摘要:

J. Chern. Soc., Furuduy Trans. I , 1987, 83, 1725-1730 The Molar Volume of a Large Polymeric Cation [ ~ l 0, ,H 481 7 + J. W. Akitt and Julie M. Elders The School of Chemistry and the Department of Inorganic and Structural Chemistry, The University of Leeds, Leeds LS2 9JT Pierre Letellier Physicochimie des Solutions, ENSCP, Universite Pierre et Marie Curie, 11 rue Pierre et Marie Curie, 75321 Paris 05, France The molar volume has been measured of the tridecameric aluminium cation [AI,,0,,H,,]7+. The volume obtained is closely similar to that calculated from the crystal structure so that this large cation appears to have negligible electrostriction on the solvent. It is also clear that the ion is large enough for any structural effects owing to the mismatch between ion size and solvent structure to be insignificant.The results show that electrostriction around an ion is short-range in nature. A major problem in the interpretation of molar volume measurements of ions is the separation of the measured value into the various contributions from electrostriction, ion intrinsic volume and solvent structural effects. The electrostriction may be calculated as arising over a large distance from the ion, when it is found to be proportional to z : , ~ or it may be limited so as to include just the first few layers of solvent molecules, when it is found to be approximately proportional to z ~ , ~ a conclusion which is in accord with many of the experimental data.4~5 The intrinsic volume of the ion is not equal to its physical volume, but contains a structural contribution arising from the way the ion fits into the solvent solvating i t 2 We considered that some insight into these problems might be obtained if we were to attempt to measure the volume of a large molecular ion. If this was sufficiently greater in radius than a water molecule, it should approximate to a smooth body immersed in a continuum fluid and have an intrinsic volume equal to its actual physical volume.It should then be possible to obtain the electrostriction contribution directly with some confidence. We have chosen for the large ion the tridecameric aluminium cation, [A113040H48]7+, which is formed during the hydrolysis of salts of AlIII and has been much studied. Its crystal structure is known,6 and recent n.m.r. studies have allowed the preparation to be monitored and have enabled us to define the conditions for the preparation of solutions in which contamination by other species is at a minimum.'^ * Experimental Unconventionally, the species of interest has to be prepared in situ with an accuracy attainable during the make up of solutions of salts of well known stoichiometry.This was achieved as follows. The salt, AnalaR AlCl, - 6H,O, was weighed out in a tared flask and a suitable volume of water added. Anhydrous sodium carbonate was weighed out separately, with sufficiently accurate adjustment of the weight taken to give a hydrolysis ratio, [OH]/[Al], of 2.46f0.01. The aluminium salt solution was heated and the hydrolysis carried out by adding the carbonate as rapidly as was consistent with full redissolution of the precipitate which formed with each addition.The solution was cooled 17251726 Molar Volume of [A113040H4,]7+ immediately after the precipitate from the last addition had dissolved so as to minimise the decomposition of the tride~amer.~ The reaction can be written: 1 3A1(H20)i+ + 16Na2C0, -+ [Al13040H4s]7+ + l 6C02 T + 32Na+ + 54H20 so that in calculating the mass balance, we have to allow in effect for the addition of Na,O rather than Na,C03. The preparation was weighed and the molality of the cation and NaCl calculated. This stock solution was then diluted by weight to obtain a series of solutions of decreasing tridecamer and NaCl concentrations, which then had their densities determined. This was repeated five times.A second set of preparations was also carried out in which the stock solutions were diluted by weight with NaCI solution of the same concentration as the NaCl in the stock. These solutions then contained decreasing concentrations of tridecamer dissolved in NaCl solution of a given fixed concentration. This was repeated three times. The densities of these solutions were measured using a Picker densimeter vibratometer type 03D, supplied by S o d e ~ , ~ and the temperature was maintained at 298 _+ 0.002 K using a thermostat manufactured by Setaram. The density of water was taken as 0.997047 g cm-,. Results and Discussion The frequency measurements were converted to volumes using a program designed for the calculation. The results are depicted in fig. 1-3. In the case where both components vary in concentration then we can consider this thermodynamically to be a single solute of molecular weight M(2) = M(Al&l,) + 32M(NaC1) = 31 57 at a molality, m, equal to its total weight in 55.51 mol of water divided by M(2).The volume of the solution may be written in two forms:l (i) V = 55.51 C+m4, (1) where 4, is the apparent molar volume of the mixture of solutes and is the quantity which is calculated from the densities, and (2) (ii) V = 55.51 q + m T where T$ is a partial molar volume and Expanding in terms of partial molar ionic volumes gives V = 55.5 1 + (Vc,+ + 32VN, + 39Fcl) m (4) where C7 + represents the large cation. We now seek to relate the quantities 4, and K. From eqn (1) and (3) we have = [ ~ m ~ 2 / ~ m ] 5 , ~ 5 1 whence giving d(m4,) = I" Kdm = mb2 0 1 m o 42 = - I" dm ( 5 ) and yZ can be expanded in terms of ionic quantities as in the brackets in eqn (4).If the solution is sufficiently dilute then we can use the Debye-Huckel limiting lawlo pi = py+ RT log mi - RTBz? ItJ . W. Akitt, J . M. Elders and P . Letellier 1200 -dpl E - 0 1200 X 1727 1100 0.1 0.2 rni/(mol kg-')f Fig. 2. Calculated molar volumes of Al1304,H4,C1,-32NaCI solutions as a function of the square root of the molality of the combined components. The straight line is drawn with a slope of 867.5 cm3 mo1-1 m-t. I I 0.1 0.2 650 ' rn+/(mol kg-')+ Fig. 3. Calculated molar volumes of All,04,H,,C1, dissolved in NaCl solution as a function of mi of A1 salt.1728 Molar Volume of [A1 30,,H4,]7 -t where pi is the ionic chemical potential.Taking the derivative of this expression with respect to pressure, we obtain where the term (aB/CIP)RT has been replaced by the constant a. If we now replace the various ionic terms in in eqn (5) by expression (6) we can express d2 as The ionic strength of our particular mixture varies with concentration and is equal to 60m, which when substituted into eqn (7) gives q52 = e-619.67amk (8) The value of a is calculated by repeating the calculations [eqn (5), (6), (7)] for a 1 : 1 electrolyte, which gives b2 = e - i a m i . It is well known that the limiting slope of a plot of d2 vs. mi in such solutions has the value 1.867,' so that a must have the value - 1.400. Thus, for the mixture of electrolytes investigated here we predict a limiting slope of the volume-concentrationi curve of 867.5.This prediction can be used to interpret our data in two ways as shown in fig. 1 and 2. (i) We can use the method proposed by Redlich,12 and calculate for each determination the quantity $v - 867.5mi and plot this as a function of m. This gives a straight line which can be reproduced to zero concentration through the rather inaccurate results obtained at the lowest concentrations and so obtain an accurate value of the intercept (1 149 cm3 mol-l, fig. 1). (ii) Conventionally, we can plot the data as a function of mi and attempt to construct the best-fitting straight-line tangent to the curved plot and which intersects the axis at the correct volume. This is best done using the Redlich plot and so gives the same intercept.Fig. 2 shows how the data behave. In order to obtain the volume of the large cation, we have to subtract from this the volume of 32 mol NaCl(16.62 x 32l) and the volume of seven chloride anions. We take this as 23.23 cm3 mol-1 using the value of - 5.4 cm3 mol-l for H+." This approach gives a volume for the tridecameric cation of 454.5 cm3 mol-l. In the case where the sodium chloride concentration is kept constant and only the concentration of the tridecamer heptachloride is allowed to vary we encounter a very different situation. The presence of a high concentration of electrolyte means that electrostriction effects are lost and that we obtain a volume equal to that of the tridecamer heptachloride at all concentrations. The molar volume thus appears as a straight horizontal line when plotted as a function of concentration, as is shown in fig.3. The intercept is 670 cm3 mol-1 and comprises the volume of cation and the anions. However, we cannot now assume that the ionic volume of these latter is equal to the volume at infinite dilution, since their electrostriction has been saturated by the background salt. Instead we will assume that we must take the intrinsic volume of the anion and that this is 1.92 times the volume occupied in the crystal, as has previously been determined for 1 mol dm-3 KCl The crystalline volume of C1- is 14.94 cm3 mol-1 so that the intrinsic volume is 28.7 cm3 m01-V~ Subtracting seven times this value from 670 gives the tridecamer volume as 469.1 cm3 mol-l. This is within 15 cm3 mol-l of the previous result; i.e.they agree to within 3.2%.J. W. Akitt, J. M . Elders and P. Letellier 1729 Table 1. Details of calculation of volume of tridecameric cation using two values of outer oxygen radius oxygen volume /A3 volume of cation/cm3 mol-l radius of sphere of 4 caps of cation calculated measured /A 455-469 1.28 1188 44 1 747 450 1.4 1254 450 804 484 The Physical Volume of the Tridecameric Cation The volume was calculated from the dimensions of the crystal structure.6 The molecule consists of a central, tetrahedral A10, unit surrounded by four groups of three A10, octahedra sharing an edge with one oxygen atom common to all three octahedra and to the A10, unit. The sets of three octahedra are bound together by OH bridges at the other end of the shared edges, and these A1,0,, units are interlinked by double OH bridges. The structural formula is written [A104Al12(OH),,(H20)12]7+.Each of the aluminium atoms in the octahedra carries a single, terminal water molecule and it is this which determines the maximum radius of the cation. The molecule is tetrahedral, with four plane faces formed by the sides of three of the A130,, units, but can be approximated by a sphere with four caps removed. Defining the physical volume of such irregular solids is not possible to any degree of precision,15 though in the present case, where the radius of the individual oxygen atoms is only one fifth the radius of the whole, one can bound the surfaces by smooth geometric constructions with relatively little error. This in effect assumes that the cation will rotate in a close-fitting cavity in the solvent. It is then necessary to obtain the radius of the circumscribing sphere and the distance between the planar faces and the centre of the molecule, The positions of the corresponding atoms is defined by the structure and is easily found. The plane faces are slightly concave and an average value was taken for the positions of the oxygen atoms.The most important approximation to be made then is the radius to be taken for the outer oxygen atoms. The lengths of some hydrogen bonds in the crystal suggest this should be 1.28 A, but in order to study the influence of this radius on the final value we also used a value of I .4 A. The results are summarised in table 1. The two calculated figures give some measure of the uncertainty with which the physical volume may be known.Their average is very close to that of the measured molar volumes and it can be concluded that any electrostriction around this cation is very small. The model in which the electrostriction i s proportional to z , ~ , predicts an electrostrictive reduction of 178 cm3 mol-l and it is clear that this cannot apply in the present case. The tridecameric aluminium cation then displaces water in accord with the principle of Archimedes. Such a conclusion is unequivocal evidence for the short-range nature of electrostrictive forces. We should, nevertheless, attempt to justify our result in terms of the short-range model developed by Conway’s group.3 We have calculated the electric field conformation at 1.4 A from the surface of an A13013 group and find this to be directed always to the central part of the cation (not the exact centre) and to have a value of 0.6 x lo6 e.s.u.V, close to that of a spherical 7+ cation. The electrostriction in the first solvation sphere would thus be expected to be of the order of 70 cm3 m01-l.~’ This is indeed an underestimate since the water coordinating the planar tetrahedral faces is closer than this. Electrostriction of this magnitude should be detectable in the present experiment. We can suggest two reasons why none is observed. (i) A large cation may fit into the water structure in such a way that the solvation sphere is already as close to the surface of the ion as is possible. Electrical forces then cannot induce any further movement in the1730 Molar Volume of [A1130,,H,,]7+ water position, the ion behaving as if it were a smooth, solid body.The electrostriction would then depend on compression in the second sphere where the electric field is too small to have any appreciable effect. (ii) The ion must rotate in its solution and since it is not a perfect sphere it must buffet the first solvation sphere molecules as it does so. These will exist in a dynamically determined position around the cation, which will be somewhat expanded and will counteract any electrostrictive effect. The magnitude of volume compensation will depend on the relative lengths of correlation times of rotation of the cation and of translation of the water molecules and will be zero if the ion rotates relatively slowly. Electrostriction might thus be detectable at a sufficiently low temperature.Our measurements allow us to choose between these possibilities. Those made in pure water should demonstrate the full electrostriction of the solute ions, whereas those made in NaCl solution should exhibit none, but should still contain some dynamic contribution. The small difference between the two measurements indicates both that the electrostriction is negligible and that any dynamic contribution is very small. It follows that the correlation time of this cation will be long, and that it behaves as a smooth, hard solute. References 1 F. J. Millero, Water and Aqueous Solutions, ed. R. A. Home (Wiley, Chichester, 1972), chap. 13. 2 E. Glueckauf, Trans. Faraday SOC., 1965,61,914. 3 B. E. Conway, R. E. Verrall and J. E. Desnoyers, 2. Phys. Chem., 1965,230,157; J . Chem. Phys., 1965, 4 A. M. Couture and K. J. Laidler, Can. J . Chem., 1956, 34, 1209; 1957, 35, 207. 5 J. W. Akitt, J. Chem. SOC., Faraday Trans. 1, 1980,76, 2259. 6 G. Johansson, Arkiv Kem., 1963, 20, 305; 320. 7 J. W. Akitt and A. Farthing, J . Chem. SOC., Dalton Trans., 1981, 1617; 1624. 8 J. W. Akitt and B. E. Mann, J. Magn. Reson., 1981, 44, 584. 9 P. Picker, E. Tremblay and C. Jolicoeur, J. Solution Chem., 1974, 3, 377. 43, 243. 10 H. S. Harned and B. B. Owen, The Physical Chemistry of Electrolyte Solutions, A.C.S. Monograph 1 1 R. Zana and E. Yeager, J. Phys. Chem., 1966, 70, 954; 1967, 71, 521. 12 0. Redlich and P. Rosenfeld, Z . Elektrochem., 1931, 37, 705; 2. Phys. Chem., Abt. A , 1931, 155, 65. 13 S. Bouguerra and P. Letellier, J . Chim. Phys., 1982, 79, 845. 14 S. Bouguerra and P. Letellier, J. Chim. Phys., 1984, 81, 55. 15 E. Ayranci and B. E. Conway, J . Chem. SOC., Faraday Trans. 1, 1983, 79, 1357. Series (Reinhold, New York, 1957). Paper 611422; Received 17th July, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301725

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987, 83, 1731-1738 Selective lH-13C and lH-lH Nuclear Overhauser Enhancement Studies of Adenosine-Thymidine Interaction in Solution Claudio Rossi,* Neri Niccolai, Anna Prugnola and Franco Laschi Department of Chemistry, University of Siena, Pian dei Mantellini 44, 53100 Siena, Italy Selective proton-carbon nuclear Overhauser effect (NOE) measurements have been widely used to analyse through-space proton-carbon interactions. Intermolecular interactions between adenosine and thymidine have been studied by means of selective {H}C-NOE spectroscopy. Moreover, the role of proton chemical exchange in determining {H)C-NOE has been considered. A combined use of homo- and hetero-nuclear selective NOE to account the chemical exchange contribution to the experimental {H}C-NOE has been proposed.The nuclear Overhauser effect (NOE) is due to changes in spin population when an irradiated nucleus relaxes by the dipolar mechanism with the observed The detected NOES are able to provide valuable information on structural and dynamic characteristics of the molecules analysed since the NOE is related to the distance between the two interacting nuclei and to the correlation time modulating this magnetic interaction. Previously, the 'H-lH NOE,2-4 and more recently, selective (H)C-NOE measurements5-11 together with heteronuclear 2D-NOE studies12-14 have been used to determine motions and structural properties of macromolecules. In particular, selective (H)C-NOE determinations can be used to obtain the following information:s-llv l5 (i) the intensity of geminal proton-carbon dipolar coupling and therefore the correlation time involved in the lH-13C magnetic interaction as well; (ii) the extent of the dipolar coupling between the irradiated proton and nearby quaternary carbon nuclei.An evaluation of the correlation time values can be obtained from protonxarbon conformationally independent interactions. Moreover, selective NOE can unequivocally identify pre- viously dubious n.m.r. signals; (iii) the proton and carbonyl carbon involved in intra- molecular hydrogen bonds. The detection of heteronuclear NOE effects allows the determination of preferential conformation in solution and of proton-carbon distances.1° The aim of this paper is to extend the application of selective (H}C-NOE studies to systems where proton chemical exchange occurs by combining selective homo- and hetero-nuclear results.In this investigation accurate measurements of the selective proton<arbon NOE arising from selective irradiation in the lH region of the adenosine- thymidine spectrum were obtained. Effects of intra- and inter-molecular proton exchange on (H)C-NOE have been considered by homonuclear investigation. Experiment a1 Adenosine (A) and thymidine (T) (fig. 1) were obtained from Sigma and were used without further purification by dissolving in [2H,JDMS0 to give 0.5 mol dm-3 solutions. The source of magnesium ions was Mg(ClO,), (Aldrich). In all measurements the temperature of the sample was 40°+ 1 "C. A Varian XL-200 n.m.r. spectrometer was used for recording lH and 13C spectra. 17311732 NOE of Adenosine-Thymidine Interaction H:Fdl, 4' H 3' 2' H ^:ko$ H 3' 2' 1' HO H OH OH Fig.1. Molecular structure and numbering of thymidine and adenosine. Spin-lattice relaxation rates of carbon nuclei were obtained using the inversion recovery (1 8O0-z-90"-t), pulse sequence. R, values were calculated by computer-fitting of the relaxation curves. NOE values were determined by using the equation: NOE = ( I z - I o ) / I o where Iz and I, represent the peak intensities measured under continuous and gated decoupling, respectively. A 5% experimental error was estimated for both R, and NOE measurements. The use of selective NOE measurements for proton-carbon distance determination requires some preliminary information on both proton and carbon spin-lattice relaxation rates.In fact, selective NOES are obtained by presaturating single proton resonances with a low-power decoupling pulse for at least 10 times the duration of the lH spin-lattice relaxation time of the irradiated p r ~ t o n . ~ Under our experimental conditions the proton T, for H,(T) and NH,(A) were 0.44 and 0.19 s, respectively. The fractional dipolar contribution to the carbon spin-lattice relaxation rates, rDD (determined by comparing the experimental non-selective broad-band carbon, NOE,,,,, with the theoretically expected value for 13C nuclei totally relaxed throughout lH-13C dipolar interactions), was used to determine the dipolar contribution to the experimental carbon spin-lattice relaxation rate R?"P. The following equation has been used : 1 (1) RDD = RexprDD where RYxP is the experimental carbon spin-lattice relaxation rate and RfD its dipolar contribution.From RfD of protonated carbon the correlation times modulating the C-H magnetic interactions were determined using standard C-H distances and equations determined previously.16- l7 Results and Discussion As previously rep~rted,~-ll selective proton-carbon NOE measurements, { H}C-NOE, can be used to determine proton-carbon internuclear distances in solution when an accurate analysis of the proton relaxation mechanism is accomplished. Nevertheless, when proton chemical exchange occurs, a full inspection of the magnetization transfer process is also required.18-21 In table 1 the observed carbon spin-lattice relaxation rates, the NOE(BB1, the fractional effectiveness of the dipolar contribution and R,Dn are reported.The adenosine-thymidine carbons appear to relax slightly outside the extreme narrowing conditions (cuH + cuc)2zE + 1, a maximum theoretical NOE of 1.93 is in general expected. As pointed out in the experimental section, the effective correlation time for all protonated carbons of the adenosine-thymidine system can be determined.ls. l7 The calculated z, values suggest that a unique correlation time can be considered to be modulating proton-carbon vectors within the adenosine-thymidine aromatic regions.C. Rossi, N . Niccolai, A . Prugnola and F. Laschi 1733 Table 1. Relaxation parameters and correlation times obtained for the adenosine-thymidine system 1 RfD TI2 Robs NOEobS ,.DD~ carbon cTa 163.62 156.10 152.40 150.32 149.12 140.10 135.97 119.14 109.24 88.12 87.12 86.00 83.73 73.61 70.8 1 70.30 61.80 61.20 12.02 0.2 0.45 4.6 0.29 0.20 4.9 4.9 0.15 0.25 4.3 3.3 4.5 3.4 4.3 3.7 2.5 6.3 5.0 1 .o 1.14 1.56 1.85 1.25 0.52 1.82 1.85 0.71 1.62 1.66 1.72 1.70 1.83 1.56 1.81 1.93 1.89 1.93 1.68 0.59 0.8 1 0.96 0.65 0.27 0.94 0.96 0.37 0.84 0.86 0.89 0.88 0.95 0.8 1 0.94 1 .o 0.98 1 .o 0.84" 0.12 0.36 4.4 0.19 0.06 4.6 4.7 0.06 0.21 3.7 2.9 4.0 3.2 3.5 3.5 2.5 6.2 5.0 0.84 - - 2.2 x 10-'0 - - 2.3 x 2.35 x !0-lo - - 1.8 x lo-', 1.4 x 10-lo 2.0 x lO-'O 1.55 x 1.7 x loplo 1.7 x 1.2 x lO-'O 1.5 x loplo 1.2 x 10-l0 1.3 x lo-" ~~ a ppm from TMS.obtained from the equation rDD = NOE(,,,/1.93. 1.99 has been considered. Fractional dipolar contribution to the carbon spin-lattice relaxation rates For methyl protons the NOE maximum of site A site B WiA Fig.2. The energy-level diagram for a spin 1 /2 exchanging between two sites. W, is the probability for unit time that a nucleus changes from A to €3. W,, and W,, are the transition probabilities of spin A and B, respectively. On the basis of the calculated z, values a specific interaction between the aromatic adenosine-thymidine regions can be suggested. The sugar carbons show shorter corre- lation times owing to internal reorientation contributions to the overall motion. More structural information on the adenosine-thymidine interactions can be obtained using selective heteronuclear Overhauser effect measurements. For the adenosine-thymidine system the selective heteronuclear Overhauser effects induced by saturation of NH,(A) protons in slow exchange conditions with H,(T)1734 NO E of Adenosine- Thymidine Znterac tion Table 2.Population fraction xa saturated by selective decoupling of exchangeable protons in thymidine and adenosine-thymidine systems proton thymidine system adenosine-thymidine system proton observed OH,,(T) OH,,(T) OH,,(T) OH,.(T) NH,(A) H,(T) - 0.60 0.64 0.15 0.17 0.5 1 - 0.06 0.07 - - 0.25 protons determine a positive NOE on both C2(T) and C4(T). The nuclear Overhauser effect observed in these conditions is due to a sum of two contributions as: (2) where n, is the direct NOE due to the dipolar NH,(A)-C 2,4(T) internuclear interaction. n,, the indirect contribution due to saturation transfer, is induced by NH,(A)-H,(T) chemical exchange.In the present paper an evaluation of the extent of both n, and n, is attempted; moreover, the direct term n, has been used for obtaining structural information on the intermolecular adenosine-thymidine interaction. In the presence of proton exchange between HA and HB, the selective proton decoupling of H, affects the spin population of the HA nuclei., In this case when the steady-state conditions are reached it is possible to calculate the fractional decrease of HA intensity. The energy levels and transition probabilities for a two-site exchanging system [as for the H,(T)-NH,(A) exchange] are shown in fig. 2. The effects induced by the saturation of H, nuclei on HA population are described (H}C-NOEexpt = n, + n, by2 (3) where RA = 2wA+ &, (see fig.2) is the spin-lattice relaxation rate of HA spin, oAB = - Wo = -k is the probability per unit time of a nucleus transferring from site B to site A. Wo can be identified with the first-order or pseudo-first-order rate constant for the exchange process. If Ikl and IR,I are similar and k is within the range 1OP2-1O s-l, k can be determined by using the steady-state I z ~ = I,( 1 - k/RA). (4) In our system a value of k = 0.8 s-l for the NH,(A)--H,(T) proton exchange has been found. Under these conditions of proton exchange the selective (H}C-NOE is determined by eqn (2). Two cases can be considered: (a) if the exchangeable proton Ha is dipolar coupled to geminal or vicinal C, carbons, then where (H,}C-NOE,x,t is the experimental NOE value due to the Ha saturation.In this case it has been assumed that by irradiating Ha in the presence of saturation transfer to H,, indirect NOE from H, to C, can be neglected. The validity of this assumption in the present system will be verified later. (b) If the saturated proton H, is in chemical exchange conditions with the Ha proton, and the Ha proton is dipolar coupled to geminal and/or vicinal C , carbons, then (HaK-NOEexp, = n1 ( 5 ) n, = (H,)C-NOE,,,, - n2 (6) where n, = x,(H,}C,-NOE.C. Rossi, N . Niccolai, A . Prugnola and F. Laschi 1735 150 100 6 ( P P d Fig. 3. 13C n.m.r. difference spectrum of the adenosine-thymidine system obtained by subtracting the on-resonance spectrum (fi), from the off-resonance one cf,). The carbon spectrum (f,) was obtained by presaturating the NH,(A) protons.T = 40 "C. (a)fo, ( b ) f , -fo. xa is the Ha population fraction that is saturated when Hb is selectively decoupled. xa can be obtained easily from homonuclear selective saturation experiments. When several protons, i, in a molecule are in the Ha condition, n2 becomes n, = E xi{Hi}Ca-NOE. (7) i The above treatment is applicable only if the (Ha)Ca-NOE term can be carefully estimated. In table 2, the saturated population fraction, xa, of an Ha spin when the H, resonance was selectively decoupled is reported. These results were obtained by selective proton saturation of exchangeable protons. Direct evidence of the exchange contribution on {H}C-NOE of both C2(T) and C4(T) carbons by NH2(A) proton decoupling is shown.In fig. 3 the 13C n.m.r. spectrum of the adenosine-thymidine system in DMSO solution (fo) and the difference spectrum (fi -fo), obtained by presaturating the NH,(A) protons (fi) and 100 Hz off-resonance (fo), are shown. The intensity of C6(A), C4(T) and C2(A) signals observed in the difference spectrum are of the same order of magnitude. This suggests the presence of both direct and indirect contributions to the observed hetero- nuclear Overhauser effect of C2(T) and C4(T) carbons. Using eqn (6) and the saturated population fraction values, xa (reported in table 2), the n, contributions for C2(T) and C4(T) were determined. The applicability of eqn (6) is limited to cases in which {H,}Ca-NOE can be determined. This is possible if by selective saturation of the Ha signal the 'indirect' term, n2, can be neglected becausexa < 1 and/or when THb-Ca 9 rHa-Ca. In the present system the NH2(A)-H3(T) proton exchange is considered.If the H3(T) proton is selectively decoupled, the experimental {Ha)C-NOE determined on C2(T) and C4(T) carbons is entirely due to the n, contribution. In fact, as previously shown,8-11 the n, contribution is related through the correlation time z, to the proton-carbon distance by1736 NOE of Adenosine-Thymidine Interaction Table 3. Experimental { H)C-NOE, direct n, contribution and protonsarbon distances for the adenosine-thymidine system proton H3(T) NH2(A) carbon observed (HIC-NOE" nlb r {H}C-NOE" nlb r - 0.60 0.60 2.09' 0.34 - 2. lod 0.62 0.62 2.2OC 0.40 0.09 3.04" 2.1 8d 0.22 0.22 2.72c 0.94 0.94 2.02c - - - - - - The error in distance determination was evaluated to be +_ 10%.a Observed {H)C-NOE. term. Direct Distances obtained from neutron diffraction Distances calculated using eqn (8). I I 11 9 7 6 (ppm) Fig. 4. (a) Frequency dependence of selective NOES observed on C4(T) resonance for adenosine- thymidine (.) and MgJ1-adenosine-thymidine (.). (a) The irradiated proton region. where RIc is the experimental carbon spin-lattice relaxation rate, T ~ - ~ is the distance between the saturated proton and the observed carbon, yH and yc are the magnetogyric ratios for proton and carbon, respectively, and z, is the correlation time modulating the proton-carbon magnetic interaction. Using eqn (8) and the correlation time value determined from the dipolar contribution to the spin-lattice relaxation rate of the aromatic moiety of thymidine, both H3(T)-C2(T) and H3(T)-C4(T) distances can be determined.These distances show within experimental error the same values calculated from Dreiding models or from neutron diffraction and point out the 'indirect contribution' as irrelevant to the observed {H}C-NOE. A different case is observed for C2(T) and C4(Tj by saturating the NH2(A) protons. In fact both n, and n, terms contribute to the observed (H)C-NOE. As pointed out in table 3, by subtracting theC. Rossi, N . Niccolai, A . Prugnola and F. Laschi 1737 h I- I - c) I l l I l l / I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 8 6 4 2 1 0 8 6 4 2 6 (PPm) 8 (PPm) Fig. 5. 'H n.m.r. spectra of 0.5 mol dm-3 thymidine (a) and 0.5 mol dm-3 thymidine-adenosine (b) systems.In the top traces the on-resonance spectra obtained after OH,,(T) irradiation (at the frequency indicated by the arrows). The lower traces are the off-resonance spectra. indirect contribution due to the H3(T)-NH2(A) exchange a residual direct effect is observed. This residual effect can be related to the proton-carbon distances using eqn (8). In table 3 the NH2(A)-C4(T), H3(T)-C4(T) and H3(T)-C2(T) distances determined from heteronuclear Overhauser measurements are reported along with H3(T)-C2,4(T) distances obtained from neutron diffraction Further evidence of the adenosine-thymidine interaction can be obtained if the selective proton-carbon Overhauser effects are determined in the presence of diamagnetic metal ions.In this case MgII has been used to study the effect of divalent metal ions on the strength of the nucleoside-nucleoside interaction. As shown in fig. 4, the addition of MgI1 to the adenosine-thymidine system induces a drastic reduction of the C4(T) heteronuclear Overhauser effect observed by saturation of the NH2(A) protons. No effect on the (H)C-NOE by H3(T) saturation is observed. As previously reported24 in metal-nucleotide complexes, the metal ion essentially interacts with the phosphate groups. In metal-nucleoside complexes, nitrogen and hydroxyl groups are the most important donor groups involved in metal coordination. The adenosine-thymidine interaction appears to be drastically reduced by the addition of MgII ions. In fact the metal coordination by aromatic nitrogen, carbonyl and aminic groups of both adenosine and thymidine inhibits the formation of ' Watson-Crick' complexes and the dipolar interaction NH2(A)-C4(T) disappears.Conclusions On the basis of our results it can be concluded that in the presence of proton chemical exchange some equations can be developed in order to account for the extent of the1738 NO E of Adenosine- Thymidine Interaction exchange contribution to the observed selective {H)C-NOE. The adenosine-thymidine system has been chosen because it is a suitable system for studying exchange problems; moreover some insights on the purine-pyrimidine interaction can be obtained. In fact, as shown in table 2, the addition of adenosine to the thymidine solution completely modifies the proton exchange pattern.The saturation of OH,(T) and OH,<(T) protons induces a strong saturation on the H3(T) proton in the thymidine system. In the presence of adenosine the saturation of the same protons does not significantly alter the H3(T) proton intensity, suggesting the involvement of this proton in intermolecular interactions. In fig. 5 the different extents of the saturation transfer phenomenon in thymidine and adenosine-thymidine systems on OH,<(T) saturation are shown. The results obtained on the MgII-adenosine-thymidine system support the previous interpretation on the adenosine-thymidine interaction and show the effect of MgII in destroying the nucleoside-nucleoside ' Watson-Crick ' interaction. References 1 A. Abragam, The Principles of Nuclear Magnetism (Clarendon Press, Oxford, 1961).2 J. H. Noggle and R. E. Shirmr, The Nuclear Overhauser Eflects: Chemical Applications (Academic Press, 3 C. R. Jones, C. T. Sikakana, S. Hehir, M. C. Kuo and W. A. Gibbons, Biophys. J., 1978, 815. 4 C. R. Jones, C. T. Sikakana, M. C. Kuo and W. A. Gibbons, J. Am. Chem. SOC., 1978,100, 5960. 5 J. Urawa and S. Tekeuchi, Org. Magn. Reson., 1978, 11, 502. 6 M. F. Aldersley, F. M. Dean and B. E. Mann, J. Chem. SOC., Chem. Commun., 1983, 107. 7 V. Leon, R. A. Bolivar, M. L. Tessayco, R. Gonzales and C. Rivas, Org. Magn. Reson., 1983,21,470. 8 N. Niccolai, C. Rossi, V. Brizzi and W. A. Gibbons, J. Am. Chem. SOC., 1984, 106, 5732. 9 N. Niccolai, L. Pogliani, E. Tiezzi and C. Rossi, Nuovo Cimento, 1984, 3D, 993. New York, 1971). 10 N. Niccolai, C. Rossi, P. Mascagni, P. Neri and W. A. Gibbons, Biochem. Biophys. Res. Commun., 1 1 N. Niccolai, C. Rossi, P. Mascagni, W. A. Gibbons and V. Brizzi, J. Chem. Soc., Perkin Trans. I , 1985, 12 P. L. Rinaldi, J . Am. Chem. SOC., 1983, 105, 5167. 13 C. Yu and G. C. Levy, J . Am. Chem. SOC., 1983, 105, 6994. 14 C. Yu and G. C. Levy, J. Am. Chem. SOC., 1984, 106,6533. 15 N. Niccolai, A. Sega, M. Scotton and C. Rossi, Gazz. Chim. Ztal., 1985, 115, 149. 16 A. Allerhand and R. A. Komoroski, J. Am. Chem. Soc., 1973,95, 8228. 17 R. S. Norton and A. Allerhand, J . Am. Chem. SOC., 1976,98, 1007. 18 I. D. Campbell, C. M. Dobson, R. G. Ratcliffe and R. J. P. Williams, J. Magn. Reson., 1978, 29, 397. 19 J. B. Lambert and J. Kepeers, J . Magn. Reson., 1980, 38, 233. 20 J. Henning and H. H. Limbach, J. Magn. Reson., 1982, 49, 322. 21 J. Henning and H. H. Limbach, J . Am. Chem. Soc., 1984,106, 292. 22 S. Forsen and R. A. Hoffman, J. Chem. Phys., 1963,39, 2892. 23 M. N. Frey, T. F. Koetzle, M. S. Lehmann and W. C. Hamilton, J. Chem. Phys., 1973,59, 915. 24 A. T. Tu and M. J. Heller, Structure and Stability of Metal-Nucleoside Phosphate Complexes, in Metal 1984, 124, 739. 239. Ions in Biological Systems, ed. H. Sigel (Marcel Dekker, New York, 1974). Paper 611457; Received 21st July, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301731

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987,83, 1739-1750 Adsorption of Organic Molecules on Titanium Dioxide (Rutile) Surface Yasuharu Suda and Mahiko Nagao" Research Laboratory for Surface Science, Faculty of Science, Okayama University, Tsushima, Okayama 700, Japan Adsorption isotherms of organic molecules, n-BuOH, n-BuC1 and n-C,H,, have been measured on rutile samples having a controlled number of surface hydroxyl groups. The number of irreversibly adsorbed organic molecules decreased almost linearly with increasing surface hydroxyl group content of the sample. The hydroxyl groups on the rutile surface inhibit the irreversible adsorption of n-BuC1 and n-C,H,,, but allow only n-BuOH to be adsorbed irreversibly in appreciable amounts. From the results of infrared spectra and gas chromatographic analysis, it was evident that the irreversible adsorption of n-BuOH occurs on the dehydroxylated rutile surface through the mechanism of dissociation and/or of coordination to the surface Ti4+ ions, and on the hydroxylated surface through such reactions as substitution for molecular water and esterification with acidic surface hydroxy groups.The number of water molecules substituted by n-BuOH was estimated to be 1.255 molecules nm-2. Non-polar n-C7HI6 molecules and polar n-BuC1 molecules were physisorbed on the rutile surface. The greater amount of n-BuCl adsorbed was interpreted in terms of the additional interaction between the dipole of n-BuC1 and the electrostatic field of the rutile surface. The presence of hydroxyl groups, which are produced by dissociative chemisorption of water, on a metal oxide surface has a significant influence on the surface properties of the solid.In previous studies the adsorption of a series of normal aliphatic (C,-C,) alcohols on zinc oxide' and titanium dioxide (rutile)2 have been investigated in order to elucidate the effect of surface hydroxyl groups on the adsorption of alcohol molecules. The results showed that the interaction of the rutile surface with alcohol molecules is stronger than that with water, in contrast to the case of zinc oxide. In the present work, the interaction of the rutile surface with three kinds of organic molecules which have different functional groups was investigated in connection with the effect of surface hydroxyls on the adsorption of these organic molecules.Experiment a1 Materials and Pretreatment The original rutile sample used in this study was supplied by Teikoku-kako Co., and was the same as that used in the previous work.2 The sample was first degassed at 600 "C for 4 h under a vacuum of 1.33 mN m-2 in order to remove surface contaminants. The sample was then kept for 15 h at room temperature in contact with saturated water vapour to promote complete surface hydroxylation. Rutile samples covered with a different number of surface hydroxyls were obtained by evacuating the fully hydroxylated sample at any desired temperature between 25 and 600°C under a vacuum of 1.33 rnNmv2. The organic compounds used as adsorbates were butan-1 -01 (n-BuOH), 1 - chlorobutane (n-BuCl), and n-heptane (n-C,HI6), all of which were guaranteed grade 17391740 Adsorption of Organic Molecules on Rutile reagents of Nakarai Chemicals.These organic liquids were purified by distillation in the usual way and stored in contact with 4A molecular sieves dehydrated at 500 "C for 5 h. The adsorbates were allowed to evaporate into gas reservoirs equipped with a trapping tube and a greaseless stopcock and were subjected to several freeze-pump-thaw cycles before use. Determination of Surface Hydroxyl Content and Surface Area The surface hydroxyl content of the rutile sample, which is the number of hydroxyl groups remaining on the surface after evacuating the fully hydroxylated sample at a given temperature, was determined by the successive ignition-loss method.l? The specific surface area of the sample degassed at 600 "C was found to be 9.18 m2 g-l by applying the B.E.T.method to N, adsorption data. Measurement of Adsorption Isotherms of Organic Molecules The first adsorption isotherm of organic molecules was measured at 25 "C for the sample having a definite number of surface hydroxyls. The sample was then exposed to the saturated vapour of the same adsorbate for 15 h to ensure complete adsorption. After evacuating this sample at 25 "C under a vacuum of 1.33 mN m-2 for 4 h, the second adsorption isotherm was determined at 25 "C. The apparatus and procedures applied have been described elsewhere. Measurement of I.R. Spectra Infrared spectra of adsorbed species were measured by the transmission method for surfaces of both strongly dehydroxylated (600 "C evacuated) and fully hydroxylated (25 "C evacuated) samples.A self-supporting sample disc of 2 cm diameter was placed in an in situ cell fitted with fluorite windows. Infrared spectra were recorded by using a Nippon-Bunko model A302 diffraction grating spectrophotometer. Measurement of Gas Chromatograms n-BuOH vapour was adsorbed on rutile surfaces having a different number of surface hydroxyls (i.e. samples treated at 25, 150 and 600 "C) until a monolayer was formed. The gas phase in equilibrium with the adsorbed phase was then analysed by gas chromato- graphy. A glass tube column was packed with Porapak-N and kept at 120 "C during measurement. A thermal conductivity detector (t.c.d.) was used at a current of 125 mA; the flow rate of He carrier gas was 40 cm3 min-l.Results and Discussion Relation between the Amount of Adsorbed Molecules and the Surface Hydroxyl Content of the Sample The surface hydroxyl content of rutile sample is shown in fig. 1 as a function of evacuation temperature. In the determination of this content it was assumed that a water molecule is liberated by condensation of the two surface hydroxyls. It is seen from fig. 1 that the number of surface hydroxyls exhibits a maximum (7.88 OH groups nm-,) for the fully hydroxylated surface evacuated at 25 "C and decreases with increasing temperature, especially in the range 25-300 "C, and that for the sample evacuated at 600 "C hardly any hydroxyls are left. Molecular water was also present on the surface treated at lower temperatures (< 150 "C), as will be discussed later.N I 5 z 2 z 0 M .Y 5 s s U c k U Y. Suda and M . Nagao f 1741 degassing temperature/"C Fig. 1. Surface hydroxyl content of rutile sample evacuated at various temperatures. Fig. 2 represents the first and second adsorption isotherms of n-BuOH on rutile samples with different hydroxyl contents. The first adsorption isotherm, especially for the sample treated at a higher temperature and hence with a lesser hydroxyl content, is close to the Langmuir type in shape. As is evident from fig. 2, a steep rise cannot be observed in the isotherm near the saturated vapour pressure, which is indicative of the absence of multilayers. In the previous work,2 the adsorption isotherms of a series of normal aliphatic alcohols from MeOH to 1-PrOH on rutile were measured.The results showed that the tendency for multilayer formation decreases with increasing alcohol chain length. It was considered that in these systems the first adsorbed layer has an autophobic n a t ~ r e , ~ on which further adsorption is restricted owing to a strong and vertical orientation of the hydrocarbon chain. The tendency for multilayer formation of n-BuOH is smaller than that of l-PrOH2 and accordingly the present result may be regarded as an extension of the previous results. Fig. 2 also shows that the amount of n-BuOH in the first adsorption increases with decreasing surface hydroxyl content, while that in the second scarcely changes with hydroxyl content of the sample. The adsorption isotherms of n-BuC1 are shown in fig.3. The shape of the isotherm for n-BuC1 is close to the type I1 in Brunauer's classification and is indicative of a multilayer formation. The amount of n-BuC1 in the first adsorption increases with decreasing surface hydroxyl content, similarly to the case of n-BuOH adsorption described above, while that in the second decreases with decreasing hydroxyl content of the sample. The shape of the adsorption isotherm of n-C,H,, with no functional groups is of type I1 and similar to that of n-BuC1, as shown in fig. 4. However, the difference in the adsorbed amounts between the first and second adsorptions is not so large for this adsorbate. In particular, for the sample evacuated at 25 "C the second adsorption isotherm is very similar to the first, indicating an entirely reversible adsorption on the fully hydroxylated surface. The effect of variation of surface hydroxyl content on the amount of n-C,H,, adsorbed is very small for both first and second adsorptions; however, it can be said1742 Adsorption of Organic Molecules on Rutile 0 0.2 0.4 0.6 0.8 1.0 relative pressure (PlP,,) Fig.2. Adsorption isotherms of n-BuOH on rutile samples evacuated at various temperatures (“C) after complete hydroxylation: ‘J, 25; 0, 100; A, 150; 0, 200; 0, 300; Q, 600. Open and filled symbols represent the first and second adsorption, respectively. 0 0.2 0.4 0.6 0.8 1.0 relative pressure (PIP,) Fig. 3. Adsorption isotherms of n-BuC1 on rutile samples evacuated at various temperatures after complete hydroxylation. Symbols are the same as those in fig.2.Y . Suda and M . Nagao 1743 1 1 I I 0 0.2 0.4 0.6 0.8 1.0 relative pressure (PIP,) Fig. 4. Adsorption isotherms of n-C,H,, on rutile samples evacuated at various temperatures after complete hydroxylation. Symbols are the same as those in fig. 2. I I I I I I 1 I I E 0 2 4 6 0 surface hydroxyl content/OH groups nm-* Fig. 5. Relationship between the monolayer capacity of adsorbed organic molecules and the surface hydroxyl content of rutile sample: 0, n-BuOH; 0, n-BuC1; a, n-C,H1,. Open and filled symbols represent V,, and Vmz, respectively. that the smaller the hydroxyl content of the sample, the larger the amount of the first adsorption and the smaller the amount of the second, as in the case of n-BuC1. This implies that the presence of hydroxyl groups appreciably affects the adsorption of organic molecules, even if the molecules had neither functional groups nor polar nature.Fig. 5 illustrates the relationship between the monolayer capacity for organic molecules1744 Adsorption of Organic Molecules on Rutile and the surface hydroxyl content of the rutile sample, the former being estimated by the B-point method for n-BuOH and by the B.E.T. method for the other two adsorbates. Here, Vml and Vm2 are the monolayer capacities based on the first and second adsorption isotherms, respectively. For the three kinds of adsorbate molecules, Vml, which involves both chemisorption and physisorption, is in the order n-BuOH > n-BuC1 > n-C,H,, and Vm2 (involving only reversible physisorption) is in the order n-BuC1 > n-C,Hl, > n- BuOH over the whole range of surface hydroxyl content. As shown in fig.5, an excellent linear relationship is established between the amount of adsorbed molecules and the surface hydroxyl content of the sample, regardless of the nature of organic molecules. The variation of monolayer capacity for n-BuOH with surface hydroxyl content can be regarded as an extension of the previous results obtained for the system of rutile and C,-C, normal aliphatic alcohols.2 It is interesting that the difference between V,, and Vm2 values for this adsorbate is significantly large on the hydroxylated surface as well as on the dehydroxylated surface and that the Vm2 value is almost constant, regardless of the surface hydroxyl content of the sample.Furthermore, the average area occupied by an adsorbed n-BuOH molecule is estimated to be 0.300-0.360 nm2 from the Vml values, i.e. very close to that for 1 -PrOH in the previous work.2 This fact seems to support the adsorption model based on the concept of autophobicity found in fig. 2, in which the n-BuOH molecule is adsorbed by directing its OH group to the surface and by keeping the hydrocarbon chain perpendicular to the surface. For n-BuC1, Vml decreases linearly with increasing surface hydroxyl content, while Vm2 increases linearly. The area occupied by an n-BuC1 molecule estimated from Vml values ranges from 0.404 nm2 on the dehydroxylated surface to 0.497 nm2 on the hydroxylated surface, the latter value being very close to the calculated value (0.500 nm2) based on the assumption that n-BuC1 molecules lie with the hydrocarbon chain parallel to the surface.4 Therefore, it seems that on the hydroxylated surface the n-BuC1 molecule is adsorbed on top of surface hydroxyls, lying flat with the hydrocarbon chain parallel to the surface.On the other hand, on the dehydroxylated surface n-BuC1 is assumed to be adsorbed by directing its negative pole, or the chlorine atom, to the surface Ti4+ ions. Taking account of the occupied area on the dehydroxylated surface described above and of the interaction between the dipole of the n-BuC1 molecule (dipole moment 1.90 Dt) and the electrostatic field of the rutile s ~ r f a c e , ~ it would seem reasonable to consider that n-BuC1 is adsorbed on the dehydroxylated surface keeping the hydrocarbon chain not parallel to the surface but tilted against it.The variation of the monolayer capacity for n-C,H,, with the surface hydroxyl content of the sample has the same tendency as that in the case of n-BuC1. Taking account of the occupied area, which is estimated to be 0.704-0.771 nm2 from Vml values, it is likely that the n-C,H,, molecule is adsorbed with the hydrocarbon chain parallel to the surface, regardless of the degree of surface hydroxylation. The difference between Vml and Vm2, denoted as Krr, is plotted against the surface hydroxyl content of the sample in fig. 6. Here, refers to the amount of adsorbed molecules remaining on the surface after evacuation at 25 "C under vacuum of 1.33 mN m+. The general trend is very clear from fig.6 ; the amount of irreversibly adsorbed organic molecules exhibits a maximum on the dehydroxylated surface and decreases linearly with increase in the surface hydroxyl content. Furthermore, it is also found that for n-BuOH is much larger than those for the other two adsorbates over the entire range of hydroxyl content. It is worth noting that on the fully hydroxylated surface Krr values for n-BuC1 and n-C7H16 are very close to zero, while that for n-BuOH shows a value of 2 molecules nmP2, in agreement with the values for C,-C, aliphatic alcohols on the same surface.2 The dependences of Krr and V,, upon the surface hydroxyl content suggest that the surface hydroxyls on rutile inhibit irreversible adsorption, but t 1 D = 3.33564 x C m.Y. Suda and M . Nagao 1745 ti I I I I I I I 1 I surface hydroxyl content/OH groups nm-2 Fig.6. Relationship between the amount of irreversible adsorption and the surface hydroxyl content of rutile sample: 0, n-BuOH; 0, n-BuC1; A, n-C,H,,. act as effective sites for the reversible adsorption of both n-BuC1 and n-C,H16. The surface hydroxyls also act as effective sites for the irreversible adsorption of only n-BuOH. Therefore, it seems that the surface evacuated at 25 "C after the first adsorption of n-BuOH is covered with C4Hg groups, regardless of the surface hydroxyl content of the sample, which results in unchanged Vm2 values for n-BuOH (fig. 5). Adsorbed State of Organic Molecules n-BuOH Fig. 7 shows the i.r. spectra of n-BuOH adsorbed on the dehydroxylated and hydrox- ylated rutile surfaces.For n-BuOH adsorbed on the dehydroxylated surface at an equilibrium pressure of 67 N m-2, three distinct absorption bands appear at 2964, 2875 and 2935 cm-l, the first two being assigned to the asymmetric and symmetric CH stretching vibrations of the methyl group, respectively, and the last to the asymmetric CH stretching vibration of the methylene group [spectrum ( t ~ ) ] . ~ The absorption bands due to the bending vibrations were also observed at 1458 and 1375 cm-l for the methyl group, although the spectra were not illustrated in the figure. In addition to these CH bands, spectrum (c) taken after subsequent evacuation at ambient beam temperature (ABT) exhibits a new band at 3655cm-l, which may be due to the free hydroxyl groups.7v8 It has been revealed that dissociative adsorption of alcohols occurs on the dehydroxylated rutile surface to produce both alkoxyl and hydroxyl g r o ~ p s .~ - ~ ~ In spectrum (b) the 3655 cm-l band is obscure, probably owing to the fact that the newly formed hydroxyl groups can also physisorb n-BuOH molecules by hydrogen bonding. As is obvious from spectrum (e), the fact that the characteristic bands of the alkyl group remain tenaciously even after evacuation at ABT substantiates the presence of a large amount of n-BuOH molecules adsorbed irreversibly. On the other hand, it is well known that the fully hydroxylated rutile surface carries both surface hydroxyls produced by dissociative chemisorption of and molecular water adsorbed as a ligand coordinating to the surface Ti4+ ions.l79 l8 In the region of the OH stretching vibrations, the fully hydroxylated rutile gives i.r.spectrum 58 FAR I1746 Adsorption of Organic Molecules on Rutile 38 36 34 32 30 20 30 36 34 32 30 28 wavenumber/ 1 O2 cm-' Fig. 7. Infrared spectra of n-BuOH adsorbed on the dehydroxylated [(a)-(e)] and hydroxylated [(ft(i)] rutile surfaces: (a) background spectrum of rutile dehydroxylated at 600 "C; (b) adsorption of n-BuOH at 67 N m-2; (c) after evacuation of n-BuOH vapour at ABT; ( d ) adsorption of n-BuOH at 670 N m-2; (e) after evacuation of n-BuOH vapour at ABT; (J) background spectrum of the hydroxylated rutile; adsorption of n-BuOH at (g) 67 N m-2 and (h) 670 N mW2; (i) after evacuation of n-BuOH vapour at ABT. (f) with three distinct bands: the first at 3655 cm-l due to the free hydr~xyls,~* the second at 3520cm-l due to the hydroxyls perturbed by hydrogen-bonded water molecules8 and the broad band in the vicinity of 3400 cm-l, which can be assigned to the vibrations either of adsorbed water molecules8* l9 or of mutually hydrogen-bonded hydroxyls.16 When n-BuOH vapour is adsorbed on the hydroxylated surface at an equilibrium pressure of 67 N m-2, the CH bands assignable to the C4H, groups appear explicitly at 2968,2940 and 2883 cm-l, accompanying a remarkable decrease in intensity of the 3655 cm-l band [spectrum (g)].Furthermore, as the vapour pressure of n-BuOH is increased to 670 N m-2 the absorption band assigned to the free OH groups disappears and there exists only a broad band centred at ca. 3375 cm-l [spectrum (h)].In spectrum (i), observed after evauation of n-BuOH vapour at ABT, the fact that the absorption bands due to the C4Hg groups remain tenaciously suggests that n-BuOH molecules are adsorbed strongly even on the hydroxylated surface. The 3655cm-1 band, which is obscure in the presence of n-BuOH vapour, appears again in spectrum (i), but the band intensity is much weaker than that in the background spectrum cf). From gas-chromatographic analysis of the vapour in equilibrium with the adsorbed phase, water vapour was proved to be involved in the gas phase. The more the surface hydroxyl content of the sample, the greater the amount of water vapour in the gas phase; the number of water molecules liberated from the rutile surface by adsorption of n-BuOH was found to be 1.53,0.275 and 0.0095 molecules nm-2 for the samples evacuated at 25, 150 and 600 "C, respectively.This result substantiates the occurrence of the substitution of alcohol for molecular water remaining on the surface as well as the esterification reaction between the surface hydroxyls and alcohol molecules.Y. Suda and M. Nagao 1747 Thus, on the dehydroxylated rutile surface, the adsorption of n-BuOH molecules seems to proceed in such ways as dissociation to produce both surface butoxyl and hydroxyl groups, formation of coordinative bonds to the surface Ti4+ ions by utilizing its free electron pair in oxygen (this mechanism includes the formation of an intermediate structure between the coardinatively and dissociatively adsorbed species20) and esterifi- cation with surface hydroxyls newly formed by the dissociative adsorption.Jones and Hockey15 assumed that most of the external surface of rutile is composed of three planes, (1 lo), (10 1) and (loo), and proposed the model for water adsorbed on these planes; adsorption on the (1 10) plane involved the dissociation of water molecules to produce two surface hydroxyls, whereas on the (101) and (100) planes water molecules are adsorbed as a ligand coordinated to a surface Ti4+ ion. By analogy with their model, the following adsorbed states could be proposed for n-BuOH adsorption: R R R (b) (c) Type (a) seems to correspond to adsorption on the (1 10) plane, and type (b) and (c) to the (101) or (100) planes. In the case of type (a), the Ti4+ site energy is so strong that the completely dissociative chemisorption of alcohol occurs on the dehydroxylated surface to produce free hydroxyl groups.On the other hand, in the cases of types (6) and (c), the Ti4+ site has not enough energy to dissociate alcohol molecules, so that the hydroxyl group of n-BuOH interacts with an adjacent surface oxide ion without complete rupture of the 0-H bond in alcohol. Assuming such adsorption mechanisms, the features of i.r. spectra shown in fig. 7(a)-(e) can be explained as follows. In the initial stages of adsorption the dissociation of n-BuOH occurs to produce free hydroxyl groups, giving the 3655 cm-l band [spectrum (c)]. As the adsorption proceeds, the broad band centred at 3400 cm-l, due to the perturbed vibration of OH groups in alcohol, increases and the free OH band disappears, resulting from a consumption of OH groups in the esterification reaction with n-BuOH molecules [spectrum (41.As in the cases of C,-C, alcohols on the hydroxylated rutile surface,2 it is conceivable that the adsorption of n-BuOH proceeds by rapid displacement for molecular water present on the surface, esterification with surface hydroxyls bearing an acidic nature and reversible physisorption on the free OH groups through the formation of hydrogen bonds. The presence of molecular water on the rutile surface treated at lower temperatures should be emphasised because these water molecules can be displaced by alcohol molecules, as confirmed by gas chromatographic analysis. For the sample treated at 150 "C, on which molecular water should be removed l8 water vapour was detected in the gas phase in equilibrium with the adsorbed phase, which suggests strongly that esterification occurs between the acidic surface hydroxyls2l* 22 and some of the adsorbed n-BuOH.It is reasonable, therefore, to assume that the difference in the1748 Adsorption of Organic Molecules on Rutile wavenumber/ 1 O2 cm-' Fig. 8. Infrared spectra of n-C7H16 adsorbed on the dehydroxylated [(a)-(d)] and hydroxylated [(e)-(h)] rutile surfaces: (a) background spectrum of rutile dehydroxylated at 600 "C; adsorption of n-C,H,, at (b) 67 N m-2 and ( c ) 930 N m-2; ( d ) after evacuation of n-C7H1, vapour at ABT; (e) background spectrum of the hydroxylated rutile; adsorption of n-C,H,, at cf) 67 N m-2 and (g) 930 N md2; (h) after evacuation of n-C,H,, vapour at ABT.number of water molecules liberated between the samples treated at 25 and 150 "C corresponds to the amount of water adsorbed: 1.255 H,O molecules nmP2. This amount of water is included in the water content of the sample treated at 25 "C (fig. 1). The i.r. spectra of n-C,H,, adsorbed on the dehydroxylated and hydroxylated surfaces are shown in fig. 8. When n-C,H,, molecules are adsorbed on the dehydroxylated surface, the CH stretching bands at 2964 and 2860 cm-l due to the methyl group and those at 2933 and 2840cm-l due to the methylene group appear distinctly and increase their intensities as more n-C,H,, molecules are adsorbed [spectra (b) and (c)]. Spectrum ( d ) , recorded after exposing the sample disc to the saturated vapour of n-C7H16 and subsequently evacuating at ABT, shows that the residual bands are just observable in the CH stretching region, which corresponds to the results for &.in fig. 6. This finding is particularly surprising in view of the fact that n-C,H,, molecules have neither permanent dipole moments nor reactive functional groups. From the most recent in~estigation,~ it can be accepted that the dehydroxylated rutile surface has an electro- static field enhanced by the exposure of higher valence ions such as Ti4+ and 02-. Thus, the irreversible adsorption of such a non-polar molecule might occur owing to a stronger interaction between the electrostatic field of the surface and the induced dipole of the molecule, in addition to the contribution of dispersion force.On the fully hydroxylated surface the CH stretching bands increase in intensity with increasing vapour pressure of n-C7HI6, which is accompanied by a shift of the free OH band to lower frequencies as much as 15 cm-1 [spectra (f) and (g)], Spectrum (h) is restored to its original pattern [spectrum (e)] after evacuation at ABT, in agreement with the observations by Graham et aZ.,23 who carried out an infrared study of the adsorption of hydrocarbons on rutile.Y. Suda and M . Nagao 1749 wavenumber/ lo2 cm-' Fig. 9. Infrared spectra of n-BuC1 adsorbed on the dehydroxylated [(a)-(d)] and hydroxylated [(e)-(h)] rutile surfaces : (a) background spectrum of rutile dehydroxylated at 600 "C ; adsorption of n-BuC1 at (b) 67 N m+ and (c) 870 N m-2; ( d ) after evacuation of n-BuC1 vapour at ABT; (e) background spectrum of the hydroxylated rutile; adsorption of n-BuC1 at df) 130 N m-2 and (g) 800 N mP2; (h) after evacuation of n-BuC1 vapour at ABT.This is also consistent with the fact that the two isotherms, the first and second ones in fig. 4, are exactly the same. These results suggest a weak interaction between the rutile surface covered with hydroxyls and the n-C,H,, molecules, probably through a dispersion force. n-BuC1 Fig. 9 shows the i.r. spectra of n-BuC1 adsorbed on rutile. On the dehydroxylated surface the features of the spectra are much like those described for n-C,H,, adsorption. The sharp bands in the CH stretching region increase in intensities as more and more n-BuC1 molecules are adsorbed [spectra (b) and (c)], In the region of lower frequencies, though not illustrated in the figure, the deformation bands appeared at 1458 and 1380 cm-l for the methyl group and at 1465 cm-l for the methylene group, and the band due to the symmetric deformational vibration of the methylene group to which the chlorine atom is attached was observed at 1433 cm-l.After evacuation at ABT these four bands disappeared, but the absorption bands in the CH stretching region still remained slightly [spectrum (41, indicating an irreversible adsorption on this surface. It is reasonable to consider that reversible physisorption is the main process occurring on this dehydrox- ylated surface. However, the dehydroxylation seems to enhance the electrostatic field strength of the surface as described above, so that the stronger interaction with the permanent dipole of n-BuC1 molecules might lead to irreversible adsorption.As is seen from fig. 9, when n-BuC1 molecules are adsorbed on the hydroxylated surface, the increase in intensity of the CH bands is accompanied by an increase in the intensity of the 3525cm-l band, which is assigned to the OH groups perturbed by hydrogen bonding [spectra (f) and (g)].8 Moreover, it can also be found that the free OH band at 3655 cm-l shifts to lower frequencies, indicative of the interaction between1750 Adsorption of Organic Molecules on Rutile the free hydroxyl groups and the adsorbate molecules, and it overlaps with the band of perturbed hydroxyls at 3525 cm-l to become a shoulder at 3640 cm-l [spectrum (g)].Taking into account the fact that the free OH band at 3655 cm-l shifts to 3640 cm-l owing to interaction with n-C,H,, molecules on the hydroxylated surface, it seems reasonable to assume that the shoulder at 3640 cm-l is assigned to the hydroxyl groups perturbed by interaction with C,H, groups of n-BuCl molecules. On the other hand, the band at 3525 cm-l can be assigned to the OH groups perturbed by the formation of hydrogen bonds to C1 atoms in n-BuC1. For the hydroxylated surface it might be considered from these i.r. spectra that n-BuC1 is adsorbed through hydrogen bonding by directing its C1 atom to the hydrogen atom of surface hydroxyls, lying C,H, group down on adjoining hydroxyl groups with the hydrocarbon chain parallel to the surface. Spectrum (h), measured after exposing the sample disc to the saturated vapour of n-BuC1 and subsequently evacuating at ABT, is almost restored to its original pattern [spectrum (e)] except that the residual CH bands are still observable in the spectrum. From the comparison between fig.8 and 9, it is found that the interaction of n-BuCl with the hydroxylated surface is stronger than that of n-C7H16 with the same surface, because the following two effects might probably contribute besides the dispersion force: the interaction between the permanent dipole of n-BuC1 and the electrostatic field of the solid surface, and the hydrogen bonding between its C1 atom and the surface hydroxyl groups. References 1 M. Nagao and T. Morimoto, J. Phys. Chem., 1980,84, 2054. 2 Y. Suda, T. Morimoto and M. Nagao, Langmuir, 1987,3,99. 3 J. Barto, J. L. Durham, V. F. J. Baston and W. H. Wade, J. Colloid Interface Sci., 1966, 22, 491. 4 T. Morimoto and Y. Suda, Langmuir, 1985, 1, 239. 5 Y. Suda and M. Nagao, to be published. 6 L. J. Bellamy, The Infrared Spectra of Complex Molecules (Wiley, New York, 1958). 7 P. Jackson and G. D. Parfitt, Trans. Faraday SOC., 1971, 67, 2469. 8 D. M. Griffiths and C. H. Rochester, J. Chem. SOC., Faraday Trans. 1, 1977,73, 1510. 9 C. M. Hollabaugh and J. J. Chessick, J . Phys. Chem., 1961, 65, 109. 10 A. A. Isirikyan, A. V. Kiselev and E. V. Ushakova, Kolloid Zh., 1963, 25, 125. 11 P. Jackson and G. D. Parfitt, Trans. Faraday SOC., 1971, 67, 2469. 12 P. Jackson and G. D. Parfitt, J. Chem. SOC., Faraday Trans. I , 1972, 68, 1443. 13 G. D. Parfitt, Prog. Surf Membr. Sci., 1976, 11, 181. 14 P. Jones and J. A. Hockey, Trans. Faraday SOC., 1971, 67, 2669. 15 P. Jones and J. A. Hockey, Trans. Faraday Soc., 1971, 67, 2679. 16 M. J. Jaycock and J. C. R. Waldsax, J. Chem. SOC., Faraday Trans. I , 1974,70, 1501. 17 M. Primet, P. Pichat and M-V. Mathieu, J. Phys. Chem., 1971, 75, 1216. 18 M. Primet, P. Pichat and M-V. Mathieu, J. Phys. Chem., 1971, 75, 1221. 19 G. Munuera and F. S. Stone, Discuss. Faraday SOC., 1971, 52, 205. 20 H. Knozinger, Z. Phys. Chem., 1970, 69, 108. 21 H. P. Boehm and M. Herrmann, 2. Anorg. Allg. Chem., 1967,352, 156. 22 H. P. Boehm, Discuss. Faraday SOC., 1971,52, 264. 23 J. Graham, C. H. Rochester and R. Rudham, J. Chem. SOC., Faraday Trans. I , 1981,77, 2735. Paper 611543; Received 28th July, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301739

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987,83, 1751-1759 Origin of Boron Mobility over Boron-impregnated ZSM-5 A Combined High-resolution-Solid-state llB Nuclear Magnetic Resonance/Infrared Spectral Investigation Moein B. Sayed Chemistry Department, Faculty of Science, Al-Azhar University, Cairo, Egypt Boron-impregnated ZSM-5 samples have been shown by llB n.m.r. spectro- scopy to have different phases, which are dependent on the zeolite pre- treatment conditions. While the original ZSM-5 sample reveals a resonance associated with tetrahedral boron (incorporated during synthesis in the zeolite lattice at a typical location of zeolitic aluminium), the other impregnated samples show a variety of spectral patterns associated with different phases of boron. For non-pretreated impregnated samples, boron exists as H,BO,, as indicated by a very broad resonance at lower field.Pretreatment at 0.1 mPa/673 K effects the condensation of H,BO, with the host zeolite Brnrnsted sites into an ESi-0-B(OH,) species. If excess of boron is present it exists as amorphous B,O,. Pretreatment at 0.1 mPa/ 1073 K assists in a further dehydration into a surface-annealed Pyrex-like structure. This structure does not exhibit the Pyrex stability, where it reverts to the less-structured phase upon zeolite rehydration. This suggests a role for water in mobilizing boron on ZSM-5 surfaces. An i.r. investigation of methanol reaction (static conditions) over samples pre- treated at 0.1 mPa/673 K reveals and confirms the boron mobility for the impregnated samples. Methyl borate ester is shown to form at 293 K in proportions increasing linearly with the impregnated boron and is identified in the gas phase.Similar vaporization of boron or boron ester is not evident for the original sample which contains framework boron. Gas chromatography of both toluene alkylation (by methanol) and disproportionationl has revealed irregular changes of zeolite activity and shape-selectivity on stream with time as induced by zeolite impregnation. Irregular changes appear to be a function of both zeolite boron content and the catalytic process involved. The less boron impregnated in a sample and the more the reaction involves water, the more irregular changes are observed. This has been explained in terms of boron mobility within the zeolite. Another unusual aspect attributed to boron mobility is the molar benzene/xylenes ratio (expected to be 1) which changes dramatically during the course of toluene disproportionation.l The reaction began yielding dominantly benzene, continued giving a ratio of ca.1 and surprisingly ended with the lower ratio of 0.5. Toluene initially reacts by dealkylation, then after being equilibrated over the zeolite disproportionation takes place, and finally owing to an excess of trapped methyl groups (from dealkylation) side alkylation occurs, resulting in the lower ratio. In order to build up a good understanding of this unusual behaviour induced by the zeolite impregnation, a study of the state of the boron modifier has been initiated. Since the modified zeolite must be pretreated before use in catalysis and the pretreatment conditions vary with the catalytic process, attention should be paid to comprehending the phase with which the modifier exists and where it might vary with the pretreatment conditions.This could be of interest in discussing the role of the modifier in altering the catalytic process. The role of impregnated boron in modifying ZSM-5 acidity has 17511752 Boron Mobility over ZSM.5 been the subject of a microcalorimetric study of ammonia sorptioa2 Kaeding3 has demonstrated the state of impregnated boron to be amorphous B203. Gabelica et u ~ . , ~ Y using a high-resolution solid-state llB m.a.s.n.m.r. have succeeded in distinguishing boron in different phases in zeolites. Boron introduced into the zeolite by impregnation occupies a different phase from that introduced during the zeolite synthesis.** Boron introduced by impregnation is believed to occupy zeolite Brarnsted sites, which shows up in i.r.spectroscopy as a linear decrease of Brarnsted site band absorbance (population) at 3610 cm-l with the zeolite boron content and as a new band at ca. 930 cm-l appears (associated with -Si-O-B= skeleton). The excess boron is believed to exist as B203.1 The aim of the present investigation is to present additional information helpful both in disclosing the state of the boron at different pretreatment conditions and in explaining the key points of observation reported above. Investigation of the modifier boron has been followed by high-resolution solid-state (h.r.s.s.) llB m.a.s.n.m.r.spectroscopy, while the study of its nature and properties was followed by i.r. spectroscopy with reference to methanol reactions, as a system involving water suitable for probing boron mobility over the zeolite particle. Experiment a1 Materials Pure and degassed (by freeze-pumpthaw cycles) methanol (Merck) was used to probe boron mobility within boron-impregnated ZSM-5. The adsorbent was HZSM-5 modified by boron impregnation. Several samples exhibiting 0.3, 0.6 and 1.6 wt % boron were prepared from original HZSM,-5 by reaction with boric acid1 and were labelled as HZSM,-5, HZSM,-5 and HZSM3-5, respectively. Methods and Equipment Zeolite samples (ca. 14 mg) were pressed (ca. 7500 kPa) into thin wafers (ca. 5 mg cm-2) and were pretreated in a greaseless infrared cell at ca.0.1 mPa and various temperatures : 293, 673 and 1073 K. The wafers were subjected to analysis by i.r. and llB n.m.r. For the catalytic methanol conversion process, wafers pretreated at 673 K were exposed to equal pressures of methanol (1.33 kPa) at 293 K within the i.r. reactor. 1.r. sampling of both the reaction gas phase and the surface was followed under static conditions for a reaction proceeding at increased temperatures and time intervals typical of the Mobil process.6 1.r. spectra (4000-1000 cm-l) were measured in absorbance mode with 2.8 cm-l resolution using a double-beam Perkin-Elmer 580 spectrometer. H.r.s.s. llB m.a.s.n.m.r. data were collected by J. B. Nagy at Namur University using methods and equipment described The chemical shifts were determined using BF, * OEt, as an external standard in a Bruker CXP-200 spectrometer.Results and Discussion The State of the Modifier Boron Boron, as previously demonstrated,l exists as reactant H3BO3 for non-pretreated impregnated ZSM-5. Zeolite pretreatment at 0.1 mPa/673 K effects condensation of occluded H3BO3 with the host zeolite Brarnsted sites into boron linked to the zeolite as illustrated in scheme 1 (later). The suggested skeleton, rSi-O-B(OH), is based on several observations, such as the linear decrease of the zeolite Brsnsted site absorbance at 3610 cm-l with increased boron content and appearance of a skeletal =Si-O-B= i.r. band at 915 cm-l which intensifies with boron content.' Consistent with this is the appearance of a broad i.r. band at ca.3660 cm-l for samples pretreated at 673 K (later), which indicates strong internal H-bonding of the type found for adjacent silanols thatM. B. Sayed 1753 P 1 1600 1200 800 wavenumberlcm-' Fig. 1. / 2.89 ppm \ - 6 (PPm) Fig. 2. Fig. 1. 1.r. spectra (160&800 cm-l) of differently pretreated ZSM-5: (a) non-impregnated HZSM,-S, pretreated at 673 K; (b) impregnated HZSM,-5, pretreated at 673 K; (c) as for (b), after pretreatment at 0.1 mPa/1073 K. Fig. 2. H.r.s.s. llB n.m.r. spectra of (a) HZSM,-5, with m.a.s., scale bar = 10 ppm; (b) HZSM,-5, with m.a.s., scale bar = 10 ppm; (c) HZSM,-5, without m.a.s., scale bar = 20 ppm. persists for silica at such a high pretreatment temperat~re.~ Fig. l(a) shows the i.r. spectrum of HZSM,-5 in the region of oxide vibrations, 1600-800 cm-l.A small contribution from lattice boron (incorporated during the zeolite synthesiss) is indicated by a trigonal B-0 stretch (1385 cm-l) and skeletal mode, ESi-O-B= (915 cm-l). These boron absorptions intensify with increasing impregnation [fig. 1 (b)] for samples pretreated at 673 K and their absorbances are greatly enhanced by zeolite pretreatment at 1073 K [fig. l(c)]. These data, interpreted independently, could be misleading, since they give the impression that heat treatment assists boron diffusion from amorphous H,BO, into the zeolitic lattice, which is not the case (see later). Consistent with the i.r. data of the zeolite, llB n.m.r. spectra reveal that boron exists as amorphous H3BO3 for non-pretreated impregnated samples and even for pretreated samples having an excess of boron.This phase [fig. 2(c)] is distinct from the more structured boron incorporated into the zeolitic lattice [fig. 2 (a)]. While the former has a broad resonan~e,~? the latter has a sharp resonance at ca. -2.89 ppm, indicative of tetrahedral Tetrahedral boron [fig. 2(a)] is the only phase found for HZSM,-S.* Additionally, HZSM,-5 and HZSM,-5 show a less-ordered phase indicated by a broader [fig. 2 (b)] resonance at higher field. This phase can be associated with ESi-O-B(OH),.1754 Boron Mobility over ZSMS , I " " ' " ' ' ~ 1 " ~ " ' I 90 0 - 90 6 (PPm) Fig. 3. H.r.s.s. llB n.m.r. spectra (90 to -90 ppm): (a) as for fig. 1 (a); (b) as for fig. 1 (b); (c) as for fig. 1 (c); ( d ) as for fig. 1 (c), but after zeolite rehydration at 293 K.HZSM,-5 exhibits the same phase together with the dominant amorphous H,BO, phase [fig. 2(c)]. Pretreatment at higher temperatures (1073 K) assists in diffusing boron into a fourth phase [fig. 3(c)] characterized by a more complex profile with resonances at ca. 77, 35, - 2.9, -43 and - 83 ppm. This may be associated with a Pyrex glass-like str~cture.~ However, the resultant phase does not exhibit similar stability, since it can easily be hydrolysed back to the less-condensed phase =Si-0-B(OH),, for which the resultant spectrum [fig. 3(d)] is a combination of those for samples pretreated at 673 K [fig. 3(b)] and at 1073 K [fig. 3(c)]. This suggests that water mobilises boron into the less-ordered structure and eventually into the amorphous phase, particularly if hydrolysis takes place at the catalytic reaction temperatures. A similar phase transformation is not evident for HZSM,-5, which confirms the immobility of lattice boron created during synthesis.8 This must be kept in mind when discussing the data of the two modifications, i.e.impregnation and lattice incorporation during synthesis. The following scheme may help to illustrate the phase transformation detected by llB n.m.r. spectroscopy [fig. 3 (bb(d)]. As shown, boron occupies different phases and therefore reveals different ll€? n.m.r. spectra. However, the boron in these different phases occupies trigonal sites, rendering absorption of i.r. radiation at similar wavenumbers in each case plausible.8 The advantage of this transformation scheme, whether forward or backward, is easily explained in the light of the present spectral data and is ideally suitedM.B. Sayed 1755 for interpreting boron mobility over the impregnated samples, which has been suggested for demonstrating modified catalytic pr0perties.l Boron Mobility/Methanol Reactions Fig. 4-6 summarize the i.r. data for the zeolites pretreated at 673 K and for sorbed methanol at different reaction temperatures. It is clear that the Brmsted band absorbance at 36 10 cm-l decreases progressively with increased impregnation, i.e. on going from HZSM,-5 [fig. 4(a)] to HZSM,-5 [fig. 6(a)]. The absorbance of the band at 3720 cm-l, associated with terminal silanols, is also lowered. The broad band structure, indicative of strong H-bonding between boranol and adjacent 0-H groups,’ =Si-0-B(OH), occurs [fig.6(a)] at 3660 cm-l. The results of methanol reaction with the surface at 293 K are shown in fig. 4(b)-6(b) for HZSM,-5, HZSM,-5 and HZSM,-5, respectively. The data for HZSM,-5 reveal an interaction intermediate between HZSMl-5 and HZSM,-5. Methanol reactions over HZSM,-5 resulted in spectral data discussed previously.6 Unusual results are detected for the present modified system. While the band associated with trigonal boron B-0 stretch at 1385 cm-l diminishes only slightly [fig. 4(b)] for HZSM,-5, it disappears for the impregnated samples. Concurrently, surfaces dominated by H-bonded and chemi- sorbed methanol6 appear [fig. 4(b)-6(b)]. In addition, new bands appear at ca. 2870,1480 and 1360 cm-l to distinguish the modified samples [fig.5(b) and 6(b)] from HZSM,-5 [fig. 4(b)]. Assignment of the new bands is necessarily based on the following findings: while two of these bands (2870 and 1480 cm-l) appear in the range of C-H stretching modes and bending deformations, the third (1 360 cm-l) is peculiar and is very intense. Surprisingly, zeolite desorption at 293 K effects a complete disappearance of these absorptions, indicating association with weakly sorbed species. These findings imply the involvement of the volatile methylborate ester, particularly if the band at 1360 cm-l is assigned to the B-0 stretching ~ibration.~? lo If the above assignment is true, a linear dependence of the band absorbance of these absorptions on the modifier boron content must then be revealed.Indeed, this is the case (fig. 7), but these lines do not extrapolate to the origin. However, investigation of the reaction gas phase could solve the problem. Fig. 8 shows i.r. spectra for the zeolite samples as measured for the gas phase after sorption at 293 K. In addition to the well known methanol spectrum,6 two other absorptions at ca. 1490 and 1365 cm-l distinguish the gas phase of the modified samples [fig. 8(b)-(d)] from that of HZSM,-5 [fig. 8(a)], where they intensify with the modifier boron content, suggesting a contribution to1756 Boron Mobility over ZSM.5 0 wavenum ber/cm-' Fig. 4.1.r. spectra (4000-1200 cm-' of CH,OH/HZSM,-5: (a) zeolite pretreated at 673 K/2 h; (b) after methanol sorption at 293 K; (c) after methanol reaction at 473 K/2 h; ( d ) after methanol reaction at 573 K/0.5 h.wave num ber/ cm - ' Fig. 5. 1.r. spectra (4000-1200 cm-l) of CH,OH/HZSM,-5: as for fig. 4 with ( d ) being after reaction at 573 K/h. gaseous boron compounds. Spectral measurement in two different phases (homo- geneous methanol and heterogeneous methanol/zeolite) would not permit summation of the relevant absorbances. However, it is obvious that if the band absorbance (fig. 7) is corrected in the light of the data of fig. 8 this could lead to a linear dependence passing through the origin. If this argument is correct, one can suggest that part of the modifier,M. B. Sayed 1757 I 1 I 4000 3000 2000 1500 1: wavenumberlcm-' Fig. 6.1.r. spectra (4000-1200 cm-') of CH,OH/HZSM,-5: as for fig. 4, with ( d ) being absent for no reaction.0.8 d I 6 0.6 0.8 3 I 0.6 5 2 0 m Y b) 0.4 2 5 2 s 2, P -a 0.2 2 Fig. 7. Dependence of the i.r. absorbance at 1360 cm-l (a) and at 1480 cm-l (b) of surface boron on the zeolite boron content. which is presumably loose, reacts with methanol at 293 K to yield borate ester characterized by low vapour pressures, This appears to be consistent with the data of zeolite characterization, where amorphous boron is present in a large excess for the more highly impregnated zeolite. The results of methanol reactions at 473 K are summarized in fig. 4(c)-6(c). All theI758 Boron Mobility over ZSM.5 w avenum ber/ cm- ' Fig. 8. 1.r. spectra of the gas phase (4000-1000 cm-l) after methanol sorption at 293 K: (a) HZSMo-5,(b) HZSM1-5,(c) HZSM2-5,(d) HZSM3-5.samples seem to exhibit similar activity for methanol dehydration and the spectra reveal surfaces dominated by dimethylether.6 The latter is more intense for HZSM,-5 [fig. 4(c)] than for the other samples [fig. 5(c) and 6(c)], reflecting an increased constraint with boron for the modified samples. Successive CH,OH addition and reaction, particularly under dynamic conditions, could reduce this artificial constraint and boron could then be cleared away. The band at 1455 cm-l is attributed to dimethylether,6 since borate ester has no absorptions near this waven~mber.~~ lo Trapping of the CH, group in the form of borate ester explains origin of the supplementary alkylation observed in the previous study.l Borate ester can easily revert to methanol by hydrolysis under acidic (zeolite) conditions.The results of methanol reactions at 573 K are summarized in fig. 4(d) and 5(d). HZSM,-5 [fig. 6 (d)] is incapable of converting (CH,),O into hydrocarbons, probably owing to the absence of strong Bransted sites. As shown, HZSM,-5 yields mixed aliphatic (1460 cm-l) and aromatic (1 505 cm-l) hydrocarbons incorporated in a spectral pattern6 characteristic of such a conversion stage. In addition, the modified samples reveal absorptions at 1710 and 1545 cm-l associated with carboxy1ates.l'- l2 Silicalite impreg- nated with 0.6 % boron reveals similar carboxylate species, suggesting activity for the modifier boron in forming carboxylates from simple oxygenates. Finally, since it has been observed that a part of the boron has reacted with methanol and departed as the ester into the gas phase, it could be of interest to measure the percentage loss of boron on the surface effected during one catalytic cycle.This could be done by evaluating loss of band absorbance (surface) at 1385 cm-l. This reveals the loss of 6, 21 and 41% for HZSM,-5, HZSM,-5 and HZSM,-5, respectively. No loss is detected for HZSM,-5, confirming the immobility of boron incorporated in the zeolitic structure .M . B. Sayed 1759 Conclusion Exploring the origin of boron mobility, which was invoked to interpret some previous data1 and has been proven by the present data, becomes of necessity so that one can be aware of the defects involved in the technique of zeolite modification. Since a high loss of boron is shown to correspond to high population of amorphous H,BO, or B,O,, it is plausible that such species are the origin of most labile boron on the surface.Conversely, the structural boron in HZSM,-5 is firmly held and is immobile under the experimental conditions used. The mobility of the less-ordered phase formed by condensation, =Si-0-B(OH),, is governed by the catalytic process: if it involves water (e.g. toluene alkylation by methanol1 or methanol self-reaction), the skeletal boron can easily be hydrolysed back to amorphous H,BO, and can therefore be labile; if it does not involve water (e.g. toluene disproportionationl), mobility is retarded. This gives clear-cut evidence that water is a mobility-determining factor. References 1 M. B. Sayed and J. C. Vedrine, J. Catal., 1986, 101,43. 2 M. B. Sayed, A. Auroux and J. C. Vedrine, Appl. Catal., 1986, 23,49. 3 W. W. Kaeding, U.S. Patent, 1977,4029716. 4 Z . Gabelica, G. Debras and J. B. Nagy, Catalysis on the Energy Scene, ed. S . Kaliaguine and A. Mahay 5 2. Gabelica, J. B. Nagy, P. Bodart and G. Debras, Chem. Lett., 1984, 1059. 6 M. B. Sayed and R. P. Cooney, Aust. J. Chem., 1982, 135,2483. 7 M. L. Hair, Infrared Spectroscopy in Surface Chemistry (Edward Arnold, London, 1967). 8 M. B. Sayed, A. Auroux and J. C. Vedrine, J. Catal., in press. 9 R. R. Servoss and H. M. Clark, J. Chem. Phys., 1957, 26, 1157. (Elsevier, Amsterdam, 1984), p. 1 13. 10 A. Rogstad, B. N. Cyvin, S. J. Cyvin and J. Brunvoll, J, Mol. Struct., 1976, 35, 121. 11 J. D. Donaldson, J. F. Knifton and S. D. Ross, Spectrochim. Acta, 1964, 20, 847. 12 J. Datka, J. Catal., 1974, 32, 183. Paper 611545; Received 28th July, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301751

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I , 1987,83, 1761-1770 Catalytic Activity and Structure of Mo Oxide Highly Dispersed on ZrO, for Oxidation Reactions Takehiko Ono,* Hisashi Miyata and Yutaka Kubokawa Department of Applied Chemistry, University of Osaka Prefecture, Sakai, Osaka 59 I , Japan The structure of Mo-Zr oxides has been investigated by X-ray diffraction and Fourier-transform infrared and laser Raman spectra. At low Mo content, MOO, is highly dispersed on ZrO,, an amorphous phase being formed which is characterized as polymolybdates. The i.r. and Raman bands due to Mo=O in polymolybdate are shifted from 930 to 970cm-' with increasing Mo content on ZrO,, suggesting that with low Mo catalysts surface Mo species are more distorted. The rates of oxidation of ethanol and propene pass through maxima around 10 atom % Mo.These results provide evidence that the rates over polymolybdate are 1&20 times higher than those over crystalline MOO,. Previously we have reported that the promoter effect observed with Mo-Ti1 and V-Ti2 oxides in the oxidation of alcohol and propene arises from the formation of an amorphous phase over the surface of titania. Such a non-crystalline phase formed at low Mo or V content has been characterized by polymolybdate or polyvanadate species on titania, their distorted structures bringing about the weakening of the Mo=O or V=O bond, i.e. enhancement of the activity for the oxidation reactions. A similar conclusion has been drawn in the case of Mo-Si oxide catalyst^.^ Jeziorowskii and Knotzinger4 and Ng and Gulari5 have also reported that such polymolybdate species are formed on Al,O, or TiO, in a monolayer and that they play an important role in hydrodesulphuriz- ation.More recently, Chan et a1.6 have also studied the nature of supported W, Mo and V oxides on TiO, and Al,O, by laser Raman spectroscopy. As regards Mo-Zr oxide catalysts, some workers have studied the reduction of NO, to N, with H, and NH, g a ~ e s . ~ - ~ Only a few papers reported oxidation reactions. lo In this work the structure of Mo-Zr oxide catalysts has been investigated by X.r.d., F.t.i.r. and laser Raman techniques. The correlation between the surface structures of Mo oxides and the catalytic activities for propene oxidation and the oxidative dehydro- genation of ethanol have been examined.Experimental Mo-Zr oxide catalysts containing x = 1, 3, 5, 10 and 20 mol % MOO, [Mo(xkZr] were prepared by impregnation of the ZrO, support with a solution of ammonium heptamolybdate (Nakarai Co.) followed by calcination at 720 K. MOO, was prepared by heating ammonium heptamolybdate at 720 K. ZrO, was prepared as follows. The Zr hydroxides were precipitated from ZrOC1, (Kishida Co.) with an ammonia solution. The precipitates were washed, dried and heated at 720 K. ZrMo,O, was prepared by heating a finely ground equimolar mixture of MOO, and ZrO, at 920 K. The B.E.T. surface areas are listed in table 1. The apparatus consisted of a closed circulation system of ca. 290 cm3. The catalytic oxidation and reduction of propene and ethanol were carried out under the circulation of a mixture of propene (or ethanol), oxygen and helium.The reaction products in 17611762 Catalytic Activity and Structure of Mo-Zr Oxide Table 1. Structure and surface area of Mo-Zr oxides surface fraction of concentration surface amorphous of amorphous Mo catalyst area/m2 g-l oxide phase Mo oxide (%) oxide/pmol m-, - ZrO, 50-70 ZrO, (monoclinic) - Mo( 1 t Z r 47 ZrO, 100 1.7 Mo( 3 )-Zr 52 ZrO, 100 4.6 Mo( 5)-Zr 70 ZrO, 100 5.8 Mo(20)-Zr 59 ZrO, + MOO, 35-65 13 Mo( lO)-Zr 70 21-0, + MOO, 70-8 5 9 propene oxidation were CO, CO,, acrylaldehyde and traces of acetone and ethanal. Products of oxidative dehydrogenation of ethanol were mainly ethanal and trace of ethylacetate. These were analysed by gas chromatography. In order to obtain initial rates, conversions were kept below ca.5% in both reactions. The X-ray diffraction patterns of catalysts were obtained using a Rigaku Denki D-3F or a RAD-rA X-ray diffractometer using Cu Kor radiation. Using a Rad-rA meter the goniometer stepper motor system and the signal were interfaced with a versatile data acquisition system. As has been reported a small amount of MOO, (0.5-1 .O mol % in supports) is detectable by using a step scanning method. The i.r. spectra of Mo-Zr oxides were recorded using a Hitachi EPI-G2 infrared spectrometer or a Shimadzu F.t.i.r. 4000 spectrometer; the sample was prepared by mixing KBr and Mo-Zr oxides (ca. 1 wt%). The F.t.i.r. spectra were taken in a thoroughly dry air-purged atmosphere. After 100 accumulations had been obtained the data were transferred on an IEEE-488 bus line to a master computer (PC 9801 M2, NEC).Details of the methods were described previously.ll The Raman spectra were recorded using a JASCO NR-1000 laser Raman spectrometer. An Ar ion laser was tuned to the 514.5 nm line for excitation. The laser power was set at 200 mW. As far as the Mo-Zr oxide catalysts were concerned, even at 200 mW power the spectra of samples did not change during the measurements. Photoluminescence and excitation spectra were measured by a Shimadzu RF-50 1 spectrofluorophotometer with colour filters to eliminate scattered light at 77 K. Results and Discussion Catalytic Oxidation of Ethanol and Propene on MeZr Oxides Fig. 1 shows the rate of oxidative dehydrogenation of ethanol and the selectivity towards ethanal over Mo-Zr oxides.All the Mo-Zr oxides showed high selectivity towards ethanal formation (> 95%). A small amount of ethylacetate and trace of CO, and ethylether were also formed. The rate of ethanal formation passed through a maximum around Mo(l0)-Zr. The rates of ethanal formation over MOO, and ZrO, were ca. 0.8 and 0.04 pmol md2 min-l, respectively. Thus, the maximum rate was ca. 4 and 100 times larger than those over MOO, and ZrO,, respectively. The rates in the absence of gaseous oxygen were 3 to 4 times smaller than those in the presence of oxygen. However, the rate passed through a small maximum at Mo( 10 j Z r and the selectivities towards ethanal in the absence or presence of gaseous oxygen are similar. Similar experiments were carried out using propene (fig.2). The initial rate of pro- pene oxidation also passes through a maximum around 10 atom% Mo. The rates of propene conversion over MOO, depended on its preparation and ranged from 1 to 0.2 pmol m-2 min-1 in this case, values which are several times smaller than that over theT. Ono, H . Miyata and Y. Kubokawa h E x C ._ 2 $? 0 z C 2 50- 5 u, 0 1763 100- w - (b ) 0 I l l l l l l l l ~ Mo( 10)-Zr catalyst. The rate over ZrO, was 0.002 pmol rn-, min-l, being extremely smaller than those over Mo-Zr catalysts. The selectivity towards acrylaldehyde increased with increasing Mo contents, while that towards CO and CO, decreased. Traces of ethanal and acetone were also formed, but no formation of propanal was observed. The selectivities at conversions below 5% on each Mo-Zr catalyst were constant under various propene and oxygen pressures (0.8-8 kPa).The selectivities towards acrylalde- hyde for the catalysts Mo( 10)-Zr to Mo(20)-Zr were nearly the same as that for crystalline MOO,. X-Ray Diffraction Table 1 also shows the crystalline phases of Mo-Zr oxides determined by X.r.d. The ZrO, showid the monoclinic phase. No diffraction lines were observed corresponding either to a new phase or to MOO, in the catalysts Mo(1)-Zr to Mo(5)-Zr, while the diffraction lines due to MOO, were observed with Mo(l0)-Zr and Mo(20)-Zr. The crystallinity of the MOO, phase in the catalysts was determined from the average intensities with Mo-Zr oxides and the corresponding physical mixtures of pure MOO, and ZrO,. The results suggest that the crystallinity of the MOO, phase is decreased and that an amorphous phase is formed in the Mo-Zr oxides.From the X.r.d. data it follows that with Mo(1)-Zr and Mo(5)-Zr ca. 100% of the molybdenum oxide is present as an amorphous phase, while with Mo(l0)-Zr and Mo(20)-Zr 80% and 50%, respectively, is amorphous (table 1). Previously it was suggested that the amorphous phase observed with Mo-Ti1 and Mo-Si3 oxides arises from the formation of Mo oxide highly dispersed1'764 Catalytic Activity and Structure of Mo-Zr Oxide Mo (atom %) Fig. 2. Rates (a) and selectivities (b) for propene oxidation over Mo-Zr oxide catalysts at 673 K, p(C,H,) = 3.3 kPa and p ( 0 , ) = 3.3 kPa. (b) A, CO; 0, CO,; 0, CH,=CHCHO. on TiO, and SO,. As described below, no MOO, is found by F.t.i.r.and laser Raman spectra with catalysts Mo( 1)-Zr to Mo(S)-Zr either. With Mo(SO)-Zr, MOO, crystals in the catalysts were somewhat oriented in a particular direction. Thus it was difficult to determine the exact percentage of crystalline MOO,. I.R. and Raman Spectra The i.r. spectra of Mo-Zr oxides are shown in fig. 3. No i.r. bands due to ZrO, in the region 1 100-800 cm-l were observed with any of the catalysts. Mo(S)-Zr and Mo( 10)-Zr show very weak bands around 960 and 910 cm-l. The i.r. spectrum of Mo(20)-Zr shows sharp bands at 995 and 885 cm-l, attributable to crystalline MOO,. The spectrum of ZrMo,O,, which was formed by heating ZrO, with MOO, above 820 K,1° shows i.r. bands at 980, 920 and 800 cm-l [fig. 3(e)], suggesting that there are no ZrMo,O, species in Mo(S)-Zr and Mo( 10)-Zr.In order to obtain quantitative information from the severely overlapping and very weak bands, 100 accumulations were obtained using F.t.i.r. for lower Mo content catalysts. Fig. 4 shows the spectra after subtracting the spectrum of ZrO, itself. With low Mo catalysts, such as Mo( 1)-Zr and Mo(3)-Zr, the bands were observed at 935 and 951 cm-l. The spectrum of Mo(l0)-Zr seems to be more complex. Therefore, bandshape analysisll was applied to this spectrum in the region 1050-850 cm-l. Gaussian bandshape was assumed. For the Mo(l0)-Zr catalyst the bands in this range are composed of five peaks (fig. 5). The broad band at 960 cm-l was separated into two peaks at 970 and 950 cm-l. It was confirmed that the subtraction spectrum of Mo( 10)-ZR- Mo(S)-Zr clearly shows the band at ca.970 cm-l. The band around 935 cm-l of Mo( 1)-Zr is shiftedT. Ono, H . Miyata and Y. Kubokawa 1765 I I 1100 1000 900 800 w avenum ber/cm-' Fig. 3.1.r. spectra of MeZr oxide catalysts. (a) ZrO,, (6) Mo(5)-Zr, ( c ) Mo( 10)-Zr, ( d ) Mo(lLO)-Zr, (e) ZrMo,O, and (f) MOO,. 2 wt % sample in KBr was used. to 950-970 cm-l with increase in Mo content. The band at 995 cm-1 is attributable to the stretching vibration of Mo=O of crystalline MOO,. Crystalline phases are considerably more Raman active than surface phases.l29 l3 Similar results were obtained with Mo(l0)-Zr [fig. 6(a)], where only 15-30% of the Mo oxide is present as a crystalline phase as shown by X.r.d. (table 1). The bands at 997 and 822 cm-l, attributable to crystalline MOO,, are very strong.Note that with catalysts with lower Mo contents than Mo(S)-Zr no bands due to the crystalline phase were observed [fig. 6 ( b b ( d ) ] , in agreement with X.r.d. results. With Mo( 1 j Z r to Mo(5bZr the band positions are shifted from 925 to 950-960 cm-l with increasing Mo content as are the F.t.i.r. bands. With Mo( 10)-Zr the band at 960 cm-l is small compared with other samples. This suggests that the Raman intensity is strongly affected by the scattering efficiency of the sample. As shown in table 2, the Raman bands of the catalyst were shifted to higher wavenumber with the calcination temperature from 720 to 1220 K. Structure of Mo-Zr Oxides With Mo-Ti oxide,l? Mo-A1 oxide4 and Mo-Si oxide,l49 l5 the Raman and i.r.bands in the range 980-910 cm-l have been attributed to the stretching vibration of Mo=O of polymolybdates. In addition, Griffith and Lesniak16 have reported that polyanions in a solution such as Mo,Oi; and Mo,O& exhibit Raman bands at 940-950 cm-l and that MOO:- has a band at 895 cm-l. Thus, the bands in the region 900-1000 cm-l are attributable to the stretching vibration of the terminal Mo=O bond of polymolybdate on ZrO,. In this work, the i.r. and Raman bands of the Mo=O stretching vibration were1766 Catalytic Activity and Structure of Mo-Zr Oxide 1100 lo00 900 800 wavenumber/cm-' Fig. 4. F.t.i.r. spectra of Mo-Zr oxide catalysts. (a) Mo(l0 j Z r , (b) Mo(5)-Zr, ( c ) Mo(3)-Zr and ( d ) Mo(1)-Zr. 2 mg of sample in 150 mg KBr was used.1050 1000 950 900 850 wavenumber/cm-' Fig. 5. Peak separation of i.r. spectra of Mo(l0)-Zr catalyst in the range 850-1050 cm-l. (a) Original spectrum, (b) separating peaks (995, 970, 950, 910 and 875 cm-l). shifted from 930 to 960-970 cm-l on going from Mo( 1)-Zr to Mo( 10)-Zr. Similar band shifts have been obtained in W-Al, Ti oxide,s Mo-A1 oxide6 and Mo-Ti ~ x i d e . ~ Iannibello et all7 have reported that the shifts in the Raman bands with W content on A1,0, are associated with a distortion of the surface W oxide species owing to lateral interaction and/or the heterogeneity of the adsorption site on A1,0,. On the other hand,T. Ono, H . Miyata and Y . Kubokawa 1767 n E: Y .C( 2 W x e Y ...I Y ...I s 5 d I I I 1100 1000 900 800 700 wavenum ber/ cm-' Fig.6. Laser Raman spectra of Mo-Zr oxide catalysts. (a) Mo(l0)-Zr, (b) Mo(S)-Zr, (c) Mo(3)-Zr, ( d ) Mo(1)-Zr and (e) ZrO,. Chan et aL6 have reported that such a shift is attributable to the coordination of water molecules on the surface oxide species, i.e. at lower coverages the surface oxide species are coordinated more water molecules than at higher coverages. Considering that in the present study the Raman data were obtained in the ambient air after the calcination, the shift to higher wavenumber with temperature (table 2) seems to arise from the chain growth of polymolybdates on ZrO,. It is interesting that the surface molybdate of Mo(3)-Zr is stable on ZrO, even after heating above 1220 K, although crystalline MOO, easily reacts with ZrO, to form ZrMo,O, around 820-920 K.Similar results with W-A1 oxide catalysts have been reported by Chan et aZ.ls It has been reported that tetrahedral molybdate species are formed on TiO, at low Mo content, while octahedral species are formed at high Mo ~ o n t e n t . ~ With Mo-A1 ~ x i d e , ~ however, the Raman bands in the region 930-960 cm-l have been reported to be attributable to octahedral molybdate species, both at low and high Mo contents. The isolated tetrahedral Mo species on SiO, exhibit an emission maximum at ca. 440 nm together with an excitation maximum at 26&280 nm by charge-transfer process.1g* 2o Therefore, similar experiments were carried out with the Mo( 1 )-Zr catalyst.1768 Catalytic Activity and Structure of Mo-Zr Oxide Table 2. Change in the Raman bands of polymolybdate in Mo(3)-Zr calcination temp./K wavenumber/cm-' 720 935 875-850 760 920 950 870 760 1020 960 875 - 1120 970 870 760 1220 960 870 760 Raman spectra were recorded in ambient air after calcination. Table 3. Rate constants in propene oxidation ratea of C,H, conversion per amorphous catalyst Mo oxide klb kZb k,/k2 Mo( 1)-Zr 0.76 4.8 0.7 6 Mo(3)-Zr 0.55 - Mo( 10kZr 0.50 1.4 10 0.1 MOO, 0.08-0.02 0.5 4 0.1 - - a The rates (fig. 2) are divided by the concentration of amorphous Mo oxide (table 1). Calculated from the equation denotes the rate (pmol rnp2 min-l) at various C,H6 and 0, pressures (1-10 kPa). See ref. (1). = k l k2p(C3H6)p(02)0.5/[k1p(C3H6) +k2p(02)0*5i, where With Mo(1)-Zr a weak emission maximum at 490 nm and an excitation maximum at 290 nm were observed, which were attributable to tetrahedrally coordinated Mo ions on 21-0,.The concentration of tetrahedral species was estimated by the comparison of emission intensity with Mo-Si oxide catalysts which contain tetrahedral Mo specie^.^ Such a comparison may be allowed under the assumption of equal probability in radiationless deactivation from charge-transfer excited states between the Mo-Si and Mo-Zr oxides. The results show that only 1 % of Mo ions in Mo(1)-Zr is present as tetrahedral, where ca. 2pmol rn-, of Mo ions are contained on ZrO,. Thus, it is concluded that the 99% of Mo species in Mo(1)-Zr are in octahedral. With Mo(3)-Zr and Mo(5kZr no or little tetrahedral species are present. Catalytic Activity and Structure of Mo-Zr Oxides As described above, with Mo(1)-Zr to Mo(l0)-Zr the majority of the Mo oxide is present as amorphous polymolybdates, while with Mo-Zr oxides containing more than 20 atom % of Mo a considerable fraction consists of crystalline MOO,.In the region of low Mo content the activity of Mo-Zr oxides for oxidative dehydrogenation increases with increasing Mo content, corresponding to the concentration of amorphous species, i.e. polymolybdates (table 1). The dehydrogenation rates per amorphous Mo (pmol mP2 min-l/pmol m-,) are as follows: 0.1 for Mo(1)-Zr, 0.5 for Mo(3)-Zr, 1 for Mo(S)-Zr and Mo(l0)-Zr, and 0.7T. Ono, H . Miyata and Y. Kubokawa 1769 for Mo(20)-Zr. These rates on the catalysts except Mo(1)-Zr are 10-20 times larger than 0.06 on crystalline MOO,, calculated using the concentration of surface Mo as ca.13 pmol rn-, with the (100) and (001) planes. Such an enhancement in the oxidative dehydrogenation over polymolybdate seems to be due to the distorted structure, i.e. the weakening of the Mo=O bond from 995 cm-l to 970-950 cm-l. The very small rate on Mo(1 j Z r seems to be due to the hydration of more water molecules on polymolybdate because of the low reaction temperature at 453 K. In the region of high Mo content (above 20%) the increase in the Mo content will cause an increase in the fraction of crystalline MOO,, which is less active than the surface polymolybdate. Thus, it is explicable that the oxidation activity passes through a maximum around 10 atom % of Mo. The rates of propene conversion per amorphous Mo oxide are shown in table 3.The rates range from 0.7 to 0.5 bmol rn-, min-l/pmol m-,) for Mo(1)-Zr to Mo(l0)-Zr and are little different, while a large difference of the selectivity towards acrylaldehyde is found on the catalysts, i.e. below 1 % for Mo(1)-Zr and 20% for Mo(l0)-Zr (fig. 2). Although we cannot tell its precise nature, such a difference seems to arise from the difference in the distorted structure of Mo oxides. The rate constants for the reduction step (k,) and the oxidation step (k,) were determined (table 3) by applying a redox mechanism.l For Mo(l0)-Zr, kJk, is ca. 0.1, which is similar to that for crystalline MOO,, while the k, for Mo(1)-Zr is remarkably larger than that for crystalline MOO,. A low value of k,/k, suggests that the reduction step is rate-determining. The promoter action in the Mo( l)-Zr oxide catalyst is attributed mainly to the increase in the rate of reduction step, while in the case of Mo(l0)-Zr this is attributed to the increase in the rates of both reduction and oxidation step.As described above, the remarkable difference of kinetic features between Mo( 1)-Zr and Mo( 10)-Zr catalysts has been confirmed in addition to that of the selectivity towards acrylaldehyde. This suggests that the polymolybdate in Mo( 10)-Zr has nearly the same character in a redox mechanism as that of crystalline MOO,. Structural studies have revealed that the bands due to Mo=O of Mo(1 j Z r are shifted extremely to lower wavenumbers owing to the low coverage (ca. 0.1) of Mo and the hydration of more water molecules.Since the reaction temperature is 673 K in propene oxidation, most water molecules should be removed from the catalyst surface. Therefore, the change of selectivity and activity in propene oxidation on Mo(1)-Zr seems to originate from the change of character of polymolybdate itself. The polymolybdate of Mo( 1)-Zr seems to have a shorter chain and a more distorted structure since the interaction between the polymolybdate and the surface of ZrO, is stronger at low Mo content than at high Mo content. This explanation is supported by the fact that the total oxidation activity from propene to CO and CO, increases with the decrease in the Mo content on ZrO,. In ethanol oxidation, however, further oxidation of ethanal to CO and CO, did not proceed on the Mo(l-3)-Zr catalysts.This seems to be due to low reaction temperature (453 K) compared with that (673 K) in propene oxidation. The authors thank Dr Masakazu Anpo for measurements of photoluminescence spectra and Dr Takashi Ohno (Kobe University) for F.t.i.r. measurements, The authors also thank Messrs Kazunori Yokochou and Hiroyuki Kamisuki for carrying out some experiments. References 1 T. Ono, Y. Nakagawa, H. Miyata and Y. Kubokawa, Bull. Chem. SOC. Jpn, 1984, 57, 1205. 2 Y. Nakagawa, T. Ono, H. Miyata and Y. Kubokawa, J. Chem. SOC., Faraday Trans. I , 1983,79,2929. 3 T. Ono, M. Anpo and Y. Kubokawa, J . Phys. Chem., 1986,90,4780. 4 H. Jeziorowskii and H. Knozinger, J. Phys. Chem., 1979, 83, 1166. 5 K. Y. S . Ng and E. Gulari, J . Catal., 1985, 92, 340.1770 Catalytic Activity and Structure of Mo-Zr Oxide 6 S.S. Chan, I. E. Wachs, L. L. Murrel, L. Wang and W. K. Hall, J. Phys. Chem., 1984,88, 5831. 7 S. Okazaki, M. Kumasaka, T. Yoshida and K. Kosaka, Ind. Eng. Chem. Prod. Res. Dev., 1981,20,301. 8 T. Iizuka, M. Itou, H. Hattori and K. Tanabe, J. Chem. Soc., Faraday Trans. I , 1982, 78, 501. 9 H. Hattori, K. Tanabe, K. Tanaka and S. Okazaki, Proc. 3rd Int. Congr. Chemistry and Uses of Molybdenum (Climax Molybdenum Co., Michigan, 1979), p. 188. 10 T. Frausen, P. C. Van Berge and P. Mars, Preparation of Catalysts I (Elsevier, Amsterdam, 1976), p. 405. 11 H. Miyata, K. Fujii, S. Inui and Y. Kubokawa, Appl. Spectrosc., 1986, 40, 1177; H. Miyata, K. Fujii, T. Ono, Y. Kubokawa, T. Ohno and F. Hatayama, J. Chem. SOC., Faraday Trans. 1, 1987,83, 675. 12 F. Roozeboom, J. Medema and P. J. Gellings, Z. Phys. Chem. N.F., 1978, 111, 215. 13 F. P. J. Kerkhof, J. A. Moulijn, R. Thomas and J. C. Oudejans, Preparation of Catalysts II (Elsevier, 14 C. P. Cheng and G. L. Schrader, J. Catal., 1979, 60, 276; H. Jeziorowskii, H. Knozinger, P. Grange 15 R. S. Seydomonir, S. Abdo and R. F. Howe, J. Phys. Chem., 1982,86, 1233. 16 W. P. Griffith and P. J. B. Lesniak, J. Chem. SOC. A, 1969, 1066. 17 A. Iannibello, P. L. Villa and S. Marengo, Gazz. Chim. Ital., 1979, 109, 521. 18 S. S. Chan, I. E. Wachs, L. L. Murrell and N. C. Dispenziere Jr, J. Cafal., 1985, 92, 1. 19 M. Anpo, I. Tanahashi and Y. Kubokawa, J. Chem. SOC., Faraday Trans. 1,1982,78,2121; M . Anpo, 20 Y. Iwasawa and S. Ogasawara, Bull. Chem. Soc. Jpn, 1980, 53, 3709; V. B. Kazanskii, Kinet. Catal., Amsterdam, 1979), p. 77. and P. Gajardo, J. Phys. Chem., 1980,84, 1825. I. Tanahashi and Y. Kubokawa, J. Phys. Chem., 1982,86, 1. 1983, 24, 1338. Paper 6/ 1546; Received 28th July, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301761

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987,83, 1771-1778 Comments on the Mechanism of MTG/HZSM-5 Conversion Moein B. Sayed Chemistry Department, Faculty of Science, Al-Azhar University, Cairo, Egypt A comparative F.t.i.r./g.c.-m.s. study of MTG conversion on a series of progressively dealuminated HZSM-5 and H-mordenite surfaces reveals the role of zeolite dealumination in modifying the zeolite pore void (i.e. catalytic volume) and therefore the shape-selectivity. This is illustrated by a different distribution of the primary alkenes formed under similar experimental conditions. G.c.-m.s. analysis of a reaction at the early stages of dimethyl- ether conversion reveals that ethene, which forms in trace quantities on HZSM-5, is dominant on H-mordenite. The propene/butenes ratio is high (ca.1.7) for H-mordenite, but is < 1.0 and decreases with increased dealu- mination for HZSM-5. Isobutene dominates for increasingly dealuminated HZSM-5. Methanol reformation proceeds concurrently with alkene forma- tion, suggesting a route involving either propagation and/or decomposition to alkenes. The mechanistic implications of the interdependence of zeolite dealumination, modified pore void and identity of alkenes are analysed. The methanol-to-gasoline (MTG) catalytic process has received much academic and industrial interest. Its mechanism, specifically the identity of the primary alkene, has been the subject of critical argument. The linear dependence of hydrocarbon formation activity on zeolite Brarnsted acidity' favours the oxonium proposal,2 the most probable rnechani~m.~-~ This has been confirmed recently.67 Experience indicates that the method of ZSM-5 preparation and pretreatment is most effective in determining catalytic activity and selectivity.*-ll Acid ion-exchange of non-calcined ZSM-5 induces extensive loss of the zeolite al~minium.~ Conversely, precalcined zeolite has fair stability upon exchange to the acid form.lo' l1 It appears that zeolite calcination prior to exchange assists in forming a surface resistant to dealumination.These differently pretreated surfaces have shown different catalytic activity and selectivity, where non-dealuminated HZSM-5 yields reaction products rich in aromatics and minor aliphatics, dominated by propene,l0 whereas dealuminated HZSM-5 yields products rich in aliphati~s,~~ l2 dominated by butenes.The aim of this work is to comment on these different characteristics. A combined Fourier-transform infrared (F. t.i.r.)/gas chromatography mass spectrometry (g.c.-m.s.) approach has been used to probe the origin of the various surface properties induced by zeolite dealumination. While i .r. spectroscopy has been employed to monitor the reaction throughout the course, g.c.-m.s. has been employed for analysing qualitatively and quantitatively the reaction content near the boundary of dimethylether to alkenes conversion. Experimental ZSM-5 preparation, characterization and pretreatment have been described Six progressively dealuminated HZSM-5 samples and dealuminated mordenite (Zeolon 100, Norton Co.) were used. The experimental methods and greaseless i.r.cell (static reactor) are described elsewhere.6 Methanol conversion was studied at low reaction level to avoid possible recyclization arising usually from static reactors. After 17711772 Mechanism of MTGIHZSM-5 Conversion 4000 2000 1000 wavenum ber/cm Fig. 1. 1.r. spectra (4000-1000 cm-l) of CH,OH/HZSM-5 (surface): (a) zeolite pretreated at 0.1 mPa/673 K for dealuminated HZSM-5, (b) sorbed methanol at 293 K, (c) after reaction at 473 K (dominated by dimethylether), ( d ) after reaction at 573 K/2 h (dominated by hydrocarbons). the primary methanol to dimethylether conversion had been completed at 473 K, the temperature was raised to 573 K and reaction was permitted to proceed just to the stage of alkene formation, as indicated by i.r.spectroscopy. 1.r. spectra of both the reaction gas phase and the surface were measured at increased temperatures and time intervals using a Nicolet MX-IE F. t.i.r. spectrometer. Intermedi- ates and products formed at or near the boundary of alkene formation were analysed both qualitatively and quantitatively using a Kratos MS-30 mass spectrometer. Helium was used as a carrier gas with a flow-rate of ca. 30 cm3 min-l. Separation of the cell contents was satisfactorily done using a programmed temperature rise (5 K min-l) from 363-473 K using a g.c. column (ca. 6 mm o.d., 1.5 m long) which was packed with 50-80 mesh Porapak Q. The m.s. data were processed using a Data General Nova 4 system with a microprocessor interface, Kratos DS-55 software and EPA NIH library search control.Quantitative analysis was achieved following single-ion separation methods, where calculations were based on the base peaks of the most detected species. E.g. the ions with the masses 41 and 43 represent the base peaks of most olefins and paraffins, respectively, and therefore their associated integrated intensities can be used in determining the proportions of relevant species.M . B. Sayed 1773 4000 2000 1000 wavenumber/cm-' Fig. 2. 1.r. spectra (4000-1000 cm-l) of CH,OH/HZSM-5 (gas phase): (a) methanol sorption at 293 K, (b) after reaction at 473 K (dominated by dimethylether), (c) after reaction at 573 K/1 h, ( d ) after reaction at 573 K/2 h (dominated by hydrocarbons). Results Infrared Spectroscopy The i.r. data of the MTG catalysis have been assigned in a previous study.6 However, a more illustrated and elaborated discussion necessitates reproduction of some important data.Fig. 1 (a) shows the spectral features of pretreated (0.1 mPa/673 K)7 dealuminated HZSM-5. Surface silanol is most sensitive to CH,OH sorption at 293 K as it reacts via strong H-bonding [3450 cm-l, fig. 1 (b)]. On the other hand, methanol sorption under such conditions effects formation of a surface methoxy (2942 cm-l) species characterized by a 30 cm-l downward shift from the dominant C-H stretch of gaseous methanol at 2972 cm-l. The persistence of the zeolite Brarnsted band at 3610 cm-l with increased absorbance indicates non-involvement of the associated site in surface methoxy forma- tion, which presumably forms on the zeolite Lewis site.13 Fig.2(a) shows the gas-phase spectrum of methanol, dominated by 0-H stretch (3675 cm-l), C-H stretch (both asymmetric at 2972 cm-l and symmetric at 2842 cm-l), C-H bend (1 458 cm-l), 0-H bend (1350 cm-l) and C-0 stretch (1033 cm-l).1774 Mechanism of MTGIHZSM-5 Conversion Table 1. Results of g.c. -m.s. analysis in terms of 100% hydrocarbons ordinary dealuminated dealuminated catalyst HZSM-5 (75%) HZSM-5 (70%) mordenite alkenes ethene propene butenes pentenes hexenes heptenes alkanes ethane propane butanes pent anes hexanes heptanes c,=/c, 57.06 5.44 20.19 21.53 7.25 2.65 42.94 0.37 2.37 19.40 15.08 5.72 0.94 48.00 4.55 10.74 23.02 5.61 3.93 0.15 52.00 2.88 19.41 15.61 11.22 2.88 0.46 - 54.44 18.55 17.20 10.24 4.13 3.61 0.71 45.56 8.89 21.20 9.79 4.58 1.10 1.70 - Reaction at temperatures below 573 K effects dehydration of methanol into dimethyl- ether, which is indicated [fig.l(c)] by a sharp absorption on the surface and is distinguished [fig. 2(b)] by its C-0 stretch at 1102 cm-l and CH, rock at 1178 cm-l in the gas phase. Reaction at 573 K effects conversion of the resultant ether into alkenes and higher hydrocarbons, indicated by aliphatic C-H deformations at 1462 and 1383 cm-l [fig. 1 (d)] on the surface, together with aromatic ring-breathing absorptions at ca. 1510 cm-'. Analysis of the gas-phase content after a relatively short time [fig. 2(c)] reveals the most useful information regarding the mechanism of dimethylether conversion. Alkenes form directly from the relevant ether, indicated by loss of the dimethylether band absorbance, particularly at 1102 and 1178 cm-l [fig.2(c)], together with the appearance of new absorptions characteristic of asymmetric alkenes at 1650 cm-l, assigned to the asymmetric C=C stretch.6 Concurrently, methanol absorp- tions (particularly at 3675 and 1033 cm-') which vanished previously reappear as an indication of methanol reformation. This cannot be attributed to an equilibrium governing the proportion of methanol and dimethylether since a reaction starting with dimethylether as a reactant (in the absence of methanol) also yields methanol over this catalyst. Methanol reformation has not been reported to occur in the course of the MTG catalysis; however, it was clearly shown in the 13C n.m.r. data [fig.4 of ref. (9)]. Methanol reformation at this critical stage is significant and is very useful in revealing the mechanism. Fig. 2(d) shows the results of reaction at longer time and the spectral pattern is characteristic of isobutene and/or isobutanen6 However, the presence of the asymmetric C=C stretch (1650 cm - I ) with considerable band absorbance favours assignment to isobutene. This could be confirmed if the spectral measurement was extended to a lower region, but this was impracticable using the CaF, windows designed for this system. Gas Chromatography-Mass Spectrometry G.c.-m.s. analysis of the cell contents was carried out when dimethylether, as revealed by i.r. spectroscopy, had fallen to a low level of concentration. The g.c. separation of the gas-phase components permitted hydrocarbons to be analysed in the absence ofM . B.Sayed 1.0 0.8 .- Y Lt c.7 u" --.- e, 0.6 24 1775 - . - Si02 /.41203 1 I I I 1 20 40 60 I I I 0 o\ 20 40 60 80 % A1 loss Fig. 3. Variation of the alkene propene/butene ratio with extent of zeolite dealumination. product H,O, CO,, CH,OCH, and traces of CH,OH. The data (table 1) therefore relate to percentage hydrocarbons. The results (computer-processed peak matching) of the g.c.-m.s. analysis of the gas contents [fig. 2(c), i.r. spectroscopy] reveal that alkenes higher than propene are branched and dominated by isobutene, and that alkanes higher than propane are also branched and dominated by isobutane, thus confirming the i.r. assignment. The presence of alicyclics and aromatics is negligible and their appearance is limited to the surface [fig.1 ( d ) , i.r. spectroscopy]. Alkenes higher than propene are but-2-ene for non- and least-dealuminated HZSM-5 and dealuminated H-mordenite (isobutene becomes in- creasingly dominant for more dealuminated HZSM-5), 2-methylbut-2-ene and branched hexene and heptene. Alkanes higher than propane are isobutane, 2-methylbutane 3-me thylpentane and 3-methylhexane. A quantitative comparison of reaction products formed over the different samples reveals that the hydrocarbons are almost all aliphatics, comprising a roughly equal mix of alkanes and alkenes. While ethene predominates (table 1) over mordenite, it forms in trace quantities on ZSM-5. Propene and butenes are dominant among other alkenes. The propene/butenes ratio is relatively high over mordenite (ca.1.7), whereas it is less than unity and decreases with increased dealumination for HZSM-5 (see fig. 3). The proportions of other alkenes are regarded as minor; they decrease further with increased molecular weight. Alkanes, on the other hand, increase in proportion to increasing molecular weight, reaching a maximum proportion for isobutane, and then decrease at higher molecular weights. The quantitative data must be regarded as relative rather than absolute, since oxygenates are not taken into account. However, this should not affect the correlation study. Discussion Two major spectral features form the bases of this discussion: the interdependence of the primary alkene identity on the zeolite effective pore volume (fig.3, g.c.-m.s.) and methanol reformation proceeding concurrently with alkene formation [fig. 2(c), i.r., and1776 Mechanism of MTGIHZSM-5 Conversion Z- non- steric + HZ + Z - + - CH3OH2 + CH3OCH3-CH3OHCH3-CH30HCH2 PCH~OHCH~CH~ + H2O + I 4 CH3OH I I-I v + CH30H2 t 4. CH3OH2 Z- non- steric _ . . + + - CH,OH, + fig. 4 of ref. (9)]. Discussion of the data of dealuminated H-mordenite must be correlated with care, taking into account the intrinsic structural differences from HZSM-5. Scheme 1 shows a rather elaborated form for the conversion stage of dimethylether into primary alkenes. As suggested,6 this proceeds via a series of ether propagation steps to form higher ethers (viz. methylethyl, methylisopropyl and methyl-t-butyl), which is followed by decomposition into the alkene derivatives (viz.ethene, propene and isobutene, respectively). Higher ethers are not easy to detect, perhaps because of their simultaneous appearance with dominant dimethylether. However, methanol reformation at this specific stage, together with varied proportions of primary alkenes under similarM. B. Sayed 1777 experimental conditions, makes their involvement (scheme 1) sensible. The role of higher ethers in the MTG process is evident from a previous study.14 In this mechanism, Brarnsted sites are the catalytic function that catalyses the present stage via the oxonium mechanism. It appears very probable that primary alkenes form directly from higher ethers (scheme l ) , since methanol is a concomitant product. Because their formation involves bulky propagated ethers, the distribution of these alkenes should be sensitive to the catalytic volume permitted.H-mordenite, having no wide intersections,15 would favour dimethyl- ether decomposition once propagated to methylethylether, which is plausible with the detection of ethene as the dominant alkene. Conversely, HZSM-5, having wide intersections,16 would favour ether propagation over decomposition, particularly if the zeolite is dealuminated, which is also plausible with the increased dominance of isobutene (fig. 3) over increasingly dealuminated HZSM-5. The effect of zeolite dealumination in increasing the pore volume is shown by data for hydrocarbon sorption, where maximum sorption capacity increases by ca. 8.2 and 8.8% for n-hexane and 3-methylpentane, respectively, over 28 % dealuminated HZSM-5.Increased sorption capacity has also been reported for other dealuminated ~ e o l i t e s . l ~ - ~ ~ The formation of isobutene, rather than propene or ethene, over dealuminated ZSM-5 surfaces is favoured by the stability of the tertiary carbonium and by the larger catalytic volume. Alkene distribution under these static conditions agrees with data obtained from a dynamic r e a c t ~ r ~ ? ~ ~ for dealuminated HZSM-5, which excludes the dependence of varied alkene proportions on the type of reactor, and the dependence on zeolite dealumination should be emphasised. The mechanism in the present form (scheme 1) demonstrates the effect of zeolite dealumination on the modified catalytic volume (and therefore selectivity) and also ex- plains the origin of reformed methanol.In principle, the mechanism agrees with that of Van Hooff and coworkers,2 with the additional advantage that the C-H bond (for induc- tive reasons) is more polar for asymmetric RQH (scheme 1) than for symmetric R,O+ (Van Hoofl), facilitating attack of the transient Z- (scheme 1) for trapping a proton from the former than the latter oxoniums. Also, RdH -CH, reaction with CHGH, is energeti- cally more probable than Steven’s rearrangement invoked by Van Hooff and coworkers.2 Finally, the back reaction of the zwitterion (e.g. CH,OHCH,) is retarded by the presence of the reactive oxonium CH,OH, (see scheme 1). + t - + References 1 M. Guisent, F. M. Cormerais, Y. S. Chen, G. Perot and E. Freund, Zeolites, 1984, 4, 108.2 J. P. Vandenberg, J. P. Wolthuizen and J. H. C . VanHooff, Proc. 5th Int. Conf. Zeolites, ed. L. V. C. Rees (Heyden, London, 1980), p. 649. 3 T. Mole and J. A. Whiteside, J . Catal., 1982, 75, 284. 4 T. Mole, J . Catal., 1983, 84, 423. 5 D. Farcasiu, J . Catal., 1983, 82, 252. 6 M. B. Sayed and R. P. Cooney, Aust. J . Cliern., 1982, 35, 2483. 7 M. B. Sayed, R. A. Kydd and R. P. Cooney, J . Catal., 1984,88, 137. 8 E. G . Derouane, Zeolite Science and Technology, NATO AS1 Ser. E80, ed. F. R. Ribeiro (Martinus 9 E. G. Derouane, J. B. Nagy, P. Dejaifve, J. H. C. VanHooff, B. P. Spekman, J. C. Vedrine and Nijhoff, Dordrecht, 1984), p. 437. C. Naccache, J . Catal., 1978, 53, 40. 10 J. R. Anderson, K. Foger, T. Mole, R. A. Rajadhyaksha and J. V. Sander, J. Catal., 1979, 58, 114. 1 1 R. A. Rajadhyaksha and J. R. Anderson, J . Catal., 1980, 63, 510. 12 P. Dejaifve, J. C. Vedrine, V. Bolis and E. G. Derouane, J . Catal., 1980, 63, 331. 13 M. B. Sayed, J. Chem. SOC., Faraday Trans. 1 , 1987,83, 1751. 14 F. X. Cormerais, G. Perot, F. Chevalier and M. Guisent, J . Chern. Res. ( S ) , 1980, 362. 15 W. M. Meier, Z . Kristallogr., 1961, 115, 439. 16 G. T. Kokotailo, S. L. Lawton. D. H. Olson and W. M. Meier, Nature (London), 1978, 272, 437. 17 R. M. Barrer and M. B. Makki, Can. J. Chern., 1964, 42. 1481. 59 FAR 11778 Mechanism of MTGIHZSM-5 Conversion 18 G. T. Kerr, J . Phys. Chem., 1967, 71, 4155; 1968, 72, 2594. 19 P. E. Eberly, C. N. Kimberlin and A. Voorhies, J . Catal., 1971, 22, 419. 20 1. M. Belenkaya, M. M. Dubinin and I. I. Kristofori, Zzv. Akad. Nauk SSSR Ser. Khim., 1973, 505. 21 N. Y. Chen, J. Phys. Chem., 1976, 80, 60. 22 J. R. Kiovsk, W. J. Goyetta and T. M. Notermann, J . Catal., 1978, 52, 25. 23 R. Berman, J . Catal., 1981, 68, 242. 24 G. A. Olah and R. H. Schlosberg, J . Am. Chem. Soc., 1968, 90, 2726; 1973, 95,4939. Paper 6/ 1607; Received 6th Augusl, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301771

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

.I. Chem. Soc., Faraday Trans. 1, 1987,83, 1779-1782 Measurements of Tracer Diffusion Coefficients of Lithium Ions, Chloride Ions and Water in Aqueous Lithium Chloride Solutions Kazuko Tanaka* The Institute of Physical and Chemical Research, Wako-shi, Saitama 351 -01, Japan Masao Nomura Research Laboratory for Nuclear Reactors, Tokyo Institute of Technology, Meguro-ku, Tokyo 152, Japan Tracer diffusion coefficients of lithium ions, chloride ions and water in aqueous lithium chloride solutions over a wide range of concentrations at 298.2 K have been determined by means of the diaphragm-cell method using 6Li, 36Cl and 2H as tracers. The experimental value of the tracer diffusion coefficient of lithium ion was equal to that of the chloride ion at 18.6 mol kg-l, although at lower concentrations the values for the lithium ion were smaller than those of the chloride ion.This fact may indicate that the direct interaction between lithium and chloride ions is significant at 18.6 mol kg-l. It is of considerable interest to study the transport properties of highly concentrated aqueous electrolyte solutions where there are fewer water molecules than are needed to satisfy the hydration requirements of cations and anions. However, the experimental data for tracer diffusion coefficients of all constituent species in such solutions are rarely found in the literature. The aim of this paper is to report tracer diffusion coefficients of two constituent ions and water in aqueous lithium chloride solutions over a wide range of concentrations up to near saturation.The data obtained are compared with published val~esl-~ and discussed in relation to the structure of the solutions deduced from X-ray4 and neutron diffraction Experimental Measurements were made with a conventional diaphragm cell which incorporated a sintered glass disc. Details of the experimental procedure were as described in a previous paper.8 The cell was calibrated by diffusion of 0.5 mol dm-3 potassium chloride solution into pure water at 298.2 K together with Stokes' data for the ~ystern.~ The concentrations of potassium chloride were determined by weighing the residue obtained from evapora- tion of experimental solutions to dryness. The enriched 6LiCl and ?LiCl were prepared from 6LiC03 and ?LiCO,, respectively (Tomiyama Pure Chemical Institute Ltd, Japan) and aqueous hydrochloric acid.36Cl was obtained from The Radiochemical Centre, Amersham. Deuterium oxide (99.75 atom % deuterium) was obtained from Junsei Pure Chemicals, Japan and diluted to 2 atom% with triply distilled normal water. The isotope ratios of 6Li/7Li were determined with the aid of a Varian MAT CH-5 mass spectrometer. The chloride ion in the sample was ion-exchanged for iodide before the measurements were taken. The radioactivity of 36Cl was counted by means of a standard liquid scintillation technique using ACSII (Amersham, U.S.A.) as a scintillant. The isotope ratios of 2H/1H were determined densitometrically by means of a DMA60 1779 59-21780 Measurements of Tracer Diflusion Coeficients Table 1. Tracer diffusion coefficients of lithium ion in aqueous 0.1 mol kg-I lithium chloride solutions at various initial concentrations of 6Li in lower (1) and upper (u) compartments of the cell 95.44 7.57 1.022 81.80 7.57 1.016 24.00 7.57 1.014 1.81 66.8 1 1.025 1.81 95.42 1.020 Anton Paar density meter.The water in the sample was separated from LiCl by distilling the solution before the density measurements. To avoid isotope fractionation during the distillation, water in the solution was first separated from LiCl by distillation to dryness, then the water collected was distilled under nitrogen. Results and Discussion The tracer diffusion coefficient of lithium ions in aqueous 0.1 mol kg-l lithium chloride at 298.2 K has been measured by varying the initial isotope ratios of 6Li/7Li in the upper and lower compartments of the diaphragm cell and is given in table 1.Each value listed is the average of 2 to 4 measurements and the experimental error is estimated to be within 0.5%. The agreement of observed values within the limits of experimental error indicates that the isotope effect is not significant. This observation is supported by the result of Kunze and FUOSS,~* who have found that at infinite dilution in water the ratio of self-diffusion coefficients of 6Li+ and 7Li+ is 1.0035, which is estimated from the limiting equivalent conductance of each isotope by means of the Nernst equation.ll Tracer diffusion coefficients for all constituent species in aqueous solutions of lithium chloride at 298.2 K over a wide range of concentrations are given in table 2.The initial isotope ratios of lower and upper compartments of the cell were 95.44 and 7.57 atom % of 6Li, respectively, for the first five cases of the measurements of tracer diffusion coefficients of lithium ions. The values obtained at 13.8 and 18.6 mol kg-l are averages of the values obtained using the solutions containing several isotope ratios. The measurements of tracer diffusion coefficients of chloride ion in aqueous solution of lithium chloride have been extended to include concentrations above 10 mol kg-l since the data obtained with medium concentrations are available in the literature and the values reported by Mills2 and Turq et a1.l are in good agreement. Our data, together with the literature values, are plotted as a function of concentration in fig.1. There is fairly good agreement between present data and Turq's data of tracer diffusion coefficients of the lithium ion at lower concentrations, although at higher concentrations the values reported by Turq et a1.l are slightly larger than our values. The disagreement between the two sets of water diffusion data is mainly ascribed to the difference in the tracer diffusion coefficient of water in pure water that we used13 and the value reported by Tamas et al.3 It is interesting to note that these tracer diffusion coefficients of component species in solution are related to the hydrated structure of lithium ion in the solution. It is generally accepted that although both lithium and chloride ions are hydrated, the lithium ion is more strongly hydrated than the chloride ion, and no significant change in the dynamical properties of water molecules around chloride ions was 0bser~ed.l~ In thisK.Tanaka and M . Nomura 1781 Table 2. Tracer diffusion coefficients of lithium ion, chloride ion and water in aqueous lithium chloride solutions at 298.2 K m D Li Dc, DH*O /mol kg-l / 1 OPg m2 s-l / I 0-9 m2 s-l / m2 s-' 0.10 1.022 - 2.183 0.50 1.005 - 2.114 1.03 0.93 1 - 1.934 2.10 0.914 - 1.701 5.60 0.6 15 - 1.144 13.8 0.232 0.321 0.400 18.6 0.141 0.143 0.241 ____ 3.r 2.1 - I v) N E 2 9 m I ---. 1 .c m I L A 4'. 14 L 1 0 10 20 rnlmol kg-' Fig. 1. Concentration dependence of tracer diffusion coefficients of Li+, Cl- and water in aqueous LiCl solutions at 298.2 K; Li+ (O), C1- (0) and water (A), filled symbols designate respective literature values.respect we assume that all the water molecules in highly concentrated lithium chloride solutions are hydrated to lithium ions. In concentrated solutions of lithium chloride the hydration number for the lithium ion is four, while in dilute solutions it is S ~ X . ~ T 7 3 l2 Up to 13.8 mol kg-l, where the ratio of lithium ions to water molecules is one to four, the hydration requirement of lithium ions can be satisfied. Over this concentration range there are insufficient water molecules to complete the hydration shells of lithium ions and the direct interaction between lithium and chloride ions becomes significant.5*1782 Measurements of Tracer Difusion Coeficients This picture is reflected in the measured values of tracer diffusion coefficients of lithium and chloride ions as shown in table 2.At 18.6 mol kg-l the value of tracer diffusion coefficient of chloride ions is equal to that of lithium ions. This may be an indication of direct interaction between lithium and chloride ions as a consequence of which they travel together in the solution. At concentrations lower than 18.6 mol kg-l the values of tracer diffusion coefficients of lithium ions are much smaller than those of chloride ions, suggesting that the diffusional behaviour of lithium and chloride ions are for the most part independent. This is to be expected if the lithium ion is surrounded by four water molecules. At lower concentrations the complete hydration of lithium ions precludes any significant direct interaction between lithium and chloride ions.We thank Prof. M. Okamoto, Prof. R. Tamamushi and Prof. I. Okada for helpful discussions. This work was partially supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan. References 1 P. Turq, F. Lantelme, Y. Roumegous and M. Chemla, J . Chim. Phys., 1971, 68, 528. 2 R. Mills, J. Phys. Chem., 1957, 61, 1631. 3 J. Tamas, S. Lengyel and J. Giber, Actu Chim. Acad. Sci. Hung., 1963, 38, 225. 4 I. Okada, Y. Kitsuno, H-Y. Lee and H. Ohtaki, Ions and Molecules in Solution, ed. N . Tanaka, 5 K. Ichikawa, Y. Kameda, T. Matsumoto and M. Misawa, J . Phys. C, 1984, 17, L725. 6 A. P. Copestake, G. W. Neilson and J. E. Enderby, J. Phys. C, 1985, 18,421 1 . 7 J. R. Newsome, G. W. Neilson and J. E. Enderby, J . Phys. C, 1980, 13, L922. 8 T. Hashitani and K. Tanaka, J . Chem. SOC., Faruduy Trans. I , 1983,79, 1765. 9 R. H. Stokes, J. Am. Chem. SOC., 1951, 73, 3527. 10 R. W. Kunze and R. M. Fuoss, J. Phys. Chem., 1962, 66, 930. 1 1 R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworths, Sevenoaks, 1965), p. 317. 12 J. E. Enderby and G. W. Neilson, Rep. Progr. Phys., 1981, 44, 38. 13 K. Tanaka, J . Chem. SOC., Faraday Trans. I , 1978, 74, 1879. 14 N. A. Hewish, J. E. Enderby and W. S. Howells, J . Phys. C, 1983, 16, 1777. H. Ohtaki and R. Tamamushi (Elsevier, Amsterdam, 1983), p. 81. Paper 6/1672; Received 15th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301779

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1987,83, 1783-1793 Internal Pressures, Temperatures of Maximum Density and Related Properties of Water and Deuterium Oxide Michael J. Blandamer,* John Burgess and Andrew W. Hakin Department of Chemistry, University of Leicester, Leicester LEI 7RH Procedures are described for calcuating internal pressures of water and deuterium oxide at given temperatures and pressures. The dependences on pressure of temperatures of maximum density are described using quadratic equations. Similar equations describe the dependences on pressure of the temperature at which internal pressures are zero. Equilibrium aqd con- figurational internal pressures are identified, the latter being linked to reorganisation of hydrogen bonds in water and in deuterium oxide. Considerable evidence1* points to the importance of solvent-solvent interactions in determining the sign and magnitude of kinetic parameters for chemical reactions involving solutes in aqueous solution and in D20.Recently we drew attention3 to the possibility of using internal pressure^,^*^ xi, of the solvent, water, in establishing a quantitative basis for treating kinetic data for reactions3$ in aqueous solution7~ (see also extension to aqueous solutions of neutral solutesg and of electrolyteslO). The change in internal pressurell for water, negative to positive,12 with increase in temperature is particularly interesting.' In conjunction with these studies,. we have established a method for calculating zi at given temperatures and pressures for both water and deuterium oxide.The method is based on a dependence of internal pressures about reference temperatures and pressures. The method differs from that adopted by Leyendekker~,'~ who based an analysis on the Tait equation written in logarithmic f 0 ~ m . l ~ The input to our analysis is the set of volumetric properties15 reported by Fine and Millerol6? l7 Further, we explore the dependence on pressure of temperatures of maximum density14 (t.m.d.) for both water and D20. For both liquids the t.m.d. decrease with increase in pressure. A similar trend is observed for the dependence on pressure of the temperature corresponding to zero internal pressure. Analysis Fine and Millerol67 l7 expressed the dependences of the volumes of 1 g of water and of D,O on temperature and pressure using equations having the form PVO/(VO- VP) = B+A1P+A,P2 (1) where P is the gauge pressure, p minus 1 atm;? Vo and V p are the volumes of liquid at gauge pressures zero and P, respectively; B, A, and A , are quartics in the temperature function (T- 273.15) K.The dependences on temperature of volume Vat 101 325 N rn-, are calculated using the equations given by Kell.149 la Therefore, based on eqn (l), the volume V can be calculated over the ranges 0 < P/bar < lo3 and 0 d (T- 273.15)/K < 100.0. Further, eqn (1) can be recast in a form which leads to the isothermal compressibilities K~ and (isobaric) expansivities a over these ranges. Fine and Millerol*V l7 provide tables of V, I C ~ and a at intervals of 100 bar and 5 K. These t 1 atm = 101 325 Pa.17831784 Internal Pressures of H 2 0 and D,O Table 1. Internal pressures of water and deuterium oxide; derived parametersa water deuterium oxide O/K ;n/bar n,(O; n)/bar a2 1 02a3 1 02a, 1 03a5 1 05a6 105a, 104a, 1 06a, olbar 323.15 500 3277.17 & 0.85 54.21 3 k 0.085 - 13.34k0.27 - 19.429 & 0.058 - 11.732&0.051 14.49 k 0.18 1.39kO. 14 - - 8.79 323.15 500 303 1.72 & 0.94 59.693 & 0.045 - 9.44 0.24 - 23.523 & 0.055 - 12.9699 & 0.046 19.41 k0.16 1.70 & 0.1 5 2.82 f 0.54 4.32 & 0.22 7.95 a D = standard error. tables can be used to construct tables of internal pressures containing 11 x 3 1 entries of ni for both water and D20: ni = T ( ~ / I c , ) - P . (2) A FORTRAN computer program was written which reproduced these tables for V, IC, and a based on eqn (1) and the parameters reported by Fine and Millero.l69 l7 The corresponding 1 1 x 3 1 internal pressures [eqn (2)] were fitted as a function of temperature and pressure using eqn (3), which is based on a Taylor expansion about internal pressure ni(6; n) at temperature T = 8 and pressure p = n.q(T;p)/bar = ni(0; n)/bar+(a2/K)(T-8)+(a,/bar)(p-n) + (a,/K2) ( T - 8)2 + (aJK bar) ( T - 8) ( p - n) + (a,/K2 bar) ( T - 8)2 ( p - n) + (a7/K bar2) ( T - 8) ( p - n)2 + @,/bar2) ( p - n)2 + (a,/K3) ( T - 8)3. (3) Internal pressures were fitted to eqn (3) using a linear least-squares procedure, the FORTRAN program providing variance-covariance and correlation matrices. Another FORTRAN program used a minimisation routine in conjunction with equationsl6* l7 describing the dependences of a on temperature and pressure [cf differential of eqn (1) with respect to T a t constant pressure].The output was the pressure at which a is zero at a given temperature, i.e. where volume V is a minimum. The same computer program was modified to report pressure p at a given temperature corresponding to the condition that ni is zero. Results In conjunction with eqn (3), the reference temperature 8 and pressure n were set at 323.15 K and P/bar = 500.0, respectively. Consequently the off-diagonal elements of the correlation matrix indicated minimum correlation between estimates of derived parameters.lg The validity of derived parameters was tested using F-tests of the variance at the 95% confidence level. The data for water required the first seven terms in eqn (3), whereas the data for D 2 0 required nine terms as judged by these F-tests. Derived parameters and standard errors are summarised in table 1.We examine in fig. 1 the pattern of residuals, being the difference between internal pressures calculated usingM. J . Biandamer, J . Burgess and A . W. Hakin 40 (a) 0 0 I - 40 0 1785 0 20 40 60 80 100 (T-273.15)/K 40 ( b ) 20 T o O’ 0 0 0 0 - * O l O 0 0 0 I I 1 1 1 0 20 40 60 80 100 Fig. 1. Dependence of residuals, An, on temperature for (a) water and (b) deuterium oxide; An, = difference in internal pressures calculated using eqn (1) and (3). - 401 (T-273.15)/K eqn (3) (with data in table 1) and those pressures calculated from eqn (1) using the parameters reported by Fine and Millero.l69 l7 The dependences of 7ci on pressure and temperature are summarised in tables 2 and 3 together with fig. 2 and 3.With increase in temperature the internal pressure increases. At low temperature ni increases, but at high temperature 7ci decreases with increase in pressure. Consequently, at ca. 323 K, the internal pressure of water is rather insensitive to pressure. The dependences on pressure of t.m.d. for water and for D,O are satisfactorily described by eqn (4) in conjunction with the parameters summarised in tables 4 and 5: (4) The results are summarised in fig. 4, which shows that t.m.d. decrease with increase in pressure with an almost linear dependence. The dependence at ambient pressure for water t.m.d. = t.m.d.(277.15 K ; P = 0) + (a,/bar) P+ (a2/bar2) P2.1786 Internal Pressures of H,O and D,O Table 2.Internal pressures for watera t 0 0.5 1 .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 P/ bar 0 600 700 800 900 1000 - 367 - 320 - 273 - 227 - 181 - 135 - 90 - 44 (0.4) 45 89 133 520 928 1319 1693 2052 2396 2727 3044 3348 3639 3917 4182 4436 4677 4906 5124 5330 5525 5709 -274 -184 -94 -64 79 -230 -140 -53 33 117 -185 -98 -12 72 155 -141 -55 29 111 192 -97 -13 69 150 230 -53 29 110 189 267 -9 71 150 227 304 34 112 189 266 340 77 153 229 304 377 120 194 268 341 413 162 235 308 379 449 205 276 347 416 485 576 632 689 746 802 970 1012 1055 1099 1143 1347 1377 1408 1440 1473 1710 1729 1750 1771 1794 2059 2069 2080 2092 2106 2395 2396 2399 2404 2410 2718 2713 2709 2706 2705 3029 3018 3007 2999 2992 3328 3311 3296 3282 3270 3615 3593 3574 3555 3538 3889 3864 3841 3818 3797 4152 4124 4097 4072 4047 4403 4372 4343 4314 4287 4642 4609 4577 4546 4516 4870 4825 4801 4768 4735 5086 5049 5013 4978 4944 5291 5252 5215 5178 5141 5484 5444 5405 5366 5327 5666 5624 5583 5542 5501 163 243 321 396 467 199 278 355 428 498 235 313 388 460 429 271 348 422 492 560 307 382 455 524 590 343 416 487 556 621 378 450 520 587 651 413 484 553 619 681 448 518 585 649 712 483 551 617 681 741 518 585 650 711 771 552 618 682 743 801 858 912 964 1015 1064 1186 1229 1270 1310 1349 1505 1537 1569 1599 1628 1816 1839 1862 1883 1904 2120 2134 2149 2162 2175 2416 2423 2430 2436 2442 2705 2705 2705 2705 2705 2985 2980 2974 2968 2962 3258 3247 3236 3225 3214 3522 3506 3491 3476 3460 3777 3758 3738 3719 3699 4023 4000 3977 3954 3930 4260 4234 4207 4181 4154 4487 4458 4429 4399 4370 4704 4672 4640 4608 4576 4910 4876 4842 4807 4773 5105 5069 5032 4996 4958 5289 5250 5211 5172 5133 5460 5420 5378 5336 5294 a t = (T-273.15)/K; internal pressures recorded in bar.is described by dp/d(t.m.d.) and its reciprocal, d(t.m.d.)/dp, being - 50.01 bar K-l and - 1.996 x A similar pattern emerges in terms of the dependence of pressure on temperature corresponding to the condition that ni = 0: K bar-', respectively. Kell reported14 - (2.00 & 0.03) x lo-, K bar-l. T(ni = O)/K = a,/K + (a,/bar) P+ (a2/bar2) P2 + (a3/bar3) P3. ( 5 ) T.M.D. - Dependence on Temperature and Pressure A numerical method was used in the previous section in order to calculate the pressure dependence corresponding to a given t.m.d.In this section we show how the calculation can be undertaken using a different procedure. The volume of a fixed amount (e.g. 1 mol) of water V* is defined by the independent variables T and p , i.e. V* = V*[T;p]. The complete differential of this equation is dV* = ( g ) , d T + ( g ) , d p .M . J . Blandamer, J. Burgess and A . W. Hakin 1787 Table 3. Internal pressures for deuterium oxidea P/bar t 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 0 -617 -115 357 804 1229 1635 2023 2394 2749 3087 3410 3717 4009 4286 4547 4794 5026 5244 5447 5638 100 200 300 400 500 600 700 800 900 1000 - 524 - 43 41 1 843 1255 1649 2389 2737 3069 3387 3690 3978 4252 4512 4758 4989 5207 5412 5604 2027 - 432 30 468 884 1283 1666 2034 2388 2727 3054 3366 3665 3950 422 1 4479 4723 4954 5172 5378 557 1 -342 -251 103 177 525 583 927 972 1314 1346 1685 1707 2044 2055 2389 2392 2721 2717 3041 3030 3348 3332 3642 3622 3924 3899 4192 4165 4448 4418 4691 4660 4921 4889 5139 5107 5345 5314 5540 5510 - 163 250 642 1017 1380 1730 2069 2397 2715 3022 3318 3603 3877 4139 4390 4630 4858 5076 5284 548 1 -75 12 97 180 262 322 394 465 535 604 700 759 817 874 931 1063 1110 1156 1202 1247 1414 1449 1485 1520 1554 1754 1779 1805 1830 1855 2084 2100 2116 2133 2149 2404 2412 2420 2428 2437 2714 2715 2716 2717 2718 3015 3009 3003 2998 2993 3305 3294 3283 3272 3261 3585 3569 3553 3528 3523 3855 3835 3815 3796 3777 4114 4091 4068 4046 4024 4363 4370 4312 4287 4263 4601 4573 4547 4521 4495 4829 4800 4773 4746 4720 5047 5018 4990 4964 4938 5255 5227 5200 5174 5149 5454 5428 5402 5378 5354 a t = (T-273.15)/K; internal pressures recorded in bar.At a minimum in V*, Eqn (7) is differentiated with respect to temperature at fixed pressure to yield This equation for the pressure dependence of the temperature at V& is similar to that derived by Kell,l5? 20, 21 except that Kell used the density, p, as the dependent variable. The approach based on eqn (8) permits an extension of the treatment to incorporate derivatives of the thermal expansivity a and isothermal compressibility I C ~ . From the definition of a, (9) Hence at Vzin, From the definition of I C ~1788 5000 4000 3000 1000 Internal Pressures of H,O and D,O P = O bar P = 1000 bar 280 300 320 340 360 i P = 0 bar P = 1000 bar Fig.2. Dependence of zi on temperature at intervals in P of 100 bar for (a) water and (b) deuterium oxide. Table 4. T.m.d. ; derived parameters parameter water deuterium oxide t.m.d. (P = O)/OC 3.9852 +0.0012 11.416 & 0.032 102a, - 1.9964+0.0018 - 1.954k0.032 1 06a, - 5.556 & 0.053 - 1.98 _+ 0.73 standard error in t.m.d./K 0.0014 0.034M. J . Blandamer, J . Burgess and A . W. Hakin 1789 I I I i I I I ~ I I L 80 100 0 -273.15)/K Fig. 3. Dependence of xi on temperature and pressure for (a) water and (b) deuterium oxide. Therefore, The two partial differentials on the right-hand side of eqn (1 3 ) can, to a fair approxi- mation, be replaced by (ALIC~IAT) and (AaIAT). At ambient pressure with AT set at 1 K about 227 K, (aT/ap) at 277 K is calculated as 2.035 x K bar1, which agrees with that obtained numerically.Discussion According to the first and second laws of thermodynamics, dU = T dS - p dV- A d t . (14) Here dU is the change in thermodynamic energy of a closed system for a change in entropy dS at temperature T, a change in volume dV at pressure p and a change in composition/organisation d[ at affinity A . In considering the properties of a liquid (e.g. water or D20), it is convenient to identify the variable < as a measure of organisation. Hence we identify two limiting types of process.22* 23 In one limit, the affinity for spontaneous change A is zero, the change being an equilibrium transformation. In the second case < remains constant, the process being frozen/instantaneous and the resulting property characterising the 'glassy' state. We recall these points in order to identify two1790 Internal Pressures of H,O and D,O Table 5.‘xi = 0’ for water and deuterium oxide; derived parameters parameter water deuterium oxide a0 277.1337 & 0.00023 284.3354 & 0.00023 1 02a, 1 06a2 - 1.988 15 & 0.00068 - 6.23 1 & 0.05 1 - 1.76302 f 0.00048 - 5.806 & 0.03 1 109a, 1.39 & 0.10 9.47 & 0.52 lo4 (standard error in T ) / X 2.47 2.70 400 t (T-273.15)/K Fig. 4. Dependence of t.m.d. on pressure for (a) water and (b) deuterium oxide. limiting internal pressures, the equilibrium ni(A = 0) and instantaneous ni(t) properties defined by eqn (1 5) and (1 6), respectively : Hence from eqn (14) in conjunction with a Maxwell relationship [(as/a v)T = (ap/aT)V1 we arrive at ‘thermodynamic equations of state,,** 25 of which there are at least two, characterising equilibrium and instantaneous internal pressures :M .J . Blandamer, J . Burgess and A . W. Hakin 1791 0.81 B I I I I I 0 40 80 0 40 00 (T-273.15)/K Fig. 5. Dependence of (a) ni(A = 0), (6) ni({) and (c) n,(relax) on temperature at ambient pressure for water (A) and deuterium oxide (B). or Also, Clearly the analysis described in the previous section (cf. tables 2 and 3, and fig. 1-3) refers to equilibrium internal pressures q ( A = 0) together with the equilibrium expan- sivity a(A = 0) and equilibrium isothermal compressibility KT(A = 0). Calculation of q ( 5 ) requires estimates of a(r) and ~ ~ ( r ) , which are not readily available, particularly so for a(().Ultrasonic absorption data can yield K,(() through the related isentropic property ~ ~ ( 5 ) if the latter is identified by limit (v + o o ) ~ , , where v is the frequency of the sound wave.26 The second law of thermodynamics requires that K,(A = 0) > ~~(0, but no condition is set on the signs and relative magnitudes of isobaric expansivities a(l) and a(A = 0). Similarly, no restriction is placed on the signs and relative magnitudes of ni(A = 0) and ni(r). and 1.03244 x K-l, respectively, for the range 273.15 < T/K < 373,15. Combination of these estimates with related estimates of ~ , ( r ) at intervals of 10 K yields ni(r) as a function of temperature at ambient pressure.The resulting patterns are shown in fig. 5 , where we have plotted for both water and D20, ni(A = 0), ni(<) and their differences as a function of temperature. We identify the difference [ni(A = 0) -xi(<)] as ;n,(relax) on the grounds that this property describes the configurational or relaxational component of equilibrium internal pressures. With increase in temperature ni(A = 0) approaches xi(<), confirming the structural distinction which is often drawn between ‘hot’ and ‘cold’ water. Internal pressures can be understood in terms of separate attractive and repulsive contribution^.^ For water (and D,O) these contributions reflect two important characteristics of hydrogen bonding. In most cases hydrogen bonding is responsible for cohesive interactions.Nevertheless, hydrogen bonding also plays a repulsive role in the sense of holding molecules at larger distances of separation than in an analogous close-packed ~ystern.~ The trends in fig. 5 can be understood if the attractive component of hydrogen bonding dominates ni(r), For water and D,O at ambient pressure, EndoZ7 estimates a(c) = 1.1822 x1792 Internal Pressures of H,O and D,O whereas the repulsive part dominates ni(relax). Relaxation to a new configuration of hydrogen-bonded water molecules involves distortion and bending of hydrogen bonds along the lines described in the Lumry-Frank model for water28 [see also ref. (2)]. This qualitative argument also accounts for the patterns shown in the n,(A = 0)-T-p dependences (fig. 2 and 3). The impact of increased pressure on water (and D,O) is highlighted by the dependence of t.m.d.on pressure. The shift in t.m.d. to lower temperature is consistent, using the Lumry-Frank with a displacement of the two-state equilibrium, favouring the distorted hydrogen-bonded, high-density, low- volume state. This trend is confirmed by an increase in nearest-neighbour 0-0 coordination number with increase in pressure,29 a pattern similar to that produced by an increase in temperature at fixed pressure.3o In the introduction to this paper we commented on our interest in understanding the role of solvents in controlling the rates of chemical reactions. Established procedures examine the dependence of rate constants on temperature and on pressure. A common criticism makes the point that water at temperature (at constant pressure p ) is a different solvent from water at temperature &.This criticism emerges in part from the observation that the extent of hydrogen bonding in water depends on temperature. A reference is desirable. One possibility involves examining the dependence of rate constants on temperature and pressure along xi-isobars. A particularly interesting n,-isobar describes the conditions under which n,(A = 0) is zero. In one sense this condition identifies states where the pressure p equals the equilibrium thermal pressure, T(dp/aT) at constant V and ‘ A = 0’. For most systems p 6 ni(A = 0) and ni(A = 0) equals the thermal pressure. The condition ‘ ( A = 0)’ throws into focus the impact of hydrogen bonding between water molecules in producing high-volume systems, i.e.a repulsion. Actually there are few data where the dependence of rate constants can be explored on T and p along the ni = 0 isobar. This is a subject of current research. We thank the S.E.R.C. for a maintenance grant to A. W.H. References 1 M. J. Blandamer, R. E. Robertson and J. M. W. Scott, Prog. Phys. Org. Chem., 1986, 15, 149. 2 M. J. Blandamer, J. Burgess, A. W. Hakin and J. M. W. Scott, J . Chem. Soc., Faraday Trans. I , 1986, 3 M. J. Blandamer, J. Burgess and J. B. F. N. Engberts, Chem. Soc. Rev., 1985, 14, 237. 4 W. Westwater, H. W. Frantz and J. H. Hildebrand, Phys. Rev., 1928, 31, 135. 5 J. H. Hildebrand and R. L. Scott, Solubility of Non-electrolytes (Reinhold, New York, 1950). 6 A. K. Colter and L. M. Clemens, J .Phys. Chem., 1964, 68, 651. 7 M. R. J. Dack, Aust. J. Chem., 1976, 29, 771; 779. 8 M. R. J. Dack, Chem. Soc. Rev., 1975, 4, 21 1 . 9 F. W. Getzen, Solutions and Solubilities, ed. M. R. J. Dack (Wiley, London, 1976), part 11, chap. 15. 10 K. Patil and A. B. Wazalwar, h d . J . Chem., 1981, 20A, 879. 1 1 J. V. Leyendekkers, J. Phys. Chem., 1983,87, 3327. 12 S. E. Wood, J . Phys. Chem., 1962, 66, 600. 13 J. V. Leyendekkers, The Thermodynamics of Sea Water (Marcel Dekker, New York, 1976), part I. 14 G. S. Kell, Waler - A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1973), vol. I, 15 G. S. Kell and E. Whalley, Philos. Trans. R. Soc. London, Ser. A , 1965, 258, 565. 16 R. A. Fine and F. J. Millero, J. Chem. Phys., 1973, 59, 5529. 17 R. A. Fine and F. J. Millero, J . Chem. Phys., 1975, 63, 89. 18 G. S. Kell, J. Chem. Eng. Data, 1976, 12, 66. 19 M. J. Blandamer, J. Burgess, R. E. Robertson and J. M. W. Scott, Chem. Rev., 1982, 82, 259. 20 G. S. Kell, J. Chem. Eng. Dafa, 1970, 15, 119. 21 G. S. Kell, J . Chem. Eng. Data, 1975, 20, 97. 22 I. Prigogine and R. Defay, Chemical Thermodynamics, transl. D. H. Everett (Longmans, London, 23 M. J. Blandamer and J. Burgess, J. Chem. Soc., Faraday Trans. I , 1985, 81, 1495. 24 R. L. Scott, J . Chem. Phys., 1948, 16, 256. 82, 2989. chap. 10. 1954).M . J . Blandamer, J . Burgess and A . W. Hakirz 25 D. D. Macdonald and J. B. Hyne, Can. J. Chem., 1971, 49, 2636. 26 M. J. Blandamer, Introduction to Chemical Ultrasonics (Academic Press, London, 1973). 27 H. Endo, J. Chem. Phys., 1982,76,4578. 28 R. Lumry, E. Battistel and C. Jolicoeur, Faruduy Symp. Chem. Soc., 1982, 17, 93. 29 G. A. Gaballa and G. W. Neilson, Mol. Phys., 1983, 50, 97. 30 P. A. Egelstaff and J. H. Root, Mol. Phys., 1983, 50, 97. 1793 Paper 6/1679; Received 18fh August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301783

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1987, 83, 1795-1804 Pulse Radiolysis Study of Salt Effects on Reactions of Aromatic Radical Cations with C1- Rate Constants in the Absence and Presence of Quaternary Ammonium Salts Yukio Yamamoto," Shoichi Nishida and Koichiro Hayashi The Institute of Scientijic and Industrial Research, Osaka Uniuersity, 8-1 Mihogaoka, Ibaraki, Osaka 567, Japan The effect of quaternary ammonium salts on the decays of the radical cations of biphenyl, trans-stilbene, anthracene and pyrene generated by pulse radiolysis in chlorohydrocarbons has been investigated. The decays, which are due to neutralization reactions with C1-, are retarded by the addition of salts having non-nucleophilic PF,, BF; and ClO;, whereas the radical cations are rapidly quenched by salts having I- and BPh;.The retarding effect of the salts is attributed to the formation of ion pairs between the reacting ions and the counter-ions from the salts. The rate constants for the neutralization reactions in 1,2-dichloroethane have been determined for the free-ion and ion-paired states; the latter state is attained by the addition of Bu,NPF,. The rate constants for the reactions of the radical cations except for Py.+ (Py = pyrene) are in the ranges (1.3-1.9) x 1O1I dm3 mo1-l s-l for the free ions and (2.8-7.0) x 1O1O dm3 rnol-' s-l for the ion pairs. The rate constant determined in the absence of the salt for Py'+ is one order of magnitude smaller than those for the others. The salt effect is also smallest on the reaction of Py'+. The charge delocalization of the large aromatic radical cation may be responsible for the exceptional results for Py'+.The largest salt effect was observed on the reaction of Ph,CH+, a charge-localized carbenium ion investigated for comparison. The solvent effect on the neutralization reactions is also discussed. The salt effects on ionic reactions in radiation chemistry have been studied by pulse radiolysis. In solvents of low polarity the reactivities of charged species are known to be affected by the addition of salts through ion pairing with unreactive counter-ions from the salts, although in solvents of high polarity the effect of ionic strength is important. The effect of ion pairing has been largely studied on reactions of negatively charged species formed in tetrahydrofuran (THF) solutions containing alkali-metal salts.l-lo We have also studied the effect of quaternary ammonium salts on reactions of aromatic radical anions in THF.l19 l2 This study is concerned with the salt effect on the decays of aromatic radical cations in chlorohydrocarbon solutions.The radical cations generated by pulse radiolysis in chlorohydrocarbons decay by the neutralization reactions with Cl-, which is formed by dissociative electron attachment to the solvents. We recently reported that the lifetime of the biphenyl radical cation (BP'+) in dichloromethane is extended by the addition of salts such as Ph3SMX,, Ph,IMX,, C,H,MX, and Bu,NMX, (MX, = PF, or BF,).13 The salt effect has been attributed to the ion-pair formation of BP'+ with the non-nucleophilic anions of the salts. This paper is an extension of the previous study to the decay kinetics of the radical cations of various aromatic compounds such as biphenyl (BP), trans-stilbene (St), anthracene (An) and pyrene (Py) in the absence and presence of quaternary ammonium salts.The rate constants for the reactions of the radical cations with Cl- are determined for their free-ion and ion-paired states in 172-dichloroethane; the latter state is attained in the 17951796 Pulse Radiolysis of Aromatic Radical Cations presence of Bu,NPF,. The salt effect on the reaction of Ph,CH+ is investigated for comparison. Comparison is also made with the previously published results for the decay kinetics of the radical anions in THF.l19 l2 The effect of ion pairing is discussed in relation to the size and charge delocalization of the radical cations.The solvent effect on the neutralization reactions is also discussed. Experimental Chlorohydrocarbons (Wako Chemicals) used as solvents were washed three times with aqueous sodium hydroxide and water and then distilled over calcium hydride. The middle fractions were stored under vacuum over calcium hydride. The aromatic compounds and the quaternary ammonium salts were the same as those used in the previous study.ll> l2 Ph,CHBr (Tokyo Kasei) was purified by recrystallization from hexane. The experimental procedures and the techniques of pulse radiolysis have been described in the previous paper.ll The pulse radiolysis experiments were carried out by using 8 ns electron pulses with dose rates of ca.2 kGy per pulse (beam diameter ca. 4 mm) at room temperature (ca. 22 "C). Results The effects of various kinds of quaternary ammonium salts were examined by the pulse radiolysis of BP in 1,2-dichloroethane. The salts employed were Bu,NPF,, Bu,NBF,, Bu,NC10,, Bu,NI, Bu,NBPh,, CeMe,NPF,, PhMe,NPF, and BzMe,NPF, (Ce = cetyl, Bz = benzyl). The transient absorption spectra obtained in the absence and presence of the salts were the same and agreed with those reported in the 1iterat~re.l~ The absorption bands at 370 and 680 nm are attributed to BP'+, although the former band includes a small absorption due to the triplet excited state of BP having an absorption band at 360 nm.15 The formation and decay processes of BP'+ are presented as follows: CH,ClCH,Cl- CH,ClCH,Cl'+ +e (1) CH2ClCH2Cl'+ + BP -+ CH,ClCH,Cl + BP.+ (2) e + CH,ClCH,Cl-+ C1- + CH,ClCH; (3) BP'+ + C1- -+ neutral product.(4) The decay rate of BP'+ was significantly affected by the addition of the salts. The decay was retarded by the salts having PF;, BF; and ClO;, but it was accelerated by the salts having I- and BPh;, indicating that these anions react with BP'+. The decay rate was independent of the cationic moieties of the salts. It should be noted that the decay of BP'- generated in THF is retarded by these salts except for PhMe,NPF, and BzMe,NPF,, whose addition results in an acceleration of the decay.ll This means that the decay of BP'- in THF depends primarily on the cationic moieties of the salts, Fig. 1 shows the second-order kinetic plots for the decay of BP'+, monitored at 680 nm, in the absence and presence of Bu,NPF,.The plots give straight lines except for the very early stage, although the plot for the salt-free solution deviates slightly from linearity in the latter stages. The rapid decay immediately after the pulse could be attributed to geminate recombination. The slope of the second-order kinetic plot based on the assumption [BP'+] = [Cl-] corresponds to k / d , where k is the second-order rate constant, E is the molar extinction coefficient and 1 is the optical pathlength. Evidence for the second-order decay was given by the dependence of the slope on the optical pathlength of the cell described below. It was also confirmed that the reciprocals of the slopes of the plots at different wavelengths around the peak position are proportional to the optical densities.Fig. 2 shows the relative values for the slopes plotted against salt and BP concentrations. The experimental error for the slopes determined by the1 5 1 2 e, 5 9 .ff 9 -2 \ d 6 3 0 Y . Yamamoto, S. Nishida and K. Hayashi 1797 1 I I I I I I I 1 200 400 600 800 time/ns Fig. 1. Second-order kinetic plots for the decay of BP'+ monitored at 680 nm in 1,2-dichloroethane: [BPI = 4 x mol dmP3. Additive: (a) none, (b) 6 x mol dm-3 Bu,NPF,. [ Bu,NPF, mol dm-j 0 2 4 6 8 10 Fig. [BP]/10-3 mol dmW3 2. Relative values for the slopes of the second-order kinetic plots for the decay of BP' 1,2-dichloroethane. 0, [BPI = 4 x loP2 mol dm-3; a, [Bu,NPF,] = 7 x mol dmA3. + inI798 0.6 0.3 3 0 2 0.9 2 8 0.6 0.3 0 Pulse Radiolysis of Aromatic Radical Cations - 400 5 00 6 00 700 800 w aveleng thl nm Fig.3. Transient absorption spectra of 1,2-dichloroethane solution containing 6 x BP and 3 x lop3 mol dm-3 St: (a) alone; (b) with 6 x mol dm-3 mol dmP3 Bu,NPF,. 1, At the end of the pulse; 2, 110 ns after the pulse; 3, 1 ,us after the pulse. consecutive measurements was < _+ 57;. The slope decreases with increasing salt and BP concentrations and attains a constant value. It is considered that the reacting ions, BP'+ and C1-, are entirely paired with the counter-ions from the salt at the high salt concentrations above ca. 6 x mol dmP3. The high BP concentrations above ca. 4 x mol dmP3 may be necessary for the complete capture of the solvent cation, resulting in the formation of BP'+, equivalent to C1-.Similar results were obtained with the other salts having PF;, BF, and C10,; the slopes of the second-order kinetic plots for the decay in the presence of these salts were similar at high salt and BP concentrations. The salt effect was studied on the decay rates of St'+, An'+ and Py'+ in 1,2- dichloroethane. To prevent the formation of dimeric radical cations,16 the pulse radiolysis was carried out at low solute concentrations, (1-3) x 1 (F3 mol dmP3, in the presence of 6 x mol dmP3 BP. The addition of an excess of BP is for the complete capture of the solvent cation, followed by the charge transfer to the solutes. Fig. 3 shows the transient absorption spectra for the BP-St solution irradiated in the absence and presence of 6 x mol dm-3 Bu,NPF,.The transient absorption spectra observed at the end of the pulse are assigned to BP'+, and the fast decay of BP'+ accompanies the formation of St.+, which has a sharp absorption band at 480 nm and a small one at 750 nm : BP'++St + BP+St'+. ( 5 )Y . Yamamoto, S . Nishida and K . Hayashi 1799 time/ps Fig. 4. Second-order kinetic plots for the decay of St.+ monitored at 480 nm in ,2-dichloroethane solution containing 6 x mol dm-3 BP and 3 x low3 mol dm-3 St. Aditive: (a) none, (6) 6 x rnol dm-3 Bu,NPF,. The yield of St.+ in the solution containing the salt is higher than that in the salt-free solution. This indicates that reaction (4) competes with reaction ( 5 ) in the absence of the salt. Fig. 4 shows the second-order kinetic plots for the decay of St'+ in the absence and presence of the salt.The slope of the plot remained constant in the salt concentration range (6-10) x Similarly, the pulse radiolysis of the BP-An and BP-Py solutions resulted in the formation of An'+ and Py++, whose yields were increased by the addition of Bu,NPF,. Fig. 5 and 6 show the second-order kinetic plots for the decays of An'+ and Py'+ in the absence and presence of 6 x mol dmP3 Bu,NPF,. It is demonstrated that the decay rate of Py'+ is less affected by the addition of the salt than those of BP'+, St'+ and An'+. It was confirmed that the slopes of the plots are unchanged at salt concentrations (6-10) x mol dm-3. As reported previo~sly,l~-~~ the pulse radiolysis of Ph,CHBr in chlorohydrocarbons results in the formation of Ph,CH+, which also decays by the reaction of C1-.The effect of Bu,NPF, on the decay rate of Ph,CH+ was also investigated in 1,2-dichloroethane. Fig. 7 shows the second-order kinetic plots for the decay of Ph,CH+, monitored at 447 nm, in the absence and presence of 6 x The absolute rate constants for the neutralization reactions were determined using irradiation cells of 1 and 2 mm optical pathlengths. It was confirmed that the slope of the second-order kinetic plot is inversely proportional to the optical pathlength of the cell according to (slope) = k / d . The data are listed in table 1; the error for the rate constants is shown by the mean deviation. The rate constants, k, and k,, determined in the absence and presence of 6 x lop3 mol dm-3 Bu,NPF, are for the reactions in the mol dm-3 examined.rnol dm-3 Bu,NPF,.1800 Pulse Radiolysis of Aromatic Radical Cations time/ns Fig. 5. Second-order kinetic plots for the decay of An'+ monitored at 724 nm in 1,2-dichloroethane solution containing 6 x mol dm-3 An. ,4dditive: (a) none, (b) 6 x lop3 rnol dm-3 Bu,NPF,. rnol dm-3 BP and 3 x free-ion and ion-paired states. The k,/k, ratios are presented as a measure of the salt effect. The molar extinction coefficients of BP'+, An'+ and Ph,CH+ were obtained from the literature.20q 22 The values for St'+ and Py'+ were determined, in the present study, by experiments with BP, BP-St and BP-Py solutions containing Bu,NPF,, based on the literature value for BP'+. The k, value for Ph,CH+ is close to the previously reported value, 9.1 x 1Olo dm3 mol-I s-1.21 The pulse radiolysis of BP was also carried out with other chlorohydrocarbon solvents, such as 1,l -dichloroethane, dichloromethane and chloroform, in the absence and presence of Bu,NPF,.The decay rates of BP'+ in these solvents were almost similar, but higher than that in 1,2-dichloroethane. The decay was similarly retarded by the addition of the salt; the decay rate was constant at salt concentrations (6-10) x mol dme3. The experimental errors for the slopes of the second-order kinetic plots determined in the absence and presence of the salt were & 25% and 12%, respectively. The relative values for k , and the k,/k, ratios are presented in table 2 together with the dielectric constants and viscosities of the solvents.The normalization of the relative k, values for the solvent properties was examined and the results are presented in table 2. It can be seen that the rate constant depends primarily on the viscosity of the solvent. Discussion The k, values for the neutralization reactions of the radical cations except for Py'+ in 1,2-dichloroethane are close to each other, lying in the rangeY. Yamamoto, S. Nishida and K. Hayashi 1801 5.01 1 I I I I I I 1 t 4.2 1 Q) 3 . 4 s -2 4 3 0 --.- 3 2.6 1.8 3 - - - - - - 1 .o I I I 1 I I I 0 2 4 6 8 t im e/p s Fig. 6. Second-order kinetic plots for the decay of Py'+ monitored at 450 nm in 1,2-dichloroethane solution containing 6 x mol dm-3 BP and I x mol dm-3 Py. Additive: (a) none, (b) 6 x rnol dm-3 Bu,NPF,. 1 (1.3-1.9) x loll dm3 mol-1 s-l (table 1).The reactions of these radical cations in the free-ion state are diffusion-controlled and the mobility of the reacting ions is increased by the attractive force of the Coulombic interaction.The slightly smaller value for Ph,CH+, 8.5 x 10lo dm3 mol-1 s-l, might be correlated with the difference in the solvent effect as described below and/or with the small effective cross-section of the charge- localized carbenium ion. The ko value for Py'+ is one order of magnitude smaller than those for the others. A possible explanation for the striking result would appear to involve the formation of a complex between Py'+ and C1-: kc kd pY'++Cl-*pY'+ ... C1- 2 neutral product. (6) In this case the k , value determined for Py'+ corresponds to k, k,/k,. The effect of ion pairing on the reaction of Py'+ is also very small, as shown by the k,/k, ratio.The charge delocalization of the large aromatic radical cation may be responsible for the exceptional results for Py'+. On the other hand, the effect of ion pairing is largest on the reaction of Ph2CH+, for which the charge is more localized than that of the aromatic radical cations. It is reasonable to consider that the salt effect is large for the charge-localized cation because of the large interaction with the anion from the salt. The k, values for BP'+, St.+ and An'+ are in the range (2.8-7.0) x 1O1O dm3 mol-l s-l. The bimolecular rate constant for diffusion-controlled reactions between neutral species can be estimated from an approximate equation k, = ~ R T / ~ v .~ ~ The k, values are larger than the calculated value, 8.1 x lo9 dm3 m o t 1 s-l. Although the effect of the Coulombic1802 8.0 6.4 p) 4.0 u s e 8 2 - 3.2 \ 1.6 O, Fig. 7. Second-order from Ph,CHBr in 1 Pulse Radiolysis of Aromatic Radical Cations I I I I I I I 1 I I I I I i I 1 400 800 1200 1600 time/ns kinetic plots for the decay of Ph,CH+, monitored at 447 nm, produced ,2-dichloroethane: [Ph,CHBr] = 5 x lo-, mol dmP3. Additive: (a) none, (b) 6 x mol dm-3 Bu,NPF,. Table 1. Rate constants for the neutralization reactions with Cl- in 1,2-dichloroethane determined in the absence (k,) and presence (k,) of 6 x mol dm-3 Bu4NPF, ~ rate constant/dm3 mol-1 s-l cation A/nm &/dm3 mol-1 cm-l k, ks kolk, BP'+ 680 1.66 x 104 (1.9k0.3) x 10" (7.0f0.5) x 1Olo 2.7 St'+ 480 6.5 x 104 (1.6k0.2) x loll (3.3f0.2) x 1Olo 5.0 An'+ 724 1.16 x 104 (1.3f0.3) x 10" (2.8k0.2) x 1Olo 4.6 Py.+ 450 3.1 x 104 (9.4k0.7) x lo9 (8.1 k0.5) x log 1.2 Ph,CH+ 447 3.8 x 104 (8.5k0.7) x 1Olo ( 1 .5 k O . l ) ~ 1Olo 5.7 Table 2. Solvent effect on the reaction of BP'+ with Cl- relative value for rate constant solvent & ("Cy r] (22 "C)b ko k o v korl kolks 1,2-dichloroethane 10.65 (20) 0.8 1 1 1 1 2.7 1,l -dichloroethane 10.0 (18) 0.48 1.4 0.78 0.83 2.0 dic hlorome thane 9.08 (20) 0.43 1.5 0.68 0.80 2.1 chloroform 4.806 (20) 0.55 1.4 0.43 0.95 2.1 - Dielectric constant. Viscosity in cP.Y. Yamamoto, S . Nishida and K . Hayashi 1803 interaction on the mobility of the reacting ions is less important for the ion pairs than for the free ions, it is expected that a dipole-dipole interaction is still present.On the other hand, the difference in the k , values among the radical cations is larger than that in the k , values; it is too large to be explained in terms of the difference in mobility among the radical cations. Therefore, it can be said that the slower reactions of St'+ and An'+ in the ion-paired state are not necessarily diffusion-controlled. Although both the radical cations and C1- may be paired with the counter-ions from the salt, the retarding effect of the salt is attributable largely to the ion-pair formation of the radical cations with the anion from the salt. This is based on the result that the k,/k, ratio apparently depends on the radical cations and is very small in the case of Py'+.That is to say, the effect of the ion-pair formation of C1- with Bu,N+ is not important, probably because of the large size of the cation. The solvent effect on the k, value for BP.+ is compared with that for Ph,CH+ reported in the literature.21 The rate constants for the reaction of Ph2CH+ with C1- have been reported to be in the range from 8.8 x 1O1O dm3 rno1-I s-l (in 1,1,2-trichloroethane) to 2.66 x lo1' dm3 mol-l s-l (in chloroform), depending on the dielectric constant and viscosity of the solvent. It has been shown that the normalization of the rate constants for the solvent properties results in the same value. However, from table 2 it can be seen that the same normalization of the k, values for BP'+, k o q , leads to the apparently different values and that the k , r values are close to each other compared with the ko&r values. These results suggest that the effect of the dielectric constant of the solvent is less important for BP'+ than for Ph,CH+.This may also be attributed to the difference in the degree of charge delocalization between the aromatic radical cations and the carbenium ions. Similarly, the k, r values are close to each other compared with the k, EV values. Comparison is next made with the previously published results for the radical anions in THF.l1? l2 The radical anions generated by pulse radiolysis in THF decay by the reactions with the solvent cation, THF(H+). The rate constants for the reactions of BP'-, St'-, An'- and Py'- determined in the salt-free solutions are in the limited range (2.6-4.0) x loll dm3 mol-1 s-l.The rate constants normalized for the viscosity of THF [7(22 "C) = 0.48 cP] are close to the k , q values for the radical cations, except for Py'+, in 1,2-dichloroethane, This suggests that the rate constants for the neutralization reactions of the radical anions in the free-ion state also depend primarily on the viscosity of the solvent. Here the reaction of Py'+ with Cl- having a much smaller rate constant is regarded as an exceptional case. On the other hand, the rate constants for the reactions of the radical anions with THF(H+) determined in the presence of Bu,NPF, are in the larger range (3.4-12) x lo9 dm3 mol-1 s-l; the k, (= 8RT/3q) values for THF solutions is 1.4 x 1O1O dm3 mol-1 s-l. Therefore, it is considered that neutralization reactions of the radical anions in the ion-paired state are not necessarily diffusion-controlled. l2 These rate constants are apparently smaller than the k, values for the radical cations in 1,2- dichloroethane despite the lower viscosity of THF.It has already been demonstrated that the salt effect on the reactions of the radical anions is due mainly to ion-pair formation of THF(H+) with the anion from the salt? Therefore the large retarding effect of the salt on the reactions of the radical anions can be attributed to the large interaction of the charge-localized THF(H+) with the anion from the salt. Thus it is concluded that the retarding effect of quaternary ammonium salts on neutralization reactions of the radical cations and anions becomes large when the reacting cations are charge-localized and have large interactions with the anions from the salts.On the other hand, the effect of the ion-pair formation of reacting anions with quaternary ammonium cations may be less important because of the large sizes of the cations. We are grateful to Mr Kunihiko Tsumori, Mr Norio Kimura, Mr Tamotsu Yamamoto, Mr Toshihiko Hori and Dr Seishi Takeda for help with the pulse radiolysis experiments.1804 References Pulse Radiolysis of Aromatic Radical Cations 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 J. H. Baxendale, D. Beaumond and M. A. J. Rodgers, Trans. Faraday SOC., 1970, 66, 1996. B. Bockrath and L. M. Dorfman, J. Phys. Chem., 1973,77, 1002. B. Bockrath and L. M. Dorfman, J. Phys. Chem., 1973, 77, 2618.B. Bockrath and L. M. Dorfman, J. Am. Chem. SOC., 1974,%, 5708. G. A. Salmon and W. A. Seddon, Chem. Phys. Lett., 1974, 24, 366. G. A. Salmon, W. A. Seddon and J. W. Fletcher, Can. J . Chem., 1974,52, 3259. J. R. Langan and G. A. Salmon, J . Chem. SOC., Faraday Trans. 1, 1982, 78, 3645. J. R. Langan and G. A. Salmon, J. Chem. SOC., Faraday Trans. 1, 1983,79, 589. M. Ogasawara, N. Kajimoto, T. Izumida, K. Kotani and H. Yoshida, J. Phys. Chem., 1985,89, 1403. For flash photolysis studies see: M. Fisher, G. Ramme, S. Claesson and M. Szwarc, Chem. Phys. Lett., 1971, 9, 306; M. Fisher, G. Ramme, S. Claesson and M. Szwarc, Chem. Phys. Lett., 1971, 9, 309; G. Ramme, M. Fisher, S. Claesson and M. Szwarc, Proc. R . SOC. London, Ser. A, 1972, 327, 467. Y. Yamamoto, S. Nishida, K. Yabe, K. Hayashi, S. Takeda and K. Tsumori, J. Phys. Chem., 1984,88, 2368, Y. Yamamoto, S. Nishida, X-H. Ma and K. Hayashi, J. Phys. Chem., 1986,90, 1921. S. Mah, Y. Yamamoto and K. Hayashi, J . Phys. Chem., 1983, 87, 297. S. Arai, H. Ueda, R. F. Firestone and L. M. Dorfman, J. Chem. Phys., 1969, 50, 1072. G. Porter and M. W. Windsor, Proc. R. Sac. London, Ser. A, 1958, 245, 238. See e.g.: B. Badger and B. Brocklehurst, Trans. Faraday SOC., 1969,65,2588; M. A. J. Rodgers, Chem. Phys. Lett., 1971, 9, 107; A. Kird, S. Arai and M. Imamura, J. Phys. Chem., 1972, 76, 1119. R. J. Sujdak, R. L. Jones and L. M. Dorfman, J . Am. Chem. SOC., 1976, 98, 4875. L. M. Dorfman, Y. Wang, H-Y. Wang and R. J. Sujdak, Faraday Discuss. Chem. Sac., 1978, 63, 149. V. M. DePalma, Y. Wang and L. M. Dorfman, J . Am. Chem. Soc., 1978,100, 5416. Y. Wang, J. J. Tria and L. M. Dorfman, J . Phys. Chem., 1979,83, 1946. K. P. Kundu and L. M. Dorfman, Radiat. Phys. Chem., 1982, 20, 247. G. A. Olah, C. U. Pittman Jr, R. Waack and M. Doran, J . Am. Chem. SOC., 1966, 88, 1488. P. W. Atkins, Physical Chemistry (Oxford University Press, Oxford, 3rd edn, 1986), p. 743. Paper 6/1680; Received 18th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301795

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I, 1987,83, 1805-1813 A Kinetic Study of the Self-reaction of Prop-2-ylperoxyl Radicals in Solution using Ultraviolet Absorption Spectroscopy John E. Bennett? Shell Research Ltd, Thornton Research Centre, P.O. Box I , Chester CHI 3SH Overall rate constants for the self-reaction of prop-2-ylperoxyl radicals in the liquid phase (cyclohexane, decane and dodecane) have been measured over the temperature range 293-373 K using ultraviolet absorption spectro- scopy to monitor the concentration of alkylperoxyl radicals. The Arrhenius constants for the self-reaction are Et = 20 f 3 kJ mol-l and 2A, = (5 f 2.5) x lo9 dm3 mol-l s-l. Above 363 K there is evidence that reaction with the solvent becomes increasingly important. In recent years kinetic spectroscopy has been used to measure the absolute rate constants for the self-reactions of hydrocarbon peroxyl radicals in both gas and liquid phases.The importance of these reactions lies in the fact that peroxyl radicals are the main chain carriers in the oxidation of hydrocarbons, and thus their reactivity has major influence on the extent and nature of the overall oxidation process. The kinetics of these reactions have been studied extensively in the gas phase by using ultraviolet (u.v.) absorption spectroscopy to monitor directly the concentration of the peroxyl radicals, [For leading references see ref. (1) and (2).] In particular the kinetics of the self-reaction of simple alkylperoxyl radicals (meth~l,~ ethyl,' pr0p-2-y1,~9 and t-buty16) have been measured using a molecular modulation technique.These measure- ments have been complemented by detailed product analyses, which have enabled the contributions of the alternative reaction pathways to be determined. On the other hand, studies of these reactions in the liquid phase have been carried out mainly by using electron spin resonance (e.s.r.) spectroscopy to monitor the concentration of the peroxyl radicals.'- In many ways e.s.r. is ideal for the liquid-phase studies, as it has high sensitivity and is specific for free radicals. However, the sensitivity for peroxyl radicals falls off rapidly with increasing temperature because of the severe line-broadening which O C C U ~ S . ~ Amongst other factors the broadening is inversely proportional to the effective radius of the radical and thus is greater for the smaller peroxyl radicals.For most peroxyl radicals the upper temperature limit at which valid kinetic measurements can be made is determined by the line broadening rather than by the rate of reaction. For example, the upper limits for the hydroxymethylperoxyl and prop-2-ylperoxyl radicals are ca. 235 and 293 K, respectively, whereas that for the 2-methylhexadec-2-ylperoxyl radical is ca. 373 K. In contrast, relatively few kinetic studies of peroxyl radicals in the liquid phase have been made by U.V. absorption spectro~copy.~? The paucity of measurements is probably due to the difficulty of obtaining efficient photoinitiators in the liquid phase which do not absorb appreciably in the U.V. region, where the absorption band of the peroxyl radical lies (240-260 nm).Most experiments have been carried out near ambient temperature by using pulse radiolysis to produce the radicals, thus avoiding the need for a photoinitiator. -f Present address: Department of Chemistry, University of York, Heslington, York YO1 5DD. 18051806 U. V . Studies of Prop-2-ylperoxyl Radicals in Solution In this paper we present kinetic measurements of the self-reaction of the prop- 2-ylperoxyl radical using U.V. absorption spectroscopy to detect the radicals. The radical was selected because detailed kinetic4 and product5 studies of the self-reaction have been made previously in the gas phase over the temperature range 300-373 K, and recently we have made comparable measurements in the liquid phase over the temperature range 173-288 K by using e.s.r.spectroscopy.1° The present measurements cover the temperature range 293-396 K, which is similar to that covered by the gas-phase studies and extends the liquid-phase measurements, made recently by e.s.r., to higher temperature. Experimental U.V. Absorption Apparatus Alkylperoxyl radicals have an absorption band with a maximum around 240-250 nm (decadic cmax z lo3 dm3 mol-l cm-l). Thus the apparatus should be capable of operating down to 200 nm and have a sensitivity of ca. The basic apparatus is shown in fig. 1. The monitor light source was a 150 W xenon lamp which was run from a stabilized power supply (Applied Photophysics model 406/01; ripple, noise and drift all < 0.5%). The output light was focused into the plane of a rotating chopper (Rofin frequency-programmable light chopper mk 11) and then collimated with a plano-convex silica lens (50 mm diameter; 140 mm focal length).The collimated beam was stopped down to a diameter of 1 mm by an iris diaphragm before passing through the sample cell (10 mm pathlength; see below). A second iris diaphragm was located after the sample cell to cut down scattered light before the beam was focused by a second silica lens (50 mm diameter; 140 mm focal length) on the entrance slit of a small monochromator (Spex 1670 Minimate). This monochromator was supplied with a range of interchangeable fixed slits, and the smallest available (0.25 mm) were used for both entrance and exit slits. The corresponding spectral resolution was 1 nm, which was adequate for the present experiments. The light from the monochromator was detected by a photomultiplier (RCA-IP28) powered by a Brandenberg power supply (model 475 R).The output from the photomultiplier was fed into a lock-in detector (EG & G Brookdeal 9503) which also received a reference signal directly from the rotating chopper. Dependent on the type of experiment, the signal from the lock-in detector was either displayed directly on a y-t chart recorder or stored in a digital signal-averager (Biomac 1000). The peroxyl radicals were formed by the photolysis of a suitable precursor in solution in an oxygenated solvent. The photolysis lamp was a 100 W high-pressure mercury lamp (Wotan HBO 100/2), and the light was focused by a silica lens (50 mm diameter; 70 mm focal length) into the plane of an electronic shutter (Uniblitz 225 L).These components were located so that the sample cell was illuminated reasonably uniformly over its full width and symmetrically about the monitoring beam in the vertical direction. For experiments at room temperature a standard fluorimeter cell (Starna type 23; 10 mm pathlength) was used. Usually the cell was only filled to a height of 10 mm so that the whole sample was illuminated. However, checks showed that for the longer-lived radicals the measured decay rate was not affected by diffusion from the non-illuminated regions, even when the cell was completely full (ca. 40 mm height). When necessary the sample was kept saturated with oxygen by slowly bubbling gas into the solution through a hypodermic needle.A slow, well-controlled flow of oxygen was provided by supplying the oxygen from a motor-driven syringe pump. For some of the experiments at elevated temperatures a jacketed cylindrical cell was used (Starna type 65; 10 mm pathlength; 15 mm internal diameter). The cell was heated by passing a stream of hot air through the outer jacket, and temperatures up to 363 K could be attained. The temperature of the sample was measured by a coppersonstantan absorbance units [i.e. log (Zo/I) =J . E, Bennett 1807 RECORDER LOCK-IN REFERENCE SIGNAL I -,- DETECTOR c MONITOR BEAM PHOTOLYSIS BEAM ES ,-#T-, FJ PL c3 Fig. 1. Schematic layout of apparatus. ML, monitor lamp; PL, photolysis lamp; RC, rotating chopper; ID, iris diaphragm; SC, silica cell; SL, silica lens; ES, electronic shutter; PM, photomultiplier; inset: orientation of light beams with jacketed cell.thermocouple immersed in the liquid and was held constant to within f 1 K during an experiment. Alternatively, for temperatures above or below ambient the standard fluorimeter cell could be placed in a silica Dewar flask. The Dewar flask was constructed with three sets of windows, two of which were oriented to permit direct transmission of the monitor beam through the sample. The third set was located orthogonally to the other sets to permit entry of the photolysis beam at right angles to the monitor beam. The temperature inside the Dewar was maintained by a flow of gas which was supplied by a modified variable-temperature accessory for a Varian e.p.r.spectrometer. Experi- ments were run at temperatures up to 398 K with this system, and the temperature could be maintained to within f 1 K during an experiment. For the experiments using the rectangular sample cell the photolysis and monitor beams were oriented at right angles to each other. However, when the jacketed cell was used the photolysis beam had to be brought in to the cell through one of the end windows at an angle of ca. 45" to the monitor beam (fig. 1). By chopping the monitor beam before it passed through the sample and then processing the signal in a lock-in detector it was possible to eliminate the effects of scattered light from the photolysis lamp and of any fluorescence from the sample. However, the time resolution is limited by the chopping rate, and to obtain an accurate time profile of the radical concentration any change should occur in a time which is long compared to the chopping interval (2 10 times longer).Thus in the present work, in which the chopping frequency was 1 kHz, significant changes in the absorbance should occur in times > 10 ms. This response time was sufficiently fast for most of theI808 U. V. Studies of Prop-2-ylperoxyl Radicals in Solution m ,E 10 n 2 . i 5 0 0 d n - 100 150 200 - " 0 50 time after end of photolysis/ms time/ms Fig. 2. Prop-2-ylperoxyl radicals in dodecane at 333 K: (a) growth and decay curve, (b) second-order plot of decay curve. experiments, although it was found that changes occurring in a few ms could be followed with reasonable accuracy.For most of the kinetic experiments it was necessary to average several individual growth and decay curves (84000) to obtain a good signal-to-noise ratio. The signal- to-noise ratio increases approximately as Ni, where N is the number of individual scans. To accumulate a number of separate scans automatically, the electronic shutter was controlled by the signal averager. At a given point in the accumulation sweep a trigger pulse from the signal averager caused the shutter to open for a preset time. The sweep time of the signal averager and the exposure time of the shutter were selected so that both the growth and decay of the radical concentration were recorded (fig. 2). A time delay (pre-selected between 0 ms and 600 s) could be inserted between successive accumulation sweeps to ensure that the radical concentration had decayed completely before the next sweep was commenced.By this procedure1' an accurate baseline was obtained at the beginning of each sweep. When the full set of scans had been accumulated the total signal was read out to a chart recorder for subsequent analysis. For some of the faster reactions a rotating sector (light: dark ratio of 1 : 2) was used in place of the electronic shutter. In this case a trigger pulse from the shutter was used to trigger the accumulation sweep of the signal averager. With this system a delay could not be inserted between the successive sweeps, and consequently an accurate baseline could not be obtained for each sweep. Thus it was necessary to make a duplicate run with the photolysis beam blocked off to obtain the baseline.An interleaved set of four runs, two with and two without photolysis, was carried out to ascertain that the baseline had not altered appreciably during the experiment.J . E. Bennett 1809 Measurement of Radical Concentration The U.V. absorption band of the peroxyl radicals is relatively broad (half-width z 50 nm) and so the Beer-Lambert law can be used to relate the absorbance of the sample to the (1) radical concentration. Thus log(I/IJ = - E d where I,, is the intensity of the incident light, I is the intensity after absorption by the radicals, E is the decadic extinction coefficient, c is the radical concentration and I is the pathlength. The incremental change in absorption (AI) caused by the peroxyl radicals is small, i.e.(2) AI = I,- I 4 I,. Hence Therefore c = AI/I,&l. (3) log (I/I,,) = - AI/Io. For the small changes involved the response of the detection system was linear and thus the concentration was directly proportional to the output signal from the lock-in detector . It is often not easy to measure the extinction coefficient of a transient species, and so the values measured previously for peroxyl radicals in the gas phase have been used in this work. As the shape and position of the absorption bands are virtually identical in the two phases, it is reasonable to assume that the extinction coefficients are also very similar. Generation of Peroxyl Radicals The prop-2-ylperoxyl radicals were generated by the photolysis of 2,4-dimethylpentan- 3-one in an oxygenated solvent [reactions (4) and (5)] : ((CH,),CH),CO -+ 2(CH,),CH + CO (CH,),CH + 0, -+ (CH,),CHOO'. (4) ( 5 ) 2(CH,),cH-+ non-radical products (6) In the liquid phase a significant fraction of the alkyl radicals may undergo self-reaction in the solvent cage: cage instead of reacting with oxygen to give the peroxyl radical.While this reaction does not effect the kinetic measurements it does reduce the efficiency of generation of peroxyl radicals. t-Butylperoxyl radicals were generated in a similar manner from 2,2,4,4- te trame t h ylpen tan-3-one. Materials Commercially available materials were used without further purification. Both ketones were Fluka purum grade (> 97%). The solvents were AnalaR or equivalent grade. t-Butylperoxyl Radical Results Kinetic measurements of the self-reactions of the t-butylperoxyl radical were made at room temperature, so that the results could be compared with existing liquid-phase'q l2 and gas-phase values6 The gas-phase value1, of the extinction coefficient ( E = 1042 dm3 mol-1 cm-l at 250 nm) was used to calculate the concentrations of the t-butylperoxyl radical. The decays of the radical in cyclohexane gave reasonable second-order plots and the overall rate constant for self-reaction was 2k, = (1.7f 1.0) x lo4 dm3 mo1-1 s1 at 293 K.60 FAR 11810 U. V. Studies of Prop-2-ylperoxyl Radicals in Solution Prop-2-ylperoxyl Radical Detailed kinetic measurements for the prop-2-ylperoxyl radical were made in three solvents (cyclohexane, decane and dodecane) over the temperature range 293-396 K.The radical concentrations were calculated using the value of the extinction coefficient which has been measured in the gas phase4t (E = 1145 dm3 mol-1 cm-l at 250 nm). At temperatures below 363 K the radical decays were second-order (fig. 2), showing that the self-reaction was the predominant route. The Arrhenius constants deter- mined from the Arrhenius plot (fig. 3) were Et = 20+3 kJ mol-1 and 2A, = (5 & 2.5) x lo9 dm3 mol-1 s-l. Above 363 K the decays were of mixed first- and second-order kinetics, becoming closely first-order at 396 K. As discussed below, this change is attributed to reaction with the solvent becoming the dominant route at high temperature. Discussion t-Butylperoxyl Radical A very wide range of values has been reported previously for the overall rate constant, 2kt, of the self-reaction of t-butylperoxyl radicals in the liquid phase.In a review article Howard7 has presented a critical assessment of the results and suggested that the ‘best’ value for 2kt is ca. 1 x lo4 dm3 mol-l s-l at 303 K. More recently12 using e.s.r. we have measured the rate constant in a number of different solvents and have obtained a value of 2kt = (1.6 f 0.8) x lo4 dm3 mol-1 s-l at 293 K in cyclohexane. Kinetic measurements have been made in the gas phase using U.V. absorption spectroscopy6 and were complemented by detailed product ana1y~es.l~ Thus these results are probably more reliable than those obtained in the liquid phase. The value of 2kt at 293 K was (1.9 & 0.6) x lo4 dm3 mol-1 s-l. The present measurement [2kt = (1.7 f 1 .O) x lo4 dm3 mol-l s-l] in cyclohexane at 293 K is in good accord with the previous values.The agreement helps to validate the present method and also shows that the overall rate constants for the self-reaction of the t-butylperoxyl radical are very similar in both the gas and liquid phases. Prop-2-ylperoxyl Radical The present results in the liquid phase may be compared directly with those made previously in the gas phase by U.V. absorption spectroscopy4~ and recently in the liquid phase by e.s.r. spectroscopy.lO With e.s.r. it was not possible to follow the self-reactions of the prop-2-ylperoxyl radical above 293 K because of the severe line-broadeningg which occurs as the temperature is increased. Thus the present results complement those made by e.s.r.and extend the temperature range to a much higher temperature (398 K). The rate constants measured by the two methods at the common temperature (293 K) are in good agreement (2kt = 1.6 x lo6 dm3 mol-1 s-l by U.V. and 2kt = 1.1 x lo6 dm3 mol-l s-l by e.s,r.). However, the activation energy obtained from the present results is higher than that measured over the lower-temperature region (173-273 K) by e.s.r. (fig. 4; Et = 20 f 3 kJ mol-1 by U.V. and Et = 10 & 2 kJ mol-l by e.s.r.), It is also higher than those (1 1-13 kJ mol-l) obtained previously for other sec-alkylperoxyl radicals. l5 The results in the gas phase4f5 for the prop-2-ylperoxyl radical cover a similar temperature range as the present work, and the overall rate constants for the self- reactions are virtually identical in both media (fig.4). In the gas phase? 2kt = 1.3 x lo6 dm3 mol-l s-l at 293 K, Et = 18.7k0.5 kJ mol-1 and t Note that in ref. (2), (4) and (5) values for k, are given. Those given here and in fig. 4 are for 2k,, to allow direct comparison with the liquid-phase results.J . E. Bennett 181 1 X I 1 I 1 2.6 2.9 3.0 3.2 3.4 103 K / T Fig. 3. Arrhenius plot of 2k, for prop-2-ylperoxyl radical in A, cyclohexane; x , decane and @, dodecane. 2A, = (2.9 f0.2) x lo9 dm3 mol-l s-l, while in the liquid phase Et = 20 f 3 kJ mol-l and 2A, = ( 5 f 2.5) x lo9 dm3 mol-l s-l. A completely independent measurement of 2kt in the gas phase has been reported recently by Adachi and Basco., Their value of 2kt = 1.6 x lo6 dm3 mol-1 s-l at 293 K is also in very good agreement with the other results.Detailed product studies have been carried out in the gas phase5 and the contributions of the two self-reactions [reactions (7) and (S)] to the overall removal of prop-2-ylperoxyl radicals were determined : (7) (8) As discussed in the paper 2k, = 2(k7 +2k,), and Arrhenius constants for the separate reactions were calculated as 2(CH3),CHOO' + (CH,),CHOH + (CH3),C=0 + 0, 2(CH,),CHOO' + 2(CH,),CHO' + 0,. E7 = 12.0 f 1 .O kJ mol-l; 2A7 = (4.9 f 0.6) x lo7 dm3 mol-1 s-l E, = 21.3+ 1.5 kJ mol-l; 2 4 = (2.8f0.5) x log dm3 mol-l s-l. and The Arrhenius plots for the separate reactions are shown in fig. 4. It is evident that above 293 K the overall kinetics will be dominated by the non-terminating reaction [reaction (S)].On the other hand, in the low-temperature region below 250 K, the terminating reaction [reaction (7)] should predominate. The e.s.r. results are in reasonable agreement with the extrapolated curve for the overall reaction, but do not decrease with temperature as rapidly as predicted from the gas-phase results (fig. 4). Furimsky et aZ.16 have also measured the overall rate constant in the low-temperature region (1 86-229 K) by e.s.r. spectroscopy. Their results fall below a linear extrapolation for the overall reaction in the gas phase, and are considerably lower (by a factor of between 4 and 16) than the more recent liquid-phase measurements. The reason for this 60-21812 7.0 6.0 4.0 3.0 U. V. Studies of Prop-2-ylperoxyl Radicals in Solution 2.0 3.0 4.0 5.0 103 K / T Fig.4. Arrhenius plots for self-reactions of prop-2-ylperoxyl radical : (- - -) gas phase, total 2k, [ref. (4)]; (-.---) gas phase, individual self-reactions 2k7 and 2k8 [ref. ( 5 ) ] ; 0, total 2kt, calculated from extrapolated values of 2k7 and 2k8; x , liquid phase, total 2kt [ref. (16)]. (a) Liquid-phase optical (this work); (b) liquid-phase e.s.r. [ref. (lo)]. discrepancy is not clear unless, contrary to their claim, there is some interference from an unstable intermediate (e.g. trioxide) which leads to the observation of an anomalously low rate constant. In view of this discrepancy in the e.s.r. measurements it would be valuable to extend the present U.V. measurements to lower temperatures. Above 373 K the decay of prop-2-ylperoxyl radicals departs from second-order kinetics, and at 396 K it has become close to first-order.The most probable causes of the first-order behaviour are either the dissociation of the peroxyl radical : (9) (CH,),CHOO' + (CH,),cH + 0, or the propagation reaction with the solvent: (CH,),CHOO* + SHkp. (CH,),CHOOH + S'. (10) The available evidence" indicates that reaction (9) should not occur appreciably at 396 K. On the other hand, the propagation reaction may compete effectively with self-reaction at temperatures above 373 K, and will lead to pseudo-first-order kinetics in the presence of a large excess of solvent. The exact relationship between the propagation rate constant, k,, and the observed rate constant, k,, will depend on the subsequent fate of the solvent radical, S'.If thisJ . E. Bennett 1813 radical (or one formed by further reaction) reacts with another prop-2-ylperoxyl radical then k,[SH] = 0.5k1. If, on the other hand, cross-reaction does not occur then k,,[SH] = k,. In practice the situation will probably lie somewhere between these two limits, i.e. 0.5k1 < k,[SH] < k,. On this basis the measured value of k, = 49 s-l leads to a value of k, in the range (6-12) dm3 mol-1 s-l in dodecane ([SH] = 3.96 mol dm-3) at 396 K. BergeP has obtained a value of k, = 21.5 dm3 mol-1 s-l in hexadecane at 403 K. The activation energylg for this reaction is ca. 69 kJ mol-l, and thus Berger's result leads to a value of k,, = 15 dm3 mol-1 s-l at 396 K, which falls slightly above the upper limit deduced from our measurements.Strictly it is more valid to compare the rate constants per active hydrogen than the overall values. For the n-alkanes reixtion will occur predominantly at the secondary hydrogens, and thus for hexadecane the rate constant, k,/active hydrogen = 0.53 dm3 mol-l s-l and for dodecane, k,/active hydrogen = (0.3- 0.6) dm3 mol-l s-l at 396 K. On this basis the agreement between the two results is satisfactory. I thank the Royal Society-S.E.R.C. for an Industrial Fellowship at the University of York during which this work was carried out. References 1 C. Anastasi, D. J. Waddington and A. Woolley, J. Chem. SOC., Faraday Trans. I , 1983, 79, 505. 2 H. Adachi and N. Basco, Int. J. Chem. Kinet., 1982, 14, 1125. 3 D. A. Parkes, Int. J. Chem. Kinet., 1977, 9, 451. 4 L. J. Kirsch, D. A. Parkes, D. J. Waddington and A. Woolley, J. Chem. SOC., Faraday Trans. I , 1978, 5 L. T. Cowley, D. J. Waddington and A. Woolley, J. Chem. SOC., Faraday Trans. I , 1982,78, 2535. 6 L. J . Kirsch and D. A. Parkes, 5th Znt. Symp. Gas Kinetics, UMIST, Manchester, 1977, paper 37. 7 J. A. Howard, Adv. Free-radical Chem., 1971, 4, 49. 8 J. A. Howard and J. C. Scaiano, in Landolt-Bornstein: Numerical Data and Functional Relationships in Science and Technology, New Series Group II; Atomic and Molecular Physics, ed. K-H. Hellwege and 0. Madelung (Springer-Verlag, Berlin, 1984), vol. 13, part d. 74, 2293. 9 J. E. Bennett and R. Summers, J. Chem. SOC., Faraday Trans. 2, 1973, 69, 1043. 10 J. E. Bennett, G. Brunton, J. R. Lindsay Smith, T. M. F. Salmon and D. J. Waddington, J. Chem. 1 1 J. E. Bennett, G. Brunton, A. R. Forrester and J. D. Fullerton, J. Chem. SOC., Perkin Trans. 2, 1983, 12 J. E. Bennett, F. V. Higgns and R. Summers, unpublished results. 13 C. Anastasi, I. W. M. Smith and D. A. Parkes, J. Chem. SOC., Faraday Trans. I , 1978, 74, 1693. 14 L. J. Kirsch and D. A. Parkes, J. Chem. SOC., Faraday Trans. I , 1981,77, 293. 15 J. A. Howard and J. E. Bennett, Can. J. Chem., 1972,50,2374. 16 E. Furimsky, J. A. Howard and J. Selwyn, Can. J. Chem., 1980,58, 677. 17 T. A. B. M. Bolsman and D. M. Brouwer, Red. Trav. Chim. Pays-Bas, 1978,97, 320. 18 H. Berger, Chem. SOC. Annual Meeting, Manchester, 1972, abstract 2.4a. 19 S. Korcek, J. H. B. Chenier, J. A. Howard and K. U. Ingold, Can. J. Chem., 1972,50, 2285. SOC., Faraday Trans. I , 1987, 83, in press. 1477. Paper 6/ 1728; Received 26th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301805

出版商:RSC    

年代:1987

数据来源: RSC