, Can. J. Phys., 1966, 44,1279. 17 M. Bk, I.L.L. Technical Report 84BE05T (1984). 18 M. W. Johnson, Atomic Energy Res. Estab. Rep. (AERE-R7682, 1974). 19 C. Poinsignon, J. Conard, H. Estrade-Szwarkopf and A. J. Dianoux, Proc. Int. Clay Con$ 1985, ed. 20 D. Eisenberg and W. Kauzmann, The Structure and Properties of Water (Oxford Univ. Press, London, 21 G. E. Walrafen, J. Chem. Phys., 1964,40, 3249. 22 M. Moskovits and K. H. Michaelian, J. Chem. Phys., 1978,69,2306. 23 M . P. Fontana, G. Maisano, P. Migliardo, M-C. Bellisent-Funel and A. J. Dianoux, J. Phys. (Paris), 24 M. H. Brooker and M. Perrot, J. Chem. Phys., 1981,74, 2795. 25 S. P. Tay, D. D. Klug and E. Whalley, J. Chem. Phys., 1985,83,2708. 26 J . R. Scherer and R. G. Snyder, J. Chem. Phys., 1977,67,4794. 2067. Liquids; Springer Tracts in Modern Physics (Springer-Verlag, Berlin, Heidelberg, 1972), vol. 64. H. van Olphen (Elsevier, Amsterdam, in press). 1969). 1984, C7, 151. Paper 6/ 194; Receiued 27th January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203447

出版商:RSC    

年代:1986

数据来源: RSC