摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 1815-1821 Electron Spin Resonance Characterization of Rotational Isomers of the n-Butane Radical Cation with Partially Deuterated Methyl Groups in Some Halogenated Matrices Mikael Lindgren" and Anders Lund Department of Physics and Measurement Technology, Linkoping University, S-58183 Linkoping, Sweden Rotational isomers of n-butane cations with partly deuterated methyl groups have been studied in CH,CCl,, CF,ClCFCl, and CFCl, matrices at 4.2 K and temperatures above 77 K. It is shown that the abundance ratio between different isomers depends on the temperature and the matrix. Temperature- induced oscillations of the partially deuterated methyl groups decrease the pertinent proton hyperfine splittings. The experimental data for the spin density of the n-butane radical cation are in excellent agreement with INDO calculations based on optimized geometries obtained from MNDO and ab initio calculations.Since the beginning of this decade, saturated halogenocarbons have been shown to be suitable for use as matrices to stabilize radical cations in e.s.r. and optical studies. The high (ca. 11 eV) ionization potential, together with the anisotropic character of the e.s.r. signal, make it possible to record rather detailed e.s.r. spectra from radical cations generated by radiation of the frozen matrix-solute. Together with the 2H-labelling technique, this can give detailed information regarding the spin density and the configuration of the orbital occupied by the unpaired electron (SOMO) and hence the ionic structure.Interestingly, the cations, when trapped in frozen halogenated matrices, decompose by mechanisms which depend on the matrix used. 1,1,2-trichlorotrifluoroethane (CF2C1CFC12) is of glassy character, and selective H-atom abstraction of n-butane cations to form 2-butyl radicals is observed when warming the sample to ca. 100 K.l A similar decomposition scheme is found with other alkane cations, such as propane2 or n-~entane,~ where 2-alkyl radicals are formed at temperatures above 100 K. With cations of cycloalkanes similar reactions take placeY4 and ring opening has been suggested in some cases.5 Polycrystalline matrices such as e.g. l , l , 1 -trichlorotrifluoroethane (CF,CCl,) and fluorotrichloromethane (CFCl,) generally give narrower e.s.r.lines, and more detailed structures can be resolved.l9 6-9 However, hyperfine interaction with the matrix (fluoro- trichloromethane) has been reported in several cases.6j lo+ l1 Also the stability of the cations increases and photolytically induced decomposition to unsaturated alkene cations by selective loss of hydrogen from the saturated alkane cations has been reported.l? 9 9 l2 The better stability also makes it possible to study the dynamic properties, such as thermally induced vibrations and oscillations in the cations. Ring-puckering motion in cycloalkanes and other dynamical effects have been observed and discussed in several papers.12-14 In a previous paper,8 spin densities and the corresponding hyperfine coupling constants have been calculated for some n-alkane radical cations with the INDO method.Other authors have used MNDOlj or ub initiol6 calculations to study the structural changes which accompany the ionization. Their results show that ionization causes considerable 18151816 E.S.R. of n-Butane Radical Cation changes in the bond lengths compared to the diamagnetic molecules. As we believed that distortions upon ionization can give spin densities different from those reported previously,* the spin-density calculations are here refined for the n-butane cation, taking into account the optimized geometries. The results are compared with those obtained from experiments. In a recent study1 we reported that rotational isomers were observed among radical cations of partly deuterated 1,l ,4,4[2H,Jn-butane and 1 ,4[2H2]n-butane in CF2C1CFl, and CF,CCl,.However, their dynamic properties and interaction with the surrounding matrix were not clarified because of the rather narrow temperature interval in which the measurements were performed. In this study we have extended the temperature range by including some experiments at 4.2 K. Another matrix, CFCl,, has been used in the identification of the rotational isomers of the n-butane radical cation. Experimental The synthesis and sample preparation of the partially deuterated n-butane have been described previously.' E.s.r. spectra were recorded using a Varian E9 spectrometer with a Bruker ER 035 n.m.r. gaussmeter and an EIP 548 A microwave frequency counter to determine the hyperfine splittings and g-factors. The equipment used for X-irradiation and measurements at 4.2 K have been described e1~ewhere.l~ The temperature above 77 K was measured with a chromel-alumel thermocouple.Rotational Isomers As mentioned in the introduction the rotational isomers were identified in a study of 2H-labelled n-butane cations. It was shown that the SOMO of the unpaired electron was confined through the carbon skeleton including two protons, one on each of the terminal methyl groups, giving rise to the largest coupling constant of ca. 60 G. A schematic picture of the n-butane molecule is shown below together with the hyperfine splittings (in G).? In the case of 2H-labelling the corresponding hyperfine couplings are decreased by a factor of 6.5 according to the smaller nuclear magnetic moment.7.5 10.4 n 5 8 When the n-butane molecule was partially deuterated at the methyl groups we observed three different spectra overlaid. Computer simulations showed that three isomers could be identified where the main difference was whether the SOMO extended to a terminating proton or deuteron at the methyl groups. It was concluded that the molecule was tightly locked between the Cl-C2 and C3-C4 carbons to prevent rotation of the methyl groups. The spin densities and molecular structure will be further described in the theoretical section. The rotational isomers can then be characterized by these terminating protons/deuterons or a mix to give a specific e.s.r. spectrum: (I) H * . - H , two identical protons with aH z 60 G; (11) H D, one proton with aH = 60 G and one deuteron with a, z 9 G; (111) D .-.D, two identical deuterons with a, z 9 G. f 10G=lmT.M. Lindgren and A . Lund 1817 Experimental Results CF,CCI, As shown in fig. 1 the 1,l ,4,4[2H4]n-butane radical cation gives in this matrix a very similar spectrum at 4.2 and 108 K, respectively. In fig. 2 a spectrum due to the 1,4[2H2]n-butane radical cation, recorded at 96 K, is shown. All spectra due to the cations were reversible with temperature to ca. 135 K. All smaller hyperfine splittings are as described in the previous section with a proper adjustment for the ,H-labeHing of the methyl protons. However, the larger coupling due to the in-plane protons differed slightly with temperature and 2H-labelling (table 1). E.g. in the 1,l ,4,4[2H4]n-butane cation the H - - - H isomer had a coupling constant of 61.0 G at 4.2 K, 61.2 G at 77 K and 59.7 G at 122 K.The H D isomer showed a similar behaviour and the coupling constants were 60.8, 61.2 and 58.8 G at 4.2, 77 and 122 K, respectively. In the l,4[,H2]n-butane cation this effect could also be observed for the H H isomer, but unfortunately the low yield of the H -.- D isomer made the hyperfine splitting measurement unreliable in this case. It is also interesting to compare the relative intensities for the different isomers. In the 1,1 ,4,4[2H4]n-butane cation the abundance ratio for the rotational isomers I : I1 ; 111 is ca. 3 : 5: 2,l while for the 1,4[2H2]n-butane cation it is ca. 6 : 3 : 1 from comparisons with computer simulations. It should be mentioned that the contribution from isomer I11 is only approximate, since the centre lines of the spectra are affected by quartz signals.CF,CICFCI, In this matrix the e.s.r. spectra are different, depending on irradiation and recording temperature (fig. 3). At 4.2 K the relative intensity for the H ... H isomer is evidently larger than that for the H .-- D isomer. The amount of the D D isomer could not be estimated at 4.2 K because of the quartz signal, which gave a large contribution in the centre of the spectrum. At 4.2 K the abundance ratio is ca. 7: 2 between the H ... H and the H -.. D isomers, and at 77 K it is 4: 2, neglecting the small difference in linewidths. CFCI, An e.s.r. spectrum for the l,4[,H2]n-butane radical cation, recorded at 77 K, is shown in fig. 4. The lines are broader than those observed with the CF,CCl, matrix and comparable to those obtained with CF,ClCFCl,.One can easily identify the two isomers H D. The spectral structure was reversible up to ca. 145 K, but with an irreversible decrease of the total intensity with increasing temperature. This indicates a decomposition of the cation to a diamagnetic product. No other signal appeared below 145 K as observed in CF,ClCFCl,, where the 2-butyl radical is formed at 100 K.l H and H Theoretical Results and Discussion The values of the geometrical parameters for the n-butane cation are given by Bellville and Bauld15 and Bouma et a1.16 from MNDO and ab initio optimization, respectively. An extended carbon chain was assumed to have the lowest total energy, and the distortion upon ionization is mainly a lengthening of the C2-C3 bond.The unpaired electron occupies a 7a, orbital (C2h symmetry) through the carbon atom framework and two in-plane protons at each methyl group. The geometries were used as input in the INDO program and the calculated spin densities and hyperfine splittings are given in table 2, together with experimental values. The agreement is very good for the protons in the molecular plane, but the smaller splittings obtained in the CF,CCl, matrix imply a further distortion from the C,, symmetry,1818 E.S.R. of n-Butane Radical Cation 100 G Fig. 1. E.s.r. spectra of the 1,1,4,4[2H,]n-butane radical cation produced and recorded in the CF,CCl, matrix at two different temperatures. The spectra are centred on g = 2.0037. 100 G Fig.2. E.s.r. spectrum of the 1,4[2H2]n-butane radical cation recorded in the CF,CCl, matrix at 96 K. The vertical bar indicates a quartz signal at g = 2.0008. Table 1. The temperature dependence for the main hyperfine splitting for the rotational isomers H * * * H and H . * - D of the 1 ,I ,4,4[2H,]n-butane cation in the CF,CCl, matrix T/K isomer splitting/G isomer splitting/G 4.2 H * - - H 61 .O H * - - D 60.8 77 H * . * H 61.2 H D 61.2 122 Ha*- H 59.7 H 1 . . D 58.8 To obtain values according to experiments, the ion was distorted by rotations around the C2-C3 bond. For each twisted structure the bond lengths were held fixed at values given by ab initio calculations, while the bond and twist angles for the out-of-plane protons were optimized by the MNDO method.The best agreement with experimental spin density was obtained when the angle between the local planes Hl-Cl-C2-C3 and C2-C3-C4-H4 was 50" (steps of 10") from the extended form. The hyperfine splittingsM . Lindgren and A . Lund 1819 100 G H....H I I U I v Fig. 3. E.s.r. spectra of the 1,1,4,4[2H,Jn-butane radical cation produced and recorded 77 K in the CF,ClCFCl, matrix. Below the spectra, stick-plot diagrams are shown for rotational isomers. at 4.2 and two of the Fig. 4. E.s.r. spectrum of the radical cation of 1,1,4,4[2H,]n-butane recorded in the CFCl, matrix at 77 K. The vertical bar indicates a quartz signal at g = 2.0008. Table 2. INDO calculated and experimental hyperfine splittings for the n-butane radical cation in CF3CCl, (optimized geometries were obtained from ab initio and MNDO results) hyperfine splittings/G atom MNDOa ab initiob experimental twisted s tructureC - Cl - 7.2 - 6.5 c2 4.6 7.3 - methyl protons HI, 66.7 61.1 61 HI2 9.0 8.7 10.4 HI3 9.0 8.7 5 H21 H2 2 4.7 0.0 methylene protons 4.7 0.0 7.5 - - 7.2 7.3 61.8 10.6 6.9 - 1.5 1.1 Taken from ref.(1 5). Taken from ref. (1 6). The twist angle is 50" (from the extended form).1820 E.S.R. of n-Butane Radical Cation are given in table 2. We also observed that the splittings due to the H1 and H4 protons remained within 60-62 G for all twist angles. Interestingly, both the MNDO and INDO calculations gave an energy minimum at ca. 90" in this approach. From MNDO calculations the total energy was 1 kcal mo1-1 above the energy of the extended structure.This indicates a possible stabilization of a twisted conformer at low temperature, especially if one takes into account the influence of the matrix. It is interesting to compare the structure corresponding to this second energy minimum with the diamagnetic molecule. In that case a gauche confomer exists18 with a slightly higher energy than for the extended structure. The twist angle would in our notation then be 120". More detailed investigations have to be undertaken to elucidate this kind of conformation in ionized molecules also. We regard our proposed structure (50" twist angle) as tentative, according to the rather simplified theoretical approach. However, it should be added that this conformation is suitable to explain the H, elimination reaction that occurs under illumination to produce the but-2-ene radical cation.l? Two adjacent methylene protons are as close as possible when the twist angle is ca.60". The distortion of the Czh symmetry as observed in the CF,CCl, matrix, the calculated spin density for the twisted structure and the second energy minimum near this conformation, support the suggestion of a twisted structure. It has been reported that in CFCl3lg and CF,CCl,9 matrices the hyperfine splitting for the in-plane protons is 75 G at 150 and 135 K, respectively. Together with the reported INDO recalculation,8 which predicted a splitting of 68 G for the extended structure, we suggested1 that the n-butane cation had a somewhat twisted structure at low temperature to give a smaller splitting, and that the values reported at the higher temperature were caused by thermally induced oscillations around the C2-C3 bond. The INDO results from the ab initio geometry show that this need not be the case, since the hyperfine coupling is almost independent of twist angle.Another explanation in terms of a thermally induced reorientation of the methyl group has been proposed.21 As observed in table 1, the main splitting measured at 4.2 K remained unchanged (or increased slightly) at 77 K and then decreased when the temperature was further raised. The same hyperfine splitting remained constant in n-butane and 2,2,3,3[2H,Jn-butane between 77 and 125 K. Similar behaviour is observed in the partly deuterated ethyl radical,20 where the P-proton hyperfine interaction decreases with increasing temperature.The proton coupling constant, given by the cos2 rule, is reduced by quantum-mechanical averaging through thermally induced rotations around the equilibrium conformation. The hyperfine coupling of a deuteron in a partially deuterated methyl group would increase with temperature.20 An estimation of the energy barrier to internal rotation of the partly deuterated methyl or ethyl groups in the n-butane cation is, however, not yet possible. For the protiated cation the measured value is 2.3-2.4 kcal mo1-1.21 Another interesting feature of the rotational isomers is the abundance ratio at different temperatures. In CF,CCl, there is scarcely no change between 4.2 and 108 K, while in CF,ClCFCl, the H - - . D isomer contributes less at lower temperature.It is, however, not clear whether the difference in abundance ratios at 4.2 K reflects the population difference among the rotational isomers in the cation or that the neutral molecule is either preferentially in the H H form and/or is preferentially ionized in this form. One can hypothesize that in the diamagnetic molecule all forms of combinations H H, H --. D and D - - . D are possible. Since the bond energy is lower for a C-H than for a C-D fragment this may explain the fact that the H H isomer is over-represented while the D D isomer is less probable in all cases. This makes sense, regarding the abundance ratios for the 2H2 and 2H, cases in the CF,CCl, matrix.M . Lindgren and A . Lund 1821 We are indebted to Dr Kjeld Rasmussen (Technical University of Denmark, Lyngby) and Dr Harald Mollendal (University of Oslo) for valuable discussions during Nordiska Molekylfysikkdager, Vettre, Asker, Norway in May 1986, where some parts of this work were presented. References 1 M.Lindgren, A. Lund and G. Dolivo, Chem. Phys., 1985,99, 103. 2 K. Toriyama, K. Nunome and M. Iwasaki, J. Chem. Phys., 1982,77, 5891. 3 A. Lund, M. Lindgren, G. Dolivo and M. Tabata, Radiat. Phys. Chem., 1985, 26, 491. 4 M. Iwasaki, K. Toriyama and K. Nunome, Faraday Discuss. Chem. Soc., 1984, 78, 19. 5 X-Z. Qin and Ff. Williams, Chem. Phys. Lett., 1984, 112, 79. 6 B. Becker, K. Plante and M. D. Sevilla, J. Phys. Chem., 1983, 87, 1648. 7 J. T. Wang and Ff. Williams, J . Phys. Chem., 1980,64, 3156. 8 K. Toriyama, K. Nunome and M. Iwasaki, J. Phys. Chem., 1981,85,2152. 9 M. Tabata and A. Lund, Radiat. Phys. Chem., 1984,23, 545. 10 M. C. R. Symons and P. J. Boon, Chem. Phys. Lett., 1983, 100,203. 11 L. D. Snow and Ff. Williams, Chem. Phys. Lett., 1983, 100, 198. 12 M. Tabata and A. Lund, Chem. Phys., 1983,75, 379. 13 K. Toriyama, K. Nunome and M. Iwasaki, J . Chem. Soc., Chem. Commun., 1984, 143. 14 K. Ohta, H. Nakatsuji, H. Kubodera and T. Shida, Chem. Phys., 1983, 76, 271. 15 D. J. Bellville and N. L. Bauld, J. Am. Chem. SOC., 1982, 104, 5700. 16 W. J. Bouma, D. Poppinger and L. Radom, Israel J. Chem., 1983, 23, 21. 17 M. Ogasawara, M. Lindgren, A. Lund and G. Nilsson, Chem. Phys. Lett., 1985, 117, 254. 18 K. Rasmussen, Lectures Notes in Chemistry, uol. 37 (Springer-Verlag, Berlin, 1985). 19 J. T. Wang and Ff. Williams, Chem. Phys. Lett., 1981, 82, 177. 20 M. J. Ramos, D. McKenna, R. C. Webster and E. Roduner, J. Chem. SOC., Faraday Trans. I, 1984, 21 M. Iwasaki and K. Toriyama, J. Am. Chem. SOC., 1986, 108, 6441. 80, 255. Paper 6/ 1737; Received 27th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301815

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1987, 83, 1823-1833 Oxygen Reduction with Hydroxy- 1,4-naphthoquinones immobilized at Carbon Electrodes Tsutomu Nagaoka," Toshiaki Sakai, Kotaro Ogura and Takashi Yoshino Department of Applied Chemistry, Yamaguchi University, Tokiwadai 255 7, 755 Ube, Japan Oxygen reduction with naphthazarin (5,8-dihydroxy-l,6naphthoquinone) has been studied by cyclic voltammetry and by flow-through electrolysis. The mechanisms of oxygen reduction were studied thermodynamically. At pH 7 the semiquinone anion is the most likely candidate for oxygen reduction among the reduced species of naphthazarin, including quinols, and the superoxide is mainly produced by t h s species. By flow-through electrolysis e.s.r. the semiquinone has been detected under anaerobic con- ditions.Its concentration depends on the potential because of dispropor- tionation. Distribution of the semiquinone has been calculated against the potential, and the results are compared with the potential dependence of the e.s.r. intensity. The redox activity of other quinones has also been studied. Naphthazarin is the most active in oxygen reduction. Dioxygen reduction with semiquinones is of special interest in connection with the drug action of anthracycline antitumour antibiotics such as adriamy~in.l-~ The semiquinones of such anthracyclines transfer an electron to molecular oxygen, leading to the formation of superoxide, from which hydrogen peroxide or the hydroxyl radical arise. These oxygen intermediates may degrade DNA and be involved in the cardiotoxicity of the anthracyclines.Many authors have discussed the redox mechanisms of these quinones electrochemically,10-18 but there have been few electrochemical studies of the direct observation of oxygen reduction with them. Here we study naphthazarin (NH,; H represents a hydroxyl proton) in this way as a model anthracycline. Anne and Moiroux17 discussed the redox mechanisms of hydroxyanthraquinones in N,N-dimethylformamide (DMF) in the presence of oxygen, and reported that electron transfer was accompanied by proton transfer between the hydroxy group and the superoxide, yielding hydrogen superoxide. Keita and Nadjo studied the catalytic synthesis of hydrogen peroxide by an anthraquinone derivative.19 Degrand observed oxygen reductionlg with an anthraquinone polymer immobilized at carbon electrodes, and suggested that electron transfer from an anionic quinol may not be the rate- determining step.20 Land et at.reported the redox potentials of naphthazarin and adriamycin at several different pH values by pulse-radiolysis te~hniques.~l-,~ Experiment a1 Reagents NH, was prepared as described elsewhere,24 and recrystallized twice from ethanol. Catechol was sublimed in vacuum at 100 "C before use. The other chemicals used were of analytical reagent grade. All experiments were done in the buffer described previo~sly.~~ 18231824 Oxygen Reduction with Naphthazarin Fig. 1. Cross-sectional diagram of the cell for flow-through electrolysis: A, Pyrex glass tubing, 0.d. 12 mm, 5 cm long; B, silicone rubber plug; C, counter-electrode; D, Nafion tubing, 0.d.3 mm; E, Pyrex tubing, 0.d. 3 mm; F, silicone rubber ring; G, glassy carbon fibres; H, inlet for sample solution; I, reference electrode; J, lead of the working electrode. Cyclic Voltammetry Voltammetric experiments were done with a laboratory-made potentiostat and a Toho-Giken model 2230 potential scanner at 25 "C. Working electrodes were short rods 3 mm in diameter (grade GC-20, Tokai) epoxied into Pyrex glass tubing or press-fitted into Teflon housing. The reference electrode used was Ag/AgCl (saturated KCl), but all the potentials reported here refer to the normal hydrogen electrode (NHE). The working electrodes were hand-polished using Sic emery paper and 10 pm alumina for 1 min. After being polished, the electrodes were sonicated for 5 min.Most quinones investigated were sparingly soluble in aqueous solution, so they adsorbed on to the surface of the carbon electrodes. To measure the redox activity of the quinones, given amounts (10-50 nm3) of a 5 mmol dm-3 DMF solution of a quinone were added to 5 cm3 of buffer in an electrolysis cell. The solution was stirred by N, or 0, gas bubbling through, and the quinones adsorbed on to the polished surface. When redox cycling between oxygen and quinones was studied, oxygen was saturated at 1 atm* (1.3 mmol dm-3). Flow-through Electrolysis Fig. 1 shows a cross-sectional diagram of the cell for flow-through electrolysis. The main body of the cell consisted of Pyrex glass tubing (0.d. 12 mm) and two silicone rubber plugs. Nafion tubing (0.d.3 mm) was used as a separator between the working and counter-electrode compartments. The working electrode was a bundle of glassy carbon fibres (Tokai) tied with a 0.2 mm Au wire, a lead of the working electrode. The lead was taken out from the flow system through a small hole in other Pyrex glass tubing (0.d. 3 mm), and the hole was sealed with an epoxy resin. The counter-electrode was a helix of a 0.5 mm Pt wire located coaxially to the working electrode. Sample solution flowed into the cell from an inlet and was electrolysed completely before it reached the outlet. Flow rates were regulated with an Atto model SJ-1211 pump. When redox cycling was studied, carrier solutions containing both NH, and oxygen were used. * 1 atm = 101 325 Pa.T. Nagaoka, T.Sakai, K. Ugura and T. Yoshino 1825 0 20 40 60 rl mol cm-2 / 'J I I I I I I 1 I I I J 0.5 0.0 -0.5 E/V us. NHE Fig. 2. Oxygen reduction mediated by naphthazarin confined to the surface of a glassy carbon electrode, and dependence of the mediated current on the surface concentration of the quinone, r. pH 7.0, 25 "C. (A) Cyclic voltammograms at the electrodes to which naphthazarin adsorbed in the presence (a) and absence (b) of oxygen, and at a bare carbon electrode in the presence of oxygen (c). (B) Plots of the current of the catalysed peak us. r in the presence of oxygen (O), and of the cathodic peak current us. r in the absence of oxygen (e). Flow-through Electrolysis E.S.R. E.S.R. spectra were recorded with a Jeol model JES-ME-1 X spectrometer. Semiquinone radicals were generated with an ex-situ flow-through electrolysis e.s.r.cell described previously.26 Within 1 s of being electrolysed quantitatively at the generator (working electrode), the sample solution flowed into the cavity. Results Cyclic Voltammetry Fig. 2 shows cyclic voltammograms of the electrodes onto which NH, adsorbed. In the absence of oxygen, one pair of adsorption peaks was observed. The peaks shifted with pH by ca. -60 mV per decade, which is characteristic of a quinone/hydroquinone (Q/QH,) reaction. The inset shows a curve of the peak current for oxygen catalysis, i,, vs. the surface concentration, r, of NH,, which was estimated from the area of the cathodic peak in the absence of oxygen. The current was the sum of current at the modified and bare surfaces, the latter being present at small coverage of NH,.Since the coverage of the modified surface increased with r at the expense of that of the bare surface, the (total) current in the inset started from a non-zero value, and increased with a relatively small gradient at small r values. At the r value of 2 x mol ern-,, the ip us. r curve levelled, showing that oxygen reduction obeyed diffusion-controlled kine tics. Fig. 3 shows the pH dependence of the catalysed current for NH, and juglone (5-hydroxy-1,4-naphthoquinone). At pH < 3, redox cycling between NH, and oxygen was not detected. At pH > 7, the amount of adsorbed NH, decreased. The first ionization constant of NH, is 7.85,,l so this decrease probably arises from desorption of the anionic1826 20 --- s +a lo 0 Oxygen Reduction with Naphthazarin 1 3 5 7 9 1 1 PH PH Fig.3. Deactivation of redox cycling at electrodes modified by naphthazarin (A) and juglone (B). Current of the oxygen peak catalysed (D) and cathodic current of quinone/hydroquinone peaks (0) in the absence of oxygen. 0 1 2 3 4 5 6 7 8 r/ 1 o - ~ mol cm-2 Fig. 4. Comparison of catalytic activity at electrodes modified by quinones at pH 7.0: A, naphthazarin; fly 9,lO-phenanthrenequinone; 0, juglone, 0, 1,2-naphthoquinone. 0.2 0.0 -0.2 - 0.4 -0.6 E/V vs. NHE Fig. 5. Apparent number, nap, of electrons transferred from adsorbed naphthazarin to oxygen at pH 7.0. Concentrations of naphthazarin in carrier solutions are in mmol dm-3: (-) 0.1, (- - - - -) 0.05, (---) 0.01 and (----) 0.0. form, NH-.Although the NH, peaks became very small at pH > 7, they were still detected. A similar decrease was also observed for juglone (JL), but not for 9,lO- phenanthrenequinone (PQ). In fig. 4, the ip vs. r plots are shown for the quinones investigated. The slope of the curve for NH, was the steepest, showing that the catalytic activity is the greatest among these quinones. 1,2-Naphthoquinone (NP) showed the smallest activity, and catalytic current was very slight.T. Nagaoka, T. Sakai, K. Ogura and T. Yoshino I 0.2 0.0 -0.2 -0.4 E/V us. NHE Fig. 6. Potential dependence of e.s.r. intensity in the absence of oxygen at pH 8.0. 1827 -0.6 I \. 0 2 4 6 8 1 0 1 2 1 4 PH Fig. 7. Potential us. pH of naphthazarin and oxygen: (a) E2, (b) E, and ( c ) Eox. Flow-through Electrolysis By examining ferrocyanide reduction with the cell shown in fig.1, we found electrolysis to be quantitative (> 95% efficiency) at flow rates up to 4 cm3 min-l, and the number of electrons transferred could be evaluated directly from the limiting current up to this flow rate. On the assumption that electrolysis is quantitative, the apparent number of electrons transferred from NH, to oxygen, nap, was evaluated as a function of the potential (fig. 5). Fig. 5 clearly shows that the overall reaction is a two-electron process. Oxygen reduction at the bare surface was more irreversible in flow-through electrolysis than in cyclic voltammetry (see fig. 2 and 5). This difference was presumably due to the difference in treatment of the carbon electrodes.It is practically impossible to polish carbon fibres with alumina; the fibres were treated with 0.1 mol dm-3 NaOH and HCl only before the experiments. Flow-through Electrolysis E.S.R. When a 1 .O mmol dmW3 aqueous solution of NH, was reduced at - 0.1 V us. NHE and at pH 8.0 in anaerobic conditions, sharp e.s.r. spectra assigned to the semiquinone anion NHL- were obtained. The semiquinone was fairly stable in anaerobic conditions as reported by Dodd and Mukherjee.’ E.s.r. intensities in the steady state were monitored at given potentials (fig. 6). The semiquinone concentration was very small at potentials more negative than -0.3 V, because of disproportionation of the semiquinone.1828 Oxygen Reduction with Naphthazarin 0 2 4 6 8 1 0 1 2 1 4 PH 0 2 4 6 8 1 0 1 2 1 4 PH 0 2 4 6 8 1 0 1 2 1 4 PH Fig.8. Families of distribution curves for the naphthazarin quinols (A), semiquinones (B) and parents (C). Equilibrium of NH, and its Reduced Species in Aqueous Solution For NH,, one-electron redox potentials of the couples, quinone/semiquinone (El) and semiquinone/quinol (E2), were evaluated from the pK, and potential values reported by Land et aZ.,21 as shown in fig. 7. E, is the redox potential when the total concentrations of the parents and semiquinones are equal, and E, is the redox potential when the total concentrations of the semiquinones and quinols are equal. Therefore, these potentials can be expressed as follows: El = E"(NH,/NH;-) - (RT/F) In (KkX,/X,) E, = E"(NH',-/NHi-) - (RT/F) In (KFQK,"Q/Kp) - (RT/F) In (X1/X2) where E"(NH,/NH;-) and E"(NH;-/NHg-) are the standard redox potentials of the NH,/NH;- and NH;-/NHi- couples.X,, X2 and X3 are defined by the following equations : Xo = [H+I2 + Kl[H+] + KIK,T. Nagaoka, T. Sakai, K . Ogura and T. Yoshino 1829 0.0 0.1 0.0 -0.1 -0.2 EIV us. NHE Fig. 9. Potential dependence of distribution of (---) NH,, (-.-.-) NH- 3 3 (------) NHi- and (-) NHH- at pH 7. In the calculation the total concentrations of the parents, semiquinones and quinols were kept constant. where K , KR and KHQ are the acid dissociation constants of the parents, the semiquinones and the quinols, respectively. Kl = [NH-] [H+]/[NH,] ; KF = [NHi-] [H’]/[NHJ; KF = [Nm3-] [H+]/[NH’2-]; K , = [N2-] [H+]/[NH-]; KF = [NHo2-] [H+]/[NH;-]; KBQ = [NH,] [H+]/[NH,]; KFQ = [NH3-] [H+]/[NH;-]; KFQ = [NHi-] [H+]/[NH,]; KFQ = [N4-] [H+]/[NH3-].Fig. 8 shows families of the distribution curves of the parents, semiquinones and quinols calculated from the pK, values. Fig. 9 shows the potential dependence of the equilibrium concentrations of the semiquinone and quinols calculated by the following Nernst equations : E = El -(RT/F) In [NS]/[NZ] E = E2 - (RT/F) In [NL]/[NS] where [NZ], [NS] and [NL] are the total concentrations of the quinones, semiquinones and quinols, respectively. In the calculation, the sum of the total concentrations of the quinones, semiquinones and quinols was assumed to be constant. The semiquinones arise from disproportionation as follows : quinone + quinol c semiquinone. K, = [NSI2/[NZ] [NL] = exp [F/RT(E, - E2)].(1) (2) The apparent formation constant of the semiquinones is defined by Discussion Many workers have found equilibria between semiquinones and oxygen using pulse- radiolysis 27-30 and between quinol and oxygen using e.s.r. 32 Land et al. mentioned that the quinol of naphthazarin could reduce oxygen.22 Accord-1830 Oxygen Reduction with Naphthazarin ingly, one must consider the semiquinones and quinols as possible candidates for oxygen reduction. In neutral solution, the following equilibria should be considered (see fig. 8): (3) (4) 0, + NH;- e 0;- + NH, 0, + NH, + 0;- + NHi+ 0, + NH; + 0;- + NH; 0, + NHi- e 0;- + NH;- (6) where NHH- is the semiquinone anion, NHj the neutral semiquinone, NHi+ the cationic semiquinone, NH, the neutral quinol, NH; the anionic quinol and NHi- the quinol dianion. Using the data reported by Land et a1.213 22 we evaluated the standard redox potentials, E", of the couples, NHJNH;, NHJNHi- and NH,/NH;- to be 0.197, - 0.157 and - 0.104 V, respectively.The E" of the NH,/NHi+ couple was estimated to be > 0.486 V. The superoxide generated is further reduced to hydrogen peroxide by receiving an electron from the reduced species of the quinone or the superoxide itself (disproportionation of super~xide~~? 34) : (7) and the overall reaction is eventually a two-electron process. In fig. 7, the pH dependence of El and E, is shown, together with that of the one-electron redox potential of oxygen, Eox. The levels in the E, and E, curves corresponded to the E" of the couples, NH,/NHi- and NH;-/NHt-.Even if EOx is more negative than El and E,, oxygen can be reduced to peroxide by the working of the irreversible reaction, (7). The equilibrium constant, K,, of eqn (n)[n = (3)-(6)] is given by 0;- + 2H+ + e -+ H,O, K, = exp [F/RT(E,", - EG)] where EG is a standard redox potential of the quinone couple involved in equilibrium. EEx is -0.155 V, where the concentration of oxygen is taken as 1 mol dm-3.27-299 35 The values of K3, K4, K5 and K, were calculated to be 0.14, < 1.4 x lo-", 1.1 x and 1.1, respectively. As seen in fig. 9, the concentrations of NH, and NH; were calculated to be larger than the concentrations of the other candidates. However, the E" values of the NHi+/NH, and NHJNH; couples were very positive, making K4 and K, negligibly small. Accordingly, one-electron reduction with NH, and with NH; is not thermodynamically favoured, and superoxide production would be negligible through these quinols.Fully protonated quinols (QH,) transfer electrons only at negligible rates in other electron- exchange systems between quinol and cytochrome c and between quinol and q ~ i n o n e . ~ " ~ ~ Since K, is large enough in neutral pH, as can be seen from eqn (2) and fig. 7, it is likely that the pathway of eqn (3) is mainly operative for oxygen reduction at pH 7. At this pH, reaction with the quinol dianion, NHi-, seems not to be important, while K6 is ca. 8 times as large as K3. This is because the calculated concentrations of NHi- were two orders of magnitude smaller than those of NH;- (fig. 9). At pH 8 the dianion route may contribute more to the formation of superoxide than at pH 7, because the dianion concentration becomes higher, as is shown in fig.8(a). However, Ks has a maximum at ca. pH 8, so the semiquinone route is still important. At higher pH, the dianion route would predominate, because the K, value decreases with pH, and because the dianion concentration increases with it. We concluded that superoxide is mainly produced by NHi- at pH 7. Deactivation of redox cycling at low pH, as seen in fig. 3, seems to arise from depletion of the semiquinones, because the Ks calculated using eqn (2) is very small there. In fig. 2 the difference in current between the curves (a) and (c) was substantial even at 0.1 V. At this potential distribution of NHi- (the ratio of the equilibrium concentration ofT.Nagaoka, T. Sakai, K . Ogura and T. Yoshino 1831 Table 1. Redox potentials of quinones and formation constants of superoxide at 25 "Ca E : b ELC compound /V us. NHE /V us. NHE Kd PKae PHRf 1.5 x 10-l 2.78 - 3.5 5.6 juglone 5.0 1,2-benzoquinone 0.210h 0.370h 6 . 7 ~ 5.0h - (0.224) 0.417 naphthazarin -0.107' - 0.062' -0.101 - - (- 0.3 15) - - -0.015 - 9,lO-phenanthrenequinone (-0.182) 0.027 (2.86) - 1,2-naphthoquinone (- 0.039) 0.164 (1.1 x lop2) 4.8h - - - - a Values in parentheses are based on estimates from eqn (9). of quinones at pH 7. of superoxide defined by exp [F(F,, - q ) / R T ] . redox cycling is no longer observed. ref. (40). First one-electron redox potentials Formation constants pK, of semiquinones. f pH values at which Data taken from Two-electron redox potentials of quinones at pH 7.Data taken from ref. (21) and (22). NHi- to the total concentration of quinoic and quinolic species) was evaluated to be 3 x which was equal to the maximum value of distribution at pH 4.5. Therefore, redox cycling should be detected at pH > 4.5, if cycling is initiated by NHB-, as has been discussed so far. This pH roughly agrees with the pH for deactivation shown in fig. 3 (a). Degrand found similar deactivation at carbon electrodes coated with a film of an anthraquinone polymer at low pH.,O Electron transfer from the anionic quinol (QH-) might be possible, since the E" of the QH'/QH- couple may be more negative than the E" of the NHJNH; couple. He suggested from analysis of electrode kinetics that electron-transfer from QH- is not the rate-determining step.Semiquinone formation was confirmed by e.s.r. (fig. 6). At pH 8 the calculated equilibrium concentration of NHi- has a maximum at ca. - 0.12 V, which is ca. 3 times larger than that at pH 7. At pH 8 a maximum in the e.s.r. intensity appeared at about the same potential as the maximum calculated, but the peak width of the e.s.r. intensity was much larger than that calculated. The reason for the difference is not clear, but electrode kinetics may be involved. It is certain by the potential dependence of the e.s.r. intensity and of the calculated results that the semiquinone resulted from disproportiona- tion. At pH 7 the semiquinone formation was not marked at potentials more negative than -0.2 V, because the concentrations of the parent quinones are very small at such potentials. Oxygen reduction by the semiquinone route is still feasible at these potentials, since the parents arise from oxygen reduction.This reasoning was supported by the nap us. potential curves given in fig. 5 , in which nap did not decrease at potentials more negative than -0.1 V. However, quinols can react with oxygen indirectly via disproportionation, eqn (l), which was observed by Land et a1.22 Svingen and Powis and also Land et al. measured the El of adriamycin using pulse-radiolysis 23 The El us. pH curve of adriamycin is very similar in shape to that of NH, except that the former curve is shifted in the negative direction by ca. 200 mV.23 The pKa of the adriamycin semiquinone is 2.9, very close to the value of the naphthazarin semiquinone.21 However, Kano et al. electrochemically measured the El and E, values of adriamycin at several pH, estimating that the pK, of the semiquinone is 6.9.In table 1 K values for equilibrium between a semiquinone and oxygen, corresponding to eqn (3), were tabulated for the quinones investigated. The catalytic activity of the quinones increased in the following sequence: NP 4 JL < PQ < NH,. No catalytic activity was detected in the case of 1,2-benzoquinone (BQ). It is possible to compare the1832 Oxygen Reduction with Naphthazurin Catalytic activity with the K values, provided the reaction proceeds through the semiquinone and the rate constant of the reaction of eqn (7) is the same for all quinones.Unfortunately we could obtain accurate E, values only for NH, and BQ in aqueous media.,’? 40 BQ had no catalytic activity, so the activity can be explained in terms of the K value for these quinones. A plot of the two-electron redox potential at pH 7, EL, against the first one-electron redox potential at pH 7 , ET, was a straight line for the eleven quinones reported el~ewhere,~~ and gave the following expression by linear regression analysis : The correlation coefficient of 0.96 was obtained for the plot. The coefficient of the Ei term deviated from unity, showing that Ez = 0.92E: -0.404. The K values for PQ and NP were calculated by the above equation and are shown in parentheses in table 1. The difference in E: and E i at pH 7 was very small for NH,, and the plot for this quinone deviated greatly from the line of regression (see table 1).This shows that the formation of the quinols is not so energetically favoured from the semiquinones owing to intramolecular hydrogen bonding. The situation is probably similar for JL. Since the E L of JL was 85 mV more positive than that of NH,, the E: of JL would be more positive than that of NH,, making the K value smaller. While the K value estimated for NP can explain the catalytic activity, the value for PQ was calculated to be larger than that for NH,. This may be attributed to uncertainty in the estimation of E:, because the Ei value estimated from eqn (9) would have an error of 100 mV, even if complexities such as intramolecular hydrogen bonding are not involved.2Ek = 1.92E: - 0.404. (9) References 1 N. Bachur, S. L. Gordon and M. V. Gee, Cancer Res., 1978,38, 1745. 2 V. Berlin and W. A. Haseltine, J. Biol. Chem., 1981, 256, 4747. 3 J. W. Lown, H-H. Chen, S-K. Sim and J. A. Plambeck, Bioorg. Chem., 1979, 8, 17. 4 J. W. Lown, H-H. Chen, J. A. Plambeck and E. M. Acton, Biochem. Pharmacol., 1982, 31, 575. 5 J. Butler, B. M. Hoey and A. J. Swallow, FEBS Lett., 1985, 182, 95. 6 B. A. Svingen and G. Powis, Arch. Biochem. Biophys., 1981, 209, 119. 7 N. J. F. Dodd and T. Mukherjee, Biochem. Pharmacol., 1984, 33, 379. 8 H. Wefers, T. Komai, P. Talalay and H. Sies, FEBS Lett., 1984, 169, 63. 9 J. W. Lown, H-H. Chen, J. A. Plambeck and E. M. Acton, Biochem. Pharmacol., 1982, 31, 575. 10 G. M. Rao, J. W. Lown and J. A. Plambeck, J.Electrochem. Soc., 1977, 124, 195. 11 G. M. Rao, A. Begleiter, J. W. Lown and J. A. Plambeck, J. Electrochem. Soc., 1977, 124, 199. 12 G. M. Rao, J. W. Lown and J. A. Plambeck, J. Electrochem. SOC., 1978, 125, 534. 13 G. M. Rao, J. W. Lown and J. A. Plambeck, J. Electrochem. Soc., 1978, 125, 540. 14 H. Berg, G. Horn, U. Luthardt and W. Ihn, Bioelectrochem. Bioenerg., 1981,8, 537. 15 H. Berg, G. Horn and W. Ihn, J. Antibiot., 1982, 35, 800. 16 A. Ashnagar, J. M. Bruce, P. L. Dutton and R. C. Prince, Biochim. Biophys. Acta, 1984,801, 351. 17 A. Anne and J. Moiroux, Nouv. J. Chem., 1984,8, 259. 18 K. Kano, T. Konse and T. Kubota, Bull. Chem. Soc. Jpn, 1985, 58, 424. 19 B. Keita and L. Nadjo, J. Electroanal. Chem., 1983, 145,431. 20 C. Degrand, J . Electroanal. Chem., 1984, 169, 259.21 E. J. Land, T. Mukherjee and A. J. Swallow, J. Chem. SOC., Faraday Trans. I , 1983, 79, 391. 22 E. J. Land, T. Mukherjee and A. J. Swallow, J. Chem. Soc., Furaday Trans. I , 1983, 79, 405. 23 E. J. Land, T. Mukherjee, A. J. Swallow and A. J. Bruce, Arch. Biochem. Biophys., 1983, 255, 116. 24 T. Y. Toribara and A. L. Underwood. Anul. Chem., 1949, 21, 1352. 25 T. Nagaoka, T. Sakai, K. Ogura and T. Yoshino, Anal. Chem., 1986,58, 1953. 26 T. Nagaoka, S. Okazaki, T. Itoh and T. Fujinaga, J. Electroanal. Chem., 1981, 127, 289. 27 Y. A. Ilan, D. Meisel and G. Czapski, Zsr. J. Chem., 1974, 12, 891. 28 D. Meisel and G. Czapski, J. Phvs. Chem., 1975, 79, 1503. 29 P. S. Rao and E. Hayon, Biochem. Biophys. Res. Commun., 1973, 51, 468. 30 L. G. Forni and R. L. Willson, Meth. Enzymol., 1984, 105, 179. 31 J. W. Lown. S-K. Sim and H-H. Chen, Can. J. Biochem., 1978, 56, 1042. 32 J. W. Lown, H-H. Chen, S-K. Sirn and J. A. Plambeck, Bioorg. Chem., 1979, 8, 17. 33 B. H. J. Bielski, J. Photochem. Photobiol., 1978, 28, 645.T. Nagaoka, T. Sukui, K. Ogura and T. Yoshino 1833 34 Z. Bradic and R. G. Wilkins, J. Am. Chem. Soc., 1984, 106, 2236. 35 J. Chevalet, F. Rouelle, L. Gierst and J. P. Lambert, J. Electroanaf. Chfm., 1972, 39, 201. 36 P. R. Rich and D. S. Bendall, FEBS Lett., 1979, 105, 189. 37 P. R. Rich and D. S. Bendall, Biochim. Biophys. Acta, 1980, 592, 506. 38 P. R. Rich, Biochim. Biophys. Acta, 1981, 637, 28. 39 P. R. Rich, in Function of Quinones in Energy Conserving Systems, ed. B. L. Trumpower (Academic 40 A. J. Swallow, in Function of Quinones in Energy Conserving Systems, ed. B. L. Trumpower Press, New York, 1982), pp. 73-83. (Academic Press, New York, 1982), pp. 59-72. Paper 611753; Received 1st September, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301823

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987, 83, 1835-1842 The Effect of Hydrogen Sulphide on the Adsorption and Thermal Desorption of Carbon Monoxide over Rhodium Catalysts S. David Jackson" ICI plc, Agricultural Division, PO Box I , Billingham, Cleveland TS23 ILB Brian J. Brandreth and David Winstanley ICI plc, New Science Group, PO Box 11, The Heath, Runcorn, Cheshire WA7 4QE The effect of hydrogen sulphide on carbon monoxide adsorption and Dice versa has been studied. Changing the metal precursor has a considerable effect on the energetics of adsorption of both adsorbates. Contrary to what might have been expected, a surface saturated with hydrogen sulphide will still adsorb carbon monoxide and, depending on the metal precursor, carbon monoxide can displace adsorbed hydrogen sulphide.On all the catalysts at least three types of site can be specified by their adsorption behaviour to carbon monoxide and hydrogen sulphide in the presence of the other. The effect of sulphur on the activity of supported-metal catalysts has aroused consider- able interest, especially those catalysts based on nicke1.l However, the interest in nickel catalysts has been so great that many of the other Group VIII metals have had little or no study. The same is true about the effect of sulphur on the adsorption/desorption characteristics of other adsorbates over Group VIII metals. This study is concerned with the adsorption of hydrogen sulphide over rhodium catalysts and the effect hydrogen sulphide has on carbon monoxide adsorption and desorption. In previous studies29 we have shown that catalysts prepared from different rhodium salts, on the same support, gave different adsorption/desorption characteristics for the same adsorbed gas.We were therefore interested in whether a similar effect would be seen with hydrogen sulphide, and whether any differences could be observed on its effect on carbon monoxide adsorption. Experiment a1 The apparatus used throughout this study was a pulse-flow microreactor system. Using this system the catalysts could be activated in flowing hydrogen, maintained in flowing helium and then covered by the adsorbate gas by injecting pulses of known size into the helium carrier-gas stream and hence to the catalyst. The amount of gas adsorbed, from any pulse, was determined from the difference between a calibration peak area and the peak area obtained following the injection of pulses of comparable size onto the catalyst.Adsorptions and desorptions were followed using a thermal-conductivity-detector gas chromatograph, fitted with a Porapak Q-S column, coupled to a mass spectrometer (Spectamass SM IOOD). Three catalysts were used in this study : rhodium trichloride/silica, rhodium nitrate/silica and rhodium oxide/silica. The rhodium trichloride and nitrate catalysts were prepared by adding silica (Davison 952, surface area 280 m2 g-l) to an aqueous solution of the compound and evaporating to dryness. The rhodium oxide catalyst was prepared by heating a rhodium nitrate/silica catalyst to 723 K in flowing air (a 100 cm3 min-l) for 2 h. After the supported salt was dried, 0.50 g was placed in the 18351836 Efect of Hydrogen Sulphide on Rhodium Catalysts reactor on a glass sinter and reduced in a flow of hydrogen (80 cm3 min-I) by heating to 573 K.The catalyst was cooled in flowing hydrogen and maintained in a flow of helium (80 cm3 min-l). Pulses of carbon monoxide (2.95 cm3, 6.67 x 103-2.00 x lo5 Pa) or hydrogen sulphide (0.2 cm3, 1.01 x lo5 Pa) were passed over the catalyst at 293 K until adsorption was complete; then the catalyst was heated in flowing helium to 573 K at 25 K min-l and the desorption noted. The catalyst was then cooled in flowing helium and the adsorption-desorption cycle repeated. Both the helium (B.O.C., 99.998 % ) and the hydrogen (B.O.C., 99.9998 % ) were further purified by passing them through activated palladium bronze to remove any oxygen impurity, and a bed of soda asbestos to remove any water impurity.The hydrogen sulphide (B.D.H., 99.6%) was further purified by freezing and pumping, while the carbon monoxide (B.D.H., 99.5%) was used as received. ESCA was carried out on the samples using a Kratos XSAM 800 spectrometer. Results Catalyst Characterisation The catalysts were examined, by X-ray photoelectron spectroscopy, both as prepared and after use. All the as-prepared samples gave Rh(3d) binding energies of 308.4k 0.2 eV, compared to 309.5 eV obtained, on the same machine, from RhIII standards. The ratio of Rh: CI in the as-prepared rhodium trichloride sample was 1 : 3. After use, the Rh(3d) binding energy shifted to 307.5 eV for both the chloride- and oxide-derived catalysts: this binding energy and the shape of the peaks are typical of rhodium metal.There was no evidence in either sample for higher valency states of rhodium. No residual chloride was detected on the sample prepared from rhodium trichloride after reduction. The catalyst derived from rhodium nitrate gave no rhodium signal after reduction, although microanalysis and chemisorption studies indicated the presence of rhodium. This lack of a signal could be explained by the rhodium particles being at the bottom of the silica pores and so outside the X.P.S. depth profile. Micromeritics analysis of the silica indicated pores in the region of 0.3-2.0 nm. The reduced samples were also analysed by transmission electron microscopy and the particle sizes estimated; the particle size was found to increase in the order nitrate < oxide < chloride, with the nitrate-derived sample having all the measured particles in a narrow range of 1-1.4 nm, while with the chloride-derived catalyst the majority of the particles were in the range 2-3 nm.Carbon Monoxide Adsorption/Desorption Pulses of carbon monoxide, at 4 min intervals, were passed over each of the freshly reduced catalysts. The results are shown in table 1. The amount of carbon monoxide adsorbed on successive adsorption/desorption cycles decreases to steady-state values also shown in table 1. We define the steady state as having been reached when two successive chemisorptions give identical results. Once saturated with carbon monoxide, the catalysts were heated from 293 to 573 K at a heating rate of 25 K min-l in the helium gas stream.The desorption products were passed through the gas chromatograph and mass spectrometer. Both carbon monoxide and carbon dioxide were desorbed. The ratios of desorbed carbon monoxide to desorbed carbon dioxide are given in table 1. The desorption profiles for carbon monoxide are detailed in fig. 1. Water was also observed as a desorption product but was not quantifiable.S. D. Jackson, B. J. Brandreth and D. Winstanley 1837 Table 1. Carbon monoxide adsorption initial steady state catalyst % Rha CO(ads)b CO/Rh CO/CO,(des)c CO(ads)* CO/Rh CO/CO,(des)c RhClJSiO, 1.6 101.9 0.7 30.2 36.2 0.2 5.2 Rh(NO,),/SiO, 0.9 71.8 0.8 5.6 19.7 0.2 11.8 Rh,O,/SiO, 0.9 85.3 1 .o 4.7 42.7 0.2 0.5 a Units:w/w.Units:pmol (g catalyst)-'. Ratio of desorbed CO to desorbed CO,. t 3 73 4 73 573 TIK Fig. 1. Carbon monoxide desorption profiles. (a) Desorption of CO from a steady-state oxide-derived catalyst; (b) desorption of CO, after H,S displacement, from an oxide-derived catalyst; ( c ) desorption of CO from a steady-state nitrate-derived catalyst; ( d ) desorption of CO, after H,S displacement, from a nitrate-derived catalyst; (e) desorption of CO from a steady-state chloride-derived catalyst; and (f) desorption of CO, after H,S displacement, from a chloride- derived catalyst. Hydrogen Sulphide Adsorption/Desorption Pulses of hydrogen sulphide, at 10 min intervals, were passed over each of the catalysts immediately after reduction. The results are shown in table 2 with those of the steady state.Hydrogen was evolved during the adsorptions but we were unable to quantify the amount. A sample of silica was reduced in an identical manner to the catalysts to investigate possible hydrogen sulphide adsorption on the support. The amount of hydrogen sulphide adsorbed was < 0.6 pmol (g sample)-l. Once saturated with hydrogen sulphide the catalysts were heated and the desorption noted. No hydrogen sulphide was desorbed from any of the catalysts.1838 Eflect of Hydrogen Sulphide on Rhodium Catalysts Table 2. Hydrogen sulphide adsorption initial steady state catalyst % Rha H , s ( a d ~ ) ~ S/Rh CO/Rh H,S(ads)b S/Rh CO/Rh - RhCl,/SiO, 1.6 94.5 0.6 0.7 9.9 0.1 0.2 Rh(NO,),SiO, 1.6 107.7 0.7 0.7 13.8 0.1 0.2 Rh,O,SiO, 1.6 85.2 0.6 0.9 12.3 0.1 nmc a Units:w/w.Units:pmol (g catalyst)-'. Not measured. Effect of Hydrogen Sulphide on Carbon Monoxide Adsorption/Desorption Once in the steady state, with respect to carbon monoxide adsorption, the catalysts were pre-covered with carbon monoxide and then subjected to pulses of hydrogen sulphide. The results are shown in table 3. When the specific behaviour of each of the catalysts to the carbon monoxide/hydrogen sulphide adsorption/displacement is examined, significant differences become apparent. When the catalyst was derived from rhodium trichloride/silica only 6.1 % of the adsorbed carbon monoxide was displaced, but the amount of hydrogen sulphide adsorbed was equal to the amount of carbon monoxide preadsorbed. With the nitrate-derived catalyst 62.9 % was displaced but the amount of hydrogen sulphide adsorbed was almost equal to that of the preadsorbed carbon monoxide. However, with the oxide-derived catalyst 27.2% was displaced, but in this case the amount of hydrogen sulphide adsorbed was considerably less than the carbon monoxide preadsorbed: only 59.4%.Clearly, changing the metal precursor has a considerable effect on the ability of hydrogen sulphide to adsorb/displace on a carbon monoxide pre-covered surface. Once adsorption/displacement had ceased the catalysts were heated to 573 K and the desorption noted. After this thermal desorption the catalysts again had carbon monoxide adsorbed and thermally desorbed. The results are shown in table 3 and fig. 1. For the chloride-derived catalyst, only 46.8% of the carbon monoxide adsorbed immediately before hydrogen sulphide displacement was readsorbed, while for the nitrate- and oxide-derived catalysts the figures are 86.6 and 71.9% , respectively .Effect of Carbon Monoxide on Hydrogen Sulphide Adsorption/Desorption Catalysts, which were in the steady state with respect to hydrogen sulphide adsorption, were pre-covered with hydrogen sulphide and then subjected to pulses of carbon monoxide. The results are shown in table 4. The catalysts were heated to 573 K once adsorption/displacement had ceased and the desorption noted. Hydrogen sulphide was displaced from only one catalyst, that derived from rhodium oxide/silica. Discussion This discussion will consider the relationship between carbon monoxide and hydrogen sulphide adsorption over the three catalysts and also the differences produced on changing the metal salt.Table 1 shows the amount of carbon monoxide adsorbed on fresh and steady-state catalysts. The adsorption and thermal desorption of carbon monoxide over these catalysts has been discussed previously.2 However, it is pertinent to the present study to reiterate some of the conclusions. Our reason for using the steady-state catalysts for the displacement reactions rather than the fresh catalyst was that we had shown thatS. D . Jackson, B. J. Brandreth and D. Winstanley 1839 Table 3. Effect of hydrogen sulphide on preadsorbed carbon monoxide co H2S gas co/ co co/ catalyst % Rha (pre-ads)b (sec-ads)" displacedd CO,(des)e (ads)f CO,(des) RhCl,/SiO 1.6 38.2 38.3 2.3 CO 56.1 17.9 20.1 Rh(NO,),/SiO, 0.9 19.5 19.0 12.3 CO 50.4 16.9 22.0 Rh,O,/SiO, 0.9 45.4 26.9 12.4 CO 10.7 32.6 1.4 0.8 CO, a Units:w/w.Amount of H2S adsorbed on CO-covered surface. Ratio of CO to CO, detected during the thermal desorption immediately following H,S adsorption. f Amount of CO adsorbed following thermal desorption. Units:pmol (g catalyst)-'. Amount of gas displaced by H,S adsorption. Table 4. Effect on carbon monoxide adsorption of preadsorbed hydrogen sulphide gas catalyst % Rh" H,S(pre-ads)b CO(sec-ads)" displacedd CO/CO,(~~S)~ H2S(adsr RhCl,/SiO, 1.6 10.2 8.6 0 12.1 9.9 Rh(NO,),/SiO, 1.6 12.9 3.8 0 3.1 13.8 Rh,O,/SiO, 1.6 15.9 4.3 0.2 H,S 0.6 12.3 a Units:w/w. Amount of CO adsorbed on a hydrogen sulphide pre-covered surface.Ratio of CO to CO, detected during thermal desorption following CO adsorption. f Amount of H,S adsorbed following thermal desorption. g 0.3 pmol (g catalyst)-' of CO, was displaced during this H,S adsorption. Units:pmol (g catalyst)-'. Amount of gas displaced by CO adsorption. the adsorption and desorption of carbon monoxide from a fresh catalyst was a poor model for a working catalyst, whereas the steady-state catalyst had adsorption and desorption characteristics similar to those of a working catalyst., Carbon dioxide is a major by-product of the thermal desorption of carbon monoxide from these catalysts, and we have shown that its production is due to a water-gas shift reaction., However, we previously believed the source of the water to be the helium carrier gas, but further studies by ourselves and other workers4 using the same support have now shown that the source of water is, in fact, the silica, which retains significant quantities of water even after drying, calcining and reduction processes.The effect of passing hydrogen sulphide over carbon monoxide pre-covered surfaces is detailed in table 3 : with each catalyst there is displacement of carbon monoxide and adsorption of hydrogen sulphide. Three sets of results support the premise that the hydrogen sulphide adsorption has taken place on the metal : (i) material is displaced, (ii) the desorption profile changes (fig. 1) and (iii) the amount of carbon monoxide readsorbed decreases (table 3). Other points in favour of the hydrogen sulphide being adsorbed on the metal are that the same support was used throughout, yet the amounts adsorbed vary from catalyst to catalyst and also reports in the literature suggest, in agreement with our own results, that silica will not adsorb significant quantities of hydrogen ~ulphide.~ We may speculate that the differing amounts of carbon monoxide displaced by hydrogen sulphide reflect the different modes of adsorbed carbon monoxide, which may vary in desorption energy.At least three different types of adsorbed carbon monoxide have been identified on rhodium by infrared spectroscopy.6 These are: bridge (Rh,-CO), linear (Rh-CO) and digeminal (Rh-[CO],). On thermal desorption from the chloride-derived catalyst, the carbon monoxide1840 Eflect of Hydrogen Sulphide on Rhodium Catalysts desorption profile alters, indicating that the energetics of the carbon monoxide adsorp- tion sites has been changed by the sulphur adsorption.This type of behaviour is not unexpected. An infrared study on the effect of sulphur on carbon monoxide adsorbed on rhodium/silica6 (apparently produced from the chloride) showed that the bridge- bonded Rht--CO band disappeared, while the intensity of the digeminal Rh-(CO), band increased by 50%. If this were the case for our system we would expect the loss of the higher-temperature part of the desorption in favour of the low-temperature part, and this is indeed observed with the desorption from the chloride-derived catalyst. Therefore, we suggest that for the chloride-derived catalyst the species displaced is the bridge-bonded Rh,-CO.However, on thermal desorption from the nitrate- and oxide- derived catalysts, no effect on the desorption profile is noted, yet a significant quantity of adsorbed carbon monoxide was displaced. One explanation for this effect would be for displacement to have occurred across the whole spectrum of adsorbed carbon monoxide. However, it can equally be explained by specific displacement of a given type of adsorbed carbon monoxide followed by a redistribution of the residual carbon monoxide over the surface. The second explanation may be slightly more favoured, as it is known that metal surfaces will reconstruct under increasing amounts of adsorbed sulphur,' although some difference in the carbon monoxide desorption profile would still be expected.A difference is noted for all three catalysts, however, in the ratio of desorbed carbon monoxide to desorbed carbon dioxide (the production of carbon dioxide was shown2 using isotopic studies to have been produced principally by a water-gas shift reaction). For all three catalysts the amount of carbon dioxide produced drops dramatically, so that even though the hydrogen sulphide only affects the carbon monoxide desorption profile of the chloride-derived catalyst, it reduces the reactivity of the adsorbed carbon monoxide on all three samples. On readsorption of carbon monoxide it was found that the adsorptive capacity had decreased for each of the catalysts (table 3). The catalyst which had the highest amount of displacement (nitrate) also has the largest amount of readsorption, while that with the smallest displacement (chloride) has the smallest percentage readsorption.In fact, the amount readsorbed for each catalyst can be described in terms of the different adsorption sites detectable by the hydrogen sulphide displacement (table 5). Using these classifications it can be seen that changing the metal salt changes either the ability of sites, or the number of sites with the ability, to adsorb or not to adsorb carbon monoxide and/or hydrogen sulphide. The CO(des) : CO,(des) ratios from the thermal desorptions are lower than those from the first desorption when sulphur was present. This is due to the production of carbon dioxide remaining constant while the amount of carbon monoxide desorbed has decreased. The amount of hydrogen sulphide adsorbed on a fresh catalyst is shown in table 2; in the case of the chloride- and nitrate-derived samples, the CO: H,S ratio for adsorption on a fresh catalyst is 1.1: 1, which suggests that the same sites are being measured.However, with the oxide-derived catalyst the ratio of CO:H,S is 1.5: 1. Therefore the hydrogen sulphide cannot adsorb on all the sites that are available for carbon monoxide adsorption. These views are supported by the results from the carbon monoxide/hy- drogen sulphide adsorption/displacement chemisorptions, where CO: H,S adsorption ratios were chloride, 1 .O : 1 ; nitrate, 1 .O : 1 and oxide, 1.7 : 1, so that the relationships found between the hydrogen sulphide and carbon monoxide adsorption sites on the fresh catalysts are also extant on catalysts which are in the steady state with respect to carbon monoxide chemisorption.The important point to note, however, is that the oxide-derived catalyst has a considerable site density, which is available for carbon monoxide adsorption but not hydrogen sulphide adsorption. It would be useful to be able to specify the site which will accept an adsorption mode of carbon monoxide but not of hydrogen sulphide. It is tempting to suggest that the site which will adsorb carbon monoxide but not hydrogen sulphide is that site which allows the digeminal carbon monoxideS. D. Jackson, B. J. Brandreth and D . Winstanley 1841 Table 5. Description by sites of carbon monoxide readsorption sites where sites where non-S(ads)" S displaces S adsorbs amount of CO catalyst sites ( A ) CO" (B) with CO" (C) total readsorbed" RhCl,/SiO, 0 2.3 35.8 17.9 (1/2C) 17.9 Rh,O,/SiO, 18.4 14.0 3 1.4 32.5 ( A + B) 32.6 Rh(NO,),/SiO, 0.5 12.3 7.2 16.4 ( A +B+ 1/2C) 16.9 a Units:pmol (g catalyst)-l. adsorption, but at present we have no conclusive evidence as to the nature of the site.In the steady state, with respect to hydrogen sulphide adsorption, there is a far smaller number of reusable sites than in the steady state with respect to carbon monoxide adsorption. The effect of carbon monoxide on preadsorbed hydrogen sulphide is shown in table 4. Clearly, carbon monoxide can adsorb on all the samples even when they are nominally saturated with hydrogen sulphide ; indeed, with the chloride-derived sample the CO: H2S ratio is 0.85 : 1.These results are in contrast to those obtained by Yamada and Tamaru8 on carbon monoxide over polycrystalline rhodium; they state that one sulphur adatom blocks the adsorption of eight carbon monoxide molecules. However, again there are considerable differences depending on which metal precursor is used. The catalyst derived from rhodium oxide was the only sample where hydrogen sulphide was displaced by carbon monoxide and for which some of the carbon monoxide was also retained through thermal desorption, only to be displaced as carbon dioxide by subsequent hydrogen sulphide adsorption. There are two possible reasons why the hydrogen sulphide can be displaced from the oxide catalyst and yet not from either the chloride- or nitrate-derived catalyst.First, it is possible that the site energetics have changed, so that the adsorbed hydrogen sulphide is held weakly enough for carbon monoxide to displace it. However, a more likely cause is that on the chloride and nitrate catalysts there is no mechanism for sulphur desorption. This is related to the mode of hydrogen sulphide chemisorption, which is usually disso~iative,~ giving 2(H-*) and S-*, but as we are using a pulse-flow method any adsorbed hydrogen will be displaced by either sulphur or carbon monoxide, hence leaving no available hydrogen for the sulphur to desorb as hydrogen sulphide. On the oxide catalyst the fact that there is displacement suggests that some of the adsorbed hydrogen sulphide may have been adsorbed as an HS-* species; sulphur can then desorb (CO + 2HS-* + H2S + CO-* + S-*).In fact such a species has been seen spectro- scopically over ruthenium.1° At this point we must consider why changing the metal precursor should have such a dramatic effect on the adsorption-site energetics. An obvious reason is that some of the counter-ion, chloride or oxide, is left on the surface. This has been suggested for variations in the frequency and mode of adsorption observed during an infrared spectroscopic study of carbon monoxide on rhodium/alumina catalysts. l1 Indeed, if these variations are caused by residual counter-ions it would appear that by choosing a suitable ion the adsorption-site energetics could be modified in such a way as to lessen the effect of any subsequent poisoning. For this to occur the residual ion must affect the ability of the rhodium to bind other adsorbates either geometrically or electronically. However, no residual chloride was detected by ESCA with the chloride-derived catalyst.Nor was there any evidence for the presence of non-zero-valent rhodium with either the oxide- or chloride-derived samples. However, another possible cause of these effects may be particle size. Variations in particle size have been shown to influence the mode of carbon monoxide adsorption,12 as different modes of adsorption have different 61 FAR 11842 Efeect of Hydrogen Sulphide on Rhodium Catalysts energetics, then a variable response to the effect of sulphur could be obtained purely by varying the particle size. For this to be the case, however, the particle size must be different for the three catalysts, and indeed the catalysts do show different particle sizes, although all are between 1 and 3 nm.However, it is in this size range that there are dramatic changes in both the number of surface atoms, relative to the bulk, and the average surface coordination number:13 e.g. a 1 nm particle has 92% surface atoms and an average surface coordination number of 5; a 1.5 nm particle has 76% surface atoms and an average surface coordination number of 6.6; by 2 nm only 63 % are surface atoms and the average surface coordination number is 7.2; and at 3 nm the particle only has 45% surface atoms and an average surface coordination number of 7.7. Certainly on changing the surface coordination number from 5 to 8 we would expect the energetics of the adsorption sites to change considerably, as this is equivalent to going from a surface of ‘corners’ to a mixed surface of corners, edges and planes. Therefore we believe that the differences in the chemisorption behaviour of these catalysts are indeed due to variations in particle size, and this view can be supported by the fact that the catalyst with the largest particle size (chloride) mostly closely resembles ‘bulk metal ’, e.g. a wire, in terms of its chemisorption properties.21 References 1 C. H. Bartholomew, P. K. Agrawal and J. R. Katzer, Ado. Catal., 1982,30, 135. 2 S. D. Jackson, J. Chem. SOC., Faraday Trans. 1, 1985,81, 2225. 3 B. J. Brandreth, S. D. Jackson and D. Winstanley, J. Catal., 1986, 102, 433. 4 S. D. Jackson, unpublished results; G. Webb, personal communication. 5 W. H. Jones and R. A. Ross, J. Chem. SOC. A, 1968, 1787. 6 C. R. Guerra, J. Colloid Interface Sci., 1969, 29, 229. 7 L. D. Schmidt and D. Luss, J. Catal., 1971, 22, 269. 8 T. Yamada and K. Tamura, Surf. Sci., 1984, 139,463. 9 J. Saleh, C. Kemball and M. W. Roberts, Trans. Faraday SOC., 1961, 57, 1771; I. E. Den Besten and P. W. Selwood, J. Catal., 1962, 1, 93. 10 G. B. Fisher, Surf. Sci., 1979, 87, 215. 11 S. D. Worley, C. A. Rice, B. A. Mattson, C. W. Curtis, J. A. Guin and A. R. Tarrer, J. Chem. Phys., 12 N. Sheppard and T. T. Nguyen, Adv. Infrared Raman Spectrosc., 1978,5,67; J. T. Yates, T. M. Duncan 13 P. Chini, Gazz. Chim. Ztal., 1979, 109, 225. 1982, 76, 20. and R. W. Vaughan, J. Chem. Phys., 1979,71, 3908. Paper 611 803; Received 9th September, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301835

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. 1, 1987,83, 1843-1849 Nuclear Magnetic Resonance Self-diffusion Studies of Methanol-Water Mixtures in Pentad-type Zeolites Jurgen Caro, Martin Bulow and Jurgen Richter-Mendau Zentralinstitut f i r Physikalische Chemie der AdW der DDR, Rudower Chaussee 5, 1199 Berlin, German Democratic Republic Jorg Karger, Michael Hunger and Dieter Freude Sektion Physik der Karl-Marx- Universitat Leipzig, Linnistrasse 5, 7010 Leipzig, German Democratic Republic Lovat V. C. Rees* Chemistry Department, Imperial College, London S W7 2A Y Molecular self-diffusion data of methanol-water mixtures sorbed in pentasil- type zeolites of two different Si/A1 ratios are reported. At medium loading and room temperature, the self-diffusion coefficients of sorbed methanol and water are found to be of the order of m2 s-l.For both sorbate molecules intracrystalline mobility is greater for the higher Si/Al ratio. While water self-diffusion is only slightly affected by sorbate concentration, in the case of methanol a decrease of approximately one order of magnitude with increasing loading has been observed. In contrast to the liquid state, the sorbed state of binary mixtures exhibited no minima of the self-diffusion coefficients. The molecular mobilities observed can be correlated with structural features of the pentasils determined from 'H n.m.r. measure- ments. One of the reactions catalysed by pentad-type zeolites (e.g. ZSMS) is the conversion of methanol to gasoline (MTG).l In addition to the hydrocarbons formed, ca. half of the weight of the reaction products is water.Investigations of both sorption equilibria and dynamics of the substances involved in the MTG reaction have attracted consider- able interest. Although single-component studies of methan01~-~ and water8-12 sorbed on pentasils have been discussed in great detail in the literature, little information on the binary mixtures has been reported. This paper deals first with results of self-diffusion measurements of both single components and binary mixtures sorbed on two pentasil-type zeolites of different A1 contents in their hydrogen forms (i.e. HZSMS of Si/A1 sz SO and an almost Al-free analogue, i.e. silicalite of Si/Al z 1230). The investigation of the intracrystalline molecular mobility has been carried out using the n.m.r.pulsed field gradient technique.l39 l4 In contrast to the sorption uptake experiments this technique provides direct information on the intracrystalline molecular mobility under equilibrium condi- tions. Therefore, problems which arise when analysing sorption uptake data can be avoided. Diffusion coefficients for the water-silicalite system have been determined which are two orders of magnitude smaller than the values for the water-ZSMS system;12 however, in these experiments the uptake rate was evidently restricted by a water film on the outer surface of the silicalite crystalsll? l2 and the diffusion coefficients derived from these experiments do not describe the intracrystalline water mobility in silicalite. In this paper the intrinsic self-diffusion coefficient of water can be measured directly by n.m.r.and has been found to be remarkably greater in silicalite than in ZSMS. 1843 61-21844 Se If-difusion Studies Secondly, owing to the relatively high sorption heat of methanol on ZSM5,15 significant temperature increases occur during the uptake experiment^.^ This additional complication affects significantly the concentration of sorbate at the crystal surface, and this must be taken into account in the determination of sorption diffusion coefficients. Thirdly, from detailed investigations of the synthesis of ZSM5 zeolite^^^-^^ it became evident that ZSMS specimens of identical crystallographic type may reveal substantial differences in their habit, depending on the method of preparation and the chemical composition.In the case of sorption uptake by pentasil crystals, in general the anisotropy of diffusion in the pentasil micropore system has to be analysed. If the crystals show different habits the diffusion anisotropy must also be correlated to the net fluxes through various crystal faces in terms of the appropriate Fick's law solution. Owing to the direct observation of molecular displacements within well defined time and space intervals, the n.m.r. pulsed field gradient technique is most likely able to overcome these influences. Fourthly, the application of n.m.r. spectroscopy to multicomponent systems has the advantage of allowing one to measure self-diffusion coefficients of each of the mixture components by using alternatively only one proton-containing component.The self- diffusion coefficient of this component is then obtained by lH n.rn.r.,O* 21 On the other hand, self-diffusion coefficients of each component of a mixture may be obtained by the pulsed field gradient method from the Fourier-transformed spectra of proton echo signals. By this method self-diffusion coefficients for the binary liquid mixture methanol- water were determined.,, Experiment a1 The two pentasil-type zeolites of Si/Al ratios 50 and 1230 have been synthesized according to established p r o c e d ~ r e s . ~ ~ After calcination of the as-synthesized zeolites at 600 "C for 2 h they were transformed into the H-form by NH,-exchange followed by a further calcination for 2 h at 600 "C. For the n.m.r. studies, the zeolite crystals were evacuated at 400 "C until a pressure < lop2 Pa was maintained.The lH m.a.s.n.m.r. spectra of the outgassed zeolites were measured using the home-built pulse spectrometer HFS-270 at the University of Leipzig at 270 MHz and a spinning frequency of 3 kHz. The total proton concentration was determined from the free induction decay after a z/2 pulse and comparing its initial value with that of a standard.,,' 25 For the lH n.m.r. self-diffusion measurements, the zeolites were loaded with sorbate by vacuum distillation immediately after sample activation. For the two-component diffusion studies deuterated compounds have alternatively to be used and the n.m.r. signal can then be attributed to the proton-containing species. For example, in the case of CH,OD + D,O mixtures the self-diffusion coefficients of methanol could be obtained. However, the analogous determination of the water self-diffusion coefficients from CD,OD + H,O mixtures is complicated by an unavoidable proton exchange between the OD-group of methanol and water.g Therefore, CD,OH + H,O mixtures have been analysed taking into account the fact that the observed n.m.r.signal is not exclusively due to water molecules, However, because of the relatively small amounts of methanol used in the water self-diffusion measurements there should be no essential influence of the methanol proton on the determination of the water self-diffusion coefficient. Furthermore, the small differences between the mobilities of both components (cf. fig. 4, later) confirm this approach. Results Characterization of the Zeolites under Study by lH M.A.S.N.M.R. and Sorption of Water The lH m.a.s.n.m,r.spectra of the unloaded silicalite and ZSM5 specimens are shown in fig. 1 . Line (A) at 2.0 ppm is either caused by non-acidic OH-groups at the outer surfaceJ . Chem. SOC., Faraday Trans. 1 , Vol. 83, part 6 Plate 1 Plate 1. S.e.m. of the silicalite crystals (a) and ZSMS crystals (b) under study. J. Caro et al. (Facing p . 1845)J . Caro et al. 1845 of the pentasil crystals (terminal OH-groups) or inside the crystals as lattice defects or associated with some amorphous material. Line (B) at 4.3 ppm is ascribed to acidic (bridging) O H - g r ~ u p s . ~ ~ > 25 Fig. 1. lH m.a.s.n.m.r. spectra of the unloaded silicalite [(a) %/A1 z 12301 and ZSMS [(b) Si/A1 z 501 used in this paper.Silicalite From the extremely high concentrations of SOH-groups [28 x (g silicalite)-'] as found by lH m.a.s.n.m.r. it follows that after calcination in air at 600 "C for 2 h ca. 3% of the Si atoms exist as SOH. For other similarly prepared pentasil samples it has been determined by 29Si m.a.s.n.m.r. that up to 7% of the framework Si is associated with OH-groups.26 These concentrations of non-acidic hydroxyl groups are indeed much higher than the value which would result from the maximum possible number of terminal SiOH groups at the outer crystal surface. Assuming an average crystal size of 19 pm [cf. plate 1 (a)] the maximum fraction of terminal OH at the crystal surface amounts to < of the total amount of Si atoms. It seems most likely, therefore, that the majority of OH groups must be present as internal silanols.Furthermore, by comparing the number of silanol groups in pentads synthesized by means of different templating organic molecules it can be concluded that only tetrapropylammonium species give rise to high concentrations of framework This result supports the assumption that siloxy bonds in pentads are broken by strong alkaline attack of the TPAOH, forming non-acidic silano1s.28 The large amount of water adsorbed by the silicalite sample considered corresponds with the extremely high concentration of SiOH groups within that sample. The saturation sorption capacity of water measured at 27 "C shows that 1 5 % of the available void volume of silicalite (0.19 cm3 g-l) is occupied by water.Owing to the relatively large size of the silicalite crystals used [cf. plate 1 (a)], the high amount of water sorbed cannot be due to a water film on the outer crystal surface, which could result from the interaction of water molecules with surface hydroxyls via hydrogen b0nding.l'. l2 It must be assumed that the high concentration of internal silanols leads to the remarkable decrease in hydrophobicity of the pentasil structure. The latter conclusion has been verified independently by another experiment. The hydrated silicalite crystals were covered by an epoxy resin containing Mn2+ ions. In complete accordance with our expectations, both the transverse relaxation times and the intensity of the spin echo in a 7r/2-7c pulse sequence remained unchanged. Such a result confirms that there is no close contact between sorbed water molecules and Mn2+ cations, and the water molecules observed in the n.m.r.experiments must indeed be situated in the interior of the crystals. On the other hand, having applied this procedure to silicalite after presorption of mesitylene, the transverse relaxation times of mesitylene molecules were shortened by more than two orders of magnitude. This result is consistent with the1846 Self-diflusion Studies fact that mesitylene is too large (van der Waals diameter 0.86 nm) to penetrate into the intracrystalline channel system. Therefore, mesitylene has to form a layer on the outer surface of the crystals being subjected to the direct influence of the Mn2+ ions. In addition to the line for the non-acidic hydroxyls [fig, 1 (a)], the lH m.a.s.n.m.r.spectrum of HZSMS shows a second line stemming from acidic OH groups. The fraction of the latter amounts to ca. 45% of the total concentration of OH groups. The smaller amount of silanols in ZSMS (24 x 1019 g-l) compared with the silicalite considered agrees with the finding26 that the concentration of internal silanols decreases with increasing aluminium content of the ZSM5. This result furthermore supports the point of view that the silanols formed by calcination have to be regarded as an inherent part of highly siliceous pentasils synthesized by means of TPA.26v 29 The concentration of bridging OH groups amounts to 19 x 1019, i.e. ca. 45% of the total number of protons. This concentration of acidic OH groups is equal to the number of aluminium atoms thus indicating the absence of large amounts of extra-framework species.In accordance with the literature,l1? 30 the amount of sorbed water increases with increasing A1 content of the pentad-type zeolite. For the HZSM5 under study a maximum sorption capacity of 45 mg water g-l at 27 "C, corresponding to a pore-filling of ca. 24%, has been found. Single-component Self-diffusion in HZSMS and Silicalite Fig. 2(a) shows that the intracrystalline mobility of water in silicalite is higher than in HZSMS. While the self-diffusion coefficient of water in HZSMS seems to pass through a slight maximum at a concentration of ca. 2.5 molecules per unit cell, the water mobility in silicalite decreases with increasing water loading.These results should be related to the different extent of specific sorbate-sorbent interaction in the pentasils considered. It can be expected that the interaction is stronger in the case of the acidic OH groups present in HZSMS owing to the electrostatic field gradients caused by the framework aluminium. One can, furthermore, assume that the slight increase in the water mobility in HZSMS (up to medium loading) is due to the predominance of these interactions at low loadings compared with mutual interactions and steric hindrances of the sorbate molecules at higher loadings. Whereas the latter process determines the decrease of water self-diffusion coefficients in HZSMS with further increase in water loading, it influences the self-diffusion coefficients of water in silicalite even at low concentration.Compared with that influence, the interaction between water molecules and internal silanol groups of silicalite is obviously small. Fig. 2 (b) shows the concentration dependence of methanol self-diffusion coefficients in silicalite and HZSMS. Unlike the corresponding water data, both systems show a monotonic decrease of molecular mobility over ca. one order of magnitude. This behaviour [analogous to the decrease in intracrystalline mobility found with benzene in both silicalite and (Na, H)ZSM5 by sorption uptake methods31] may be explained by either a corresponding decrease of the molecular free volume with increasing sorbate loading or an increase of the mean residence time of molecules between succeeding molecular jumps.21 For ethane and propane in the pentad channel it could be deduced from n.m.r.measurements that it is the latter process that controls the molecular mobility. In this case the reduction of the intracrystalline diffusivity with increasing loading arises predominantly from a decrease in the jump rate due to increasing sorbate-sorbate interference effects. At high concentration of sorbed ethane and propane the jump lengths turned out to be reduced also. Assuming that a similar mechanism probably operates in the case of methanol, i.e. that molecular migration proceeds most likely by jumps between adjacent sorption sites, e.s.r. studies suggest thatJ . Caro et al. 6 - wl 4 4 - -E ", \ a 2 - 9 1847 1 I I I I 1 ( a ) a . @ a @ a @ ( p O O O E 000p - I I I I I 0 1 2 3 4 6 4 7 2 E N 1 s u 2 W 9 6 4 I =.T I (6) 0 0. ~~ 0 1 2 3 4 5 methanol molecules per 4 unit cell Fig. 2. Self-diffusion coefficient of water (a) and methanol (b) in silicalite (filled symbols) and HZSMS (open symbols) at 27 "C vs. sorbate concentration. (The error bars indicate the uncertainty of the absolute values; the relative errors between the values for different concentrations are represented by the diameters of the symbols.)1848 Self-difusion Studies the methanol molecules are probably localized in the channel intersections rather than in the pore segments of the pentasil framework.6 In recent quasi-elastic neutron scattering experiments' the jump distance of sorbed methanol has been determined to be ca. 0.5 nm and the self-diffusion coefficient of methanol in HZSMS at 22 "C has been found to be of the order of 10-l' m2 s-l.This value is ca. one order of magnitude lower than those reported here, even for high sorbate concentrations. However, before too much significance can be attached to these differences, the experiments should be repeated with identical samples, especially in the case of pentasils. Self-diffusion of Methanol and Water in their Binary Mixtures in HZSMS In fig. 3 self-diffusion coefficients of water and methanol in their binary mixtures for two different total loadings in HZSM5 are plotted against the mole fraction of methanol. 2.0. CI 'rA "E 2 ' 1.5. 0, \ 0 25 50 75 100 mol % methanol Fig. 3. Self-diffusion coefficients of methanol (squares) and water (circles) in their binary mixtures sorbed in HZSMS for constant total sorbate concentrations of 35 mg g-l (open symbols) and 50 mg g-l (full symbols) at 27 "C and comparison with the data of the liquid mixtures:33 ( a - * * * .) D(CH,OH) and (- - - -) D(H,O).When the molecular mobility of the sorbed single components are compared with the data of the pure liquids, especially at the loading of 30 mg sorbate (g zeolite)-' the different behaviour of the components becomes evident. Whilst the mobility of sorbed water was found to be enhanced with respect to the bulk liquid phase of water, the mobility of sorbed methanol is slightly reduced. This change in behaviour may again be explained by differences between the cross-sections of the molecules and the channel diameter, respectively. At a total loading of 45 mg g-l the self-diffusion coefficient of water has also been found to be lower than the values for the free liquids.As seen from fig. 3, no minima in the self-diffusion coefficient of the sorbed mixtures are found. The influence of the sorption potential of the zeolite channel surfaces and the limited diameterJ . Caro et al. 1849 of the intracrystalline channel system must destroy the highly structured methanol-water complexes assumed to be present in the bulk liquid-phase mixtures. These complexes have been proposed to explain the minima found in the diffusion coefficients of the liquid 33y 34 The continuous support and many valuable discussions afforded by Professor Dr H. Pfeifer (Leipzig) during the work are gratefully acknowledged. Further, the authors wish to thank Dr U.Lohse (Berlin) and B. Zibrowius (Berlin) for stimulating discussions and D. Prager (Leipzig) for experimental assistance. Dr W. Heink (Leipzig) is thanked for the development of the sophisticated pulse spectrometer suitable for self-diffusion measurements. The authors thank Dr S. A. Barri (London) for the synthesis of the two pentasil specimens used. References 1 S. L. Meisel, J. P. McCullough, C. H. Lechthaller and P. S. Weisz, Chemtech., 1976, 6, 1986. 2 H-J. Doelle, J. Heering, L. Riekert and L. Marosi, J. Catal., 1981, 71, 27. 3 S. J. Kmiotek, Wu Pingdong and Y. H. Ma, AIChE Symp. Ser., 1982, 219, 82. 4 H. Negishi, M. Sasaki, T. Iwaki, K. F. Hayes and T. Yasunaga, J. Phys. Chem., 1984,88, 5564. 5 H. Lechert, J. Wienecke and W.D. Basler, Proc. Int. Symp. Zeolite Catal., Siofok, Hungary, 1985, 6 C. S. Narasimhan, M. Narayana and L. Kevan, J. Phys. Chem., 1983,87, 984. 7 H. Jobic, A. Renouprez, M. Bee and C. Poinsignon, J. Phys. Chem., 1986,90, 1059. 8 H. Nakamoto and H. Takahashi, Zeolites, 1982, 2, 67. 9 A. Ison and R. J. Gorte, J. Catal., 1984, 89, 150. p. 681. 10 G. Debras, A. Gourgue, J. B. Nagy and G. De Clippeleir, Zeolites, 1985, 5, 377. 11 P. H. Kasai and P. M. Jones, J. Mol. Catal., 1984, 27, 81. 12 S. G. Hill and D. Seddon, Zeolites, 1985, 5, 173. 13 H. Pfeifer, in NMR-Basic Principles and Progress, ed. P. Diehl, E. Fluck and R. Kosfeld (Springer- 14 J. Karger, H. Pfeifer and W. Heink, Proc. 6th Int. Conf. Zeolites, ed. D. Olson and A. Bisio (Butter- 15 U. Messow, K.Quitzsch and H-J. Herden, Zeolites, 1984, 4, 255. 16 G. Debras, A. Gourge, J. B. Nagy and G. De Clippeleir, Zeolites, 1985, 5, 369. 17 V. N. Rommanikov, V. M. Mastikhin, S. Hocevar and B. Drzaj, Zeolites, 1983, 3, 31 1. 18 Z. Gabelica, E. D. Derouane and N. Blom, A.C.S. Symp. Ser., 1984, 248, 219. 19 R. Mostowicz and L. B. Sand, Zeolites, 1982, 2, 143. 20 J. Karger, M. Lorenz and M. Bulow, J. Colloid Interface Sci., 1978, 65, 181; Izv. Akad. Nauk SSSR, Ser. Chim., 1980, 1741 ; Colloid Surf., 1984, 11, 353. 21 A. Germanus, J. Karger and H. Pfeifer, Zeolites, 1984, 4, 188. 22 J. Kida and H. Uedaira, J. Magn. Reson., 1977, 27, 253. 23 R. J. Argauer and G. R. Landolt, U S . Patent 3, 702, 886, 1972. 24 H. Pfeifer, D. Freude and M. Hunger, Zeolites, 1985, 5, 274. 25 D. Freude, M. Hunger and H. Pfeifer, Z . Phys. Chem. N.F., in press. 26 G. L. Woolery, L. B. Alemany, R. M. Dessau and A. W. Chester, Zeolites, 1986, 6, 14. 27 D. Freude, M. Hunger, H. Pfeifer and W. Schwieger, Chem. Phys. Lett., 1986, 128, 62. 28 S. G. Fegan and B. M. Lowe, J. Chem. Soc., Faraday Trans. I , 1986,82, 785; 801. 29 A. W. Chester, Y. F. Chu, R. M. Dessau, G. T. Kerr and C. T. Kresge, J. Chem. SOC., Chem. Commun., 30 D. H. Olson, W. 0. Haag and R. M. Lago, J. Catal., 1980, 61, 390. 31 A. Zikanova, M. Bulow and H. Schlodder, Zeolites, in press. 32 J. Caro, M. Bulow, W. Schirmer, J. Karger, W. Heink, H. Pfeifer and S. P. Zdhanov, J. Chem. Soc., 33 Z. J. Derlacki, A. L. Easteal, A. V. J. Edge, L. A. Woolf and Z. Roksandic, J. Phys. Chem., 1985, 89, 34 H. G. Hertz and H. Leiter, 2. Phys. Chem. N . F., 1982, 133, 45. Verlag, Berlin, 1972), vol. 7, p. 53. worths, Guildford, 1984), p. 184. 1985, 289. Faraday Trans. I , 1985, 81, 2541. 53 18. Paper 6/ 1848 ; Received 17th September, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301843

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987,83, 1851-1858 Synthesis of Montmorillonite-Viologen Intercalation Compounds and their Photochromic Behaviour Hirokatsu Miyata, Yoshiyuki Sugahara, Kazuyuki Kuroda and Chuzo Kato* Department of Applied Chemistry, Waseda University, Ohkubo-3, Shinjuku-ku, Tokyo 160, Japan The photochemical properties of four different viologen di-cations (V2+, 1,l'- dipropyl-, 1,l '-diheptyl-, 1,1 '-diphenylmethyl- and 1,l '-diphenylethyl-4,4'- bipyridinium di-cations) intercalated into the interlayer space of montmorillonite have been studied. The colour of the intercalation com- pounds which were co-intercalated by poly(viny1 pyrrolidone) (PVP) changed to blue when irradiated by a mercury lamp, showing the formation of the radical cations. Their characteristic absorptions appeared at ca.610 and 400nm in their visible spectra. Since the viologens were present in their cationic forms in the interlayer space it was thought that PVP acted as an electron donor for the photochemical reduction of the viologens. The intensity decay of the radical cations was very slow in comparison with the species in PVP matrices, indicating their remarkable stability. Photochromism has been widely investigated because of its chemical interest and industrial app1ications.l Since the colour development of viologens (V : 1,l'-dialkyl- 4,4'-bipyridinium salts) by electrochemical reduction is well known, viologens as electrochromic substances have been actively studied in recent years.2 Viologens can also be reduced by U.V. light in the presence of electron donors: and there have already been some reports on the photochromism of viologens.Kamogawa el aZ.3? investigated the photoreduction of viologens in various matrices. They found that the photoreduction of viologens was stimulated in polar aprotic matrices or solvents and that it is affected by both counter-anions and N-substituents on the viologen, They proposed the mechanism to be electron transfer from the viologen counter-anions to the viologen di-cations. The influence of solid matrices such as cellulose and gelatin on the photoreduction of viologens has also been rep~rted.~? On the other hand, inorganic layered materials have attracted considerable attention because some of these materials can form intercalation compounds which accommodate inorganic or organic guest molecules (or ions) in their interlamellar spaces.However, there have been relatively few studies on the photochemical properties of intercalation compounds. As to the photochromism of intercalation compounds, only a study using mica-AgNO, complexes has been reported.' There are no studies of the photochemical properties of intercalation compounds consisting of an inorganic layered host and an organic photochromic guest molecule reported so far, although organic photochromic substances such as viologens and spiropyrans have the advantage of high-contrast colour changes under illumination. Therefore, in this paper we report the synthesis and photochromic behaviour of inorganic-organic intercalation compounds. Montmorillonite, which is a swelling clay mineral, was suitable as the inorganic layered 18511852 Mon tmor il loni te- Viologen Intercalation Compounds host because it is well known that various organic cations can be intercalated into the interlayer space and that viologens as di-cations can also be intercalated within it.8,9 With respect to the intercalation of methyl viologen into clay minerals, Raupach et a1.l0 studied the state of the intercalated methyl viologen in montmorillonite and vermiculite. In montmorillonite the viologen lies flat in the silicate layer, while it is in a twisted form and inclined with a definite angle in vermiculite.However, there are no reports on their photochemical behaviour. Only a study of the influence of the inorganic host using MPS, on the stability of the viologen radical cation has been reported.ll Therefore in the present study we synthesized montmorillonite-viologen intercalation compounds co-intercalated by poly(viny1 pyrrolidone) (PVP), measured visible spectral changes, colour development and fading processes by the use of visible spectroscopy, and discussed their photochromism.Since viologens are adsorbed in the di-cationic form, PVP is considered to be an electron donor, and this is also discussed, Experimental Materials Montmorillonite [ideal formula (NaOe33) (A15,3Mgl,3) (Si,)O,,(OH), - nH,O] from the Ater- azawa mine, Yamagata Prefecture, Japan, was used after purification by elution. Four different viologens were employed having n-propyl, n-heptyl, phenylmethyl and phenyl- ethyl groups as the N-substituents.The bromide salts of these viologens were synthesized by refluxing 4,4’-bipyridine and the corresponding bromides in DMF under an N, atmosphere. The crude products were recrystallized three times in methanol before use except for phenylethyl viologen, which is rather difficult to recrystallize. PVP (Tokyo Kasei Co.; = 10000) was used as received. Synthesis of Intercalation Compounds Homo-ionic Na-montmorillonite was prepared by treating the starting montmorillonite with 3% NaCl aqueous solution for 24 h three times. Na-montmorillonite was stirred in PVP aqueous solution with a weight ratio of PVP to Na-montmorillonite of 1.5, so that an Na-montmorillonite-PVP intercalation compound was formed. After being washed thoroughly with methanol, the compound was dried, ground and dispersed in water again.An aqueous solution of the viologen was then added to the suspension and the mixture was stirred for 48 h. The amount of viologen added was equivalent to the cation exchange capacity of the original Na-montmorillonite (1 19 mequiv. per 100 g clay). Thus the viologen was intercalated into the interlamellar space of the montmor- illonite. These viologen-montmorillonite-PVP intercalation compounds were washed with methanol in order to remove the viologens adsorbed on the outer surface until the washings were nearly free of viologen. Analysis All synthesized samples were characterized by powder X-ray diffraction and i.r. spectroscopy. X-Ray diffraction analysis was performed on a Rigaku Denki RAD 11-A instrument (Fe Ka radiation, Mn filter).Infrared spectra were recorded on a Shimadzu IR-400 spectrometer using a KBr disc. Photochromic Behaviour Self-supporting films of viologen-montmorillonite-PVP were prepared by drying the samples spread on acrylic plates. The visible spectra of these films were recorded on a Shimadzu UV-210A spectrometer before and after irradiation by a 100 W Hg lampH. Miyata, Y. Sugahara, K. Kuroda and C. Kato 1853 (Rikoh Kagakusangyo). An acrylic plate was also set at the reference beam line. The spectral changes were then observed. Irradiation was performed without using filters and the distance from the light source was 10 cm. In addition, the variation of the absorbance with time in both colour development and fading processes was observed by the use of an n-heptyl viologen-montmorillonite-PVP film.For the colour-developing process, the increase in the absorbance at 607 nm was measured as a function of the irradiation time every 10 s for 5 min. For the colour-fading process the irradiated film was left in the spectrometer and the decrease in the absorbance was recorded with time. This colour development and fading process was repeated three times. Results and Discussion Characterization of the Intercalation Compounds Na- Mon tmor illon ite- P VP Intercalation Compound The basal spacing (do,,,) of the Na-montmorillonite-PVP intercalation compound prepared in this study was 23.4 A, indicating an expansion of the interlayer space by 13.8 A. (The basal spacing of Na-montmorillonite was 9.6 A.) In addition, the i.r.spectrum of the intercalation compound showed characteristic bands due to PVP (2955 cm-l, C-H stretching; 1425-1465 cm-l, C-H bending; 1290 cm-l, C-N stretching),12 even though the sample had been thoroughly washed with methanol [fig. 1 (a) and (b)]. These results confirmed the formation of an Na-montmorillonite-PVP intercalation compound. Judging from the expansion of 13.8 A of the interlayer region, it was thought that the polymer chains of PVP were probably adsorbed in a bimolecular layer because the molecular thickness of the vinyl pyrollidone monomer is 5.6 A and the polymer chain is usually bent. The ratio of PVP in the intercalation compound was 30 wt % , and was determined by thermogravimetry . Vio logen- Mon t m o r illon it e- P VP In t e r cala t ion Compounds The X-ray diffraction patterns of the synthesized viologen-montmorillonite-PVP inter- calation compounds were very diffuse for all four different samples in comparison with that of the Na-montmorillonite-PVP compound.Their doe, values are shown in table 1, and are lower than that of the Na-montmorillonite-PVP compound. Nevertheless, they are larger than those of four corresponding viologen-montmorillonite compounds. The numbers of guest molecules in the four synthesized intercalation compounds are summarized in table 2. The numbers of viologens in the intercalation compounds were calculated by subtracting the amount of viologen contained in the washings from that of the original viologen. The amount of viologen in the washings was measured by the use of a u.v.-visible spectrometer.In addition, the total amount of organic substances was obtained from the results of thermogravimetry. By using these data, the amount of PVP in the interlayer space and the ratio of PVP to viologen were calculated. As table 2 shows, the amount of viologen adsorbed reached the level of at least 70-75% of the cation exchange capacity of montmorillonite. Table 2 also shows that the amount of PVP decreased on complexation with the viologens. If the viologens were adsorbed only on the external surface of the montmorillonite and were not intercalated into the interlayer space, the large amount of viologen adsorbed in this experiment cannot be explained. Furthermore, we have found that PVP could not be de-intercalated from the Na-montmorillonite-PVP intercalation compound by washing with methanol, which is a good solvent of PVP.All these results support the hypothesis that the viologens were intercalated into the interlamellar space of the montmorillonite. The broadening of the powder X-ray peaks and the decrease of the do,,, values can be explained by the partial desorption of PVP accompanied by the intercalation of the viologens.1854 Montmorillonite- VioIogen IntercaIation Compounds Table 1. Basal spacings of the complexes sample propyl viologen-montmorillonite heptyl viologen-montmorillonite phenylmethyl viologen-montmorillonite phenylethyl viologen-montmorillonite Na-montmorilloni te-PVP propy 1 viologen-mon tmorilloni te-PVP heptyl viologen-montmorillonite-PVP phenylmet hyl-mon tmorilloni te-PVP phenylethyl viologen-montmorillonite-PVP ~~ 13.2 3.6 14.2 4.6 15.1 5.5 15.1 5.5 23.4 13.8 22.4 12.8 19.2 9.6 21.5 11.9 22.6 13.0 a Adool = doo, -9.6 A.Table 2. Amounts of adsorbed organic substances sample viologen PVP /mmol per 100 g clay /g per 100 g clay Na-mon tmorillonite-PVP I heptyl viologen-montmorillonite-PVP 42 propyl viologen-montmorillonite-PVP 43 phenylmethyl viologen-montmorillonite-PVP 45 phenylethyl viologen-montmorillonite-PVP 43 43 24 22 24 32 The i.r. spectra of these compounds showed additional absorption peaks ascribed to viologens (1640 and 1560 cm-l, ring vibration of pyridine ring) in addition to the absorption peaks of the Na-montmorillonite-PVP intercalation compounds, even though the ternary intercalation compounds were washed sufficiently.The i.r. spectrum of the compound using n-heptyl viologen is shown in fig. 1 (d). Other compounds showed similar spectra. From these results it was apparent that viologen-montmorillonite-PVP intercalation compounds had been formed. Photochromism of the ViologewMontmorillonite-PVP Intercalation Compounds The colour of the intercalation compounds changed to blue on irradiation by a mercury lamp. When the montmorillonite-viologen intercalation compound which did not contain PVP was irradiated, it did not show a colour change. The degree of colour development depended on the form of the sample; oriented films turned blue very homogeneously, while colour development was not remarkable in the powder or bulk samples. Fig. 2 shows variations in the visible spectra of a film of the n-heptyl viologen- montmorillonite-PVP intercalation compound before and after colour development.After colour development the characteristic absorption bands due to the viologen radical cation were observed around 610 and 400 nm. The formation of the radical was also confirmed by e.s.r. spectroscopy. The absorption spectra of the four different com- pounds after irradiation did not greatly differ. As described above, in the photochromism of viologen in a PVP matrix the radical cation is thought to be formed by electron transfer from the counter-anion to the viologen di-cation. However, in our study the intercalated viologen was adsorbed as the di-cationicH. Miyata, Y. Sugahara, K. Kuroda and C. Kato 1855 4000 2400 1400 650 Fig.1. 1.r. spectra of (a) Na-montmorillonite, (b) Na-montmorillonite-PVP intercalation com- pound, ( c ) n-heptyl viologen and ( d ) n-heptyl viologen-montmorillonite-PVP intercalation compound. wavenum ber/ cm -l form in the interlayer space. Therefore, it is not possible to explain the formation of the radical cation by electron transfer from the counter-anion. Thus it was thought that PVP acted as an electron donor. The absence of Br- in the intercalation compounds was confirmed by X-ray fluorescence spectroscopy in almost all compounds, while only a trace amount of Br- was detected in the phenylmethyl viologen system. However, the precise mechanism is still ambiguous. In order to check the effective wavelengths of the light emitted from the high-pressure mercury lamp, the u.v.-visible spectrum of n-heptyl viologen in a PVP matrix was measured.The spectrum showed a maximum at 268 nm. However, the absorption edge nearly reached the visible region. Therefore, every wavelength present in the U.V. light from the mercury lamp seemed to be responsible for the photochemical reaction. A broadening of the absorption range has been reported when viologens were incorporated into the gelatine and cellulose matrices.5* The p ho t ochromic be haviour of the n-heptyl viologen-mon tmorillonite-PVP inter- calation compound with time was also observed (fig. 3). In the colour-development process a rapid increase in absorbance just after irradiation was seen and the absorbance became constant after 5 min even though irradiation was continued.Fig. 4 shows the variation of absorbance over the first 7 min. The time required to reach a constant value depended on the thickness of the film. In the colour-fading process the absorbance of the sample began to decrease when the light irradiation was stopped. However, the rate1856 1.4- 1.2 1.0- gd8- -e s 2 Mon tmorillonite- Viologen Intercalation Compounds , I I) 11 I ' I1 I 1 1 1 . ..; j ; I I I I I I I I I I I I I \ I I I i A 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 time/h Fig. 3. Photochromic behaviour of n-heptyl viologen-montmorillonite-PVP intercalation compound: 0, lamp on; A, lamp off.H. Miyata, Y. Sugahara, K. Kuroda and C. Kato 1857 OAO - d) 5 % 4 0.30 - 0.181 1 , 60 120 180 240 300 360 420 time/s Fig. 4.The colour-development process of n-heptyl viologen-montmorillonite-PVP intercalation compound: A, lamp off. of colour fading was remarkably slower than that of the colour-development process. It took ca. 4-6 h to obtain the completely faded sample even in the thin film used in this measurement. The thicker films required much more time for the colour fading, and sometimes the blue colour remained even after several days. The rate of colour fading of the viologen in the intercalation compound was very slow compared with that in the PVP matrix. This can be explained by the prevention of contact between the viologen radical cation formed in the interlamellar space and the oxidizing agent (atmospheric oxygen), since the colour-fading process in the PVP matrix was explained by the oxidation caused by oxygen in air.The rate and yield of reduced viologen may be influenced by the amount of co-intercalated PVP. However, the amount of PVP could not be controlled experimen- tally because the intercalated PVP was partly de-intercalated by the insertion of viologen and also could not be re-intercalated. Conclusion Intercalation compounds of montmorilloni te containing both PVP and viologens in the interlayer space have been synthesized. These intercalation compounds showed photochromism from colourless to blue by irradiation by U.V. light. A mechanism has been considered in which PVP acted as an electron donor for the reduction of the viologens, since bromide ion in the original viologens was not present in these intercalation compounds. The present work was partially supported by a Grant-in-Aid for Special Project Research in Molecular Assemblies (no. 582 18033, 601 04004) from the Japanese Ministry of Education, Science and Culture.1858 Mon tmoril1o;i ite- Viologen Intercalation Compounds References 1 G. H. Brown, Photochromism (Wiley-Interscience, New York, 197 1). 2 M. Yamana and T. Kawata, Nippon Kagaku Kaishi, 1977,941. 3 H. Kamogawa, T. Masui and M. Nanasawa, Chem. Lett., 1980, 1145. 4 H. Kamogawa, T. Masui and S . Amemiya, J . Polym. Sci., Polym. Chem. Ed., 1984, 22, 383. 5 M. Kaneko, Y. Imamura, K. Hayashi and A. Yamada, Kobunshi Ronbunshu, 1982,39, 665. 6 M. Kaneko and A. Yamada, Makromol. Chem., 1981, 182, 1 1 1 1 . 7 N. Kambe and T. Yamada, Muter. Res. SOC. Symp. Proc., 1984, 24,467. 8 0. D. Philen Jr, S. B. Weed and J. B. Weber, Clays Clay Miner., 1971, 19, 295. 9 Y. Soma and M. Soma, Kokuritsu Kogai Kenkyusho Kenkyu Hokoku, 1982, 36, 227; Chem. Abstr., 97, 168632j. 10 M. Raupach, W. M. Emerson and P. G . Slade, J . Colloid Interface Sci., 1978, 69, 398. 1 1 0. Poizat, C. Sourisseau and Y. Mathey, J . Chem. Soc., Faraday Truns. I , 1984, 80, 3257. 12 C. W. Francis, Soil Sci., 1973, 40, 1 15. Paper 6/ 188 1 ; Received 22nd September, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301851

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Furuduy Trans. I, 1987,83, 1859-1868 Reactive Intermediates for the Ethene Homologation Reaction on Molybdena-Silica Catalysts Katsumi Tanaka" Research Institute for Catalysis, Hokkaido University, Kita-ku, Sapporo 060, Japan Ken-ichi Tanaka The Institute for Solid State Physics, The University of Tokyo, 7-22-1 Roppongi, Minato-ku, Tokyo 106, Japan Ethene has been selectively homologated to propene (3C=C + 2C=C-C) on reduced molybdena-silica at room temperature, The initial homologation activity was enhanced by grafting the methylene species on the surface with SnMe,. Deuterated methane and cyclopropane were formed by reacting ,H, with the surface on which the ethene homologation had been carried out. These results clearly suggest that the ethene homologation proceeds through metal methylidene and metallocycle intermediates.The homologation ac- companied the hydrogen exchange of ethene, which was considered to be a similar function to Q-elimination of hydrogen from the metallocycle. [13C,]propene was dominantly formed by the reaction of [13C,]ethene on the CH, furnished surface. Taking account of the fact that the ethene metathesis proceeds lo3 times faster than the homologation, it was concluded that the propagating species for the homologation is different from that for the metathesis. The homologation of ethene giving propene was discovered on Mo(CO),-derived y-Al,O,-supported catalysts by O'Neill and Rooney.' Since then the reaction has been assumed to be caused by the mononuclear metal alkylidene and metallocyclobutane,2* which are accepted as the key intermediates for the alkene metathesis rea~tion.~ In the field of organometallic chemistry, d i - i r ~ n , ~ diosmium6 and dicobalt p2-CH2 complexes7 react with ethene to form propene selectively. These reactions also prove the important role of metallocycle intermediates. The methylidene species (CH,) has been implicated in the Fischer-Tropsch (FT) hydrocarbon synthesis reaction,8-10 and the comparison of this reaction with homologation37 9 9 l1 is an interesting subject.From the spectroscopic point of view, bridged and terminal CH, species were detected on Ru(O01)12 and Fe,13 respectively, which are related to the FT reaction. It is noted that l4? l5 and ruthenium-117 l5 based heterogeneous FT catalysts also cause alkene homologation. A similar mechanism is proposed to the reactions on the basis of a high normal to is0 ratio in the butene pr0duced.l' Compared to metallic catalysts, isobutene is selectively formed in the homologation of propene on supported oxide catalysts.ls It is interesting to point out that the catalytic alkene homologation proceeds in the presence of hydrogen on Fe and Ru catalysts,ll* l7 which is compared to the self ethene homologation in the absence of hydrogen on these ~ata1ysts.l~ Recently it was suggested that the active species for ethene metathesis is different from that for homologation.This is based on the fact that the addition of KOH to an MoO,/SiO, catalyst enhances the ethene homologation activity more than 10 times without affecting metathesis activity and selectivities.ls In addition, it was directly proved that ethene homologation proceeds via a different intermediate from the metathesis by detecting 1 ,3-[13C,]propene in the reaction of [13C,]ethene with the CH, grafted surface.lg In this paper, first the requisite function of 18591860 Ethene Homologation catalyst for the homologation and secondly the reactive intermediate species for the reaction are reported.Experimental The starting material of molybdena-silica was obtained by immersing silica (Kiesel Gel 60, Merck) into an aqueous solution of ammonium paramolybdate. The amount of Mo was 2.8 atom % to silica. Then it was heated on a hot plate while stirring to remove water, and was dried in an oven at 120 "C for 12 h. The powder was heated slowly from room temperature to 500 "C under vacuum and then subjected to either oxidation or reduction.The fully oxidized Mo03/Si0, was obtained by oxidizing it with 0, at 500 "C for 1 h. Reduced MoO,-,/SiO, and MoO,/SiO, catalysts were obtained as follows. The molybdenum-silica evacuated at 500 "C was reduced with H, (ca. 100 Torr, 1 Torr z 133.3 Pa) at 500 "C for 1 h to obtain the fully reduced MoO,/SiO, and water was removed with a trap at - 196 "C. While the partially reduced MoO,-,/SiO, was prepared by reoxidizing the MoOJSiO, with a 1 : 1 mixture of N,O and H, (ca. 150 Torr) at 200 "C for 1 h. The same treatment was carried out to obtain the M o O ~ ~ ~ - ~ ~ ~ / T ~ O , which catalyses alkene metathesis without the accompanying hydrogen exchange and the double-bond shift reactions.20 Finally, each catalyst was evacuated at 500 "C for 1 h and cooled to room temperature under vacuum.The catalyst at this stage is denoted as the virgin catalyst. The vapour of tetramethyltin, SnMe, ( 5 Torr), was diluted with He (ca. 60 Torr) and the gaseous mixture was reacted on 0.5 g of virgin catalyst at room temperature normally for 30 min, then evacuated at the same temperature for 30 min. A glass circulation system with a volume of ca. 260 cm3 was used for the catalyst treatment and the ethene homologation. Analysis of products was performed with an on-line gas chromatograph (Hitachi 163 with f.i.d. and t.c.d.), while ,H and 13C distributions were measured with a mass spectrometer (Hitachi RMU-6) with ionization voltages of 70 V for methane and 10 V for ethene and propene.Reactant gases were all commercially available (ethene Takachiho; [2H,]ethene, MSD; [13C,]ethene, Amersham Int.) with purity > 99.9%, and were used without further purification. Hydrogen was purified through Pd-Ag thimbles and deuterium was used after passing through molecular sieve 5A which had been activated by evacuation at 350 "C. Results and Discussion The Function of the Catalyst for Ethene Homologation As shown in fig. 1, when an equimolar mixture of [l2CC,] and [13C,]ethene was added to an MoO,/SiO, catalyst at room temperature, [l3C1]ethene and propene were formed by metathesis and homologation, respectively? Propene was formed after an induction period and was composed of [l3C0], [l3C1], [l3CC,] and [13C3] isomers.The induction period was not observed when the homologation was repeated on the catalyst following evacuation at room temperature. Therefore it is concluded that the induction time is attributed to the period in which an active intermediate is produced. The partially 2C=13C met at hesis propene homologation. / c=c + 13c=13c reduced MoO,-,/SiO, catalyst also showed both ethene homologation activity and metathesis activity at room temperature. Note that the homologation rate is ca. lo3 times slower than the metathesis rate and that homologation on the MoOJSiO, catalyst (0) proceeded twice as rapidly as on the MoO,-,/SiO, catalyst (a); however,K. Tanaka and K. Tanaka 1861 c 5 % B E 00 a 0 time/min Fig. 1. Homologation and metathesis of ethene using a 1 : 1 mixture of [l3C,,] and [13C2]ethene on the MoOJSiO, catalyst (0, .), the MoO,-,/SiO, catalyst (0, m) and on the MoO,/SiO, catalyst in the presence of 6 Torr CO (a, a) at room temperature.Catalyst, 0.5 g; total ethene pressure, 10-15 Torr. metathesis on the Moo,-,/SiO, catalyst (H) occurred faster than on the MoOJSiO, catalyst (a). A fully oxidized MoO,/SiO, catalyst did not catalyse either homologation or metathesis at room temperature. From these results it can be concluded that homologation activity is increased as the catalyst is reduced, whereas metathesis activity reaches a maximum on the partially reduced molybdena-silica. It is reported that metathesis activity of propene is not obtained on the MoOJTiO, catalyst at room temperature; however, the activity has a maximum on the M00,~,~,~,/Ti0, catalyst and decreases as the reduction of molybdenum oxide proceeds.18 The result obtained on molybdena-silica holds for molybdena-titania.The distribution of products formed in ethene homologation on the MoOJSiO, catalyst at room temperature is shown in table 1. After 180 min, propene was formed selectively. As the reaction proceeded the amount of propene increased and a small amount of isobutene and but-2-ene was found (2700min). Note that formation of but-2-ene was limited during ethene homologation because the MoOJSiO, catalyst is quite active for the metathesis of propene.18 It should be mentioned that isobutene is formed by the homologation of propene since the MoOJSiO, does not catalyse the normal to iso-carbon-chain isomerization at room temperature.The formation of CH * c=c-c c-c-c CH, I I I II i c=c I I C - homologation + metathesis c=c+c-c-c-c - - - - - - - isobutene strongly suggests that a metallocycle intermediate is involved in alkene homologation on the MoO,/SiO, catalyst. This is because the titanium-methylene complex in the presence of trimethylaluminium, the so-called Tebbe complex, reacts with propene to form isobutene via the metallocycle intermediate.21 If this is the case, it is interesting to know whether but-1-ene or isobutene is formed in the homologation reaction of propene, because but-1-ene is produced by puckering of the metallocycle with1862 Ethene Homologation Table 1. Distribution of the product formed in ethene homolog- ation on the MoOJSiO, catalyst at room temperature - composition (% ) time/min C,H, C,H, l-C,H, iso-C,H, t-C,H, c-C,H, 40 99.9 0 0.1 0 0 0 180 96.6 3.1 0.3 0 0 0 2700 51.6 44.4 0.1 1.0 2.2 0.8 Catalyst, 0.5 g; ethene pressure, 14.5 Torr.a 1 -methyl substituent as suggested in Fe/MgO catalysts,14 while isobutene is produced by puckering of the metallocycle with a 2-methyl substituent. To obtain information about the active site for ethene homologation, a CO adsorption experiment was carried out. When an equimolar mixture of [l2CC,] and [13C2]ethene was added on the MoO,/SiO, catalyst in the presence of CO, the metathesis proceeded (a) with the same rate as observed in the absence of CO (a), whereas the formation of propene (0) was retarded entirely (fig. 1). Accordingly, we can conclude that the active site for homologation differs from that for metathesis.The requisite function of the homologation-active site is considered. When a 1 : 1 mixture of [,H,] and [,H,]ethene reacted on the MOO,-,/SiO, catalyst at room temperature, propene (0) was slowly formed and [,H,]ethene (a) was formed rapidly by metathesis (fig. 2). Deuterium exchange occurred simultaneously and an equilibrated mixture of [,H,], [2HI], [,H,], [2H,] and [,H,]ethene was obtained in 40 min. When the same reaction was performed on the MoOJSiO, catalyst at room temperature hydrogen exchange occurred so rapidly that equilibrated deuterium isomers of ethene were obtained in less than 1 min. These results suggest that homologation accompanies hydrogen exchange of ethene, while metathesis does not necessarily proceed concurrently with hydrogen exchange. Ethene metathesis did not proceed on MoO,/SiO, at room CH,=CH, [,H,] + CD,=CD2 [,H4] temperature.However, when a 1 : 1 n CH,=CD, [,H2] CHD=CH, [,HI] CHD=CHD [,H2] CHD=CD, [2HS’ Kture of [,H,] and [,€ metathesis hydrogen exchange lethene was contactecl wit h the MoO,/SiO, catalyst treated with SnMe, at room temperature only [ ,H ,] ethene was formed as shown in fig. 3. This reaction proceeds catalytically since ,H was not diluted in the reaction and metathesis activity was also seen when the reaction was repeated. It is significant that ethene metathesis (CH,=CH, + CD,=CD, -+ CH,=CD,) proceeds without the accompanying hydrogen exchange of ethene on the MoO,/SiO,-SnMe, catalyst.Tanaka and Tanaka reported that the entirely metathesis-inactive MoO,/TiO, catalyst is changed into a super-active alkene metathesis catalyst when it is treated with SnMe,; however, it does not catalyse hydrogen exchange and double-bond shift reactions.22 It is assumed that the activity is caused by the methylene species grafted on the surface from SnMe, by the reaction 2CH, -+ CH, + CH,.,, If this is the case in the Mo03/Si02-SnMe, system, the result of fig. 3 implies that the methylene-grafted surface can participate in the propagation step of ethene metathesis in which Mo=CH, (Mo=CD,) reacts with [,H,]ethene ([,H,]ethene) to form metallocyclobutane and itK. Tanaka and K. Tanaka 1863 time/min Fig. 2. The reaction of a 1 : 1 mixture of [,H,] and [2H,]ethene on the Moo,-,/SiO, catalyst at room temperature.Catalyst, 0.5 g; total ethene, 12.8 Torr; reactant ethene composed: [,H,], 56.3%, [,H3], 2.6% and [,H,], 41.1%. I 60 IhO Id0 5* time/min Fig. 3. Ethene metathesis using a 1 : 1 mixture of [,H,] and [,H,]ethene on the MoO,/SiO, catalyst activated with SnMe,. Catalyst, 0.5 g ; total ethene, 12.0 Torr; reactant ethene composed: [,H,], 45.4% ; [",I, 1.8% and [,H,], 52.80/, . decomposes to give [,H,]ethene and Mo=CD, (Mo=CH,) repeatedly. In fact, a small amount of methane was formed during the reaction of SnMe, on the MoO,/SiO, Mo=CH, + CD,=CD, L 7 Mo=CD, + CH,=CD, catalyst. In conclusion, metathesis does not need to accompany hydrogen exchange in the ethene molecule and homologation should take place concomitantly with hydrogen exchange via a metallocycle intermediate.This is because the reaction of CH, with [,H,]ethene must yield CH,=CD-CD, and CD,=CD-CH,D through p-elimination (1,2-hydrogen shift) of the metallocycle as mentioned later.1864 Ethene Homologat ion tirne/min Fig. 4. Ethene homologation on the SnMe,-treated MoO,/SiO, catalysts at room temperature. Catalyst, 0.5 g; ethene reacted, 24.5 Torr; methane formed during the activation with SnMe,, 0.216 Torr (O), 0.083 Torr (A) and 0.014 Torr (m). The Reactive Intermediate for Ethene Homologation Ethene homologation was carried out on the MoOJSiO, catalyst treated with SnMe,. A small amount of methane was formed by reaction with SnMe, and the homologation activity was enhanced as the amount of methane increased (fig.4). In general, the formation of methane from metal methyl reagents suggests that a-hydrogen abstraction of a methyl group23. 24 occurs to give CH, and the hydrogen formed subsequently reacts with another methyl group to form CH, on the surface.25 If this is the case for the MoO,/SiO, catalyst, the enhancement of initial homologation activity may be attributed to the methylene species grafted on the surface. It is consistent with the experimental result that the induction time disappeared on the SnMe,-treated MoOJSiO, catalyst. The following experiment was carried out to determine the species formed by contacting the surface with SnMe,. The SnMe,-treated MoOJSiO, catalyst was evacuated at 400 "C for 1 h and was repeatedly reacted with D, at 80 "C, then the deuterium distribution of methane formed was measured.The formation profile of [2H]-containing methane isomers is shown in fig. 5. [2H,]methane was formed rapidly in the early stages of the reaction and reached saturation, while E2H,]methane was formed slowly and the amount increased linearly as the reaction proceeded. This result implies that both CH, and CH, species are present on the surface and the reactivity of CH, (+ ,H,) is higher than that of CH, (+ ,H) with respect to the reaction with deuterium. The formation of [,Holmethane infers that a-hydrogen abstraction of methyl groups still takes place during the reaction. The presence of and [,H,]methane shows that the CH, species formed on the surface are easily changed to CHD and CD, by exchange reaction with deuterium, and they pick up D, to form the corresponding deuteromethane isomers.The reactive intermediate formed during ethene homologation on the MoO,/SiO, catalyst was studied. To detect such species, D, was added to the MoO,/SiO, catalyst at 80 "C on which ethene homologation had been carried out for 12 h at room temperature and which had been evacuated at 80 "C for 1 h. The distribution of products and that of deuteromethane isomers formed in the reaction with D, are shown in table 2. When D, was reacted for 1 h (run l), methane, cyclopropane as well as C, and C, compounds were formed. It is significant that the methane formed was composed of [,H,], [,H3] and [2H4] isomers. During the reaction, desorbed cyclopropane, C , and C , compounds were removed from the gas phase with a liquid-nitrogen trap, so that methane is considered to be a primary product formed by the hydrogenation of carbonaceousK .Tanaka and K. Tanaka I b 8 12 16 20 1865 time/h Fig. 5. The reaction of D, on the SnMe,-treated MoO,/SiO, catalyst at 80 "C. Prior to the reaction, the catalyst had been evacuated at 400 "C for 1 h. Catalyst, 1.0 g ; pressure of D,, 8.4 Torr. Table 2. Distribution of the products and of deuteromethane isomers formed in the reaction with D, on the MoO,/SiO, catalyst at 80 "C, on which ethene homologation had been carried out for 12 h at room temperature and the surface had been evacuated at 80 "C for 1 h amount ,H distribution (% ) produced product /Torr [",I [2H,] [2H2] [,H,] [,H,] run 1 : 14.2 Torr D, was contacted on the surface for 1 h methane 2 x 0 0 51.8 29.5 18.7 cyclopropane 7 x loe2 - - - - - - - - I c, 2x10-2 - c, 2x10-, - - - - - run 2: 28.5 Torr D, was contacted for 15 h after run 1 methane 1 x lo-, 0 0 57.3 26.7 10.6 cyclopropane 3 x lop3 - - - - - c, 2 x lo-, - - - - c, 5~ 10-3 - - - - - - species on the surface.A similar result was obtained in run 2, where D, was reacted at 80 "C for 15 h. Note that cyclopropane was not detected during ethene homologation but was formed by contact with D, on the homologation-active surface. This result implies that a metallocycle species exists on the surface. In addition, the metallocycle species can be conceived as the reactive intermediate in the homologation of ethene. However, it is crucial to rule out the possibility of cyclopropane formation related to the metathesis-active intermediate, because mononuclear alkylidene complexes yield cyclopropanes in metathesis 27 The formation of [,H,], [,H3] and [,H,]methane (table 2) indicates that a similar reaction sequence occurs, CH, -+CHD-+CD,, as was observed in fig.5 and they react with D, to give the respective deuterium-containing methane isomers. This result clearly proves the presence of CH, species on the surface. Note that a mononuclear alkylidene complex reacts with H, producing the corresponding alkane,28 while a p2-CH, complex also shows reductive1866 t ' Ethene Homologation '0 0.2 0.4 0.6 0.8 1.0 conversion of homologation (%) Fig. 6. The fraction of 13C in propene formed in the homologation of 13C-labelled ethene on the SnMe,-treated molybdena-silica catalysts.0, 9 Torr of [13C,]ethene on the MoO,/SiO, catalyst treated with 50 Torr SnMe, for 90 min (4.3 x lo-, Torr CH, formed). @, 12 Torr of a 1 : 1 mixture of [l3Cz] and [12C,]ethene on the MoO,-,/SiO, catalyst treated with 5 Torr SnMe, for 240 min (7.4 x Torr CH, formed). elimination with H, to form methane.5g 29 Accordingly it can be concluded that the formation of methane and cyclopropane implies the presence of metal methylidene and metallocycle species as the reactive intermediates for both homologation and metathesis of e t hene. To discriminate significant species for homologation and metathesis, [13C,]ethene was reacted on the SnMe,-treated MoOJSiO, catalyst at room temperature and the 13C fraction in propene was compared to that in ethene.As shown in fig. 6, the 13C fraction in propene gradually increased and reached the same level in ethene as the reaction proceeded. The 13C content extrapolated to zero conversion was 60%, which means that one 12CH, species grafted from SnMe, is involved in propene. A similar result was obtained in the reaction of a 1 : 1 mixture of [12C2] and [13C,]ethene on the SnMe,-treated MoO,-,/SiO, catalyst at room temperature, The 13C content in propene was 33 % in the initial stage of the reaction and it slowly increased to that in ethene as the reaction proceeded. The increase of 13C fraction in propene is interpreted as follows: the CH, species supplied from SnMe, on the surface initially participate in the homologation and 13CH, species can be supplied on the subsequently formed vacant site from [13C,]ethene, which is followed by the successive reaction with [13C,]ethene to give [13C,]propene.It should be remembered that ethene metathesis proceeds more than lo3 times faster than homologation. Therefore, the metathesis-active Mo=CH, species supplied from SnMe, should be changed instantaneously to Mo=13CH, species by the contact with [WJethene, and at the same time the 13C content of Mo methylidene should soon be equilibrated to be equal to that in ethene. If the Mo=CH, species which participate in metathesis are also involved in homologation, the 13C fraction in propene should be equal to that in ethene from the initial stage of the reaction. Accordingly, it can be concluded that the reactive intermediate for ethene homologation is clearly different from that for metathesis.In conclusion, a probable mechanism for the ethene homologation may be as follows, although the initiation step, methylene formation, is uncertain at the present time. Methylidene species are not necessarily limited to p2-CH,. However, the latter is thoughtK. Tanaka and K. Tanaka 1867 to be a reasonable species according to the stoichiometric reaction in organometallic complexe~.~-~ The dislocation of the methylidene carbon has been proved in the same D D I CD,=CD-*CH2D Mo / x I Mo Mo + CD,=CD, -* Mo I I Mo CD,=*CH-CHD,. Mo system.ls In the present paper, it is established that the homologation of ethene proceeds via a metallocycle intermediate. Hydrogen must transfer through 8-elimination when the metallocycle decomposes to give propene,'? 1 4 9 21 and this is probably related to the function of catalyst that causes the hydrogen exchange of ethene.References 1 P. P. ONeill and J. J. Rooney, J. Am. Chem. SOC., 1972,94,4383. 2 C. O'Donohoe, J. K. A. Clarke and J. J. Rooney, J. Chem. Soc., Faraday Trans. I, 1980,76, 345. 3 J. J. Rooney, J. Mol. Catal., 1985, 31, 147. 4 K. J. Ivin, Olefin Metathesis (Academic Press, New York, 1983); I. J. Rooney and A. J. Stewart, Catalysis (Specialist Periodical Reports, The Royal Society of Chemistry, London, 1977), vol. 1, p. 277; N. Calderon, J. P. Lawrence and E. A. Ofstead, Ado. Organomet. Chem., 1979,17,449; R. H. Grubbs, Comprehensive Organometallic Chemistry, ed. G. Wilkinson (Pergamon Press, Oxford, 1982), vol.8, p. 499; C. P. Casey, Reactive Intermediates, ed. M. Jones and R. A. Moss (John Wiley, Chichester, 1985), vol. 3, p. 109. 5 C. E. Sumner Jr, P. E. Riley, R. E. Davis and R. Pettit, J. Am. Chem. SOC., 1980, 102, 1752. 6 K. M. Motyl, J. R. Norton, C. K. Schauer and 0. P. Anderson, J. Am. Chem. SOC., 1982,104,7352. 7 K. H. Theopold and R. G. Bergman, J. Am. Chem. SOC., 1980, 102, 5694; 1983, 105,464. 8 R. C. Brady I11 and R. Pettit, J. Am. Chem. SOC., 1980, 102, 6181; 1981, 103, 1287. 9 K. Tanaka, I. Yaegashi and K. Aomura, J. Chem. SOC., Chem. Commun., 1982,938. 10 W. A. Hermann, Angew. Chem., Int. Ed. Engl., 1982, 21, 117. 1 1 M. Leconte, A. Theolier, D. Rojas and J. M. Basset, J. Am. Chem. SOC., 1984, 106, 1141. 12 P. M. George, N. R. Avery, W. H. Weinberg and F. N. Tebbe, J. Am. Chem. SOC., 1983,1oS, 1393. 13 S-C. Chang, Z. H. Kafafi, R. H. Hauge, W. E. Billups and J. L. Margrave, J. Am. Chem. SOC., 1985, 14 F. Hughes, B. Besson and J. M. Basset, J. Chem. SOC., Chem. Commun., 1980, 719. 15 Y. Iwasawa and M. Yamada, J. Chem. SOC., Chem. Commun., 1985, 675. 16 T. Yamaguchi, S. Nakamura and K. Tanabe, J. Chem. SOC., Chem. Commun., 1982,621. 17 R. A. Strehlow and E. C. Douglas, J. Chem. SOC., Chem. Commun., 1983, 259. 18 K. Tanaka and K. Tanaka, J. Chem. SOC., Chem. Commun., 1984, 1626. 19 K. Tanaka, K. Tanaka, H. Takeo and C. Matsumura, J. Chem. SOC., Chem. Commun., 1986,33. 107, 1447.1868 Ethene Homologation 20 K. Tanaka, K. Miyahara and K. Tanaka, J. Mol. Catal., 1982, IS, 133. 21 F. N. Tebbe, G. W. Parshall and G. S. Reddy, J. Am. Chem. SOC., 1978, 100, 361 1 . 22 K. Tanaka and K. Tanaka, J. Chem. SOC., Chem. Commun., 1984, 748. 23 N. J. Cooper and M. L. H. Green, J. Chem. SOC., Chem. Commun., 1974, 761. 24 J. K. Kochi, Organometallic Mechanisms and Catalysis (Academic Press, New York, 1978), chap. 12. 25 E. L. Muetterties, Inorg. Chem., 1975, 14, 951. 26 C. P. Casey and T. J. Burkhardt, J. Am. Chem. SOC., 1974, 96, 7808. 27 R. H. Grubbs, D. D. Carr and P. L. Burk, Organotransition Metal Chemistry, ed. V. Ishi and 28 C. P. Casey and S. M. Newmann, J. Am. Chem. SOC., 1977,99, 1651. 29 W. A. Hermann, Adv. Organometal. Chem., 1982, 20, 159. M. Tsutsui (Plenum Press, New York, 1975), p. 135. Paper 611955; Received 3rd October, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301859

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 1869-1877 An Electron Spin Resonance Study of Triplet Radical Pairs in Single Crystals of X-Irradiated L-Ascorbic Acid at 77 K Jerzy T. Masiakowskit and Anders Lund" Department of Physics and Measurement Technology, Linkoping University, S-58183 Linkoping, Sweden Single crystals of L-ascorbic acid have been irradiated and investigated at 77 K. The signals due to triplet radical pairs were observed. Three radical pairs, separated within the pair by 5.07, 6.28 and 8.46 8, were interpreted. The radicals which form the pairs are of the same type as those studied previously at room temperature as free radicals and published elsewhere. The location of the three types of radical pairs in the crystal was found by the comparison of the experimental results with crystal data.There are some signals arising from other radical pairs. It was not possible to follow the lines over a sufficient range of orientations to make an interpretation. Radiation-induced radicals in single crystals of L-ascorbic acid have been studied previously at room temperat~rel-~ and at 77 K.l It was found that at least four different types of radicals form under radiation in this temperature range. In the present work we have observed signals arising from triplet radical pairs. The pairs were formed by X-irradiation at 77 K. Three of them are interpreted as being due to similar radical pairs which differ only in relative positions and separations in the crystal. All pairs are formed by two identical paired radicals, the same for all three cases.The radicals exhibit hyperfine splittings and spin-flip satellites. The spin-flip-forbidden transitions in the case of triplet radical pairs are discussed. The trapping sites of radical pairs in the crystal were found by comparison of the experimental results with the values expected from the crystal data. Experimental Single crystals of L-ascorbic acid were grown by slow evaporation from aqueous solution at 5 "C. The crystal structure is monoclinic with space group P21.4v There are four molecules per unit cell: two AB dimers related by a twofold axis of symmetry. Within the dimer, the A and B molecules are related by a pseudo-axis of symmetry parallel to the twofold axis. The furan rings of all four molecules are almost exactly coplanar and parallel to the ac crystal plane.For e.s.r. measurements an a*bc(a* = b x c ) reference axis system, the same as in the earlier study,3 was chosen. Irradiation was made with an X-ray tube with an Au anode operated at 70 kV and 20 mA. The dose rate at the position of the sample was ca. 1 Mrad h-l and the samples were irradiated for ca. 30 min at 77 K. Then the crystals were placed into the measurement quartz dewar filled with liquid nitrogen. The crystals were mounted for each experimental plane separately. The e.s.r. spectra were recorded on a Bruker ER 200 D spectrometer. t Permanent address : Department of Physics, Adam Mickiewicz University, PL-60780 Poman, Poland. 18691870 X-Irradiated L-Ascorbic Acid ' b Fig. 1. The e.s.r. spectrum of a single crystal of L-ascorbic acid X-irradiated and observed at 77 K.The magnetic field is applied along the c axis. The arrows point out an example of a radical pair signal. Results Fig. 1 shows the e.s.r. spectrum measured at 77 K immediately after irradiation. The central part of the spectrum of 50 G range is a result of overlapping of three or four spectra due to different types of radicals having similar g factors. The structure of those radicals is not discussed in the following report. Three types of radicals observed at 77 K were proposed by McDearmon and Mou1ton.l The outer lines are identified as being due to different triplet radical pairs. The large separation of lines arises from the dipole-dipole interaction of two unpaired electrons.* The intensity ratio of radical pairs to monoradical signals was better for lower radiation dose;' however, the radiation was continued for a longer period in order to obtain a better signal-to-noise ratio.The angular variations of the electron-electron dipolar splitting are shown in fig. 2 for three patterns, for which it was possible to follow the lines over a range of orientations to make a reliable tensor analysis. The spin Hamiltonian for the triplet radical pair (S = l), in which rapid spin exchange takes place, may be expressed8 by &' = /?HS*g-n+S-D*S (1) g = (81+g2)/2. (2) where n is the unit vector along the external magnetic field and The hyperfine coupling is neglected here. For the allowed transitions, IMs) c.+ IM, - 1 ), the resonance field is given by (3) H = hv/ga - (3/2) (n* D n) (2Ms - l)/g/? - (1 /4H) [TrD - (3/2) (n D ~t)~]/(gS)~.The separation of the two lines is given by dhh = 3(n-D*n)/gp (4)J. T. Masiakowski and A . Lund 1871 - - b C a" a* b Fig. 2. Angular dependences of the electron-electron dipolar splitting for the three radical pairs. (a) Radical pair 1, (b) radical pair 2, (c) radical pair 3. Table 1. The D and g tensors of radical pairs observed in irradiated L-ascorbic acid at 77 K ~~~~~~~~ ~ direction direction radical pair DIG a* b c R / A g a* b C 1 -142.5 0.201 55.0 0.783 87.5 0.589 32.7 0.723 2 -74.9 -0.072 42.2 -0.687 3 -30.7 -0.111 13.0 0.597 17.7 -0.795 0.838 0.174 0.517 -0.517 -0.562 - 0.646 0.384 -0.712 - 0.588 0.507 0.621 0.853 0.402 0.333 0.9 17 0.371 0.150 -0.597 2.0037 5.07 2.0064 2.0092 2.0017 6.28 2.0043 2.007 1 2.0019 8.46 2.0050 2.0073 - 0.008 0.944 0.155 0.828 -0.330 - 0.539 -0.131 0.956 - 0.262 0.975 - 0.066 -0.212 0.983 -0.183 0.001 0.99 1 0.125 - 0.037 0.222 0.323 0.920 0.098 0.530 0.843 0.003 0.265 0.964 The D tensors were adjusted to give Tr (D) = 0. The parameter Tr (D)/d/[Tr (D"] was taken as a measure of non-zero trace before adjusting.It was equal to 0.009 for radical pair 1,0.022 for 2 and 0.053 for 3.1872 X-Irradiated L-Ascorbic Acid radical pair 2 50 G c--------) radical pair 3 Id ! 1 !i 1 ; : J : Fig. 3. The e.s.r. spectra of single crystals of L-ascorbic acid irradiated and observed at 77 K. The magnetic field is applied (a) at an angle of 10" from the a* axis in the a*b plane, (b) at an angle of 70" from b in the bc plane.The expanded portions of the spectra showing the hypefine structure of the three radical pairs were accumulated 10 times. The stick diagrams show 12.3 G hyperfine splitting and spin-flip satellites (broken lines) separated by 1 1.1 G. and the centre of the two lines is H, = (hv/gp> - ( 1 /4H) [Tr D2 - ( 3 / 2 ) (n D n)2]/(g/l)2. ( 5 ) Since the D tensor of radical pairs is usually nearly axially symmetric the average separation R within the pair can be estimated from the maximum separation dhh of the two lines by the point dipole approximation from the relation dhh = 3gp( 1 - 3 COS2 @)/2R3. (6) The D tensors determined from eqn (4), g tensors determined to the second order from eqn ( 5 ) and R distances calculated from eqn (6) are listed in table 1 for all three radical pairs in fig.2.J . T. Masiakowski and A . Lund I873 Discussion As reported bef~rel-~ L-ascorbic acid forms under radiation different types of mono- radicals. In order to discuss the results it is essential to determine the structure of paired radicals. In all three cases the spectra show hyperfine structure (fig, 3). The lines are weak and are visible only at orientations in which no overlapping occurs. We assume, however, that the hyperfine structure is identical for all three radical pairs. The pairs differ only in relative positions and distances, but not in structure. As shown in fig. 3 the hyperfine structure always exhibits a strong central line. This made it easier to follow the fine splitting. Based on the earlier study we conclude that the radical pairs consist of two identical monoradicals with the structure (1).The structure of this monoradical was examined carefully at room temperature and was published previously3 (radical 11). 0 H 'H k Since this radical showed strong spin-flip lines [fig. 1 and 2 in ref. (3)] it is necessary to consider spin-flip-forbidden transitions in the case of triplet radical pairs. According to the theory developed by Minakata and Iwasakis the hyperfine separation of allowed and forbidden transitions in triplet radical pairs is as follows: for I - 1) c-) 10) da = A+/2 - V , df = A + / 2 + vn and for 10) c-) I + 1 ) d" = v n - A - / 2 df = v n + A - / 2 where vn = g n Pn H l h ( 1 1) (12) and The spin-flipping can be treated as an extreme case for which A,, << 2v,lo and hence eqn (12) can be expressed as A - + = v'[( f A,,+ 2 ~ ~ ) ~ + c2] c2 = A:, + AEy.A = 2vn 2 A,, + c2/4vn. (13) From eqn (7), (8) and (13) one obtains d" = A,,/2+c2/8vn x 0 df = 2vn + A,,/2+ c2/8vn x 2vn 62 FAR 11874 X-Irradiated L-Ascorbic Acid Table 2. The parameters of the monoradical forming the radical pairs [from ref. (3)] directions isotropic tensor eigenvalues value a* b C 0.020 0.999 0.002 0.987 -0.020 0.157 -0.158 0.001 0.988 -0.806 0.035 -0.591 g A(H& 23.2 1 24.8 0.054 0.998 -0.014 26.5 J 24*8 1-0.590 0.043 0.807 spin-flip line separation = 12.2 G Fig. 4. The projection of the molecular structure along the b axis. The numbers in parentheses are the averages of five ring atoms b coordinates (in A). The arrows show the experimental D,, directions for radical pairs 1, 2 and 3.for the I - 1 ) c-) 10) transitions, and from eqn (9), (10) and (13) da = A,,/2 - c2/8vn z 0 df = 2vn - A,,/2+ c2/8v, x 2vn (16) (17) for the 10) c-) 1 + 1) transition. Eqn (1 5) and (1 7) give the same result for the separation of spin-flip lines (2gn Bn H / k ) for both transitions, as the doublet separation between satellite lines observed for monoradicals.10J . T. Masiakowski and A . Lund 1875 "t, Fig. 5. The relative positions of molecules projected in the bc plane for (a) radical pair 1, (b) 2 and ( c ) 3. The experimental and calculated D,, directions are compared. The final interpretation of the hyperfine structure is shown in stick diagrams under the spectra in fig. 3. The paired radicals have structure (1).The hyperfine splitting is one-half of that in the monoradical because of rapid spin exchange and is 12.3 G. The lines are surrounded by spin-flip satellites separated by 1 1.1 G. The spectral parameters of the monoradical at room temperature, taken from ref. (3), are listed in table 2. Radical Pair 1 The comparison of experimental results with crystal directions allows us to assume that radical pair 1 is observed when the radicals trapped on molecules A and A' (fig. 4) are paired. To calculate the expected fine coupling between A and A' we assumed a simple four-point spin distribution Lp = 0.6 on C(2) and p = 0.4 on O( 1) in both molecules; these p values were obtained from the semiempirical INDO calculations performed for the monoradica13] and used the point dipolar approximation.The calculated 0 tensor (in G) is then (- 168,80, 88) with the Dll direction (0,0.78,0.63). The experimental distance R in table 1 is calculated by assuming that the spins are localised at two points [eqn (ti)]. By applying the same equation to the theoretical data, comparable values of R should be obtained. This turns out to be the case and the distance R calculated from eqn (6) (4.80 A) is in good agreement with experiment (table 1). The same calculation for the 62-21876 X-Irradiated L-Ascorbic Acid B-B’ radical pair (fig. 4) gave results different from the experimental values: the D tensor (-340, 126, 214) with the Dll direction (0.07, 0.99, 0.10) and the distance R = 3.79 A. The fact that we do not observe the signal due to the B-B’ radical pair is in agreement with the paper by Nunome et aZ.,s who did not observe radical pairs with the electron-electron dipolar coupling along the axis of the 2p atomic orbital forming the n molecular orbital of the monoradicals.In addition to fig. 4, fig. 5(a) shows the projection of the A-A’ radical pair on the bc plane, on which the experimental D,, direction is compared with that expected from the point dipolar approximation. The calculated g tensor of radical pair 1 (table 1) has significantly higher eigenvalues and greater anisotropy than the g tensor of the monoradical (table 2). The directions are, however, quite comparable. Radical Pair 2 Radical pair 2 is observed when the radicals are produced in the B and B” molecules (fig. 4)- The D-tensor calculated by the point dipolar approximation under the assumption of a four-point spin distribution is (- 80, 38, 42) G with the Dll direction along (0, 0.56, 0.80) and the distance R = 6.15 A.These values agree well with the experimental results. In fig. 5(b) the experimental and calculated D,, direction are compared in the projection on the bc plane. The g tensor has the same anisotropy as in the radical pair 1 case, but the eigenvalues are shifted somewhat to lower values. The giso = 2.0043 here is nearly equal to the giso = 2.0041 of the monoradical. The directions are similar. Radical Pair 3 The weakest observed fine coupling corresponds to the A-A” (fig. 4) trapping site of radical pair 3 . The separation of the molecules is so large that we assume here that p = 1 on C(2) atoms.Taking R = 8.96 A from the crystal data we obtained the axial D tensor (- 26, 13, 13) G. The direction joining the C(2)-C”(2) atoms is (0.02,0.35,0.94). Fig. 5 (c) shows the difference in the projection on the bc plane of the C(2)-C”(2) and observed Dlr directions. The g tensor of radical pair 3 is very similar to the g tensor of radical pair 2. This agrees with the fact that the B and B” molecules as well as A and A” are much more separated in space than the A and A’ molecules (fig. 4 and 5). Other Radical Pairs There are some other lines due to unrecognised radical pairs. An example is the signal pointed out by arrows in fig. 1. It was visible only around H 11 c orientation (ca. & 10’) with the maximum of fine splitting exactly along the c axis and equal to dhh = 213 G.Because we did not observe this signal at other orientations we assume that it is the maximum electron-electron dipolar splitting corresponding to the eigenvalue D = -71 G along the c axis. Using eqn (6) the distance between the radicals in the pair is 6.40 A, which equals almost exactly the unit cell parameter c = 6.41 A. The hyperfine structure of the signal is a 1 : 2: 1 triplet (fig. 1) of 5.1 G splitting with no spin-flip satellites or site splitting. The very rough interpretation of this signal is that it might be due to the radical pair formed by monoradical (2) [I11 (in ref. 3)] trapped on molecules related by single translation along the c axis (e.g. A’-”’ or B’-B” in fig. 4).J . T. Masiakowski and A . Lund 1877 \ 0 H J. T. M. thanks the Swedish Institute and the Department of Physics and Measurement Technology, Linkoping University, for financial support. References 1 G. F. McDearmon and G. C. Moulton, Radiat. Res., 1982, 89, 468. 2 L. Rytka, Masters Thesis (Adam Mickiewicz University, Poznan, 1982). 3 J. T. Masiakowski, A. Lund and M. Lindgren, J. Chem. SOC., Faraday Trans. I , 1987,83, in press. 4 J. Hvoslef, Acta Crystallogr., Sect. B, 1968, 24, 23. 5 J. Hvoslef, Acta Crystallogr., Sect. B, 1968, 24, 1431. 6 Y. Kurita, J. Chem. Phys., 1964,41, 3926. 7 S. Ya. Pshezhetskii, A. G. Kotov, V. K. Milinchuk, V. A. Roginskii and V. 1. Tupikov, EPR of Free Radicals in Radiation Chemistry (John Wiley, New York, 1974), pp. 78-82. 8 K. Nunome, K. Toriyama and M. Iwasaki, J. Chem. Phys., 1975, 62, 2927. 9 K. Minakata and M. Iwasaki, Mol. Phys., 1972, 23, 11 15. 10 W. Gordy, Theory and Application of Electron Spin Resonance (John Wiley, New York, 1980), pp. 179-196. Paper 6/21 11 ; Received 30th October, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301869

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1987, 83, 1879-1883 Infrared Spectroscopic Studies of Hydrogen Bonding in Triethylammonium Salts Part 4.-Rearrangement of Hydrogen-bonded Ion Pairs of Triethylammonium Salts caused by Interaction with Tetrabutylammonium Salts in Solution Alexander A. Mashkovsky," Ahat A. Nabiullin and Stanislav E. Odinokov Pacijic Institute of Bio-organic Chemistry, Far-East Scientific Centre, U.S.S.R. Academy of Sciences, Vladivostok-22, U.S.S.R. Rearrangement of triethyl- and tetrabutyl-ammonium salts in chloroform solution has been revealed from the characteristic features of the v(N+H) bands caused by Fermi resonance interaction. Temperature changes of the i.r. spectra show this process to be reversible. Rearrangement constants Kl and K-, and enthalpies -AH (for some cases only) of this process have been determined by measuring the total integrated intensities of the v(N+H) bands of complexes which are in equilibrium.Kl values of the tetrabutylam- monium salts increase with decreasing hydrogen bonding strength in triethylammonium salts. The measured enthalpies of rearrangement and those calculated from the v(N+H) bands of hydrogen-bonded complexes studied are in agreement. It has also been shown that tetrabutylammonium salt anions can participate in the rearrangement of hydrogen-bonded ion pairs like the organic bases studied previously. It has been rep~rtedl-~ that parameters (the frequency of the centre of gravity and the integrated intensity) of the v(N+H) stretching modes in the i.r. absorption spectra of triethylammonium salts in chloroform or organic bases provide information on the structure and hydrogen-bonding energy of these complexes.Hydrogen-bonding energies of the triethylammonium ion pairs with a greater number of anions have been estimated.lY4 Association of these with various organic bases has also been investigated.2* In ref. (5)-(7) formation of the ion-molecule hydrogen-bonded com- plexes of tetrabutylammonium salts with neutral proton donors (alcohols, phenols, carboxylic acids) has been studied. In this context it is of interest to study hydrogen bonding interactions between ' charged ' acids and ' charged' bases (using as examples the triethylammonium cation and anions of tetrabutylammonium salts). Experimental Triethylammonium salts and their solutions were prepared as described in ref.(l), (4) and (7). Commercial tetrabutylammonium salts were used. Before recording the i.r. spectra, freshly prepared solutions were kept for 1-2 h over 4A molecular sieves to remove traces of water. 1.r. spectra were recorded using a Specord 75 IR spectrometer (Carl Zeiss, Jena) fitted with a logarithmic recorder. Matched cells with CaF, and quartz windows were used. Variable temperature spectra were recorded in a special box thermostatted to an accuracy of k0.5 "C with heated or cooled air streams. Integrated intensities were determined by numerical integration of the spectrograms. Errors in measuring the integrated intensities did not exceed 2-3 % . 18791880 Hydrogen Bonding in Triethylammonium Salts 2200 2600 3000 ulcm-' Fig. 1.1.r. spectra of triethylammonium salts dissolved in CDCl, (C = 0.05 mol dm-,): (a) Et,N+H - - * ClO;; (b) ( - * - ) Et,N+H . . Pic-, (-) + Bu,N+C10; (1 : 3); (c) ( - - a ) Et,N+H - - I-, (-) + Bu,N+ClO; (1 : 3); ( d ) ( - a ) Et,N+H - - - Br-, (-)+ Bu,N+CIO; (1:3); (e) ( - -) Et,N+H * - - C1-, (-)+Bu,N+ClO; (1:6). The v(CH) bands of cation are hatched. Results and Discussion In this work interaction between tetrabutylammonium perchlorate (Bu,N+ClO;) and triethylammonium salts (Et,N+H - - . X-, X- = C1-, Br-, I-, Pic-) as well as triethyl- ammonium perchlorate and tetrabutylammonium salts (Bu,N+Y-, Y- = C1-, Br-, I-, Pic-) in chloroform solution [see reaction (l)] have been studied by i.r. spectroscopy: Et,N+H - - - X-+ Bu,N+Y-A Et,N+H * * * Y-+ Bu,N+X-.(1) I1 X-1 I 1.r. absorption spectra of some triethylammonium salt solutions in the v(N+H) stretching mode region are given in fig. 1. In all cases the v(N+H) band contours have a complicated structure owing to Fermi resonance perturbations.', 4 9 To study equilibrium (1) we examine the number of observed peaks of the v(N+H) multiple-band structure, their frequencies and peak intensities as well as the integrated intensity ( A ) of the whole v(N+H) band, which characterizes hydrogen-bonded complexes I and 11. As shown in fig. 1, addition of tetrabutylammonium perchlorate to triethylammonium salt solutions always gives a maximum in the i.r. absorption band at ca. 3080 cm-l, which is also typical forA . A . Mashkovsky et al. 1881 Table 1. The basicity of anions and thermodynamical parameters of rearrangement of the triethyl- and tetrabutyl-ammonium ion pairs in chloroform solution ~~ ~ ~ Et,N+H * * - X- Bu,N+Y- X- Ex- -AHsa Y- q 5 O C b p5 -1 O C b Ey- ClO; 1.01 k0.04 22.0 CIO, 1 .o 1 .o 0.85 & 0.04 Pic- 1.40 f 0.08 30.5 Pic- 7.2 k 0.7 0.130 f 0.010 1.10 & 0.08 I- 1.43k0.12 32.1 I- 10.5f 1.2 0.090f0.010 1.16f0.05 Br- 1.57f0.15 35.2 Br- 49.6 2 4.5 0.020 & 0.004 1.29 & 0.05 C1- 1.69k0.12 37.3 C1- 265.0 f 30.0 0.004 f 0.001 1.45 f 0.10 ~~~ a -AHs (see text) in kJ mo1-1 P5 OC in dm3 mol-l.Basicity factors of tetrabutylam- monium salts anions (Ey-) are calculated from the spectral data for the ion-molecule hydrogen bonds from ref. (7). - 1 - .u N” 2 - -2 * I- * 25 35 -AH/kJ mol-’ Fig. 2. Relationship between values of log e5 OC for rearrangement (1) and energies of hydrogen bonds of triethylammonium salts Et,N+H * - X-.the i.r. spectrum of Et,N+H - - - Cloy solution [see fig. 1 (a)]. Increasing the temperature of the solution results in reversible enhancement of this band intensity and in decreasing band intensity of the initial hydrogen-bonded ion pair Et,N+H - - - X-. This demon- strates the existence of dynamic equilibrium (1) in solution, which shifts towards the complex with the weaker hydrogen bond on temperature increase. In table 1 the enthalpies of Et,N+H - - X- hydrogen bonds from previous works37 are shown. The values of equilibrium constants for rearrangement (1) calculated by measuring the total integrated intensity A of the v(N+H) bands of complexes I and 11, as described,2p3 are also shown.We suppose that valuf of e5 OC is equal PJ1 OC = 1 .O, when X- = Y- = Cloy. As may be seen in fig. 2, log q5 values decrease regularly with increasing hydrogen- bond strength in the ion pairs Et,N+H - * - X-. Thus, rearrangement (1) becomes less advantageous thermodynamically as the ecthalpy of the hydrogen bond in the initial ion pair I increases. On the other hand, P1 values of the reverse equilibrium increase1882 3 - 2 - 1 - V 0 & 0 - Hydrogen Bonding in Triethylammonium Salts 1.5 2.0 2.5 log K::i$ Fig. 3. Correlation between values of log K250c for rearrangement (1) and logK$!G,C, for association of PhOH with anions: (a) K,; (b) K-,. 1.5 - 0.75 - 0.95 % - L I I I 1 3.1 32 3.3 103 KIT Fig. 4.Examples of Van't Hoff plots for the determination of the enthalphy of rearrangment (I): (a) X- = ClO;, Y - = Br-; (b) X- = Pic-, Y - = ClO,. regularly for Pic- < I- < Br- -c Cl- as tetrabutylammonium salts Bu,N+Y- (Y- = Pic-, I-, Br-, Cl-) are added to triethylammoniuq perchlorate solution (see table 1). Note that values of G5 OC are equal to values 1 (within the limits of experimental error) for all systems studied here. Investigation of the proton-accepting properties of tetra-alkylammonium anion salts in methylene chloride solutions with the standard proton donor phenol6 has shown that association constant of the complexes PhOH - - * Y- also increase forA . A . Mashkovsky et al. 1883 Table 2. Enthalpies of some ion-pair rearrange- ments in chloroform X- Y- Q 5 0 c -AHt A(-AHs)a C10, C1- 265.0 17.6k1.7 15.3 C10; Br- 49.6 13.0f1.5 13.2 Br- C10, 0.02 14.2k1.3 13.2 Pic- C10; 0.13 8.4f0.8 8.5 a A(-AH,) is the difference in hydrogen- bonding energies of complexes I and I1 [see reaction (l)], which are determined from spectral parame ters.C10; < Pic- < I- < Br- < C1-. The same dependence of association order of anions on their basicity has also been obtained in ref. ( 5 ) and (7). In fig. 3 the values of log e5 OC and log P!:oc measured in this work have been correlated with log K$O,",c, values from ref. (6). The linear correlation obtained here shows that interaction both of the neutral (PhOH) and of the 'charged' (Et3N+H) proton donors with the anions studied is of the same character. This makes it possible to estimate the basicity factor of anions using of the Iogansen factor rules (AXij = AXoo Pi Ej) and the known acidity factor of triethyl- ammonium cations3 (PEt3N+H = 0.97).The basicity factor of anions (Ex-) estimated in this way can be used to predict the hydrogen bond enthalpy of the anions with any charged proton donors. The basicity factor of anions (Ey-) with any neutral proton donors can be used likewise. In some cases, enthalpies of the rearrangement (1) have been determined by Van't-Hoff plots (see, e.g. fig. 4 and table 2). According to these data the measured -AHt values agree rather well with the - AHs values3i calculated using Iogansen's intensity rule9 from the v(N+H) spectral parameters. As may be seen in table 2, electrostatic interactions do not make any contribution to the intensity of the v(N+H) oscillation and are excluded when accounting for thermodynamic -AHt values taken from i.r. spectral data.The results obtained here and the previous data [see ref. (1)-(4) and (7)] do not reveal any specific differences in the spectral manifestations of hydrogen bonding in ionic pairs or intermolecular and ion-molecule hydrogen-bonded complexes. We thank N. M. Shepetova for helpful remarks and for the translation of the text into English. References 1 V. P. Glazunov, A. A. Mashkovsky and S. E. Odinokov, J . Chem. SOC., Faraday Trans. 2, 1979, 75, 2 A. A. Mashkovsky, A. A. Nabiullin and S. E. Odinokov, Dokl. Akad. Nauk. SSSR, 1982, 265, 1427. 3 A. A. Mashkovsky, A. A. Nabiullin and S. E. Odinokov, J. Chem. Soc., Faraday Trans. 2, 1983, 79, 4 S. E. Odinokov, V. P. Glazunov and A. A. Nabiullin, J. Chem. Soc., Faraday Trans. 2, 1984, 80, 899. 5 I. S. Perelygin, V. Ya. Melnichenko and A. M. Afanasieva, Zh. Prikl. Spektrosk., 1979, 30, 517. 6 R. P. Taylor and I. D. Kuntz Jr, J. Am. Chem. SOC., 1972,94, 7963. 7 S. E. Odinokov, Ph.D. Thesis (Moscow, 1983). 8 A. V. Iogansen, Theor. Exp. Chem. (USSR), 1971, 7, 302. 9 A. V. Iogansen, Dokl. Akad. Nauk. SSSR, 1965,164, 610. 629. 951. Paper 6/1689; Received 19th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301879

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faradaj7 Trans. I , 1987, 83, 1885-1892 Electrostatic Interactions between Organic Ions - Part 2.-Phosphates with Amines Henry R. Wilson and Robert J. P. Williams" Inorganic Chemistry Laboratory, South Parks Road, Oxford OX1 3QR Association constants of organic cations with organic phosphate anions have been determined using an n.m.r. procedure. The constants are com- pared with expectation based on the Bjerrum theory of ion-pair formation. The values of the constants are rather different from those for simple and complex inorganic ions with the same organic phosphates. The importance of the difference for biological systems is stressed. In a recent publication which we shall refer to as Part 1,l we have analysed the binding of organic molecules carrying carboxylate anions to a wide variety of cations including simple spherical cations, such as Ca2+, some more or less spherical complex inorganic ions such as [Co(NH,)J3+ and complex organic multiamines of various structures.We studied these interactions between rigid-frame organic molecules, anions and cations, between flexible frame molecules, anions and cations, as well as between mixed systems of rigid and flexible molecules. In a review2 we have indicated the relevance of such data for the understanding of electrostatic forces in biological systems. In this paper we shall extend our observations to phosphate-containing anions including polyphosphates. In general the cations examined are similar to those used in Part 1. In the previous work the main method used for the analysis of complex ion formation was the conductivity of solutions.This method is not easily employed when the molecules under study have a series of acid dissociation constants, pK, values, over the region of pH in which association is studied. For this reason in the analysis of the binding of phosphates to polyamines we have used n.m.r. procedures following both proton and phosphorus signals. Unfortunately there are two difficulties with this method: (i) D,O has to be used as a solvent and corrections made to give values in H 2 0 ; and (ii) relatively high concentrations of species have to be used, and therefore extrapolation of the binding constants to zero ionic strength is not likely to be very accurate. The data we present are unlikely to be more than semi-quantitative guides to stability constants at zero ionic strength.The general conclusions are not in jeopardy, however, since we have managed to study a very wide range of charge types from a negative charge of six to a positive charge of four. This means that the association constants vary over ten orders of magnitude. Experiment a1 Analytical-grade anhydrous sodium pyrophosphate (PP) and sodium tripolyphosphate (PPP) were supplied by B.D.H. Adenosine di- (ADP) and tri-phosphate (ATP) were obtained as their sodium salts of the highest grade from Sigma. Pure sodium hexa- metaphosphate (HMP) flake was supplied by Koch-Light Laboratories. Triethylene- tetramine (trien) was bought from Aldrich and its density regularly checked. Nane, cyclic triethylene triamine, was kindly donated by Prof.Karl Wieghardt, Ruhr Univer- sitat, Bochum, Federal Republic of Germany. The 31P spectra were recorded on a Bruker AM-250 spectrometer operating at a I8851886 Electrostatic Interactions between Phosphate Ions frequency of 101 MHz. Typical operating parameters were a 40" pulse angle, together with a relaxation delay of 0.7 s; 32 accumulations over a 3 kHz sweep-width, a 4K memory size and a line-broadening of 9 Hz. Chemical shifts were measured against an external standard of trimethylphosphate (6 = 0 ppm), and upfield shifts are given a negative sign. The proton spectra, measured in D,O, were recorded on a Bruker WH300 spec- trometer, using a pulse angle of 60", a sweep-width of 3 kHz and a memory size of 8 K, 32 accumulations being made for each spectrum.Homonuclear pre-saturation was used for suppression of the HDO peak, continuous-wave decoupling being applied at a power level of 6 mW (15 L) for 0.25 s. Resolution enhancement of the spectra was achieved using Gaussian multiplication, with a Gaussian broadening of 0.1 and a line- broadening of - 10 Hz. 31P n.m.r. samples contained 5 mmol dm-3 of the phosphate-containing molecule made up in 99.8% D,O, with 0.1 mmol dm-3 EDTA present to remove any line- broadening metal-ion impurities. The phosphate samples analysed by pH titration in the presence of an excess of base contained 50 mmol dm-3 amine in addition to the 5 mmol dm-3 phosphate. Similarly, the 'H n.m.r. samples analysed by pH titration contained 5 mmol dm-3 base with or without 50 mmol dm-3 of the phosphate-containing amine; 1 mmol dm-3 TMS was present as an internal reference standard.Every sample was taken to pH 12 by addition of 1 mol dmP3 NaOD. In the course of a titration, the pH was lowered step by step with 0.1 mol dm-3 DCI to a value less than zero. At each stage the chemical shift of each peak was recorded. All experiments were reproduced and several were repeated at different concentrations. Determination of Thermodynamic Constants pKa values for any given anion or amine were determined by direct use of the Henderson-Hasselbach equation if it was obvious by inspection that the successive pKa values were at least two units apart. Where this was obviously not the case a method of successive approximation was employed using the projection-strip a p p r ~ a c h .~ The pKa values found for individual compounds using the n.m.r. method were always in reasonable agreement with literature values ( f: 0.2) after corrections were applied for ionic strength (table 1). The pK, values found in mixed solutions of base and anion were used to determine the anion-cation binding constants after a single binding constant had been determined at a single pH, chosen so as to be remote from any influence of an acid-base dissociation. These single binding constants were obtained using mass-action considerations from an appropriate analysis of the concentration of all coexisting species, i.e. free and bound species as determined by n.m.r., at concentrations which were varied over a wide range. At the highest concentration complexes were some SO-90% fully formed, so that the values of binding constants were readily obtained (table 2).The constants were corrected for ionic strength as above.' Given one value of an association constant and pK, values of both components in the presence and absence of the other and at the same ionic strength, it is a straightforward matter to evaluate the binding constant for all other anion-cation associated species of a given anion-cation pair, but of different charge type from the equation ApK, = log Kl -log K, where ApK, is the shift of a pKa in a particular associated species (H,X. Y 3 H,-,X*Y) from that of the free component (H,X+H,-,X) in the absence of the' second component, Y. LogK, is the association constant for a known complex H,X-Y, and log K, is that for H,-,X.Y.This simple equation applies only to fully formed species.Table 1. Change in pK, observed for the interaction of phosphates and amines (all measurements at 300 K, corrected to zero ionic strength) phos- trien4+ trim3+ trien2+ tried+ trienO phate pK,-free ApKaa charge pKa-free 2.57 6.34 9.16 10.34 HMP PP PPP ADP ATP HMP 2.96 - 3.03 7.02 - 2.60 19.59 - 3.41 (6.78 - 3.23 19.68 -4.16 7.49 - 2.68 7.40 - 2.86 5 - TL 6- 2- TL 3 - TL 4- 3 - TL 4- TL 5 - 2- TL 3- 3- TL 4- ~ APKa +4.78 + 3.57 + 1.83 + 1.47 APK, + 4.24 + 3.04 +1.51 + 1.17 APKa + 4.64 + 3.63 + 1.93 + 1.55 APK, +2.51 ApK, +1.95 + 1.04 +0.59 + 4.04 + 2.89 + 1.51 + 1.46 nane3+ nane2+ nanel+ naneO + 4.23 + 3.37 + 1.981888 Electrostatic Interactions between Phosphate Ions Table 2.Association constants, IogK, of part- icular ion pairs at 300 K and I - , 0 species logK pH trien3+. HMPs- trien4+. PP2- trien4+ . PPP3- trien3+ * ADP2- trien3+ * ATP3- nane+ HMP6- 6.12 8.0 3.25 3.0 4.64 2.4 3.33 4.8 4.98 2.4 5.07 8.5 Table 3. Association of phosphates and amines: log Kb ( I -+ 0, T = 300 K)a phosphate charge trien4+ trien3+ trien2+ tried+ HMP hexametaphosphate PP pyrophosphate PPP tripolyphosphate ADP adenosine diphosphate ATP adenosine triphosphate 5 - (20) 2- (8) 3- (12) 4- (16) 3- (12) 5 - (20) 2- (8) 3- 3- (12) 4- (16) 6- (24) 4- (16) ~~~ - - - 7.87 10.90 (18) 6.12 (12) 2.55 (6) 1.12 - - - 3.25 5.85 - 9.26 (12) 5.02 (8) 2.05 (4) 1.01 4.64 7.87 - - - - - - - - 12.03 (15) 7.39 (10) 3.76 ( 5 ) 1.93 3.33 - (9) 3.33 (6) 1.57 (3) 1.01 4.98 7.84 (12) 3.80 (8) 1.24 (4) 0.55 - - - - - - phosphate charge nane3+ nane2+ nanel+ HMP 6- - - (18) 9.30 (12) 5.07 (6) 1.82 - - - - - - hexametaphosphate - - - a Charge product in parentheses.For weaker complexes an operational value of log K,, which derived from too small a value of ApK,, was first determined using the above equation as if the system were fully formed. Then the system was reanalysed using new concentrations of the species derived from these values of K2, and so on by successive approximation until cycling had little effect, i.e. < 0.1 log units change, on the logK, values. Results Observed binding constants were determined by the above methods at the particular ionic strength of the medium in individual experiments. There is a wide variation of logK.The conditions of measurement, near-millimolar concentrations of reagents and high ionic strength are not too dissimilar from the conditions in biological solutions, where several phosphates such as phosphate itself and ATP may reach a concentration > mol dme3, magnesium ions are at least loh3 mol dmP3, and the ionic strength is ca. 0.2 mol dm-3. We shall use the data at these high ionic strengths and the conditional stability constants at pH 7 later to discuss fully the biological relevance of ourH . R. Wilson and R. J. P. Williams t I 12 1889 1 I 1 8 16 24 lz*zzl Fig. 1. Association constants of different phosphates with trien plotted as a function of charge product, z , z , : (i) for the present series of complexes, points on curved lines to right-hand side; (ii) for Mg2+ and Ca2+ with phosphates (Sillen and Martell),6 points to left-hand side; (iii) ‘reference line’, calculated from Bjerrum’s equation for ion-pair formation, see text.Key: CI, HMP; *, PP; +, PPP; v, ADP; a, ATP. The vertical lines connect points of the same zlz2 but of different charge distribution to show the loss of stability as charge is spread out. observations. Before we do so we wish to make a more fundamental analysis of the ion pairing by extrapolating all constants to the same, zero, ionic strength. This has been done (see table 3) by applying the conventional correction for log K values obtained at high ionic strength.’ The data are undoubtedly open to an error which could be as large as a factor of ten for the highest constants, since here the extrapolations are long owing to the contributions to the ionic strength from the highly charged species.Discussion The binding constants at zero ionic strength for various inorganic ion pairs, given in table 3, are plotted against the product of the charges of the counter-ions in fig. 1. The graph also shows the results taken from the literature for Mg2+ and Ca2+. The features of fig. 1 considering small spherical inorganic cations first are as follows. (1) The agreement between observation and theoretical expectation using Bjerrum’s approach for simple cations such as Mg2+ and Ca2+ is good up to quite high charge products, ca. 8. In the Bjerrum equation we have used a distance of 5 A for the closest distance of approach to obtain the line marked ‘reference’ in the figure.(2) At very high anionic charge for each group of small cations, monovalent, divalent and trivalent, there is a fall-off (not shown) from theoretical expectation in log K. This could be due to the inability of the cations to match the spread of charges inevitable with highly charged anions.1890 Electrostatic Interactions between Phosphate Ions (3) The values for binding the cyclic phosphate, HMP, are lower than those for linear phosphates. The former have a lower charge density, since the linear phosphates of the same charge type contain doubly charged individual phosphate groups. However, the configurational entropy of the cyclic phosphate is restricted. It is not too surprising that the two groups are similar.When we compare these observations with those for the carboxylate anion complexes with the same inorganic cations as described in Part 1, we see that there are some similarities, but also some major differences. The similarities are found for the carboxyl- ates, which have the anions very close together in space. Differences occur (a much lower stability for carboxylates of high charge) where the carboxylates are spread out on a framework. It is unfortunate that we have no data for organic phosphate anions with the phosphate groups spread out on such frameworks, e.g. inositol polyphosphates. We do know, of course, that Mg2+ and Ca2+ have only a low affinity for polynucleotide phosphates, DNA and RNA for example, which are polyanions of very considerable negative charge but of low charge density.Considering the nature of the structures of the anions, it is easily understood that polyphosphates are the most powerful complexing agents for inorganic cations. The binding constants of the organic amines to all phosphates are weak (fig. 1) until the charge product exceeds ca. 12. The binding strength falls far below that expected from the charge product on the basis of Bjerrum theory, and there is a noticeably much weaker binding to diester phosphates (anhydride phosphate) than to terminal, doubly charged phosphate centres even for the same charge product. Both these changes reflect the weakening of interaction as charge is spread out. This was also found, but to a less degree, for carboxylates in Part 1.l The overall result is that organic cations do not compete effectively against inorganic cations for phosphate ligands up to a charge product of ca.12, when the anionic charge is beginning to encompass a large spatial region. When the charge on the polyamine exceeds that usually found on inorganic cations, i.e. > 3, and when the anion/cation pair is such that the charge product exceeds 12, amines bind equally or better than the inorganic cations to a linear polyphosphate. This means, for example, that an amine must have at least four amine groups before it is an effective competitor against magnesium for a polyphosphate grouping of charge 4 - . If we compare the polyphosphate anions with the polycarboxylates of Part 1 we see that for crowded sets of carboxylates or phosphates the binding always favours the inorganic cations, but that the situation is very different as the charge density on the polyphosphate is lowered, e.g.in cyclic phosphates which can be likened to long-chain diester polyphosphates (see the squares in fig. 1). Although binding of a diamine to hexameta- phosphate is very weak (IogK z 2), binding increases very rapidly between a charge product z , z2 = 12 and z , 2, = 24. This implies that the spacing of the positive charges on triethylene tetramine (4+) can match the spacing of the negative charges in, for example, hexametaphosphate. The possibility of matching patterns of opposite charges was noticed in Part l1 and leads to preferential association, particularly at high charge, for organic cations relative to inorganic cations, especially to a pattern of negative charges of low density.To what extent forces other than those of electrostatic origin contribute to the binding is difficult to assess. We notice that other studies have led to parallel conclusions, but in them special attention was paid to template structures, which we have not analysed.** Biological Relevance It is noticeable that biological systems provide both a set of anions of very high local charge, as in the polyphosphates including nucleotide compounds such as adenosine triphosphate, ATP, as well as a series of compounds which have diester phosphate groups rather distantly spaced on a frame, e.g. DNA. All these phosphates are as fully ionisedH. R. Wilson and R . J . P. Williams 1891 Table 4.Effect of interaction on chemical shift (a) Change in proton n.m.r. chemical shift of amine in the presence of excess phosphate phosphate amine signal HMPg+ PP4- PPP5- ADP2-/ADP3- ATP4- trien4+ A +0.075 f0.02 + 0.04 + 0.035 + 0.04 B + 0.05 + 0.02 - + 0.055 + 0.035 C +0.035 f0.005 - + 0.035 + 0.025 B C - + 0.03 trien3+ A + 0.075 + 0.035 - - - + 0.065 - - - - - nane3+ - -0.13 nane2+ - -0.01 - - - - (b) Change in 31P n.m.r. chemical shift of phosphate in the presence of excess trien4+a charge phosphate state HMP PP PPP(a) ADP(B) AWY) protonated ( 5 -) (not (2-?+0.3 (3-) +0.1 (2-)O.O (3-) 0.0 available) ?I deprotonated ( 6 - ) -0.6 (4-) -0.7 ( 5 - ) -0.1 (3-)O.O (4-) -0.65 a Phosphate charge is indicated. as possible at pH 7. Structures such as tRNA and other polynucleotides with a complex folding pattern can provide either regions of low anionic charge density or, now and then, regions of relatively high charge density through folding, which is equivalent to an organic template construction.The nature of biological cations is now extremely interesting. The most prominent cations are the simple Mg2* and Ca2+ ions, and they are in competition with a limited set of special polyamines such as spermine and spermidine. Both of these polyamines have well-spaced amino groups separated by at least three carbon atoms. This spacing ensures that all their pK, values are > 7, i.e. above biological pH, and their charge is given by the number of amino groups. The total number of groups in both amines is four. Using fig. 1 we see that this type of amine does not have a very high binding constant for the pyrophosphates of high charge density, such as P,O;- and ATP5-.As a consequence they cannot compete with Mg2+ .for these sites. However, Mg2+ does not bind to the spread charge pattern of DNA and RNA, and it is to such polyanions that the biological polyamines are known to bind. We consider therefore that these particular polyamines have evolved so as to avoid the ionisation constants of protons, to provide a binding constant of reasonable strength to certain phosphate polymers of low charge density and to be used there in control while overcoming the competition from Mg2+. Higher binding constants may not be desirable, since they would be likely to generate slower reactions making for less rapid response to changes in environmental factors.In a subsequent article we shall consider the interaction of basic surfaces of proteins, such as histones, with the above phosphates, since here we have an even more widely spread, very mobile pattern of positively charged lysines and arginines which can match spread-out polyanions of different kinds. Once more competition from inorganic ions is important. An example is the competition between histones, polyamines and magnesium for binding to DNA. An interesting but speculative additional point is brought out by the consideration of competition between inorganic ions, which are not usually found in a biological1892 Electrostatic Interactions between Phosphate Ions environment, and amines for phosphorylated side-chains of proteins.Aluminium ions are a particularly relevant case, in view of the apparent connection between certain kinds of dementia related to Alzheimer’s disease and aluminium in water supplies. A glance at fig. 1 shows that A13+ ions would be expected to bind strongly to doubly charged phosphates of proteins, especially if these were clustered. The onset of dementia has been suggested to be connected with aberrant phosphorylation of side-arms of the proteins of neurofilaments, the aberrations producing observed filament tangles. Obviously cations will assist the cross-linking between such phosphorylated proteins, thus assisting the formation of such tangles. A13+ ions will be expected to do this very well indeed (see fig. 1). N.M.R. Shifts in Complexes The above study yields values for the n.m.r. chemical shift in phosphate resonances when the organic phosphate molecule is bound by amines. This could be a useful diagnostic tool in the recognition of enyme binding to phosphates. Table 4 shows that 31P signals of monoester phosphates suffer appreciable shifts of ca. -0.6 ppm when bound to an ammonium group at high pH. We shall use these shifts to diagnose certain types of phosphate/protein associations in a later paper. References 1 S-C. Tam and R. J. P. Williams, J. Chem. Soc., Faraday Trans. I , 1984,80, 2255. 2 S-C. Tam and R. J. P. Williams, Struct. Bonding (Berlin), 1985, 63, 103. 3 H. Rossotti, The Study of Ionic Equilibria (Longmans, London, 1978). 4 B. Dietrich, D. L. Fyles, T. M. Fyles and J-M. Lehn, Helv. Chim. Acta, 1979, 62, 1763. 5 E. Kimura, A. Sakonaka, T. Yatsunami and M. Kodama, J . Am. Chem. Soc., 1981, 103, 3041. 6 L. G. SillCn and A. E. Martell, Stability Constants of Metal-ion Complexes, Special Publications 17 and 25 (The Chemical Society, London, 1964 and 1971). Paper 611948; Received 2nd October, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301885

出版商:RSC    

年代:1987

数据来源: RSC