摘要:

J. Chem. SOC., Faraday Trans. 1 , 1988, 84(6), 1835-1845 The Behaviour of Encapsulated Non-polar Gases in Cs,Na-A Zeolites Dan Fraenkel," Baruch Ittah and Moshe Levy Department of Materials Research, The Weizmann Institute of Science, Rehovot, Israel Zeolitic encapsulation of non-polar permanent gases in Cs,Na-A is compared and discussed. H,, Ar, 0,, N, and CH, are shown to be trapped by a cages whose 8R windows are blocked by at least 2.3Cs' ions per pseudo-cell (zcs 2.3). Below this value, trapping is restricted to the j? (sodalite) cages. As shown for the large gases, encapsulation in u and p cages can be distinguished during temperature-programmed diffusion (t .p.di.) which releases the trapped gas and a transition region of combined a- and /?-cage encapsulation is clearly observed.The activation parameters for diffusion, E and Do, for these two types of cages were derived both by the graphical method (g.m.) and the peak-shape method (p.s.m.) (see preceding paper). The simpler and more convenient p.s.m. is shown to be generally applicable and not less reliable than the g.m. for pure t.p.di. peaks; in fact, it should be the method of choice at high-E-value cases. j?-cage decapsulation and a-cage decapsulation above a degree of exchange, z,, of 3.5 exhibit constant E while at the region 2.3 < zcs < 3.5 E for the cy: cages increases sharply with zcs. The pre-exponential factor, Do, depends predominantly on the type of window (i.e. size of ring, blocking cation) and decreases as the window becomes more widely open. Zeolitic encapsulation is a method by which small non-polar molecules are introduced into zeolitic cages whose windows have an effective opening that is formally smaller than the critical molecular size (kinetic diameter, 0); the molecules are forced into the cages under high external pressures at elevated temperatures and become trapped in these cages upon cooling the system prior to depressurisation. Zeolitic encapsulation of hydrogen gas in Cs,Na-A has been reported previously.'y2 In this paper we report the encapsulation of several non-polar permanent gases in the same zeolite system.Preliminary communications on this study3, emphasized the ability to distinguish between a-cage and /3-cage encapsulation experimentally when studying gases larger than hydrogen. The main aim of this paper is to present and discuss results obtained with different gases in a comparative manner and draw general conclusions relating to the diffusional behaviour of these gases in zeolitic cages and channels as reflected by the corresponding activation parameters and their dependence on the degree of Cs+ exchange, zcs.In deriving the activation parameters from experimentally measured quantities we make use of the peak-shape method (p.s.m.) developed and discussed in the preceding paper,5 as well as the previously employed graphical method (g.m.). Experimental Encapsulation and temperature-programmed diffusion (t.p.di.) experiments were performed as described earlier.4 Results given herein are based only on integral (i.e. U us. T ) t.p.di. curves.Half-height widths (co) were taken as the distance between upper and lower half-maximal tangents of the integral curve. Tables 1-6 summarize 18351836 Encapsulation of Non-polar Gases in Zeolites Table 1. Encapsulation of H, (series 1). 0 1.8 2.2 2.6 3.0 0.085 0.132 0.240 0.077 0.092 0.133 0.142 0.247 0.248 0.083 0.130 0.252 0.080 0.082 0.125 0.135 0.232 0.260 0.082 0.083 0.132 0.143 0.240 0.248 413 418 467 437 443 452 443 508 47 1 44 1 452 474 420 429 43 5 443 458 463 442 456 459 447 50 1 480 ~ 42 40 43 39 39 39 39 41 46 45 44 45 73 75 78 71 87 78 86 73 81 87 89 88 120.2 129.3 150.1 144.9 148.8 155. I 148.9 186.3 142.7 127.9 137.4 147.8 71.5 72.6 71.8 81.8 71.4 81.4 67.2 84.3 77.0 68.0 83.5 77.5 - 1.2 x lo1, 1.3 x lo1' I .8 x 1013 1.6 x 1014 3.4 x 1014 1.1 x 1015 1.4 x 1013 4.2 x 1013 2.7 x 105 2.4 x 105 2.1 x 105 1.1 x 105 2.6 x 104 4.5 x 104 4.3 x 105 2.4 x 105 5.3 x 10l4 3.6 x 10l6 9.7 x 10" 8.6 x 10l2 2.7 x lo6 1.6 x lo6 1.7 x lo6 3.0 x lo5 a Encapsulation temperature, Tc,, was 573 K ; encapsulation pressure, P,, was in the range 12.0-13.4MPa.* From p.s.m., see text. experimental conditions and results of selected encapsulation-decapsulation runs. Experiments with H, were performed to allow a comparison between the performance of a newly constructed encapsulation apparatus and that of the one used before.' Two series of experiments were carried out with H,, referred to hereinafter as series I (table 1) and series I1 (table 2), before and after the addition of improved temperature detection and heating rate control, respectively.Thus, series TI is a priori credited as the better of the two. Nevertheless, series I is of sufficient value, we believe, to be presented also. Results and Discussion Encapsulation Capacity Maximal encapsulation capacity values, V,,, for the various gases investigated are given in tables 2-6 and shown to be practically independent of the rate of decapsulation. Average values normalized to 10 MPa encapsulation pressure are plotted against zcs in fig. 1 . As seen, there are two distinct regions of encapsulation regardless of the type of gas, that of the p cages occurring at 0 < zcs < 2.3, in which low and constant capacity is observed, and that of the a cages which occurs above zcs z 2.3 and is characterized by a much larger capacity that strongly decreases toward higher zcs values. The transition from p- to a-cage encapsulation at zcs zz 2.3 has already been explained satisfactorily by the percolation theory.2 The a-cage capacity drop can to some extent be interpreted as before2 by considering the parallel decrease in the available voidD.Fraenkel, B. lttah and M. Levy Table 2. Encapsulation of H, (series 1Qu 1837 zcs cage P/K s-' VH2' T,/K w/K E/kJ mo1-1 (Do/ri)c/s-l 0 2.2 2.6 3.0 0.082 0.123 0.207 0.08 1 0.107 0.138 0.167 0.198 0.205 0.049 0.049 0.064 0.082 0.085 0.128 0.128 0.178 0.182 0.222 0.223 0.067 0.083 0.130 0.131 0.163 0.223 3.3 3.0 3.1 3.4 2.9 3.4 2.9 2.9 3.4 8.4 8.4 8.0 8.4 8.0 8.6 8.4 8.4 8.4 8.7 8.6 6.5 6.5 5.9 16.5 16.5 15.7 422 428 436 429 433 43 5 438 440 439 400 400 404 408 407 418 416 424 424 428 426 426 432 440 440 445 452 45 42 46 45 39 48 43 48 53 75 80 91 78 89 88 79 87 91 83 96 83 82 93 89 91 95 117.1 129.1 122.3 121.1 142.3 116.7 132.1 119.4 107.6 63.1 59.2 53.1 63.2 55.1 58.8 64.8 61.2 58.5 65.3 56.0 64.7 67.4 61.6 64.4 64.4 63.6 2.1 x loll 6.3 x 10l2 7.3 x 10" 3.7 x 10'l 1.1 x lo1' 8.3 x 10l2 2.2 x 10" 6.3 x 1013 8.7 x 109 3.6 x 104 1.0 x 104 1.6 x 103 4.ox 104 3.4 x 103 9.8 x 103 7.0 x 104 2.1 x 104 9.7 x 103 7.8 x 104 5.1 x 103 2.1 x 104 4.4 x 104 8.8 x 103 2.0 x 104 2.0 x 104 1.6 x 104 a Encapsulation temperature, T,, was 573 K; encapsulation pressure, P,, was 13.3 MPa.(s.t.p.) per gram of dehydrated zeolite. cm3 From p.s.m., see text. volume (per gram of dehydrated zeolite) as more Na+ ions are substituted by the bulkier and heavier Cs+ ions.However, even then a-cage capacity is far from being constant at this region. As mentioned previously,* some deviation to lower value can be expected at high zcs because of the increasing difficulty of obtaining true equilibrium values when the blocking efficiency of Cs+ reaches its maximum. This increasing efficiency is reflected by a dramatic rise in the t.p.di. TM Decapsulation Kinetics The kinetics of the diffusion-controlled decapsulation of the investigated gases from the respective zeolites was followed using the temperature-programmed heating mode (linear schedule). The t.p.di. equation on which the graphical method (g.m.) is based relates the activation parameters to T, and p:5 The corresponding Arrhenius-like plots of selected cases based on data listed in tables 2 4 are presented in fig.2. All lines obtained are indeed straight, as expected. For hydrogen, lines of different zcs lie all at a reciprocal maximal temperature between1838 Encapsulation of Non-polar Gases in Zeolites Table 3. /?-Cage encapsulation of O,a 0 0.077 0.077 0.123 0.130 0.198 0.202 1.8 0.082 0.117 0.175 2.2" 0.084 0.130 0.195 2.3 0.08 1 0.148 0.205 2.4 0.213 2.6f 0.202 2.7 583 580 595 2.5 598 3.4d 608 612 2.3 588 2.6 599 3.2d 608 2.8 589 2.7 60 1 3.2d 609 2.3 603 3.3d 619 2.3 630 3.8d 624 3.9 61 5 - - - 79 79 75 67 80 78 90 79 83 87 93 88 90 92 95 127.3 126.0 139.7 158.0 136.8 142.1 113.7 134.4 131.8 118.0 115.0 124.8 119.6 123.3 123.7 8.5 x 107 7.4 x 107 1.05 x 109 1.2 x 109 4.4 x 1O1O 4.8 x lo8 3.8 x lo6 2 .7 ~ lo8 1.5 x lo8 9.4 x lo6 4.5 x lo6 6.8 x lo6 3 . 7 ~ 107 1.35 x 107 1.3 x 107 a Encapsulation temperature, T, was 693 K unless mentioned otherwise ; encapsulation pressure, P,, was in the range 11.3-12.3 MPa. cm3 (s.t.p.) per gram of dehydrated zeolite; values relate to dehydrated 6R,, windows unless otherwise specified. From both dehydrated and rehydrated 6R,, window^.^ From p.s.m., see text. T, = 703 K. f T, = 623 K. 2.2 x lop3 and 2.5 x lop3 K-'. For the other gases lines for a-cage encapsulation group at a lower reciprocal TM, while for P-cage encapsulation (zCs < 2.3) they are well separated and drift to much lower Tii ranges. Thus, diffusion rates of the various gases studied, except hydrogen, are similar for a-cage decapsulation but differ markedly for P-cage decapsulation, and there is clear indication that in the latter case a simple relationship exists between these rates and critical molecular size (see below).Activation parameters based on the peak-shape method (p.s.m.) were calculated for each heating rate. This was done using the following equations which relate these parameters to TM and ~ 0 : ~ D, = ?-!! exp (3.6 T'/c;o). ri n2to (3) Tables 1-6 list the results obtained. As seen, good agreement is normally obtained between parameters of the same system at different /3 values and in most cases the deviation from mean values is < 10% for E and less than an order of magnitude for Do/?-:. Mean p.s.m. values are presented and compared with the corresponding g.m. values in table 7. The two H, series compare well in the p.s.m.but p.s.m. and g.m. results compare satisfactorily only for series 11. For 0, and N,, values of E for P-cage decapsulation are substantially smaller in the g.m., apparently owing to experimental difficulties encountered at high temperatures and the inherent inaccuracy of the method at high E (large slope of the Arrhenius line). We thus tend to accept the p.s.m. results of P-cage decapsulation as being more reliable. Excellent agreement between the two methods is obtained for Ar at zcs = 2.6 but not at zcs = 3.0, and it is believed that theD. Fraenkel, B. Ittah and M . Levy 1839 Table 4. a-Cage encapsulation of O,a zcs P/K s-' Vo,' T,/K co/K E/kJ mol-l (Do/r;y/s-l 2.2d 2.3 2.4 2.6" 3.0" 3.8 4.5 5.4 0.195 0.148 0.095 0.155 0.213 0.060 0.130 0.202 0.078 0.136 0.208 0.082 0.138 0.208 0.075 0.134 0.197 0.083 0.140 0.203 0.75 397 2.8 373 32.3 394 32.3 405 32.3 413 15.1 450 18.0 466 19.3 476 15.8 500 15.2 511 15.1 523 11.0 620 11.4 636 11.6 644 8.5 691 7.7 710 8.8 719 6.8 690 6.8 708 6.9 717 76 77 91 82 86 93 101 113 105 137 136 142 132 132 132 133 145 139 60.5 63.0 55.5 73.1 74.7 72.1 73.3 68.4 77.1 83.3 88.0 86.5 107.1 113.0 115.9 106.0 102.3 109.5 4.0 x 104 8.5 x 104 7.4x 103 7.0 x 104 1.1 x 105 5.5 x 104 1.1 x 104 3.6 x 103 3.1 x 104 1.8 x 103 5.2 x 103 4.6 x 103 2.2 x 104 1.2 x 105 2.05 x 104 1.05 x 104 4.3 x 104 6.6 x lo4 a Encapsulation temperature, T,, was 693 K unless mentioned otherwise ; encapsulation pressure, P,, was in the range 11.3-12.2 MPa.'cm3 (s.t.p.) per gram of dehydrated zeolite.CFrom p.s.m., see text. T, = 703 K. T, = 623 K. Table 5. Encapsulation of N," cage P/K s-' Er/kJ mol-1 (Do/ri)c/s-l 0 2.2 2.6 3 .O P 0.085 0.140 0.203 a: 0.083 0.133 0.197 0.077 0.129 0.192 a 0.085 0.133 0.210 01 0.067 0.1 16 0.203 D 0.067 0.1 16 0.203 P 2.8 657 80 2.4 670 68 2.2 677 80 7.2 390 7.5 397 - 7.3 402 - 2.7 657 73 2.7 670 85 2.5 681 73 19.4 468 90 19.6 476 87 19.1 486 90 14.8 505 89 14.8 518 94 14.8 530 99 - - 3.0 2.9 2.9 - - - - - 159.7 195.4 169.6 175.0 156.3 188.0 72.0 77.1 77.7 84.8 84.5 84.0 1.9 x 109 1.3 x lo1' 1.1 x 1O'O 3.1 x 1O1O 8.1 x 10' 2.6 x 10" 3.2 x 104 1.4 x 105 1.6 x 105 1.4 x 105 1.3 x 105 1.2x 105 ~~ ~ a Encapsulation temperature, c, was 693 K for z,, = 0 and 2.2, and 633 K for zcs = 2.6 and 3.0; encapsulation pressure, P,, was 1 1.8 MPa.From p.s.m., see text. cm3 (s.t.p.) per gram of dehydrated zeolite.1840 Encapsulation of Non-polar Gases in Zeolites Table 6. Encapsulation of Ar and CH,a ZCS gas T,/K cage P/K s-' Vgasb T'/K w/K Er/kJ mo1-1 (Do/ri)c/s-l 0 1.8 2.6 3.0 4.5 CH, 823 a P P CH, 803 01 Ar 623 a CH, 693 a Ar 623 a CH, 693 a CH4 693 a 0.208 0.210 0.235 0.225 0.243 0.083 0.127 0.258 0.125 0.233 0.083 0.133 0.267 0.08 1 0.134 0.232 0.233 0.228 4.5 472 487 6.0 742 10.9 427 3.4 834 22.3 459 21.8 466 21.5 488 22.5 487 21.3 497 12.8 505 12.7 525 13.1 559 17.3 513 17.3 526 18.5 540 538 9.2 749 - - 129 182 92 118 81 107 106 104 100 108 82 82 88 98 101 122 111 - 51.1 38.6 177.1 45.7 254.2 58.3 60.6 67.8 70.2 67.7 92.1 98.4 105.1 79.5 81.1 70.7 77.2 - 2.2 x lo2 4.5 2.2 x lo2 9.5 x 10l2 2.6 x 109 1.0 x 103 2.3 x 103 1.4 x 104 1.3 x 104 8.6 x lo3 1.1 x lo6 3.7 x lo6 6 .6 ~ lo6 3 . 2 ~ 104 4.7 x 104 4.0 x 103 2.0 x 104 - ~ a Encapsulation pressure, P,, was 11.8 MPa for Ar and between 9.6 and 9.9 MPa for CH,. ' cm3 (s.t.p.) per gram of dehydrated zeolite. ' From p.s.m., see text. lower g.m. values are less accurate. The disagreement between the g.m. and p.s.m. in the case of 0, at z,, = 3.0 and 3.8 can be explained by peak overlapping producing p.s.m. values that are too low (see below). Activation Energy Evalues for Cs,Na-A(O,), as obtained from both the g.m. and p.s.m., are plotted in fig. 3 against zcs. In the P-cage region E appears constant, and for the reasons mentioned above, the upper p.s.m. line is considered to be more reliable.In the a-cage region, E initially increases, then at zcs x 4 levels off with a value close to that of P-cage decapsulation. The agreement between the two methods is excellent except at zcs = 3.0 and 3.8. Here the a-cage decapsulation peaks are apparently contaminated with (residual) P-cage decapsulation peaks, and therefore the measured half-height widths, m, are larger than the true ones and produce erroneously smaller E values. Although both a- and P-cage decapsulation occur also below zcs = 3.0, the corresponding t.p.di. peaks there are easily measured. The fact that excellent agreement between the two methods for a-cage encapsulation is restored at zcs = 4.5 and above indicates that P-cage encapsulation practically no longer exists above zcs x 4, as we indeed predicted previously in view of an encapsulation capacity analysis of the Cs,Na-A(H,) system.2 Thus in general there seems to be a good agreement between E values derived from the g.m.and those obtained using the p.s.m. This is further confirmed in fig. 4, where E values from different sources and as obtained by the two methods are plotted against zcs for the H, case. There is a nice resemblance between the behaviour of Cs,Na-A(H,) and Cs,Na-A(02), suggesting that the nature of activated diffusion in the systems studied does not depend on the choice of gas. The constant value for P-cage decapsulation is attributed to the fact that cation exchange in the low zcs region isD. Fraenkel, B. Ittah and M. Levy Table 7. List of activation parametersa 1841 0 1.8 2.2 2.3 2.4 2.6 3 .O 3.8 4.5 5.4 H, (series I) (series 11) 0 2 N2 CH4 0 2 CH4 0 2 N2 0 2 0 2 H, (series I) H, (series I) (series 11) H, (series I) (series 11) Ar 0 2 NZ CH4 H2 (series I) (series 11) Al- 0 2 N2 CH4 0 2 0 2 0 2 P U p 1 2 0 ' P p 108 P P P P 1 1 7 p 102+10 P 89f 10 P 118+6 - 9 4 f 8 - - a 7 9 f 8 P 117+4 P 102i-5 a 5 3 f 3 a 41 (?) a - 71 01 - 67 a 74+ 1 a 9 0 f 8 a 82 a - 64 a 43 (?) p 1 B g l - a a 9 5 j i 2 a 84+8 a 8 2 f 4 a 112f9 a 115f7 a 115f8 133 & 17 123&6 138 f 12 175f20 177 155f 12 126 + 13 254 138 & 10 123 & 15 119f5 173& 17 122f3 60+4 7 5 f 7 6 0 f 6 62+6 7 3 f 1 75 f 4 69+ 1 76+9 6 4 f 3 98+7 7 3 f 5 84.4 & 0.4 7 7 f 6 8 6 f 3 112f5 106_f4 - - 8 x lo1' 3 x lo1' 2 x 104 5 x 105 2 x 109 4 x 104 - - - 7 x 1O'O 2 x 105 1 x 107 4 x 105 1 x 105 3 x 103 12 (?) 3 x 1 0 5 1 x 104 7 x 104 2 x 107 3 x 105 I x 104 3 (?) 2 x lo6 1 x 105 5 x 104 5 x 105 7 x 104 8 x 104 3 x lo6 - I x 1013 1 x 10l2 7 x los 3 x 1O1O 3 x loR 5 x 1014 5 x 107 1 x 1013 I x 107 1 x 107 3 x 104 4 x 105 I x 104 3 x 103 8 x 104 9 x 104 1 x 104 2 x 105 2 x 104 1 x 104 2 x 104 3 x 103 2 x 104 7 x 10l2 6 x 10" - 2 x 1O1O - 3 x lo6 1.3 x 105 6 x lo4 ~ a Best values underlined, see text.values. 93 and 146. Mean values, see tables 1-6. Average of two experimental Table 8. Activation parameters for the Na-A/CH, system method E/kJ mol-' Do/cm2 s-' source sorption 24.2-34.3 6 x 1OP8-3.6 x Yucel and encapsulation ca. 45" 2 x 10-6a.b this work proton spin-spin 21 1.6 x 10-7 Allonneau and relaxation Volino' Rut hven lo a Average value based on data from table 6.' Assuming ro = 2 x cm.1842 Encapsulation of Non-polar Gases in Zeolites 0 1 2 3 4 5 6 Fig. 1. Vgas [in cm3 (s.t.p.) per g of dehydrated zeolite] as a function of zCs. For symbols see fig. 2. ZCS 1.4 1.6 6.5 6.3 6.1 - ( u ) 509 6.9 I ' . I ' ' ' ' 6.7 6.5 6.3 6.1 5.9 6.1 , I 1 I /, 1 5.9 1.4 1.6 1.8 2.0 2.2 2 . 4 103 K I T , Fig. 2. Arrhenius-like plots for t.p.di. of H, series I1 (O), 0, (O), N, (H), Ar (A) and CH, (a) in Cs,Na-A. (a) P-cage, zcs = 0; (b) P-cage, z,, = 2.2; (c) cr-cage, zcs = 2.6; ( d ) a-cage, zcs = 3.0.D. Fraenkei, B. Ittah and M. Levy I843 Fig. 3. 3 -.. I z 2 \ 43 - I p - cage +I+ d - cage ! ! 50 - I I I I I 1 0 1 2 3 4 5 6 zcs Dependence of activation energy on the degree of Cs' exchange 15C 100 " - 1 z 2 G 50 I I I 4 v-\ I I I for Cs,Na-A(0,) 0 1 2 3 4 *cs Fig.4. Dependence of activation energy on the degree of CS' exchange for Cs,Na-A(H,). [O, Series I (p.s.m.); A, series I1 (p.s.m.); 0, from ref. (1) (g.m.); V, other resultsg (p.s.m.).] restricted to the 8R windows, and a possible long-range interaction between Cs' at site I1 and its electronic environment and the 6R,, window is not expected to be very influential on the blocking efficiency of this window. The initial increase in E in the a-cage region has been previously associated with a small shift of 11-site Cs+ from an off-plane position to a substantially more effectively blocking on-plane position, as zcs is varied from ca. 2.3 to ca. 4.l 68 1 F A R I1844 300 200 - I d 2 2 \ 4 100 Encapsulation of Non-polar Gases in Zeolites 2.5 3.0 3.5 4 .O olA Fig.5. Dependence of the activation energy on the kinetic diameter at zcs “N 2 (e) and 3.0 (0). As mentioned above, except for hydrogen there appears to be a clear correlation between critical molecular size of the non-polar gas and E for 8 cages. This is shown in fig. 5. E increases apparently linearly with the kinetic diameter of the gas.6 Similar behaviour was reported for the diffusion of various monatomic and diatomic gases in K-A, where the diffusion process is associated with the a-cages (i.e. passages through 8R, window~).~,~ A shallow linear increase in E is indicated in fig. 5 for a-cages having 8R,, windows. From the slopes of the straight lines the dependence of E on 0 appears to be ca. 10-fold stronger for diffusion through 6R,, as compared to 8R,6*7 and 8 R,, windows. Pre-exponential Factor As shown in table 7, within 1-2 orders of magnitude error (which, admittedly, is to be expected when the error in activation energy is ca.10-20%), different zeolite windows have discrete Do values. Thus, to a first approximation, for a certain decapsulation mode (a-cage or 8-cage) Do does not depend on the degree of Cs+ exchange. It also appears almost independent of the type of diffusing gas molecule. The only influential factor is apparently the type of window, i.e. (1) the number of ring oxygens and perhaps also their spatial arrangement (circular or elliptic ring) and (2) the size of blocking cation ; its exact location (on-plane, off-plane) is of less (or no) importance if the above explanation for the initial rise in E in the a-cage region is correct, since there seems to be no parallel increase in Do.These conclusions concerning the relationship between a zeolite window and Do transcend the particular physical situation discussed in this paper, since the activation parameters for diffusion should not depend on the type and details of the system studied or the method by which they are measured. In other words, we can compare our values with those obtained in ordinary constant-temperature sorption studies or from spectroscopic measurements, as reported in the literature. A list of sorption-based results given by Barrers for a variety of gases in Ca-A (free 8R), Na-A (8R,,) and K-A (8R,)D. Fraenkel, B.Ittah and M. Levy 1845 also suggests that within 1-2 orders of magnitude error Do is constant for a given window and does not depend on the gas. The average Do value for K-A is in accord with that obtained in decapsulation of hydrogen from K-A,' and the average Do value for Na-A agrees nicely with the value we obtained for a-cage decapsulation of methane from Na-A. [Unlike the other gases studied by us, methane is sufficiently large to be trapped within the a-cages of Na-A, although the blocking efficiency of the 8R,, windows with respect to this gas is rather poor (table 6).] Table 8 compares E and Do values for the Na-A(CH,) system as obtained by different groups and techniques. The present results are shown to be in reasonable agreement with other sorption data as summarized by Yucel and Ruthven,lo as well as with recent results obtained by Allonneau and Volino from n.m.r.-derived self-diffusivity." Work is in progress in our group on the correlation between Do and E and between Do and the type of zeolite window for various geolite-gas systems.This includes a search for a suitable theoretical basis for explaining experimentally found behaviours. Glossary number of CS' cations per pseudo-cell of zeolite A a zeolite 1-membered oxygen window occupied by an exchangeable cation of metal Me (e.g. 8Rc,) encapsulation temperature encapsulation pressure decapsulation heating rate (t.p.di.) temperature of maximal rate of release of encapsulated gas during t.p.di. half-height width of differential t.p.di. curve (i.e. dU/dT us. T ) volume at s.t.p. of total amount of thermally released encapsulated gas decapsulation time (t.p.di.) decapsulation temperature (t. p. di .) fraction of (encapsulated) gas diffused out at time t (temperature T ) gas constant activation energy for diffusion radius of zeolite particle (assuming uniformity) pre-exponential factor of the diffusion coefficient kinetic diameter This work was supported by the Fund for Basic Research administrated by The Israel Academy of Sciences and Humanities. References 1 D. Fraenkel, J. Chem. SOC., Faraday Trans. I , 1981, 77, 2029. 2 D. Fraenkel, J. Chem. SOC., Faraday Trans. I , 1981, 77, 2041. 3 D. Fraenkel, B. Ittah and M. Levy, J. Chem. SOC., Chem. Commun., 1984, 1389. 4 D. Fraenkel, B. Ittah and M. Levy, Stud, Surf. Sci. Catal., 1985, 24, 459. 5 D. Fraenkel and A. Levy, J. Chem. SOC., Faraday Trans. I , 1988, 84, 1817. 6 D. W. Breck, Zeolite Molecular Sieves (Wiley, New York, 1974). 7 P. L. Walker Jr, L. G. Austin and S. P. Nandi, in Chemistry and Physics of Carbon (Dekker, New York, 1966), vol. 2, pp. 257-371. 8 R. M. Barrer, Zeolites and Clay Minerals (Academic Press, London, 1978), p. 292. 9 D. Fraenkel, unpublished results. 10 H. Yucel and D. M. Ruthven, J. Chem. SOC., Faraday Trans. I , 1980, 76, 60. 1 1 J. M. Allonneau and F. Volino, Zeolites, 1986, 6, 431. Paper 71632; Received 9th April, 1987 61-2

ISSN:0300-9599    

DOI:10.1039/F19888401835

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1988, 84(6), 1847-1852 Solvatochromatic Indicator Study of Silicalite and Zeolite ZSM-5 G. Paul Handreck and Thomas D. Smith* Chemistry Department, Monash University, Clayton, Victoria 3168, Australia Reflectance spectrophotometric measurements on the solvatochromic indicator 4-nitroanisole adsorbed from cyclohexane solution by silicalite and ZSM-5 zeolites containing various amounts of aluminium ion have been made and used to calculate the parameter n*, which is a single-value measure of polarity and polarizability properties for each solid material. Adsorption isotherms concerned with the uptake of the dye by the solid materials have been determined. The values of ll* for silicate are less than half of those determined from similar measurements on amorphous silica.Compared with silicalite, the values of H* for the zeolites ZSM-5 are somewhat greater, showing a levelling-off with increasing content of aluminium ion, in keeping with the conservation of hydrophobicity with increasing acidity of the zeolite surface. The catalytic properties of the zeolite ZSM-5, which has gained technical importance from its role in the conversion of methanol to ga~oline,l-~ stems from a combination of its microporous structure and its acidity, which arises from the incorporation of aluminium ions into the neutral silicalite framework.8 An added feature of the zeolite ZSM-5 which is of importance in its catalytic behaviour is its hydrophobicity. The hydrophobic properties of silica surfaces' which include zeolites'* play a major role in their ability to selectively abstract organic materials from aqueous and non-aqueous solutions.''.l2 Silicalite, which is generally regarded as a hydrophobic molecular sieve, l3 has been shown to exert adsorbate-adsorbent dispersion interactions which are influential in the uptake of non-polar mate~ia1s.l~ The hydrophobic character of zeolite ZSM-5 has been shown to increase linearly with increase of the SiO,/Al,O, ~ a t i 0 . l ~ ' ~ ~ Earlier work17 on the polarity of silica surfaces using the Kamlet-Taft solvatochromic indicator method18-22 has shown that the dipolarity/polarizability (n*) is the paramount contributor to the polar properties of the surface, with hydrogen-bond donation (a) making a minor contribution. Measurements of this sort were of assistance in understanding the role played by the amorphous silica surface on the course of several reaction^.^^-^' There is no comparable information on the polarity of zeolite ZSM-5.The present study deals with the application of the solvatochromic indicator method to the determination of changes to the polarity of the surface of zeolite ZSM-5 which may occur as a result of the incorporation of different amounts of aluminium ion into its structure. Experiment a1 X-Ray diffraction measurements of microcrystalline zeolite materials were made using a Siemans D500 diffractometer. A Nikon VM5 optical microscope was used for the scrutiny of zeolite samples. Spectrophotometric measurements of dye solutions were recorded on a Pye Unicam SP8-100 spectrophotometer using 1 cm silica cells, while reflectance measurements on zeolite samples were made using the same instrument fitted with a barium sulphate-coated spheroid.Chemical analysis of the aluminium and sodium contents of the zeolites was accomplished using a Varian-Techtron atomic 18471848 Sohatochromic Indicator Study Table 1. Chemical composition of silicalite and ZSM-5 zeolite samples ~~~~ sample anhydrous unit cell Si02/Al,03 silicalite 1 Na0.01A10.01s195.990192 19 198 AIZSM-5 2 Na0.02A11.20S194.800192 158 AlZSM-5 3 Na0.02A11.36si94.640192 139 AlZSM-5 4 Na0.02A12.20s193.800192 85 AlZSM-5 5 NaO. 02A15. OSS19O. 92'1 92 36 absorption spectrophotometer. All the chemical reagents were obtained from com- mercial sources. Zeolite Preparations Silicalite was prepared by heating a thoroughly stirred reaction mixture of fumed silica (16.3 g, 0.272 mol), sodium hydroxide (7.6 g, 0.19 mol), recrystallized tetra-n-pro- pylammonium bromide (tpa) (7.05 g, 0.0265 mol), concentrated sulphuric acid (4.9 g, 0.0499 mol) and water (229 g, 12.7 mol) with a molar composition of 9.7 tpa: 34.9 Na,O: 100 SO,: 4679 H,0.309 31 The heating (448 K) was carried out in a stainless-steel autoclave for 48 h with continuous stirring (150 rev min-').The product, separated by filtration, was washed thoroughly with distilled water and dried (383 K, 2 h). To prepare zeolite ZSM-5 an aqueous solution (40cm3) of hydrated aluminium sulphate (7.45 g, 0.01 12 mol) containing concentrated sulphuric acid (4.09 g, 0.042 mol) was added with vigorous stirring to an aqueous solution (202 cm3) containing recrystallized tetra-n-propylammonium bromide (8.8 g, 0.033 1 mol), sodium silicate (SiO, 20.2 g, 0.336 mol; A1,0, 0.21 g, 2.06 x mol; Na,O 6.4 g, 0.014 mol) and sodium hydroxide (1.3 g, 0.0325 mol).The resulting gel (molar composition 2.9 tpa: 10.7 Na,O: 1 .O A1,03 : 30 SiO, : 1437 H,O) was transferred to a stainless-steel autoclave and heated (448 K, 24 h) with continuous stirring (150 rev min-'). The thoroughly washed product isolated by filtration was dried (383 K, 2 h). Other samples of zeolite ZSM-5 containing different amounts of aluminium ions were prepared by the same procedure with the amount of sulphuric acid present in the reaction mixture being adjusted to ensure the pH of the gel was in the range 10-1 1 as the amount of hydrated aluminium sulphate varied.Quality control of the zeolites prepared was ensured by recording their X.r.d. patterns. All the &spacings and relative intensities of peaks were characteristic of zeolite ZSM-5. Furthermore, material from each preparation was scrutinised under the optical microscope for any sign of unreacted gel material, which appears as a brown opaque solid in polarized light and does not exhibit birefringence under crossed polarized light.32 All the material examined was entirely crystalline. The zeolites were converted to their protonic form by refluxing in dilute hydrochloric acid (0.5 mol dm-3, 150 cm3 g-l zeolite) for 16 h. After filtration and thorough washing, the solid materials were oven dried (383 K, 2 h).Chemical analyses were carried out on the protonic forms of H-ZSM-5 using atomic absorption spectrophotometry for sodium and aluminium content and gravimetric analyses for silica after acid digestion of the zeolites. The anhydrous unit cell compositions for the zeolites used in this study are summarized by table 1.G. P. Handreck and T. D. Smith 1849 Solvatochromic Indicator Measurements A stock solution (0.026 mol dmP3) of 4-nitroanisole (ethanol-recrystallized m.p. 53 "C, literature33 54 "C; A, = 292 nm, literature 291.5 nm21) in cyclohexane (redistilled 81 T) was used to prepare dilute solutions in the range 1.67-3.36 x lop5 mol. In preliminary experiments it was shown that uptake of 4-nitroanisole by the dried zeolite samples (473 K, 2 h) from cyclohexane solutions was complete in 2 h.In subsequent measurements of dye uptake the amount of dye remaining in solution was determined spectrophotometrically using calibration curves (292 nm). Reflectance spectral meas- urements were made on zeolite samples after separation from the dye solution, rinsing with cyclohexane and drying (293 K, 20 mmHg,? 20 min). Three spectra were recorded on three different samples (10 mg) of zeolite from each equilibrium run. The peak positions were determined by measurement of the mid-point at half height. The position of the peak (vimax) recorded as wavelength was converted to frequency units and used to solve the regression equation vimax = vio + S i l l * for n*. Results and Discussion Adsorption isotherms for the uptake of 4-nitroanisole from cyclohexane by silicalite and zeolites ZSM-5 containing various amounts of aluminium ion were established, a typical result being shown by fig.1. Samples of silicalite number 1 and zeolites ZSM-5 2-5 show a steady increase in dye adsorption as the dye concentration in solution increases, with little tendency to reach monolayer adsorption capacity. In detail in each case the uptake of dye is characterized by a curved isotherm which is similar in nature to the L4 class of isotherm described by G i l e ~ , ~ ~ who attributed the second plateau to the reorientation of the dye molecules on the surface. The adsorption isotherms were not pursued beyond these concentration ranges since the amount of dye adsorbed by the surface of the zeolites did not show any effect on the dye absorption band in the region of visible light as measured by the reflectance spectra.The solvatochromic comparison method1* involves the use of indicator molecules, which are not hydrogen-bond donors and solvents which are not hydrogen-bond donors or acceptors, in the construction of a scale of dipolarity and or polarizability (ll*). To calculate an individual II* value, n,*, the shift in the absorption maxima of a particular solvator chromic indicator, in the material being studied, is used to solve the regression equation for that particular indicator as follows: vim,, = vio+SiII* where Si is the susceptibility of vim,, to changing values of ll* and vio is the peak position recorded in cyclohexane where n* + 0.00. In the case of 4-nitroanisole, vi,, = 34.17 and Si = 2.410.18 A typical example of the reflectance spectra for 4-nitroanisole adsorbed on silicalite is depicted by fig.2, where increased adsorption of the dye leaves the position of the reflectance spectrum largely unchanged. Fig. 3 shows the mean n* values calculated from the reflectance measurements of samples of the silicalite and zeolites ZSM-5, each with various amounts of 4-nitroanisole adsorbed from cyclohexane solution plotted against the aluminium content of each zeolite. Compared with samples of silica studied previously [n* = 2.68, ref. (1 7)] the value of n* is much reduced. This reduction in the dipolarity-polarizability is in keeping with the coricept that the surface is homopolar in nature and points to a quantitative measure of the hydrophobic nature of silicalite with its capacity to retain non-polar materials within its porous structure.The catalytic properties of the zeolite ZSM-5 stem from a combination of its extended surface area shape-selective channel structure like that possessed by silicalite with the surface acidity of the outside and channel surfaces t 1 mmHg = 101 325/760 Pa.1850 4.0- - I om “0 3.0- E 0 2 g 2.0- 1 ..3 c, f Sohatochromic Indicator Study Y 0 I I I I 200 300 400 500 wavelength/ nm Fig. 2. Typical reflectance spectra of silicalite with the following quantities of 4-nitroanisole adsorbed: (a) 0.8 wt %, (b) 1.44 wt %, (c) 2.48 wt %, (d) 3.52 wt %, (e) 4.00 wt %, cf) 4.32 wt %. Each reflectance spectrum was recorded over the absorbance range 0.6-1.4 and was offset from the previous spectrum for the purposes of clarity.G.P . Handreck and T. D. Smith 1851 n* I I 1 1 I I 1 2 3 4 5 6 0.91 A1 per unit cell Fig. 3. The ll* values for silicalite and ZSM-5 zeolites plotted against their respective A1 contents. The vertical lines represent the range of values obtained for different amounts of 4-nitroanisole adsorbed by the samples. brought about as a result of incorporation of aluminium into the framework structure. While the amount of aluminium ion present in the zeolite ZSM-5, even at the higher level of composition used in catalytic studies, is quite small the value of II* is sensitive to the lower range of aluminium ion content in the zeolite ZSM-5 as shown by fig. 3, which depicts a measurable rise in the value of II* for quite small contents of aluminium in the zeolite ZSM-5 with a levelling-off for increasing amounts of aluminium ion present.Thus, while the higher content of aluminium ion will endow the ZSM-5 zeolite with increased acidity after an initial rise in dipolarity-polarizability, this property remains largely unchanged and is a measure of the conservation of hydrophobicity; a fortunate combination of properties for catalytic conversion of hydrocarbons. A study of the hygroscopic nature of H-ZSM-5 suggests that the exterior of the zeolite is hydrophilic and readily forms a bound aqueous layer, whereas the uptake of water by silicalites is much less.35 Presumably the uptake of water occurs as a result of the hydration of acid sites, arising from the incorporation of aluminium into the zeolite, on the outside surface.The present results are in keeping with an increased level of polarity of the outside surface of zeolite ZSM-5 compared with silicalite, but the increased level of the surface as a whole rather than individual sites is such that the surface would be more fittingly described as being largely hydrophobic. References 1 S. L. Meisel, J. P. McCulloch, C. H. Lechthaler and P. R. Weisz, Chemtech., 1976, 86. 2 C. D. Chang and A. J. Silvestri, J. Cutal., 1977, 47, 249. 3 C. D. Chang, J. C. W. Kuo, W. H. Lang, S. M. Jacob, J. J. Wise and A. J. Silvestri, Ind. Eng. Chem. 4 S. Yuchak, S. E. Voltz and J. P. Warner, Ind. Eng. Chem. Process Des. Dev., 1979, 18, 527. 5 S. L. Meisel, Phiios. Trans. R. SOC. London, Ser.A , 1981, 300, 157. 6 A. L. Titchener, Chem. Ind., 1982, 841. 7 C. D. Chang, Catal. Rev. Sci. Eng., 1983, 25, 1. 8 N. Y. Chen and W. E. Garwood, Catul. Rev. Sci. Eng., 1986, 28, 185. 9 G. J. Young, J. Colloid Sci., 1958, 13, 67. J O N. Y. Chen, J. Phys. Chem., 1976, 80, 60. 11 N. B. Milestone and D. M. Bibby, J. Chem. Tech. Biotechnoi., 1981, 31, 732. 12 G. M. W. Schultz-Sibbel, D. T. Gjerde, C. D. Chriswell, J. S. Fritz and W. E. Coleman, Talunta, 1982, 13 E. M. Flanigen, J. M. Bennett, R. W. Grose, J. P. Cohen, R. L. Patton, R. M. Kirchner and J. V. Process Des. Dev., 1978, 17, 255. 29, 447. Smith, Nature (London), 1978, 271, 512.1852 Solvatochromic Indicator Study 14 H. Thamm, H. Stach and W. Fiebig, Zeolites, 1983, 3, 95. 15 D. H. Olsen, W. 0.Haag and R. M. Lago, J. Catal., 1980, 61, 390. 16 H. Nakamoto and H. Takahashi, Zeolites, 1982, 2, 67. 17 S. M. Lindley, G. C . Flowers and J. E. Leffler, J. Org. Chem., 1985, 50, 607. 18 M. J. Kamlet, J. L. Abboud and R. W. Taft, J . Am. Chem. SOC., 1977, 99, 6027. 19 M. J. Kamlet, T. H. Hall, J. Boykin and R. W. Taft, J . Org. Chem., 1979, 44, 2599. 20 M. J. Kamlet, J. L. Abboud, M. H, Abraham and R. W. Taft, J . Org. Chem., 1983, 48, 2877. 21 M. J. Kamlet and R. W. Taft, J. Am. Chem. Soc., 1976, 98, 377. 22 R. W. Taft and M. J. Kamlet, J. Am. Chem. Soc., 1976, 98, 2886. 23 J. E. Leffler and J. J. Zupancic, J . Am. Chem. Soc., 1980, 102, 259. 24 J. E. Leffler and J. T. Barbas, J. Am. Chem. Suc., 1981, 103, 7768. 25 J. E. Leffler, J. T. Barbas and G. C . Flowers, J. Org. Chem., 1982, 47, 2286. 26 M. Hudlicky, J . Org. Chem., 1974, 39, 3460. 27 Z. Cohen, E. Keinen, Y . Mazur and T. H. Varkony, J . Org. Chem., 1975, 40, 2141. 28 D. Avnir, P. de Mayo, and I. Oho, J . Chem. Soc., Chem. Commun., 1978, 1109. 29 P. de Mayo, A. Nakamura, P. W. K. Tsang and S. K. Wong, J . Am. Chem. SOC., 1982, 104, 6824. 30 R. J. Argauer and G. R. Landolt, U S . Patent 3702886, 1972. 31 N. Y. Chen, S . N. Miale and N. Y. Reagan, U S . Patent 4112056, 1978. 32 R. von Ballmods, in The "0-Exchange Method in Zeolite Chemistry : Synthesis, Characterization and 33 A. Groggin and T. Striton, Ind. Eng. Chem., 1936, 28, 1051. 34 C. Giles, J. Chem. SOC., 1960, 3973. 35 S. G. Hill and D. Seddon, Zeolites, 1985, 5, 173. Dealumination of High Silica Zeolites (Salle and Shuerlander, Frankfurt, 198 1). Paper 7/650; Received 13th April, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401847

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I, 1988, 84(6), 1853-1861 Prediction of Excess Volumes of Ternary Liquid Mixtures Jata D. Pandey," Rajiv K. Shukla, Arun K. Shukla and Rishi D. Rai Department of Chemistry, University of Allahabad, Allahabad-21 I 002, India Considering a ternary liquid mixture to be made up of three binary mixtures, Flory's statistical theory has been extended to obtain a relation for the excess volume. In the light of above extension, the excess volumes of 15 different ternary liquid mixtures have been predicted at 25 and 30 "C over a wide range of composition. There is excellent agreement between the experimental and theoretical excess volumes, both in magnitude and sign. For many practical purposes it is necessary to predict the properties of a multicomponent liquid mixture from the properties of pure components and from the data of binary systems.A sophisticated treatment of the liquid mixture is given by the refined version of the cell-model theory of Prigogine,' which requires various parameters for computational purposes and has poor agreement with experiment. To our knowledge, few investigations have been carried out on multicomponent liquid m i x t ~ r e s . ~ - ~ Marsh" outlined the recent developments in the techniques for measuring the excess functions of liquid mixtures as well as properties of aliphatic hydrocarbons and complex organic mixtures in terms of various equations of state, interaction parameters and chemical equilibria. Rastogil' independently reviewed the thermodynamic properties of ternary mixtures and their molecular interactions.Lark et a1.l2 developed a new batch dilatometer and determined the ternary volume effect. A critical review of the various properties of multicomponent systems has been given by Rowlinson and Swinton.' In most of the theories, the properties of the multicomponent system are determined with the help of the properties of their binary s o l ~ t i o n s , ~ ~ * ' ~ but only Flory's statistical theoryl5, l6 can be successfully used to predict the properties of the multicomponent system from those of pure components. Most of the work on excess volumes for binary systems has been carried out by McGlashan and co~orkers,'~ Patterson and co- workers,lg, l9 Benson and coworker^,^^-^^ Street and coworkers2* and Marsh and coworkers.25-29The theoretical prediction of the excess volumes of multicomponent liquid systems from Flory theory has not been done so far except some work on excess heat, activity ~oefficient,~' surface tension3' and total and preferential in ternary systems. After an exhaustive consultation of literature it appears to us that nobody has predicted the excess volumes of ternary liquid mixtures using Flory's statistical theory. In this paper 15 different types of ternary liquid mixtures have been examined using Flory theory. Theoretical The reduced equation of state derived from the resulting partition function is given by (1) P V - P3 1 - - ~ - - F p 3 - 1 V F 18531854 Excess Volume of Ternary Liquid Mixtures where p, P and f a r e the reduced pressure, volume and temperature, respectively. These are given by - P 2PV*, p = - - -- P* sq * v = u/u* = v/ v * . The reduced equation of state at zero pressure is T * = - = - T T P4I3 p 3 - 1 and ( 5 ) where V = rv is the molar volume and a is the coefficient of thermal expansion at P = 0.Thus the reduced and characteristic volumes and temperatures can be calculated using the experimental value of a and eqn (5) and (6). While predicting the surface tension of ternary liquid mixtures, Pandey and Pant3, extended Flory theory and defined an element (or segment) as an arbitrary chosen portion of the molecule, and again defined the segment as well as site fractions of molecule by the relations ry2 = X2/[X, + X3( c/ C) + XI( v / C)I (7) ry3 = X3/CX3 + X2( el + XI( v/ c>1 (8) v/i = [1-v'2-V31 (9) 8, = w2/[w2 + w3( w c)1/3 + v,( w w / 3 1 (10) O3 = ry3/[w3 + w2( C / W3 + v/,( C/ V)1/31 ( 1 1) el = ( 1 -o,-e,) (12) where v/,, ry2, ry, and 8,, e2, 8, are the segment and site fractions of components 1-3.In the light of the above relations, the excess volumes of ternary liquid mixture can be obtained using the equation where is the ideal reduced volume given by & = v1 E-k v2 <+ v3 c* Substitution of eqn (14) into eqn ( 1 3 ) gives the excess volume of the ternary liquid system which can be written as VE = (Xl v + x , c + x 3 G)[P-(ry1 c+v, g+ry, <)] (15) where VE is the excess molar volume and p i s the reduced volume of the ternary liquid mixture. Assuming the volume reduction parameters of the ternary mixtures to be linear in mole fraction, one obtains - (16) V V = Xl V + X z G + X 3 Jq where V is the molar volume of the liquid mixture, given by M , X , + M , xz + M , x3 V = P, where p, is the density of the mixture.J .D. Pandey, R. K. Shukla, A . K. Shukla and R. D. Rai 1855 Excess volumes and reduced excess volumes are correlated with the equation as mentioned below. This equation can also be-used to predict the excess volume from the knowledge of the reduced excess volume, I/”, i.e. From eqn (9, and are correlated with each other. reduced temperature of the mixture given by Pandey and Pant31 have formulated the equations for the characteristic pressure and where P* is the characteristic pressure of the ternary system and can be expressed as p* = “+!‘I p;” -k w 2 pz* vy,p,* - (v/i 0 2 xi2 -k w 2 O3 x23 + Y / 3 6ix3i)l (21) where X12, X23 and adopting the familiar Berthelot relationship, vij = (vii qij)lj2.are the interaction parameters obtained from eqn (22)-(24) by The application of the above relation makes it possible to predict the excess volume, using vE as a second approach of Flory theory to evaluate the volume of mixing in the multicomponent liquid mixture. Results and Discussion The excess volumes of ternary liquid mixtures can be studied by Flory theory in two different ways, first, by calculating the excess volume (V”) directly from the characteristic and reduced volumes and the segment fraction using thermal expansion coefficient (a) only and secondly by calculating VE using eqn (18) in terms of pE, considering all the interaction parameters on the basis of theoretical and experimental studies.It has been found that the strength of interaction between two components is weakened by the third component, showing the nearly ideal behaviour of a ternary system. The extent of this weakening of the interaction can be assessed roughly by the expression where AE12,1iq denotes the interaction between 1 and 2 in the presence of component 3, AE,,, vacuum denotes the interaction between 1 and 2 in vacuum and Py = P3p3 No/M3. The validity of eqn (25) for a number of ternary systems has shown that the percentage decrease in E12, vacuum is quite significant. The change in the excess volume in the ternary liquid system with respect to the binary system can be expressed by where AVE is the change in excess volume due to the interaction in the ternary system, showing the weakening of interactions.Here, we illustrate the validity of the extended form of Flory theory with the help of the first method described above, using the1856 Excess Volume of Ternary Liquid Mixtures Table 1. Parameters of the pure component liquids at 298.15 K molar molecular density volume a/ 1 OP3 v v* components weight /g cm-3 /cm3 mol-' K-' /cm3m01-l /cm3 mol-' n- hexadecane 226.4493 0.7707 293.7886 0.8677 1.2199 240.8231 n-tetradecane 198.3952 0.7603 260.9364 0.9422 1.2357 21 1.1576 4-methyl- 112.1730 0.91 14 123.0709 0.8793 1.2224 100.6778 n-hexane 86.1784 0.6551 131.5340 1.3896 1.3224 99.4591 carbon 153.8231 1.5840 97.1068 1.2171 1.2905 75.2422 benzene 78.1 147 0.8735 89.422 1 1.2265 1.2923 69.1921 2-bromobutane 137.0203 1.2530 109.3538 1.1519 1.2780 85.5633 cyclohexanone tetrachloride Table 2.Parameters of the pure component liquids at 303.15 K molar molecular density volume a/ 1 0-3 v V* components weight /g cm-3 /cm3 mol-' K-l /cm3 mo1-1 /cm3 mol-' cyclohexane 84.1625 0.7692 109.4043 1.2250 1.2959 84.4185 benzene 78.1 147 0.8684 89.9483 1.2098 1.2930 69.5623 tet rachloro- 165.8343 1.6064 103.2335 1.0246 1.2573 82.1049 carbon 153.823 1 1.5748 97.6778 1.2266 1.2962 75.3523 toluene 92.1418 0.8577 107.4289 1.0804 1.2675 84.7523 chloroform 119.3781 1.4706 81.1764 1.2820 1.3067 62.1189 ethylene tetrachloride experimental values of 15 ternary liquid mixtures in two sets. The experimental values of 10 ternary liquid mixtures at 25 "C are taken from the work of Heric and Brewer33 as a first set, and are presented in table 1, which contains the values of the reduced and characteristic parameters of the pure components, while table 3 contains their experimental and theoretical excess volumes.Scrutiny of table 3 reveals that the theoretical values of excess volumes for ternary systems have been found to be very close to the experimental values for systems 1, 3, 5, 8 and 10. In these systems the theoretical excess volumes also agree well in both magnitude and sign with the experimental excess volumes. The systems 2 , 4 , 6 , 7 and 9 show good agreement with the experimental excess volumes. Some deviations in magnitude (but not in sign) have been observed in these systems. The experimental values of excess volume for the second set, which contains five ternary liquid mixtures, are taken from the measurements of Rastogi et a1.34-36 at 30 "C.The reduced and characteristic parameters for the pure components are recorded in table 2, while their theoretical and experimental excess volumes are given in table 4. In most cases no deviation was found between theoretical and experimental excess volumes. As far as interaction is concerned, it need not be explained further, since an explanation has been given in the paper dealing with the experimental measurements. These results validate us of the extended form of Flory's statistical theory. We are extremely grateful to the U.G.C. (India) for financial assistance.J. D. Pandey, R. K. Shukla, A . K. Shukla and R. D.Rai Table 3. Predicted excess volumes of different ternary systems at 298.15 K 1857 0.4489 0.2523 0.2646 0.1551 0.1484 0.0625 0.0670 0.0704 0.0723 0.3318 0.2072 0.1063 0.1056 0.1062 0.5390 0.4512 0.3679 0.2829 0.2355 0.2185 0.2247 0.1596 0.1640 0.1594 0.1 176 0.1 151 0.1 162 0.4597 0.2886 0.2868 0.1999 0.1798 0.0897 0.0844 0.0790 0.0764 0.3904 0.2345 0.1344 0.1201 0.1137 1. n-hexadecane-carbon tetrachloride-benzene 0.2860 0.85 0.2780 0.94 0.4474 0.83 0.4100 0.71 0.6257 0.59 0.1453 0.42 0.3242 0.44 0.5477 0.39 0.7648 0.34 0.28 17 0.92 0.3805 0.83 0.2240 0.61 0.4362 0.56 0.6374 0.49 0.2386 0.02 0.2676 - 0.03 0.2532 -0.18 0.3659 -0.1 1 0.1579 - 0.3 1 0.3936 -0.19 0.5961 +0.10 0.2203 - 0.34 0.4218 - 0.22 0.6152 - 0.06 0.1552 -0.28 0.3404 - 0.27 0.5470 -0.19 0.2910 0.10 0.3465 0.03 0.4576 0.24 0.2 122 -0.21 0.4855 0.14 0.1824 -0.11 0.4337 0.03 0.6400 0.18 0.8002 0.23 0.3706 0.18 0.4277 0.08 0.269 1 -0.15 0.5150 0.12 0.7014 0.25 2. n-tetradecane-4-methylcyclohexanone-n-hexane 3.n-hexadecane-carbon tetrachloride-n-hexane 0.84 0.94 0.83 0.71 0.59 0.42 0.44 0.39 0.34 0.92 0.82 0.60 0.55 0.49 0.0 1 - 0.04 - 0.20 -0.14 -0.33 -0.21 + 0.07 - 0.36 -0.24 - 0.09 - 0.28 - 0.29 - 0.23 0.10 0.03 0.23 -0.21 0.14 -0.11 0.03 0.18 0.24 0.19 0.08 -0.14. 0.1 1 0.251858 Excess Volume of Ternary Liquid Mixtures Table 3. (cont.) 0.5434 0.3271 0.2610 0.1914 0.1890 0.1338 0.0919 0.0864 0.0854 0.7166 0.5342 0.5718 0.3570 0.4172 0.1919 0.1976 0.2544 0.6331 0.4643 0.3215 0.241 5 0.508 1 0.3 167 0.3481 0.2006 0.1996 0.1460 0.1563 0.0953 0.1018 0.0635 0.0889 0.1071 0.1439 0.1631 0.0976 0.21 14 0.3055 0.1232 0.2344 0.3556 0.1 197 0.2265 0.3794 0.5 152 4.n- hexadecane-4-methylcyclo hexanone-n- hexane 0.2260 - 0.08 0.2734 -0.12 0.3656 -0.11 0.4028 - 0.22 0.6346 0.15 0.2296 -0.31 0.1563 -0.32 0.3793 -0.30 0.5480 -0.23 5. carbon tetrachloride-n- hexane-benzene 0.1 195 0.14 0.1531 0.2 1 0.2210 0.25 0.1237 0.22 0.4249 0.24 0.2414 0.34 0.4133 0.35 0.5706 0.27 0.1482 0.23 0.1965 0.20 0.4557 0.28 0.1371 0.2 1 6. n-tetradecane-n-hexane-2-bromobutane 0.2082 0.08 0.2528 0.20 0.3469 0.05 0.1524 0.23 0.3585 0.10 0.1960 0.20 0.5875 - 0.09 0.1450 0.16 0.7088 -0.11 7. n-hexadecane-n-tetradecane-2- bromobutane 0.0791 0.33 0.0966 0.39 0.1579 0.42 0.1067 0.5 1 0.1740 0.52 0.33 15 0.55 0.2210 0.56 0.1082 0.6 1 0.402 1 0.58 0,2714 0.51 0.1414 0.58 0.5444 0.47 0.3893 0.52 0.2805 0.46 0.1200 0.53 - 0.09 -0.13 -0.13 - 0.24 0.10 - 0.33 - 0.33 -0.32 - 0.26 0.13 0.20 0.25 0.21 0.23 0.34 0.34 0.26 0.23 0.20 0.28 0.21 0.06 0.16 0.03 0.18 0.06 0.14 -0.11 0.10 -0.13 0.26 0.32 0.42 0.43 0.46 0.51 0.52 0.57 0.54 0.47 0.54 0.44 0.47 0.41 0.5 1J.D. Pandey, R. K. Shukla, A . K , Shukla and R. D. Rai Table 3. (cont.) 1859 0.0650 0.1028 0.1046 0.1521 0.1650 0.0878 0.2118 0.3274 0.1297 0.2648 0.3995 0.1 105 0.2522 0.5182 0.5022 0.3808 0.4 167 0.3 187 0.2188 0.2503 0.3010 0.1506 0.1832 0.2044 0.0953 0.1 180 0.1 177 0.1460 0.53 15 0.2957 0.2691 0.1850 0.1873 0.1557 0.0798 0.0757 0.0707 0.0686 0.2238 0.1281 0.1221 0.1074 8. n- hexadecane-n- tetradecane-n- hexane 0.0796 - 0.39 0.1119 - 0.49 0.1763 -0.55 0.1 151 - 0.53 0.1831 -0.56 0.365 1 - 0.56 0.2258 -0.55 0.0947 -0.61 0.4208 - 0.50 0.2924 -0.54 0.1514 - 0.49 0.5396 - 0.40 0.3998 - 0.40 0.1243 - 0.42 9.n-hexadecane-n-tetradecane-4-methylcyclohexanone 0.1640 0.67 0.1570 0.56 0.2428 0.67 0.2235 0.63 0.0959 0.62 0.2438 0.67 0.4590 0.46 0.1 188 0.59 0.276 1 0.64 0.4558 0.54 0.0897 0.45 0.2057 0.55 0.3866 0.6 1 0.5968 0.41 10. n- hexadecane-benzene-n- hexane 0.2836 0.45 0.3586 0.29 0.494 1 0.62 0.2552 0.11 0.4706 0.47 0.7030 0.75 0.2586 0.20 0.4724 0.42 0.6666 0.50 0.8205 0.49 0.4657 0.48 0.3120 0.18 0.5406 0.50 0.7202 0.56 VtEheor -0.38 - 0.48 - 0.54 -0.52 - 0.56 -0.55 -0.56 -0.61 -0.50 -0.54 - 0.48 -0.38 - 0.39 - 0.4 1 0.65 0.53 0.65 0.60 0.58 0.63 0.45 0.53 0.60 0.50 0.38 0.50 0.58 0.40 0.45 0.29 0.61 0.12 0.47 0.74 0.20 0.42 0.50 0.49 0.48 0.19 0.50 0.561860 Excess Volume of Ternary Liquid Mixtures Table 4.Predicted excess volumes of different ternary systems at 303.15 K 0.0962 0.2586 0.3016 0.3382 0.5706 0.6495 0.5012 0.771 8 0.5022 0.5605 0.5564 0.3 156 0.2901 0.2732 0.3571 0.1610 0.2523 0.2939 0.3468 0.3970 0.4893 0.523 1 0.5789 0.6 103 0.6487 0.1575 0.1738 0.2097 0.2309 0.23 14 0.3227 0.048 1 0.0486 0.206 1 0.2964 0.328 1 0.3745 0.3857 0.4375 0.5388 1 1. tetrachloroethylene-benzene-cyclohexane 0.7438 0.392 0.3936 0.580 0.6214 0.338 0.3724 0.542 0.2925 0.369 0.2630 0.295 0.0878 0.497 0.0895 0.313 0.0900 0.496 0.1 198 0.463 0.1236 0.460 0.2369 0.346 0.2948 0.304 0.3547 0.271 0.45 17 0.249 0.6575 0.295 0.5559 0.312 0.4461 0.382 0.4478 0.321 0.3736 0.361 0.2627 0.376 0.2568 0.341 0.1817 0.367 0.I 548 0.363 0.0543 0.435 12. cyclohexanesarbon tetrachloride-toluene 13. chloroform-toluene-cyclohexane 0.391 0.579 0.337 0.542 0.368 0.294 0.497 0.313 0.496 0.463 0.460 0.346 0.304 0.27 1 0.249 0.295 0.312 0.382 0.321 0.361 0.376 0.341 0.367 0.363 0.435 14. tetrachloroethylene-carbon tetrachloride-cyclohexane 0.6242 0.155 0.154 0.3756 0.27 1 0.27 1 0.3833 0.286 0.286 0.4336 0.248 0.247 0.2643 0.322 0.321 0.4520 0.208 0.207 1 5. tetrachloroet hy lene-toluene-cyclohexane 0.6356 0.457 0.5664 0.512 0.5986 0.340 0.5033 0.362 0.554 I 0.257 0.5201 0.248 0.3733 0.398 0.5292 0.150 0.4218 0.150 0.457 0.512 0.340 0.362 0.257 0.248 0.398 0.150 0.150J. D. Pandey, R. K. Shukla, A . K. Shukla and R. D. Rai 1861 References 1 I. Prigogine, V.Mathot and A. Bellmans, The Molecular Theory of Solutions (North Holland, 2 D. R. Massengil and R. C. Miller, J. Chem. Thermodyn., 1973, 5, 207. 3 W. Brostow and J. S. Sochnski, J. Muter. Sci., 1975, 10, 2134. 4 D. A. Armitage and C. G. Osborne, IUPAC Conference on Chemical Thermodynamics, Section 5 5 D. Naumann, IUPAC Conference on Chemical Thermodynamics, Section 5 (Pergamon, London 6 A. Nissema, P. Kokkonen and H. Arrola, IUPAC Conference on Chemical Thermodynamics, Section 5 7 M. Mickeleit, and R. Lacmann, IUPAC Conference on Chemical Thermodynamics, Section 5 8 R. S . Rowlinson and F. L. Swinton, Liquids and Liquid Mixtures (Butterworths, London, 1982). 9 E. L. Heric and J. G. Brewer, J. Chem. Eng. Data, 1970, 15, 379. Amsterdam, 1957). (Pergamon, London, 1982).1982). (Pergamon, London 1982). (Pergamon, London 1982). 10 K. N. Marsh, Pure Appl. Chem., 1983, 55, 467. 11 R. R. Rastogi, J . Sci. Ind. Res., 1980, 39, 480. 12 B. S. Lark, S . Singh and S. Kaur, J . Ind. Chem. SOC., 1986, L X I I I , 499. 13 G. L. Bertrand, W. E. Acree Jr and T. E. Bruchfield, J . Solution Chem., 1983, 12, 327. 14 W. E. Acree Jr and G. L. Bertrand, J . Solution Chem., 1983, 12, 755. 15 P. J. Flory, J. Am. Chem. SOC., 1965, 87, 1838. 16 A. Abe and P. J . Flory, J. Am. Chem. Soc., 1965, 87, 1833. 17 1. Cibutka, M. B. Erving and McGlashan, J. Chem. Thermodyn., 1983, 15, 49. 18 H. T. Van and D. Patterson, J . Solution Chem., 1982, 11, 793. 19 M. Costas and D. Patterson, J . Solution Chem., 1982, 11, 307. 20 B. Luo, S . E. M. Harnam and G. C. Benson, J. Chem. Thermodyn., 1986, 18, 1043. 21 D. Zhang and G. C. Benson, J. Chem. Thermodyn., 1986, 18, 697. 22 F. Kumara, C. J. Halpin and G. C. Benson, J. Chem. Thermodyn., 1983, 15, 503. 23 A. J. Treszczauowicz and G. C. Benson, J . Chem. Thermodyn., 1980, 12, 173. 24 S. F. Barrereiros, Jorge C. J. Clado and W. B. Street, J . Chem. SOC., Faraday Trans. I , 1983, 79, 25 M. B. Ewing and K. N. Marsh, J. Chem. Thermodyn., 1978, 10, 267. 26 R. P. Tomlins and K. N. Marsh, J . Chem. Thermodyn., 1977, 9, 651. 27 M. B. Ewing and K. N. Marsh, J. Chem. Thermodyn., 1977, 9, 357. 28 M. B. Ewing and K. N. Marsh, J. Chem. Thermodyn., 1977, 9, 863. 29 K. N. Marsh and P. P. Organ, J. Chem. Thermodyn., 1985, 17, 835. 30 W. Brostow and J. S. Sochanski, J . Muter. Sci., 1975, 10, 2134. 31 J. D. Pandey and N. Pant, J . Am. Chem. SOC., 1982, 104, 3299. 32 L. C. Cesteros, C. Strazielle and I. Katime, J . Chem. SOC., Faraday Trans. I , 1986, 82, 1321. 33 E. L. Heric and J. G. Brewer, J. Chem. Eng. Data, 1969, 14, 55. 34 R. P. Rastogi, J. Nath and S. S. Das, J. Chem. Eng. Data, 1977, 22, 249. 35 R. P. Rastogi, J. Nath, B. Singh and S . S . Das, Ind. J . Chem., 1977, 15A, 1012. 36 R. P. Rastogi, J. Nath and M. L. Yadava, J . Chem. Thermodyn., 1975, 6, 1179. 1869. Paper 7/864; Received 18th May, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401853

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J’. Chem. Soc., Faraday Trans. I, 1988, 84(6), 1863-1870 Characterization of Ni-exchanged Montmorillonites by X-Ray Photoelectron Spectroscopy Michael Stocker,* Klaus- Joachim Jens and Trygve Riis Department of Hydrocarbon Process Chemistry, Centre for Industrial Research, P.O. Box 124, Blindern, N-0314 Oslo 3, Norway Jostein K. Grepstad National Laboratory for Surface Studies (NALOS), Norwegian Institute of Technology, N-7034 Trondheirn-NTH, Norway X.P.S. binding energies and Auger-electron kinetic energies of some Ni- exchanged montmorillonites have been measured. The results are discussed in terms of the bonding states of the elements and their relation to the montmorillonite structure. It is shown that the preparation procedure influences the chemical and structural environment of both the interlamellar and the constituent clay elements.In recent years the silicate minerals known as clays have become of industrial importance as catalysts for a wide range of chemical reactions and processes. Clays are components of many soils and sediments and are often found in nature as large, mineralogically pure deposits.’ They have been applied as heterogeneous catalysts mainly because of their special properties, e.g. cation-exchange capability, layer dimensions, shape-selectivity, small particle size and unusual intercalation capability. These properties explain why clays afford an appreciable surface area for adsorption of organic molecules.’ A new class of clays, the so-called ‘pillared clays’, allows the reaction of large molecules, e.g. the conversion of heavy oil fractions into gasoline.A second type of clay minerals incorporate metal complexes between the silicate layers, presenting a new development in the intercalation chemistry of clay catalysts. The applicability of montmorillonite clays has been demonstrated by Swift and coworkers2g and Sohn et ~ l . ~ among others, who used nickel-substituted synthetic mica-montmorillonite (Ni-SMM) and Ni-montmorillonite (Ni-MM) in the olig- omerization of olefins. The two catalysts differ in that SMM has non-expansive layers compared with the common montmorillonite. The ability of ion-exchanged mont- morillonites to catalyse numerous chemical reactions because of their acidity and shape- selectivity has led to the extensive use of these catalysts in the last few decades, including some processes on an industrial scale.5* In spite of the technological importance of montmorillonites, the fundamental physical properties of these materials are still poorly known.Some investigations have been carried out in order to determine the elemental composition and the structure of clays;7 however, relatively little is known about the bonding states of the elements and their relation to the clay structure. X-Ray photoelectron spectroscopy (X.P.S. or ESCA) is a favoured technique for probing the bonding states of the elements in solids. Chemical shifts in the core-level binding energies provide direct information about charge transfer, oxidation state and coordination of the sample constituents.8 In X.P.S., irradiation of a solid under high vacuum by soft X-rays results in the production of photoelectrons from some core levels o f all the elements present (except for H and He).9 X.P.S.data for silicate minerals have been published previously;8,10-21 however, so far only a few results have been reported for Ni-exchanged clays. 22 18631864 X.P.S. of Ni-exchanged Montmorillonites In this paper, X.P.S. photoelectron binding energies and Auger-electron kinetic energies of some Ni-exchanged montmorillonites (including some Ni/Al,O, and Ni/SiO, systems for comparison) are reported. The purpose of this investigation was to obtain information on the bonding states of Ni and the layer elements (Si, Al, Mg) and how they relate to the chemical environment and structure of the clays.Nickel was chosen as the interlamellar cation because of its great importance as an active catalyst in industrial processes. Experimental X.P.S. Measurements The X.p. spectra were recorded in a Vacuum Generators ESCALAB Mk 11, using A1 K, (1486.6 eV) and Mg K, (1 253.6 eV) X-rays. Measured photoelectron binding energies and Auger-electron kinetic energies were corrected for sample charging by referring all energies to the C 1s binding energy of adventitious surface carbon (Eb = 284.6 eV). This carbon contamination derives from hydrocarbon species in the residual background of the u.h.v. surroundings. The background pressure was kept in the lO-'mbar range during measurement. The spectra were acquired by multichannel analysis, controlled from a Digital PDP- 1 1 /23 computer, in order to obtain good signal-to-noise ratios for the weak-intensity lines.The samples were ground to powders with an agate mortar and fixed on to suitable sample holders of 10 mm diameter using stainless-steel press equipment. Sample Synthesis Samples 1 and 2 were commercially available montmorillonites from Fluka AG (K10) and Sudchemie AG, respectively. Sample 3 was prepared by ion exchange of sample 2 with aqueous Ni(NO,), at 100 "C for 4 days under a pressure of 10 bar. The sample was then dried at 120 "C. Sample 4 was prepared by calcination of sample 3 at 450 "C for 13 h in an N, atmosphere. To prepare sample 5, sample 2 was extracted with water for 24 h, followed by a drying step at 120 "C and ion exchange with aqueous Ni(NO,), at 100 "C for 4 days and under a pressure of 10 bar. The sample was then dried at 120 "C.To prepare sample 6, sample 2 was suspended in an aqueous Ni(NO,), and Ni was precipitated at pH 8.6. The sample was filtered, washed and dried at 120 "C, followed by calcination at 450 "C for 13 h in an N, atmosphere. For sample 7, sample 2 was extracted with water for 24 h, followed by a drying step at 120 "C and ion exchange with aqueous Ni(NO,), at 70 "C for 1 day under atmospheric pressure. The sample was then dried at 120 "C. For sample 8, sample 2 was extracted with water for 24 h, followed by a drying step at 110 "C and ion exchange with aqueous Ni(NO,), at room temperature for 1 day under atmospheric pressure. The sample was then dried at 110 "C. Sample 9 (Ni/A1203): alumina was suspended in an aqueous solution containing Ni(NO,),. The mixture was stirred at 70 "C for 8 h, then dried at 100°C for 24h and calcined at 450 "C for 6 h in an N, atmosphere.Sample 10 (Ni/SiO,-Al,O,) was prepared as described in the patent.23 Sample 11 (Ni/SiO,) was synthesized according to the procedure described in the patent.,* Results and Discussion Photoelectron binding energies and Auger-electron kinetic energies of all the samples investigated are summarized in table 1, along with some reference data. Chemical analysis, loss of weight and specific surface area data are collected in table 2. Fig. 1 shows a typical wide-scan X.p. spectrum of the Ni-exchanged montmorillonite, sample 5.M. Stocker et al. 1865 Table 1. Photoelectron binding energies and Auger-electron kinetic energies for samples 1-1 1 and some reference compounds sample 1 '! :r 4 :i 0 7 8 9 1 0 11 NiO NiF, Ni(OH), Ni-spinel Ni-olivine Na-talc Ni-Y zeolite N1 O SiO, A403 Al,O, (crystalline) A1 ,03 (amorphous) MgF, MgC1, * 6H,O MgO MgBr, .6H,O olivine talc exchangeable Mg in montmorillonite phlogopi te photoelectron binding energies/eV Auger-electron I_-- - kinetic energies __ Ni 2p3,, A1 2p Si 2p 0 1s Mg 1s (MgKL,,,L,,,)/eV ref.- - 857.1 856.6 857.3 856.4 856.9 856.1 856.2 858.0 859.2 854.6 857.6 855.8 856.2 856.3 857.0 856.85 852.8 - - - - - - - - - - - - 74.7 74.8 74.6 74.2 74.8 74.2 74.0 74.6 74.3 - - - - - 74.6 - - 74.8 - - 75.0 76.0 75.6 - - - - - - - - 103.2 103.2 102.9 102.6 103.2 102.6 102.3 102.9 102.9 103.7 - - - - - 103.1 103.3 103.0 104.0 - - - - - - - - - - - - 532.5 532.4 532.2 532.0 532.5 531.8 53 1.4 53 1.9 531.5 532.9 533.4 528.4 53 1.4 53 1.6 532.1 532.5 532.1 533.1 53 1.8 533.5 533.2 - - - - - - - - - - 1303.7 1303.4 1303.2 1303.9 1303.8 1303.7 1302.9 1 303 .O - - - - - - - - __ - - - - - - 1306.3 1304.6 1303.7 1305.1 1303.9 1303.7 1305.1 1303.5 1180.2 1180.3 1 180.4 1179.4 1 180.0 1 180.0 1 180.9 1180.5 - - __ - - - - - - _- - - - - - 1177.0 1 180.4 1181.5 1180.9 1180.8 1180.7 1 179.2 1180.7 U U U a a U U U U U U a a 22 22 22 22 22 25 22 22 29 29 10 10 10 10 8 8 10 8 a This work, experimental error: & 0.2 eV.A collapse of the clay structure ( i e . a decrease of the spacing between the layers) was observed for the calcined samples 4 and 6. The loss of weights (see table 2) observed for these samples reflects to a large extent the process of structure collapse (due to calcination), whereas the surface-area data support this view only to a limited extent.Nickel From the recorded X.P.S. binding energies of the Ni 2p3/2 core levels (see table l), the synthesized montmorillonites 3-7 can be separated into two groups. For the uncalcined samples 3, 5 and 7 we found Ni 2p,/, binding energies of ca. 857 eV, close to the values reported for Ni-talc and Ni-Y zeolite. The corresponding data for the calcined samples 4 and 6 (ca. 856.5 eV) agree with those obtained for Ni in Ni-spinel and Ni-olivine. This indicates a more electron-rich environment for Ni in the calcined samples as compared with the uncalcined clays. However, the measured Ni 2p,,, core-level binding energies of1866 X.P.S.of Ni-exchanged Montmorillonites Table 2. Chemical analysis, loss of weight and specific surface area measurement (B.E.T.) of samples 1-1 1 NiO A1,0, SiO, loss of weight (B.E.T.) sample (YO) (YO) (YO) at 800 "C (YO) /m2 g-l 1 - 14.9 77.9 2 - 15.5 67.8 3 1.6 15.9 65.8 4 1.5 17.7 73.4 5 1.4 17.1 68.8 6 1.7 13.7 73.9 7 1.3 17.3 69.9 8 2.8 16.6 65.7 9 ca. 4 ca. 96 - 10 0.6 6.8 92 11 1.2 3.8 95 - 9.7 11.8 1.1 8 .O 0.8 7.4 9.1 - 206 246 287 278 295 267 317 178 182 289 33 1 I ~~~~ 0 250 5 00 750 1000 12 50 binding energy/eV Fig. 1. Widescan X-ray photoelectron spectrum of Ni-exchanged montmorillonite 5 (A1 K, excitation). all the investigated montmorillonites (3-8) are clearly different from the values reported for Ni-metal, NiO and the very ionic NiF,.From this observation we infer that Ni- exchanged montmorillonites tend to form a nickel-layered silicate structure rather than an NiO-like phase. The Ni 2p core-level spectra of the montmorillonites 3, 5, 8 and the alumina-based sample 9 exhibit strong 'shake-up' satellites at an energy separation of ca. 6.5 eV from the leading photoelectron peak. We believe these satellites to be of the same origin asM. Stocker et al. 1867 those commonly observed in the core-level spectra of certain open-shell transition-metal and rare-earth compounds.26 The observed ' shake-up' satellites are a strong indication of tetrahedrally2' or octahedrally28 coordinated nickel(I1). Moreover, the measured Ni 2p3,, photoelectron binding energies are in close agreement with those previously reported for octahedral nickel(I1) by Matienzo et aZ.28 and Koppelman and Di1lard.l' The Ni 2p3,, core levels of the silica-based samples 10 and 11 were recorded at noticeably higher photoelectron binding energies (858-859 eV), indicating a different bonding state for Ni in these samples compared with the montmorillonites (3-8) and the alumina-based sample 9. Aluminium The measured A1 2p photoelectron binding energies display only minor differences through the montmorillonite series, including the reference compounds 1 and 2.The lower values (ca. 74.2 eV) were obtained for the calcined samples (with one exception, 7, see table l), indicating a slightly more electron-rich environment for A1 in these samples as compared with the uncalcined clays. The measured A1 2p binding energies were all lower than those reported for pure A1,03 (75-76 eV).22929 The photoelectron binding energies measured for our montmorillonite samples correspond to the values for A1 in sixfold coordination. However, the lower binding energies obtained for the calcined samples 4 and 6 indicate possibly a small fraction of fourfold-coordinated Al.These findings were confirmed by 27Al m.a.s.n.m.r. measurements. Silicon The measured Si 2p photoelectron binding energies show the same trend as that observed for the A1 2p core levels. Comparatively low values (ca. 102.5 eV) were obtained for the calcined samples 4 and 6 and for the uncalcined sample 7. The montmorillonite Si 2p binding energies were all lower than the value reported for pure SiO, (104.0 eV).22 The recorded value for the Ni/SiO, sample 11, on the other hand, is in close agreement with that reported for pure silica.The recorded X.P.S. binding energies of aluminium and silicon indicate possibly a change in the bonding environment of A1 and Si due to calcination (samples 4 and 6). One might speculate whether a collapse of the interlayered structure (i.e. reduced spacing between the layers) appears to diminish the net positive charge on the A1 and Si species by reduced charge transfer to the oxygens, which reside closer to the cations in these calcined samples as compared with the uncalcined clays. Oxygen When comparing the recorded 0 1s photoelectron spectra of the montmorillonite samples with those of related compounds, one should bear in mind that the spectral linewidth for minerals with more than one bonding state of oxygen will in general be broader than that obtained for compounds with only one type of oxygen bonds.Minor differences were observed in the 0 1s binding energies of the synthesized clays. For the calcined montmorillonites 4 and 6 we obtained values close to those reported for Ni-Y zeolite and Ni-olivine,22 whereas the uncalcined sample 7 exhibited a lower 0 1s binding energy, in agreement with the values reported for Ni(OH), and Ni-spine122 (see table 1). The measured 0 1s binding energies of the other uncalcined montmorillonites, including the reference compounds 1 and 2 all agree within experimental error with the value claimed for Ni-talc.,, The higher 0 Is binding energy recorded for Ni/SiO, (1 1) is in close agreement with the values reported for pure oxide.22 A noticeable spread in reported values for the1868 X.P.S.of Ni-exchanged Montmorillonites 1182 1181 2 .: + 1180 . x C EJ c 3 .* k 9 1179 m N 4 m 1178 2 1177 11 76 MgO MgBri6H20, olivine MgCI:6H,O ~l 6f3 phlo opite A% 0 exchangeable Mg in montmorillonite 1307 1306 1305 1304 1303 1302 Mg 1s photoelectron binding energy/eV Fig. 2. Chemical-state plot for Mg: A, this work; ref. (8) and (10). 6 0 4 Si 4 0 ,20H 4 Al.Mg 4 Si Fig. 3. Structure of montmorillonite :32 T, tetrahedral ; 0, octahedral.M. Stocker et al. 1869 0 1s binding energy of pure alumina22’29 (see table 1) renders this comparison more difficult for the Ni/Al,O, sample 9. Magnesium The ‘chemical-state plot’ introduced by Wagner30g31 provides a useful format for displaying X.P.S.reference data, and thus increases the utility of core-level electron spectroscopy for identification of chemical states. In this format the kinetic energy of the most prominent X-ray-excited Auger transition is plotted us. the binding energy of the leading core-level photoelectron peak. Fig. 2 shows a ‘chemical-state plot’ of Mg for the different montmorillonites investigated and a selection of reference compounds. It appears that the recorded binding and Auger-electron kinetic energies of the uncalcined sample 7 compare closely with those reported for MgO’O and are well separated from the values obtained for the calcined samples 4 and 6. In the montmorillonite layered structure (see fig.3) Mg is surrounded by four oxygen and two hydroxide ions in an octahedral coordination. The comparatively low A1 2p and Si 2p X.P.S. binding energies recorded for the montmorillonites 4 and 6 could be attributed to a collapse of the clay structure. The converse effect (comparatively high binding energies) is observed for Mg, however. This could possibly be explained by differences in the formation of bonds involving s orbitals (Mg) and p orbitals (Al, Si), respectively (different spherical charge distributions). Note also that the ‘chemical-state plot’ for Mg indicates a close relationship between the montmorillonites and other phyllosilicates (talc and phlogopite). Conclusions From close inspection of our X.P.S. data we conclude that calcination of the montmorillonites modifies the atomic coordination and chemical environment of the constituent clay elements.The measured X.P.S. binding energies of sample 7 differ noticeably from those obtained for the other uncalcined clays. The X.P.S. binding energy of the interlamellar Ni cation (2p,,, level) was obtained at values considerably higher than that measured for NiO and close to the values reported for compounds with octahedrally coordinated Ni”. This indicates formation of a nickel- layered silicate structure. Lastly, the recorded Mg Auger (KL,,, L,,,) and photoelectron (1s) spectra unveil substantial influence by the specimen-preparation procedure on the bonding state of this alkaline-earth metal. The authors are indebted to Anne Andersen for preparing samples 9-1 1.References 1 T. J. Pinnavaia, Science, 1983, 220, 365. 2 P. G. Bercik, K. J. Metzger and W. E. Swift, Znd. Eng. Chem. Prod. Res. Dev., 1978, 17, 214. 3 H. E. Swift and E. R. Black, Znd. Eng. Chem. Prod. Res. Deu., 1974, 13, 106. 4 J. R. Sohn and H. B. Park, J . Korean Chem. SOC., 1982, 26, 282. 5 J. M. Adams, T. V. Clapp and D. E. Clement, Clay Miner., 1983, 18, 411. 6 M. P. Atkins, D. J. H. Smith and D. J. Westlake, Clay Miner., 1983, 18, 423. 7 M. H. Koppelman, in Advanced Chemical Methods for Soil and Clay Minerals Research, ed. J. W. 8 H. Seyama and M. Soma, J. Chem. Soc., Faraday Trans. I , 1985, 81,485. 9 J. M. Adams and S. Evans, Clays Clay Miner., 1979, 27, 248. 10 H. Seyama and M. Soma, J. Chem. SOC., Faraday Trans. I , 1984, 80, 237.1 1 H. Seyama and M. Soma, Chem. Lett., 198 1, 1009. Stucki and W. L. Banwart (D. Reidel, Dordrecht, 1980).1870 X . P.S. of Ni-exchanged Montmorillonites 12 B. Carriere, J. P. Deville, D. Brion and J. Escard, J. Electron Spectrosc. Relat. Phenom., 1977, 10, 13 M. H. Koppelman and J. G. Dillard, Clays Clay Miner., 1980, 28, 21 1. 14 M. H. Koppelman, A. B. Emerson and J. G. Dillard, Clays Clay Miner., 1980, 28, 119. 15 M. H. Koppelman and J. G. Dillard, J. Colloid Interface Sci., 1978, 66, 345. 16 M. H. Koppelman and J. G. Dillard, Clays Clay Miner., 1977, 25, 457. 17 M. H. Koppelman and J. G. Dillard, ACS Symp. Ser., 1975, 18, 186. 18 J. W. Stucki, C. B. Roth and W. E. Baitinger, Clays Clay Miner., 1976, 24, 289. 19 J. W. Stucki and C. B. Roth, Soil Sci. Soc. Am., J., 1977, 41, 808. 20 S. Evans and E. Raftery, Clay Miner., 1982, 17, 477. 21 S. Evans and E. Raftery, Clay Miner., 1980, 15, 209. 22 P. Lorenz, J. Finster, G. Wendt, J. V. Salyn, E. K. Zumadilow and V. I. Nefedov, J. Electron Spectrosc. 23 G. Wendt, D. Hentschel, R. Schoellner, J. Finster, K. Becker, M. Weser and J. Welker, DDR-Patent 24 K. G. Allum, Deutsche Auslegeschrijt, 1979, 2029624. 25 S. Narayanan, Zeolites, 1984, 4, 231. 26 C. S. Fadley, in Electron Spectroscopy, ed. C . R. Brundle and A. D. Baker (Academic Press, New 27 L. J. Matienzo, W. E. Swartz jr and S. 0. Grim, Inorg. Nucl. Chem. Lett., 1972, 8, 1085. 28 L. J. Matienzo, L. I. Yin, S. 0. Grim and W. E. Swartz Jr, Inorg. Chem., 1973, 12, 2762. 29 E. D. Johnson, Ph.D. Thesis (Cornell University, 1985). 30 C. D. Wagner, L. H. Gale and R. H. Raymond, Anal. Chem., 1979, 51,466. 31 C. D. Wagner, J. Electron Spectrosc. Relat. Phenom., 1977, 10, 305. 32 G. Brown, X-Ray IdentiJication and Crystal Structures of Clay Minerals (Mineral SOC., London, 85. Relat. Phenom., 1979, 16, 267. 16 00 37 (1983). York, 1978), vol. 2. 1961). Paper 71865; Received 18th May, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401863

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

1. Chem. SOC., Faraday Trans. I, 1988, 84(6), 1871-1878 Inhibition of the Thin-film Oxidation of n-Dodecane by Diphenylamine Adiele D. Ekechukwut and Robert F. Simmons* Department of Chemistry, The University of Manchester Institute of Science and Technology, P.O. Box 88, Manchester M60 lQD The inhibition by diphenylamine (DPA) of the oxidation of a thin film of n-dodecane has been studied over the temperature range 180-230 "C by following the consumption of gaseous oxygen, the disappearance of the inhibitor and the formation of peroxides. At the start of the reaction there is an induction period which increases linearly with the initial concentration of inhibitor. During the first part of this induction period there is apparently no change in the concentration of DPA, but it has all been consumed by its end, although there is only a negligible decrease in the absorption at 1600 and 3400 cm-' due to N-H.Thus it is concluded that abstraction of the amino hydrogen cannot be involved in the inhibition mechanism under the present experimental conditions, and it is suggested that the first step is probably an electron-transfer reaction between an alkyl radical and the inhi bit or. Phenols and amines are known to be effective inhibitors of the liquid-phase oxidation of hydrocarbons and are widely used as antioxidants.' It is generally agreed that they act by reacting preferentially with chain centres in the oxidation reaction, but whereas the essential features of the inhibition mechanism for phenols are well established' there is still some uncertainty about the way in which the arnines act.One view is typified by Hammond et aZ.,2 who studied the oxidation of tetralin initiated by a,a-azobisisobutyronitrile (AIBN). They examined the effect of a range of inhibitors and found that with N-methylaniline the rate of uptake of oxygen was proportional to [AIBN]0.5/[In]0.5 (where In is the inhibitor), and that there was no effect on the rate when deuterium was substituted for the amino hydrogen in diphenylamine. Thus they concluded that abstraction of the amino hydrogen was not involved in the rate-determining step of the inhibition mechanism, and they interpreted their results in terns of the formation of a complex between the alkylperoxy radical and the inhibitor, which then reacted with a further alkylperoxy radical to give chain termination.This conclusion was supported by Thomas and T ~ l m a n , ~ who confirmed that inhibition by DPA had the same kinetic characteristics as N-methylaniline, that there was no isotope effect when the amino hydrogen was replaced ' with deuterium, and that diphenyl nitroxide was formed in the inhibition mechanism. This is also an inhibitor in its own right. Although there is direct experimental evidence for the formation of such cornplexe~,~ the evidence is less conclusive in the case of DPA than for amines which have no amino hydrogen. In contrast, Brownlie and Ingold' found completely different kinetic behaviour for the oxidation of styrene (also initiated by the decomposition of AIBN) when it was inhibited by DPA. The reaction was first order in AIBN, while the order with respect to DPA was (- 1); in addition, a kinetic isotope effect was observed when deuterium was substituted t Present address : Nigerian Defence Academy, Kaduna, Nigeria.18711872 Oxidation of n-Dodecane by Diphenylamine for the amino hydrogen. As a result, they concluded that the inhibition arose through the abstraction of a hydrogen atom from the inhibitor by an alkylperoxy radical: initiation = R (1) R + 0 , = RO, (2) ( 3 ) (4) ( 5 ) A + A = inert products (6) RO, +A = inert products (7) A+RH = AH+R. (8) RO, + RH = ROOH + R RO, + RO, = inert products RO, +AH = ROOH + A Brownlie and Ingold5 interpreted their results in terms of the same basic mechanism as that for the inhibition by phenols (see above, where AH represents the inhibitor).They also showed that the kinetic results obtained by Hammond et aL2 are explained if reactions (- 5) and (8) are important, and the absence of a deuterium isotope effect was explained by the occurrence of a rapid hydrogen-deuterium exchange between the deuterodiphenylamine and tetralin hydroperoxide. Support for this interpretation comes from the observation6 that the inhibition of the oxidation of tetralin by phenol also shows the same kinetic characteristics as observed for diphenylamine by Hammond et al., While these conflicting results can be reconciled in this way, and the interpretation of Brownlie and Ingold5 gives a much more logical picture of the inhibition by amines, neither styrene nor tetralin can be considered as representative of the hydrocarbon structures which constitute the bulk of a lubricating oil.In addition, all the above work was carried out at relatively low temperatures, and the use of an initiator to increase the rate of reaction may well have produced conditions which are not representative of those found when a lubricant is used in an engine, It is also known that tertiary amines are good antioxidants,' so that inhibition does not necessarily require the abstraction of an amino hydrogen atom, and hence it is possible that more than one mechanism can operate. Thus caution is required when extrapolating the results from such systems to the behaviour of inhibitors in an actual lubricating oil. This view is supported by two pieces of work. Pederson8 examined the effect of antioxidants on petrol at 100 "C and found no kinetic isotope effect when the amino hydrogen of some p-phenylenediamine derivatives was replaced by deuterium.Particular care was taken in this work to minimise any possible hydrogen-deuterium exhange by using a fuel from which -OH, -SH and -NH groups had been removed. He found that the efficiency of the additive was directly related to the electron density of the arylamino group, and he favoured an electron-transfer reaction as the first step in the inhibition mechanism. In the other study the inhibition of the oxidation of a thin film of n-dodecane by p-methoxyphenol (PMP)' was studied in the temperature range 180- 230 "C. This showed that the primary inhibition step at this higher temperature was the reaction of the inhibitor with alkyl radicals as they were formed in the chain initiation reaction, and not the removal of alkylperoxy radicals, as concluded from the studies at lower temperatures, where AIBN had been used as an initiator.The aim of the present work was to examine a model system under conditions more representative of the high temperatures encountered in a petrol engine. This has been achieved by examining the effect of DPA on the oxidation of a thin film of n-dodecane over the temperature range 180-230 "C. As with the inhibition by PMP,' there was an induction period at the start of the reaction which increased with increasing [DPA].A . D. Ekechukwu and R. F. Simmons 1873 Whle the inhibitor was consumed during the induction period, its removal did not follow a simple rate law and there was no significant reduction in the absorption at 1600 and 3400 cm-l due to N-H during the induction period.As a result, the inhibition could not have arisen from the abstraction of the amino hydrogen from the inhibitor by a chain centre, and it is suggested that under the present experimental conditions the first step is probably an electron-transfer reaction between an alkyl radical and DPA. Experiment a1 Details of the experimental method have been reported previously,' and so only a brief outline is repeated here. The required amount of a solution of DPA in n-dodecane (normally 2 cm3) and oxygen was sealed into a thick-walled Pyrex reaction vessel and heated for a known time in a thermostatically controlled furnace.During this period the reaction vessel was rotated horizontally within the furnace to ensure that a thin film of liquid was maintained over its inner surface. After quenching, the amount of residual oxygen was determined and, in some experiments, the liquid phase was analysed for DPA and total peroxides. The non- condensible gas at - 196 "C was transferred to a gas burette to determine the total amount, and the oxygen content was determined by gas chromatography using a 10 ftj. column of silica gel (30/60 mesh) with hydrogen as the carrier gas and a katharometer detector. The analysis for DPA was made using a capillary column with a wall coating of free fatty acid (FFAP), a temperature programme of 4"Cmin-1 up to 200 "C, nitrogen as the carrier gas and a flame ionisation detector.The total peroxide content was measured using an adaptation of the iron(I1) thiocyanate method.g* lo Materials Gas-chromatographic analysis of n-dodecane (B.D.H. Chemicals, > 99 YO purity) showed a trace of impurity with a slightly longer retention time, but since the extent of reaction was limited by the relatively small quantity of oxygen it was not purified further. Diphenylamine (B.D.H. Chemicals, > 99 % purity) and oxygen (B.O.C.) were also used without further purification. Results Inhibition of the oxidation of a thin film of n-dodecane by diphenylamine has been studied over the temperature range 180-230 "C using an initial pressure of oxygen of 600mmHg,S previous work' having shown that the inhibition is independent of the initial pressure of oxygen above 200 mmHg.Fig. 1 gives the results for a temperature of 210 "C; each data point comes from a separate kinetic experiment, and one indication of their reproducibility comes from the scatter about the mean line. Although straight lines have been drawn in fig. 1, a steep sigmoid curve could be drawn as expected for an autocatalytic reaction. It will be seen that there was an induction period during which the consumption of oxygen was negligible. Fig. 2 shows that this induction period increased linearly with [DPA], although at very low concentrations the inhibitor was less effective than expected from such a relationship. At 210 "C, the subsequent rate of reaction was almost the same as that in the absence of DPA, but at 180 "C it was significantly slower.The total peroxide content of some reaction mixtures was also determined and the results are shown in fig. 3. These data are expressed as an equivalent amount of hydrogen peroxide since the method of analysis did not allow a distinction between 1 ft = 3.048 x lo-' m. $ 1 mmHg = 1.33322 x lo2 Pa.1874 Oxidation of n-Dodecane by Diphenylamine Fig. 1. Effect of diphenylamine (DPA) on the oxidation of n-dodecane. Temperature, initial pressure of oxygen, 600 mmHg. [DPA]/10-2 mol dmd3: 0, 0; 0, 1.0; T7,2.0; A, 0,4.0. I 20 LO 60 LO 20 210 "C, 3.0 and I I I 0.5 1.0 [DPA]/10-3 mol dm-3 Fig. 2. Variation in the induction period for the oxidation of n-dodecane with concentration of diphenylamine. T/"C: A, 180; 0, 190 and 0,210. hydrogen peroxide and organic peroxides, It will be seen that there was a steady build-up of peroxide to a maximum concentration by the end of the induction period, even though there was no significant consumption of oxygen.As with PMP as inhibitor, the build-up in peroxide concentration occurred at a reduced rate when DPA was present. This maximum concentration of peroxide was approximately constant during the consumption of the oxygen, and it then began to decrease again. The consumption of DPA during the induction period was followed by gas chromatography. Fig. 4 shows that [DPA] remained initially constant and then decreased linearly with time to zero. For example, with [DPA], = 4.0 x lo-* mol dm-3, [DPA] did not change for 30min at 210°C, and then fell to zero by the end of the induction period, i.e.at ca. 50 min. Gas-chromatography-mass-spectrometry was used to confirm that the species eluted at that particular retention time was DPA. It will also be seen that the period during which [DPA] remained constant increased in an approximately linear manner with [DPA],. This is in marked contrast to the behaviour observed with PMP as inhibitor, where there was a short period during which theA . D. Ekechukwu and R. F. Simmons 1 I 1875 Fig. 3. Variation in the total amount of peroxide (expressed as H20J during the oxidation of n-tlodecane. Temperature, 210 O C , initial pressure of oxygen, 600 mmHg, [DPA] = 4.0 x mol dm-3. 0, Oxygen consumed; A, hydrogen peroxide. time/min Fig. 4. Disappearance of diphenylamine (DPA) in the inhibited oxidation of n-dodecane. Temperature, 210 "C, initial pressure of oxygen, 600 mmHg.[DPA],/10-2 mol dm-3: A, 1.0; V, 2.0 and 0, 4.0. inhibitor concentration remained unchanged, which corresponded to the heating time of the reaction vessel, and the subsequent consumption was zero order with respect to the inhibitor, In addition, further information about the removal of DPA was obtained by following the change in the absorbance due to N-H at 3400 and 1600 cm-l. In particular, spectra were obtained for the initial mixture, a mixture with a reaction time within the induction period, and reaction mixtures with 30 and 100% consumption of oxygen. The absorbance due to N-H was effectively the same in the first three cases, although there was some reduction in the absorbance when the oxygen consumption was complete.In addition, the spectra showed no absorbance due to carbonyl compounds (C=O stretch at 1700 cm-') during the induction period, but a steady increase in this absorbance thereafter. The behavioural pattern of 230 "C was different from that at the lower temperatures, in 62 FAR I1876 Oxidation of n-Dodecane by Diphenylamine that oxygen consumption ceased after a small amount of reaction. For example, with [DPA], = 4 x mol dm-3, oxygen consumption stopped when only ca. 24% had reacted and no further consumption occurred during the next 30 min. This is indicative of a dramatic change of mechanism on raising the temperature from 210 to 230 "C, and it was noted that it was accompanied by an increased formation of methane.It also implies the formation of a product which is itself a more effective inhibitor of the oxidation process than the original inhibitor. Discussion The results with diphenylamine as inhibitor were very similar to those obtained with PMP, so that a similar basic mechanism was probably operative, For example, there was a negligible consumption of oxygen during the induction period and, for temperature in the range 180-210 "C, this induction period increased linearly with initial concentration of inhibitor; at the end of the induction period, [DPA] had fallen to zero. This type of behaviour is analogous to that observed for the inhibition by benzoquinone of the photochemical polymerisation of vinyl acetate" and the thermal polymerisation of styrene,12 where the accepted explanation is that the inhibitor reacted with chain centres as they were formed in the initiation reaction. Provided such a reaction was fast under the experimental conditions used in this work, the consumption of oxygen would be negligible until all the DPA had been removed.The possibility that any significant gas-phase reaction occurred has been eliminated,13 and the liquid-phase oxidation of a hydrocarbon is normally discussed in terms of a basic mechanism comprising reactions (1)-(4). In the previous low-temperature studies the rate of consumption of oxygen was only reduced by the presence of inhibitor, and the kinetic characteristics of the inhibition suggested that alkylperoxy radicals were being removed under those conditions. Since reaction was completely suppressed at the higher temperatures used in the present study, i.e.the kinetic characteristics were completely different to those obtained at the lower temperatures, it follows that another species must have been removed under the present experimental conditions. In the absence of initiator, chain initiation in the liquid-phase oxidation of hydrocarbons occurs by reaction (lo), and the resulting HO, radicals are removed by reactions (1 1) and (12) : RH+O, = R+HO, (10) HO, + RH = H,O, + R HO, + HO, = H,O, + 0,. (1 1) (12) Any removal of HO, radicals by reaction with DPA can be eliminated, as it would have only a small effect on the rate of formation of hydrogen peroxide, whereas experimentally it was found that the total rate of formation of peroxides was markedly reduced during the induction period. In addition, it would not effect the initial rate of consumption of oxygen. Removal of alkylhydroperoxide seems equally improbable, as it requires that all RO, radicals are removed by reaction (3) under the present experimental conditions. With styrene as the hydrocarbon and DPA as i n h i b i t ~ r , ~ k5/k3 = 100 at 65 "C and, if reactions (- 5 ) and (8) are important, as claimed by Brownlie and I n g ~ l d , ~ k5/k, must be of this order of magnitude.There seems to be no reason why this ratio should be markedly different for any other hydrocarbon and, in the unlikely event that E5 = 0, the ratio k,/k, will only decrease by a factor of 200 from 65 to 160 "C (assuming E3 is the same as for the gas-phase reaction of an alkylperoxy radical with a secondary paraffinic C-H bond14).This factor of 200 must be an upper limit, as it assumes E5 = 0, which seems most unlikely, and that the concentration of dissolved oxygen in the dodecane is unchanged, whereas in practice it will be lower at the higher temperature. Thus, even allowing for the uncertainty in the value of k,/k, when n-dodecane is the fuel, this isA . D. Ekechukwu and R. F. Simmons 1877 hardly a sufficient change for reaction (3) to predominate over reaction (5) at the temperatures used in the present work. In addition, if reaction (3) predominates over reaction (5) when DPA is inhibitor, the same should be true for PMP as inhibitor, as it is a less efficient inhibitor than DPA. This requires that PMP also acts by the removal of alkylhydroperoxide, whereas a much more logical picture of the inhibition process with PMP is given by the removal of alkyl radicals.8 It must also be noted that the removal of PMP is zero order, and there is no indication of an acceleration in its rate of removal during the induction period.Thus there is no evidence for any autoxidation through the thermal decomposition of an alkylhydroperoxide. While each of the above points on its own is by no means decisive, their cumulative effect must lead to the conclusion that removal of alkylhydroperoxide by DPA is not a plausible first step in the inhibition mechanism. The remaining possibility is that DPA reacts with alkyl radicals as they are formed in the initiation reaction. As pointed out earlierg for PMP as inhibitor, it is quite plausible that under the present experimental conditions the reaction of alkyl radicals with the inhibitor should predominate over their reaction with oxygen until the inhibitor has been removed, as the solubility of oxygen in hydrocarbons decreases with increasing temperature.Thus both the present results and the change in kinetic characteristics from those observed at lower temperatures can be explained very simply and logically by the removal of alkyl radicals under the present conditions. A comparison of the present analysis for DPA and those reported for PMP' shows that the former inhibitor is the more effective, but the overall situation is complex since the relative efficiency of the two inhibitors varies with temperature. For example, at 180 "C the ratio of the rate of removal of DPA to that for PMP is 1 :3.9, while at 210 "C this ratio is only 1 : 1.25.The simplest interpretation of this ratio, in terms of the ratio of the number of chain centres removed by each molecule of the two inhibitors, implies that the number of radicals removed by diphenylamine decreases with increasing temperature. The present results show that the removal of DPA was much more complex than the consumption of PMP, which followed a zero-order rate law after a short initial heating- up period. Fig. 4 shows that first there was a period during which [DPA] did not change, and the length of this period increased approximately linearly with increasing initial inhibitor concentration. After this period, however, [DPA] decreased linearly with time and was effectively zero by the end of the induction period; i.e.during this part of the induction period the removal was zero order with respect to DPA. In addition, the i.r. spectra showed that the absorbance due to N-H did not change during the induction period, and it was only towards the end of the reaction that any decrease was apparent. Thus an exactly analogous mechanism to that for PMP can be rejected, namely the abstraction of the amino hydrogen atom by a chain centre, since this would produce a zero-order removal of DPA throughout the induction period and a corresponding reduction in the absorbance due to N-H. The only alternative that can be envisaged is that there is an electron-transfer reaction between the alkyl radical, and DPA and such a mechanism has been suggested previously by Pedersoq8 although the detailed chemistry involved in the subsequent steps is by no means clear.The occurrence of such a reaction, however, affords a more satisfactory explanation of the present results than a mechanism involving abstraction of the amino hydrogen atom, and it also has the merit of explaining inhibition by tertiary aromatic amines. The results at 230 "C revealed another difference in the kinetic behaviour of DPA and PMP as inhibitors. The change in kinetic behaviour between 210 and 230 "C when DPA wa.s the inhibitor suggests that an effective inhibitor was being continuously regenerated in the system. Nitroxides are known to be formed in the oxidation of an a~lline,~ but they are less effective than the parent amine, so that they are unlikely to have been responsible for the observed effect.'The mechanism of the inhibition of the oxidation of n-dodecane by diphenylamine is 62-21878 Oxidation of n-Dodecane by Diphenylamine thus more complicated than that for p-methoxyphenol. It is concluded that in the temperature range 180-210 "C the primary inhibition reaction does not involve the abstraction of the amino hydrogen, and it seems probable that a mechanism is operative in which alkyl radicals are removed by an electron-transfer reaction between the radical and DPA. Further work is required, however, to elucidate the detailed chemistry involved. We thank Dr B. L. Booth and the referees for their very helpful comments. References 1 K. U. Ingold, A h . Chem. Ser., 1968, 75, 296. 2 G. S. Hammond, C. E. Boozer, C. E. Hamilton and J. N. Sen, J. Am. Chem. Soc., 1955, 77, 3238. 3 J. R. Thomas and C. A. Tolman. J. Am. Chem. SOC., 1962, 84, 2930. 4 A. MacLachlan, J. Am. Chem. Soc., 1965, 87, 960. 5 1. T. Brownlie and K. U. Ingold, Can. J. Chem., 1966, 44, 861. 6 J. A. Howard and K. U. Ingold, Can. J. Chem., 1965, 43, 2724. 7 J. R. Thomas, J . Am. Chem. Soc., 1963, 85, 593. 8 C. J. Pederson, Znd. Eng. Chem., 1956, 48, 1881. 9 A. D. Ekechukwu and R. F. Simmons, J. Chem. Soc., Faraday Trans. I , 1986, 82, 1965. 10 Sir A. C. Egerton, A. J. Everett, G. J. Minkoff, S. Rudrakanchandra and K. C . Salooja, Anal. Chim. Acta, 1954, 10, 422. 11 G. M. Burnett and W. H. Melville, Proc. R. Soc. London, Ser. A, 1947, 189, 456. 12 S. G. Foord, J. Chem. SOC., 1940, 48. 13 N. J. Royster and R. F. Simmons, unpublished results. 14 S . W. Benson, Prog. Energy Combust. Sci., 1981, 7, 125. Paper 71909; Received 21st May, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401871

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1988, 84(6), 1879-1887 Thermodynamics of the Zinc Sulphide Transformation, Sphalerite + Wurtzite, by Modified Entrainment Peter J. Gardner" and Peter Pang Roqial Holloway and Bedford New College, Egham Hill, Egham, Surrey TW20 OEX The dissociative sublimation of both a- and p-zinc sulphide, ZnS(c) = Zn(g) + is&), has been studied by modified entrainment in the temperature range 1010-1445 K. The following free-energy equations were derived : AG"(a)/J mol-' = 376 700- 191.9T/K and AG*@)/J mol-1 = 374 200 - 190.4 T/K from which the enthalpy and entropy changes of transformation (p -, a ) at the transition temperature (1293 K) were -2.5+ 1.5 kJ mo1-1 and - 1.5& 1.3 J K-' mo1-l. The corresponding AH"@ -, a ) at 298.15 K obtained by the third-law method was -2.3 f0.9 kJ mol-'.This enthalpy change is at variance with the currently accepted value of ca. 13 kJ mol-'. Zinc sulphide exists in two principal forms, sphalerite ($, cubic, or 3C) and wurtzite (a, hexagonal or 2H). The low-temperature form is the /3-phase, the transition tempera- ture' is 1293 _+ 10 K, and the a-form is metastable at room temperature. Numerous intermediate forms (polytypes or superstructures), with crystal structures closely related to the principal forms, have been characterized. Most compilations of thermodynamic data3-5 quote standard enthalpies of formation that imply AH@ + a) 298 z 13 kJ mol-l. This is based on an early solution calorimetric study6 at 293 K. More recent heat- capacity studies7$ suggest that the transition is nearly athermal.There is commercial interest in the use of molten zinc chloride as a hydrocracking s~lvent.~, lo Sulphur in the feedstock causes a steady accumulation of zinc sulphide in the melt. A proposed regeneration procedure involves conversion of the zinc sulphide via ZnS(c) + 2HCl(g) = ZnC1,(1) + H,S(g) followed by separation of the zinc chloride in a fluidized-bed combustor. The efficiency of the above conversion depends, inter alia, on the equilibrium constant and its temperature dependence. To estimate the latter, reliable thermodynamic data for the heats of formation of zinc sulphide in its common forms are necessary. The modified entrainment method has been used to study the reductive transport [eqn (l)] of ZnS@ in the range 1012-1271 K and ZnS(a) in the range 1300-1443 K: (1) Heats of formation of ZnS(a,p) were obtained from the temperature dependence of K , in combination with well established ancillary thermodynamic data.ZnS(a, B) + H&) = W g ) + H,S(g). Apparatus Experimental The modified entrainment method (MEM) has previously been described in detail.", l2 In summary, ZnS (ca. 300 mg) is contained in a spherical silica capsule, the orifice of which comprises a vertical capillary (ca. 2 cm long, 1 mm i.d.). The capsule is suspended from one arm of a recording microbalance and the transporting gas (hydrogen) flows 18791880 Thermodynamics of Zinc Sulphide downwards (ca. 80 cm3 min-l) over the capsule and flushes away the gaseous products of equilibrium (1). The capsule is positioned in the region of a temperature plateau generated by a vertically mounted tube furnace.The experimental parameter is the rate of mass loss (&/kg s-l) of ZnS at a preset temperature, T. Sample temperatures were measured with a calibrated Pt-Pt/ 13 YO Rh thermocouple to a precision of f 2 K. The design of the previously used capsules were improved in this work by insetting the capillary by ca. 2 mm into the capsule body rather than joining it via a short flared section; this enabled the diffusive resistance of the capillary (E/nr2: I = length and Y = radius) to be determined with greater precision. Three different capsules were used with capillary resistances from 1.912 x lo4 to 2.132 x lo4 m-'. For ZnS(P), 92 data points were recorded with ci, ranging between 7.4 x lo-'' kg s-' (at 1012 K) and 2.8 x lo-'' kg s-' (at 1271 K).For ZnS(cc), 199 data points were noted from 1300 K (ci, sz 3.6 x 10-l' kg s-l) to 1433 K (03 z 1.6 x lo-' kg s-l). Differential Scanning Calorimetry A Stanton Redcroft Model 1500 was used with a scan speed of 5 K min-l from 1073 to 1423 K with the sample (43 mg) in a flowing nitrogen atmosphere. Materials Zinc sulphide (Aldrich Gold Label, 99.99%, batch no. 0301 JK) was used without further purification. Surprisingly, it was common to lose several mass per cent of volatile material during initial heating. The sample was characterized by emission spectroscopy, X-ray diffraction (X.r.d.), thermogravimetric analysis (tga.), mass spectrometry (m.s.) and elemental analysis. Trace metallic impurities were determined by emission spectroscopy and were present at the following concentrations: Cd, 10 ppm; Na, 10 ppm; Pb, 2 ppm; Mo, 1 ppm; Ca, < 1 ppm; Mg, < 1 ppm.X.r.d. showed the sample to be 95 + YO sphalerite. It has been reported" that ZnS is slowly oxidised to ZnSO, in moist air and that the freshly precipitated material adsorbs water, thus giving a continuous dehydration loss on drying. Sulphate was shown to be absent from both the stock material and heated samples. T g a . (Perkin-Elmer TGS-2) in flowing nitrogen from 300 to 973 K at 20 K min-l showed a near-linear mass loss of 2.95 mass YO, terminating between 920 and 970 K. The dissociation pressure above ZnS at 1000 K is ca. lop2 Pa so no contribution was expected from this source. M.s. (AEI MS 30) with a solid injection probe was used to identify the volatile species.The sample was heated from ca. 350 to 623 K while continuously monitoring the mass spectrum in the m/e range 14-220. Water was noted at all temperatures, the maximum evolution being at ca. 450 K and decreasing thereafter. H,S was observed from 480 K, rising to a maximum at 540 K and decreasing thereafter. SO,, in smaller quantities, was also observed at temperatures in excess of 480 K. No other volatile material that originated from the sample was detected. Clearly the stock material was significantly contaminated with moisture which is evolved continuously. The H,S probably arises from the reaction 3 s + 2H,O = 2H,S + SO, which is known to proceed slowly on boiling S in H,O (the sample was slightly sulphur- rich, see below).Elemental analysis showed the stock material to have a Zn:S ratio of 0.993 kO.002. It has been suggested14 that the stability of sphalerite depends on it being slightly S-rich but the non-stoichiometry range is small, 0,998 & 0.00 1 . In summary, the stock material was contaminated with both moisture and S and has a composition ZnSl,,,,,,~,,,~0.168 H,O. These impurities have no influence on the MEM results as the minimum temperature for the MEM study was 1010 K, by which temperature the impurities were lost by volatilization. This was confirmed by elemental analyses on heated samples (see table 1).P. J. Gardner and P. Pang 1881 Table 1. The treatment of ZnSO and its subsequent analysis by X.r.d. treatment ambient (stock sample) 700 "C/24 h/H, 800 "C/24 h/H, 900 "C/24 h/H, 900 "C/32 days/Ar" + 9013 "C/ 10 min/H, I 100 "C/ 14 days/Ar to 1 130 "C/ 10 K min-'/N, + to 900 "C/ 10 K min-l/N, to 1230 "C/20 K min-l/N, P a! Zn: S ratio dominant dominant dominant dominant dominant minor trace trace < 5 % < 5 % % 5 % 5-10 YO < 5 % dominant dominant dominant 0.993 & 0.002 0.998 & 0.002 0.999 f 0.002 - 0.999 & 0.002 a Hydrogen cannot be used for long periods at high temperatures because of transport by reductive sublimation [eqn (l)].Hydrogen was purified by passage through a palladium diffuser (model H28/1, Johnson Matthey). Argon was purified using a commerical rare-gas purifier (model Results and Discussion Differential Scanning Calorimetry An attempt was made to measure the enthalpy of transition directly using d.s.c.The results were ambiguous; there was a slow endotherm with an onset at 1173 K that merged into a more rapid exotherm, which was complete at 1333 K. It was not possible to deconvolute these two features. These results suggest the data points from modified entrainment above 1173 K with the P-phase should be treated with caution. An X-ray diffraction study of the P -+ a phase transition has been reported by Baars and Brandtl' who steadily heated a 'weakly disordered cubic' phase from 293 K while monitoring the diRraction pattern. At ca. 1173 K, the residual disorder decreased until at 1273 K it had been eliminated. Above 1273 K there was a rapid, quantitative conversion to the hexagonal phase, with complete conversion at 1400 K.The implication of this study on our reported enthalphy and entropy of transition is considered below. RGP-4, B.O.C.). Modified Entrainment The dissociative sublimation [eqn (2)] of zinc sulphide is known to be congruent,16 and the concentration of molecular ZnS(g) in the equilibrium vapour over our temperature range is negligible :17 (2) Also there is insignificant contribution from S,,(g) (n > 2 or n = 1) in this temperature range.18 Initial experiments performed with Ar as carrier gas in the temperature range where the P-phase is stable (< 1293 K) gave extremely low rates of mass loss. All subsequent experiments were performed in hydrogen, where the participation of the eq uili bri um enhanced the rate of mass loss by ca. three orders of magnitude.The conversion of the P-phase to the a-phase has been extensively investigated and the effects of impurity levels,2*19 particle size," mechanical stress2' etc. on both the rate and temperature of conversion are well documented. The conversion is facile and it was clearly necessary to establish the identity of the material used in modified entrainment, especially in the region ca. 100 K below the transition temperature, where a slow ZnS(c) = Zn(g) + +S2(g). H,(g) + tS,(g) = H a g ) (3)1882 Thermodynamics of Zinc Sulphide conversion to the a-phase could invalidate the experiments. Accordingly, batches of stock ZnS were heated in flowing H,, N, or Ar for at least 24 h to simulate their behaviour in the modified entrainment rig, quenched rapidly and examined by X.r.d.Samples sizes and irradiation times were such that the level of discrimination between the phases was ca. 5 %. Table 1 summarises the sample treatment, the X.r.d. results and elemental analyses on the heated samples. Two separate modified entrainment experiments on the P-phase were completed with different initial samples and different temperature sequences. In one set the temperature was increased steadily from 1120 K to within 50 K of the transition temperature and decreased to 950 K, recording 55 data points [h(T)] at regular intervals. For the second set, the temperature was decreased from 11 50 K to 910 K, then increased to 1250 K, noting 37 values of h( T ) . An analysis of covariance on the regression lines (A Gi/J mol-' us. T / K ) showed no significant difference.For experiments with the a-phase, zinc sulphide was maintained at 1440 K in H, for 2 h to ensure complete conversion. Only two of eqn (1)-(3) are independent and it is convenient to write the flux equations in terms of eqn (2) and (3) and so extract an expression for the equilibrium constant for dissociative sublimation (K,), where ci> is the rate of mass loss of ZnS, l / A is the length over the cross-sectional area (the diffusive resistance) of the capillary, M , is the molecular mass of ZnS, Di are quaternary diffusion coefficients, Ki are equilibrium constants, pe is the standard-state pressure (101 325 Pa) and P is the working pressure in the MEM rig. The following notation is used throughout: 1 = H,, 2 E H,S, 3 = S,, 4 = Zn and 5 = ZnS, The diffusion coefficients (Di) include a power-law dependence on temperature and are formulated in terms of the corresponding binary diffusion coefficients (gi1) into the majority component (H,) and a multicomponent diffusion correction term (y,).Di = g i l / y i = 53:1( T/273.1 5)'+'/yi. The derivation of eqn (4) and the form of yi are outlined in the Appendix. Use of eqn (4) implies a knowledge of Kc3 and its temperature dependence. The thermodynamic parameters for the components in eqn (3) have been assessed by the compilers of the JANAF tables," and values of AG; were calculated at 100 K intervals from 900 to 1500 K from AG; = AGg8- TA(fef) (6) where (fef) = - ( G g - e g 8 ) / T (7) (8) using JANAF data for A G g 8 and A(fef). Regression of these results with Tyielded AG;/J mo1-l = (- 90 333 f 1 10) + (49.3 f 0.1) T; (900-1 500 K).The uncertainty intervals are standard deviations of the regression coefficients. A recommended expression22 for equilibrium (3) over a much wider temperature range (298-1750 K) agrees with eqn (8) to within 0.2% at 900 K and to within 3.3 YO at 1444 K, the upper temperature limit in this work. The results from two independent experiments with ZnS(B) using capsules with different channel dimensions were pooled, K, extracted from eqn (4), and AG; regressed with T to yield an equation in the form of eqn (8). A statistical routine was used which identified outliers (data pairs for which the standardized residual > 12 I), which were then rejected, the rejection sequence terminating when r2 = 0.998 (n = 79).The initial r2 was 0.995 for all data points ( n = 92), where r = correlation coefficient. AGi(J)/J mol-1 = (374200f 1200)-(190.4+ 1.1) T ; 1010-1270 K (9)P. J . Gardner and P. Pang 1883 A similar procedure was adopted for the experiments with the a-phase when the results from three separate experiments with the same capsule were pooled (n = 199, r2 = 0.997), the final regression equation being based on n = 192. AGi(a)/J mol-1 = (3767OOk900)-(191.9kO.6) T ; 1300-1440 K. (10) In both cases the rejected data points were randomly distributed throughout the temperature range; the uncertainty intervals in eqn (9) and (10) are standard deviations of the regression coefficients. Assuming that these equations may be extrapolated to the transition temperature (1 293 K) which is equivalent to the approximation, AC$(lOl0-1293 K) = AC",(1293-1440 K) w 0 we derive AW(a+cr) 1293 = -2.5k1.5 kJ mo1-1 1 ASOW 3 a) 1293 = - 1.5 f 1.3 J K-l mol-'.J The heat capacities for ZnS(a) and ZnSW) at both high8 and lowi temperatures have been reported by the Bureau of Mines (U.S.A.) group (BM). The high-temperature work revealed no enthalpy of transition and the results in combination suggest (12) ASo(p -+ a ) 1300 z 0.4 J K-l rno1-l Our results are in satisfactory agreement with the BM study. for the reversible cell reaction, A recent of a high-temperature solid electrolyte cell yielded Gibbs free energies ZnS<a) + 3Cu,O(c) = ZnO(c) + 6Cu(c) + SO,(g) which, in combination with ancillary thermodynamic data, gave Over the common temperature range our results for AG; are ca.2 YO low compared with those from eqn (13), but the constants are quite different, suggesting internal compensation. A comparison of AG; from several sources is given in table 2. The constants A and B in eqn (9) and (10) correspond, in a second-law analysis, to A% and A S , respectively, at the mid-points of the experimenta1 temperature ranges. They may be converted to apply to the standard formation reaction (298 K), (14) using ancillary thermodynamic data for zinc, 26 sulphur'' and zinc sulphide.8 These second-law conversions yield Aq4(D> = Aq[ZnS,fl298 = - 190.2k4.6 kJ mol-' AHY4(a) = Aq[ZnS, a1298 = - 196.4's 4.0 kJ mo1-1 where the uncertainty intervals are f 2 standard deviations, the usual practice for The results were also processed by the third-law method.Here, the complete data set of h ( T ) was used (92 points for the p-phase and 199 points for the a-phase), each data pair yielding a value for K , from eqn (4). Eqn (1 5 ) was then used to derive a set of values for AH&': (1 5) Gibbs free energy functions for ZnS were calculated from the heat capacities and derived functions reported in the BM study.',' Corresponding functions for Zn(g) were taken from ref. (26) and for S,(g) from the JANAF compilation.18 Finally, combining with A%[&, g]298 = 128.49 & 0.291R kJ mol-I and AHp[Zn, g1298 = 137.74 & 0.4226 kJ rnol-', AG;W)/J mol-1 = 387000- 199T; 1180-1210 K. (13) Zn(c) + S(rhomb) = ZnS(a, D) A%. A%(a, p, 298) = TA[ - (G; - GQ8)/ TI - RT In K2.1884 Thermodynamics of Zinc Sulphide Table 2.A comparison of literature data for the equilibrium ZnS(a, p) = Zn(g) + $S,(g), AGi/J mol-I = A - BT ~ ~ - ~ _ _ _ _ _ ~ ~ ~ - ~~ -___~_ - temperature A B phase range/K /J mol-1 /J K-l m o t ' ref. 1180-1210 387000 199 23 1180-1210 391 200 20 1 24 P < 1293 374 300 193 25 P 1010-1270 374 200 I90 this work a 1323-1473 360 900 183 25 G! I 300-1440 376 700 192 this work P B Table 3. Comparison of second-law and third-law enthalpies of formation at 298 K ZnSO ZnS(a) ~ ~ ___ ______ ~ ~. - ~ - - ~ AHr/kJ rno1-l' 2nd law - 190.2 4.6 - 196.4 & 4.0 A q / k J mol-I, 3rd law - 194.k0.5" -196.3k0.5" AH"@-, a)/kJ mol-', 2nd law - 6.3 6.1 AHo(P -+ a)/kJ rnol-', 3rd law - 2.3 0.7 a The uncertainty intervals quoted for the third-law results are k 2 standard deviations. standard enthalpies of formation corresponding to eqn (14) were derived. These are compared with the second law values in table 3.by X.r.d. (see table l), Samelson and BrophyIg maintain that chemically pure ZnSV) always contains stacking faults, mostly arising from twinning. Aminoff and B r ~ o m P : ~ ~ have shown that a-ZnS exists in the vicinity of the interfacial planes of this cubic twin. The experiments of Baars and Brandt," referred to earlier in the discussion of our d.s.c. results, suggest that the concentration of the a-phase in a weakly disordered cubic phase could be ca. 20%. Our d.s.c. results show a feature that occurs in the same temperature interval (900-1020 "C} in which Baars et al.note annealing of the residual disorder in their weakly disordered cubic phase. Together, this evidence suggests that our transition data could refer to the process (O.SP+0.2a) -+ a rather than B + a. This introduces an additional uncertainty into the transition data and all results specific to the /?-phase. The error limits quoted reflect this additional uncertainty. The second and third law values for the a-phase are in good agreement, but for the P-phase the agreement is less good. The transition enthalpies agree within the quoted uncertainty intervals. The principal conclusion from this work is that the transition ZnS(P) + ZnS(a) is nearly athermal, both at the transition temperature and at 298 K. This is in agreement with the BM result^,'^^ but at variance with Kapustinskii and Chentzova's solution calorimetric study.6 Briefly, this Russian work consisted of measuring the enthalpies of reaction of the a- and P-phases in concentrated HCl using an adiabatic calorimeter.With the same final thermodynamic state in each case, the difference in the heats of reaction is equal to the heat of transition. The solution of the /?-phase was complete within lOmin, but the solution of the a-phase was so sluggish that the experiments were truncated when disolution was ca. 90% complete, the extent of reaction being determined by analysis. The uncertainty attendant on solution calorimetry with slow reactions is well established,28 and the truncation procedure must be regarded as of Notwithstanding the identification of our low-temperature phase as 95 + %P .J . Gardner and P . Pang 1885 doubtful validity. Also, when. this work was repeated by the BM group,29 it was the p- phase that proved difficult to and no results were reported for it. Their result by reaction calorimetry for Aq[ZnS, a1298 = 19 1.9 0.8 kJ mol-' differs by 4.4 kJ mo1-I from our third-law result. Appendix Under conditions of modified entrainment (see Experimental section) the partial pressure, pi, of a single substance, i, contained in the capsule and using a single carrier gas is given" by Pi = P(1 -exp (-ti)] (A 1) where ti = ci,RTI/(AMi P 9 J . (A 2) The term [ 1 - exp (- ti)] in eqn (A 1) may be approximated to ti to an accuracy of 5 '/o when t i < 0.1. Mass transport in the capillary is by diffusive and convective flow, and Cus~ler'~ has shown that this exponential approximation is equivalent to the assumption that transport is entirely diffusive (the dilute-gas approximation). Extension3' of the treatment to multicomponent systems is straightforward, and equilibrium constants may be formulated in terms of the partial pressures of each gas- phase component, the exponential approximation being invoked for algebraic tractability.For zinc sulphide entrainment only two of the eqn (1)-(3) are independent and it is convenient to write the mass balance and flux equations in terms of eqn (2) and (3). A coupling parameter, 6, is defined as the flux ratio in the capillary of Zn(g) to S,(g) (for transport in argon, 6 = 2 as the sublimation is congruent). Hence K, = ( S ) ' / ( D 4 [ D 3 6]0*5) The term ci, is readily eliminated from these equations, leaving a cubic in 6.Using eqn (8) to obtain K,, and literature data2, for K2, the cubic which has only one real positive root was solved exactly giving 6 = 1.4 x lo8 at 1000 K to 6 = lo3 at 1425 K. Consequently 6-2 was approximated to 6 in eqn (A 4) and then 6 is elimated from eqn (A 3) and (A 4) to give which is identical to eqn (4). The physical significance of the large values of 6 is that because equilibrium (3) lies to the right in our temperature range, the S, is transported mainly as H,S. In a system of more than two components the effects of multicomponent diffusion3, must be included. A second-order treatment of this effect3' permits the multicomponent diffusion coefficient (DJ to be written in terms of the corresponding binary diffusion coejficient (9J with respect to the majority component (hydrogen) D, = D(H,S-H,, S,, Zn) = 5B(H,S-H,)/y2 D, = D(S,-H,, H,S, Zn) = 9(S2-H,)/y3 D, = D(Zn-H,, H,S, S,) = g(Zn-H,)/y, (A 6) (A 7) (A 8)1886 where and Thermodynamics of Zinc Sulphide 6D( 1,4) D(3,4) D( 1,2) 0(3,2) (6- 2) (6+ 1) + (6+ 1) Y3 = I+[( D(k, I ) = (ML/M$-- 1 hl?TI(G+ 1) r. = a DiAM,P6' The expressions for yi contain both 6 and ti, which themselves are functions of Di.As Di are evaluated from a knowledge of yi, a reiterative procedure is used in which yi is set initially to unity, then refined until successive estimates differ by < 1 in lo4. Some typical values of yi are shown in table Al. Table A 1. Correction factors for multicom- ponent diffusion [see eqn (A 6) and (A 8)] 1010 1.0001 0.9998 1129 1.0006 0.9988 1225 1.0019 0.9962 1320 1.0048 0.9901 1428 1.0130 0.9735 Binary diffusion coefficients (9ij) may be calculated with good precision from a knowledge of Lennard-Jones potential parameters and kinetic theory.34 Lennard-Jones constants are available34 from experimental viscosity results for H,S and H,.The usual combining rules were used. For B(Zn-H,), independent experimental data exist35 in the range 690-1 120 K, and these were assumed to apply in the overlapping experimental temperature range of this work, 1010-1445 K. Estimates of 9(S2-H,), required for preliminary calculations, were obtained from Svehela's estimated36 Lennard-Jones potential parameters for S,. The following equations were derived as described above and refer to atmospheric pressure and 1000-1450 K.B(Zn-H,)/m2 s-l = 8.109 x lop5 (T/273.1 5)1.53y 9(H,S-H2)/m2 s-l = 6.004 x lop5 (77273.1 5)1.671 9(S2-H,)/m2 s-l = 3.961 x lop5 (T/273.15)1.702 The authors are indebted to Dr E. Charsley (Stanton Redcroft Ltd) for assistance with d.s.c., to Mr K. Knight (British Petroleum plc) for X.r.d. studies, to Mr C. Whitehead (City University) for m.s., to Dr K. Holder (B.P. plc) for helpful discussion and to B.P. plc for financial assistance.P. J. Gardner and P. Pang 1887 References I V. G. Hill, Can. Miner., 1958, 6, 234. 2 I. T. Steinberger, Progress in Crystal Growth and Characterization (Pergamon, Oxford, 1983), vol. 7, 3 D. D. Wagman, W. H. Evans, V. B. Parker, R.H. Schumm, I. Halow, S. M. Bailey, K. L. Churney and R. L. Nuttall, The NBS Tables of Chemical Thermodynamic Properties, J . Phys. Chem. Ref. Data, 1982, 11, suppl. no. 2. 4 F. D. Rossini, D. D. Wagman, W. H. Evans, S . Levine and I. Jaffe, ‘Selected Values of Chemical Thermodynamic Properties’, Natl Bur. Stand. Circular 500 (U.S. Govt Printing Office, Washington, D.C., 1952). 5 K. C. Mills, Thermodynamic Data jor inorganic Sulphides, Selenides and Tellurides (Butterworths, London, 1974). 6 A. F. Kapustinskii and L. G. Chentzova, C. R. Dokl. Acad. Sci. URSS, 1941, 30, 489. 7 J. M. Stuve, US. Bureau of Mines Report of Investigation 7940, 1974. 8 L. B. Pankratz and E. G. King, U.S. Bureau of Mines Report of Investigation 6708, 1965. 9 R. T. Struck and C. W. Zielke, Fuel, 1981, 60, 795.p. 7. 10 C. W. Zielke and W. A. Rosenhoover, U.S. Patent 44241 11, 1984. 11 D. Battat, M. M. Faktor, I. Garrett and R. H. Moss, J. Chem. Soc., Faraday Trans. I , 1974, 70, 12 B. de Largy, A. Finch and P. J. Gardner, J. Cryst. Growlh, 1983, 61, 194. 13 M. Farnsworth and C. H. Kline, Zinc Chemicals, Zinc Development Association, 1973. 14 S . D. Scott and H. L. Barnes, Geochim. Cosmochim. Acta, 1972, 36, 1275. 15 J. Baars and G. Brandt, J. Phys. Chem. Solids, 1973, 34, 905. 16 G. de Maria, P. Goldfinger, L. Malaspina and V. Piacente, Trans. Faraday Soc., 1965, 61, 2146. 17 P. Goldfinger and M. Jeunehomme, Trans. Faraday Soc., 1963, 59, 285 1. 18 D. R. Stull and H. Prophet, JANAF Thermochemical Tables, NSRDS-NBS 37, Catalogue No. 013.48.37 (U.S. Govt Printing Office, Washington, D.C., 1971) and supplements published in J. Phys. Chem. Ref. Data, 1974, 3, 311; 1975, 4, 1 ; 1978, 7, 793; 1982, 11, 695. 2267. 19 H. Samelson and V. A. Brophy, J . Electrochem. Soc., 1961, 108, 150. 20 V. L. Tanson and M. G. Abramovich, Mineral Zh., 1982, 4, 35. 21 K. Imamura and M. Senna, Mater. Res. Bull., 1984, 19, 59. 22 0. Kubaschewski and C. B. Alcock, Metallurgical Thermochemistry (Pergamon, Oxford, 1979). 23 S. C. Schaefer and N. A. Gokcen, High Temp. Sci., 1982, 15, 225. 24 A. W. Richards, J . Appl. Chem., 1959, 9, 142. 25 T. Rosenqvist and K. Tungesvik, Trans. Faraday Soc., 1971, 67, 2945. 26 R. Hultgren, P. D. Desai, D. T. Hawkins, M. Gleiser, K. K. Kelley and D. D. Wagman, Selected Values of the Thermodynamic Properties of the Elements (American Society for Metals, Metals Pk, Ohio, 1973). 27 G. Aminoff and G. Broome, 2. Kristallogr., 1931, 80, 355. 28 G. I. Birley and H. A. Skinner, Trans. Faraday Soc., 1968, 64, 3232. 29 L. H. Adami and E. G. King, U.S. Bureau of Mines Report of Investigation 6495, 1964. 30 E. L. Cussler, Diflusion : Mass Transfer in Fluid Systems (Cambridge University Press, Cambridge, 31 D. Battat, M. M. Faktor, I. Garrett and R. H. Moss, J . Chem. Soc., Faraday Trans. I , 1974. 70, 32 E. L. Cussler, Multicomponent Difusion (Elsevier, Amsterdam, 1976). 33 D. Battat, M. M. Fattor, I. Garrett and R. H. Moss, J. Chem. Soc., Faraday Trans. I , 1974, 70, 34 R. C. Reid, J. M. Prausnitz and T. K. Sherwood, The Properties of Gases and Liquids (McGraw-Hill, 55 P. J. Gardner and P. Pang, unpublished work. 36 R. A. Svehla, NASA Technical Report, R-132, Lewis Research Centre, Cleveland, Ohio, U.S.A., 1984), chap. 3. 2302. 2293. Orlando, 3rd edn, 1977). 1962. Paper 71977 ; Receiued 4th June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401879

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chern. Soc., Faraday Trans. I, 1988, 84(6), 1889-1896 Pairwise Interaction Parameters for Sodium, Potassium and Halide Ions in Aqueous Solutions Andres C. R. Antonini, Michael J. Blandamer,* John Burgess,* Andrew W. Hakin and Neil D. Hall Department of Chemistry, University of Leicester, Leicester LEI 7RH Alan H. Blandamer Bendix Aerospace, Arlington, Virginia, U. S. A . Practical osmotic coefficients for mixed aqueous solutions at 298.15 K and ambient pressure containing alkylammonium and alkali-metal salts are examined from the standpoint of a composite salt having the stoichiometry (1 - y ) M+yR,N+X-, where X- represents an anion. Practical osmotic coefficients for these aqueous solutions are estimated assuming that ion-ion interactions are satisfactorily accounted for in terms of charge-charge interactions.Differences between observed and calculated practical osmotic coefficients are assigned to interactions between cospheres surrounding ions in solutions. For alkylammonium and substituted alkylammonium ions these interactions are expressed in terms of group interaction parameters. Cosphere interaction parameters are reported for Na+, K+, Br- and C1- ions, together with interaction parameters for OH groups in substituted alkyammonium ions. Derived group interaction parameters are used to comment on trends in Setschenow coefficients for amides and volatile hydrocarbons in aqueous salt solutions. According to Gurney,' hydration of ions in aqueous solutions can be understood in terms of cospheres of solvent around each ion.Further, the structure of water in the cosphere around a given ion differs from the structure of bulk liquid water at the same temperature and pressure. A broad distinction2 is drawn between two classes of ions. In one class there is strong interaction between an ion and neighbouring water molecules, the interaction being electrostrictive in some cases and hydrogen bonding in others. In the second class water-water interactions are enhanced within cospheres around hydro- phobic ions. These considerations apply to ions in 'infinitely dilute' solutions, where ion-ion interactions are absent. With increase in concentration of salt so the mean distance of separation, d, between ions decreases3 For solutions containing a 1 : 1 salt having concentration mol dm-3, d is 94 x lO-'O m.In these dilute solutions the cospheres are not thought to interact. Therefore the impact of ion-ion interactions on the thermodynamic properties of these solutions can be understood in terms of charge-charge interaction^.^ With increase in concentration of salt, the distance d decreases until at lo-' mol dmP3 the mean ion separation is ca. 20 x m. Simul- taneously the thermodynamic properties of aqueous salt solutions reflect the growing importance of contributions from overlap of cospheres. 2, Friedman and Krishnan described5 the effects of overlap in diagrammatic form, contrasting the impact of cospheres where the structures are compatible and where they are incompatible. The above considerations have a direct bearing on our interests6 in the kinetics of chemical reactions in aqueous salt solutions, including chemical reactions involving ionic and neutral substrates.In the latter case we reported7 striking dependences on added salt of Setschenow coefficients for two neutral metal complexes, [Fe(bipy),(CN),] 18891890 Pair wise Interact ion Parameters and [Fe(5NO-phen),(CN),]. In order to understand from a quantitative standpoint these solubility and kinetic data we derived' pairwise interaction parameters characterising cospherexosphere interactions in aqueous solutions. As a starting point we examined* the properties of dilute aqueous solutions containing alkylammonium salts. Within a given salt solution containing a single salt MX (where M = cation and X = anion) there are cation-cation, anion-anion and cation-anion interactions, each interaction being characteristic of the ions involved.Two contributions to each thermodynamic property are identified : (i) chargexharge (or electrical) and (ii) cosphere overlap. We identify contributions (# - 1) (elect) and (#- 1) (cosph) to (4- 1) for the solution, where # is the practical osmotic coefficient. Previously* we used Pitzer's equationsg to calculate the contribution to (# - 1) from charge-charge interactions. The residual, [(# - l)(expt) - (# - l)(elect)] was used to calculate the p" parameter defined by P i t ~ e r , ~ which was then expressed in terms of a pairwise interaction parameter g,(salt) for a given salt MX. The analysis was applied* to thermodynamic data describing the properties of aqueous solutions containing alkylammonium salts. A detailed treatment of cosphere interactions was based on the Savage-Wood scheme.lo-12 In other words interaction between cospheres around two solute molecules X and Y is described by Gibbs function pairwise interaction parameters between group i and group j and the number of such groups in the two solute molecules.11, l2 Interaction parameters involving alkylammonium ions were described in terms of pairwise Gibbs function parameters for methylene groups and other constituent groups. Based on an incremental pattern in the number of methylene groups in the alkyl chains of R,N+ cations, related pairwise interaction parameters were calculated.' The latter included G(Br- t) Br-), G(Br- t) CH,) and G(CH, - N+). Further group inter- action parameters are reported here involving OH groups based on the thermodynamic properties of substituted alkylammonium salts, e.g.(HOCH,CH,),N+Br-. Limitation of the analysis' to alkylammonium salts is obviously unsatisfactory. In view of our aim to establish links with kinetic data, we were anxious to draw into consideration pairwise interaction parameters involving alkali-metal cations, e.g. G(K+t)K+) and G(K+t+Br-). A direct approach is not possible, there being more unknowns that knowns in the associated equations. In this paper we show how pairwise interaction parameters can be calculated from the thermodynamic properties of mixed aqueous salt solutions. The nub of the argument is centred on an aqueous solution prepared using, for example, NaBr and Bu,NBr.The solute is described in terms of a composite yBu,N+(l -y)Na+Br-, where y is the salt fraction (0 < y < 1.0). Analysis of Thermodynamic Properties for a Composite Salt The following treatment13 of mixed salt solutions is convenient for the present purposes. In effect a mixed aqueous salt solution is treated in terms of an aqueous solution containing a single composite salt. A given salt solution is prepared using 1 kg of water and rn, moles of a salt v:) MZ(m) vy) Xz(s) and m2 moles of a salt v',") NZ(n) vp) WU). The solution contains the composite salt, yv:) Mz(m) yv:) Xz(z)( 1 - y ) v:) NZ(n)( 1 - y ) v:) Uz(u), where y = m,/rn and rn = (m, +m2). The stoichiometric coefficient v for the composite salt is defined by v = yv:'+yvy'+(l -y)v'n2'+(1 -y)vE'.The chemical potential of the composite salt j in aqueous solution, pu,(aq;y; T ; p ) , at temperature T, pressure p and salt fraction y is given by eqn (2), in which we assume that ambient pressure is approximately the standard pressure, p" :A . C. R. Antonini et al. 1891 where v In Q = yv:) In by:)) +yvL1) in (yv:)) + (1 - y) vC) In [( 1 - y ) v:)] +(1 - y ) vf) In [(l - y ) vL2)]. (3) v In ( y * ) = yv:) In (y,+) +yvF) In (yx-) + (1 - y ) vf) In (7%') + (1 - y ) vf) In (7,-). (4) By definition, at fixed T, p and y , lim (m -+ 0) y+ = 1 .O. Similarly &' (aq;y; T ) is the standard chemical potential of the composite salt in aqueous solution at standard pressure, p , temperature, T, and constant salt fraction, y , where mo = 1.0 mol kg-l and y + = 1 .O.The chemical potential of water (= substance 1) in the aqueous solution is gken by ( 5 ) Following the procedures described previously,' practical osmotic coefficients of water in aqueous solutions are related to the molality of a 1 : 1 salt using the equation Also p@i;y; T ; P ) = k@; T ) - V#RTM, m. #- 1 = f4+PmXm,+C$,m,2. (6) The first term F+ in this virial equation7 for (q5 - 1) is related to the ionic strength I using the equation The second term, Pmx, is related by the equation. f4 = -A? (I/mo)i//[l + b(~/rn~);]. * (7) Pmz = P(O) + pl) exp [ - a(mj/mo)i] (8) to the molality mj through two quantities, PcO) and pcl); interactions between ions in solutions are characterised by pcO). The quantity pcO) was related' using the equation g(sa1t) = 2RT'O) (9) to a painvise interaction parameter for the salt, g(sa1t).In the next stage, g(sa1t) was expressed' in terms of group interaction parameters along the lines suggested for neutral solutes by Savage and Wood1°-12 and by Lilley and c o ~ o r k e r s . ~ ~ - ~ ~ The calculations were carried out using a FORTRAN program in which separate subroutines assembled data files appropriate to the calculation of different sets of interaction parameters. Results Substituted Alkylammonium Ions The form of the analysis is illustrated by reference to practical osmotic coefficients for aqueous solutions containing substituted alkylammonium ions. The input to the analysis comprised practical osmotic coefficients at 298.15 K and ambient pressure over the range 0 < m(salt)/mol kg-l < 2.0 for the bromide and fluoride salts of (HOCH,CH2),N+.l7 These data were combined with practical osmotic coefficients over the same range of molalities and the same conditions for aqueous solutions containing the chloride and bromide salts of (HOCH2CH,)(CH3)3N+;18 these (P-hydroxyethy1)tri- methylammonium salts are sometimes referred to as choline salts.The derived g(sa1t) parameters were expressed in terms of group interaction parameters.' The scheme for (HOCH,CH,),N+Br- is set out in table 1, which is based on the lower left-hand part of a symmetric interaction matrix. A similar table was constructed for choline salts in which terminal methyl groups were expressed as 1.5 x methylene groups, following the1892 Pair wise Interaction Parameters Table 1.Pairwise group interaction parameters for (HOCH,CH,),N+Br- basis : (HOCH,CH,),N+Br- = (4 x OH) + (8 x CH,) + N' + Br- g(sa1t) = 16 x G(OH ++ OH) + 64 x G(OH t-) CH,) + 64 x G(CH, t) CH,) + 8 x G(OH ++ N') + 16 x G(CH, * N') + G(N+ ++ N') + 8 x G(OH ++ Br-) + 16 x G(CH, ++ Br-) + 2 x G(N+ ++ Br-) + G(Br- ++ Br-) Table 2. Pairwise group interaction parametersa calculated from the thermodynamic properties of substituted alkylammonium salts HO CH2 " F- C1- Br- - - - - - HO -23b -280 (30) 255' -871" N+ F- - 54 (23) 216' - 1153" - 576" C1- -214 (40) 83" - 498" Br- - 12 (40) 69" - 669' - - - - 29b - 34b - - - CH2 - - - - 249" - - 225" - - a In J kg-'; standard errors are reported in brackets after derived parameter. " Ref. (8). Ref. (1 1). Table 3. Mixed aqueous salt solutions composite salt = (1 - y) K+ yPr,N+ Br- g(sa1t) = (1 -y),G(K+ t) K+) = (1 - y ) K++(14 x y x CH,)+O, x N+)+ Br- + 28y( 1 -y) G(K+ ++ CH,) + 196y2G(CH, ++ CH,) + 2y( 1 - y ) G(K+ ++ N+) + 28y2G(CH, * N+) +y2G(N+ +--) N+) + 2( 1 - y)G(K+ ++ Br-) + 28yG(CH, * Br-) + 2yG(N+ t--) Br-) + G(Br- c) Br-) suggestion of Wood and co-workers.10-12 The estimates reported by the latter authors were used for the pairwise interaction parameters G(OH c-) OH) and G(OH c-) CH,).Pairwise group interaction parameters involving N+, Cl-, Br- and F- groups were taken from previously reported calculations. Consequently these remained four unknowns to be calculated from the properties of four salt solutions. The derived parameters are G(OH - N'), G(OH c-) Br-), G(OH +-, F-) and G(OH c-) Cl-). These estimates are summarised in table 2, which includes, for comparison, several previously determined interaction parameters.Mixed Salt Solutions A limited but important set of informationlg> 2o refers to the practical osmotic coefficients for aqueous solutions containing mixtures of alkylammonium and alkali-metal salts. These mixtures comprising 1 : 1 salts having a common anion can be described in terms of the composite salt yR,N+-(l -y)M+Br-. The starting point is the dependence of (q5 - 1) on the molality of the mixed salt at fixed temperature, pressure and salt fraction y . The derived [cf. eqn (6)--(9)] pairwise interaction parameter, g(salt), is expressed inA . C. R. Antonini et al. 1893 Table 4. Pairwise interaction parametersa G(K+ - j ) involving K' cations CH2 N+ K+ F- c1- Br- 9.4 (32) 295 (401) 196 (127) 323 (91) 174 (95) 190 (57) ' Standard errors are given in parentheses.Table 5. Pairwise interaction parameters,' G(Na+ WJ'), involving Na' cations -5.3 (28) 339 (320) " Na+ 31 (110) Br- 401 (56) CH2 ~~ ~ a Standard errors are given in parentheses. terms of pairwise group interaction parameters along the lines set out in table 3 for the system yPr,N+ * (1 - y ) K+Br-. In this particular case, Wen and co-workers2' reported practical osmotic coefficient for eleven solutions where 0 < y < 1 .O at intervals of 0.1. These data were combined with comparable data for aqueous salt solutions containing K+Br- + Me,N+Br-, Et,N+Br- + K+Br-, Bu,N+Cl- + K+Cl- and Bu,N+F- + K'F-. For the latter five systems, Wen and coworkers2' reported data for solutions where y = 0.5.In the calculations we used data for solutions where the ionic strength c 2.0 mol kg-'. Within the equations for g(salt), many of the pairwise interaction parameters are knowns, " 9 l2 (cf. table 2). The unknowns emerging from the data reported in ref. (20) are the pairwise interaction parameters involving K+ ions. In effect there are at this stage in the calculation 17 equations involving six unknowns. The latter are listed in table 4, together with their estimates and standard errors calculated in the manner previously described.' The calculations were repeated using the data from ref. (19) (contained in the deposited supplementary material) describing the thermodynamic properties of mixed aqueous solutions containing alkylammonium and sodium salts, i.e.Me,N+Br- + Na+Br- and Pr,N'Br-+Na+Br- solutions. The outcome was 14 equations in four unknown pairwise group interaction parameters involving Na+ cations. The derived parameters are summarised in table 5 . Discussion The main thrust of this paper has been to establish procedures for using thermodynamic properties of mixed salt solutions in calculating pairwise group and ion-ion interaction parameters for salt solutions. We have assumed that at constant ionic strength, the contribution made by chargexharge interactions to deviations from ideal are constant.1894 Pairwise Interaction Parameters 230 0 2 5 021 0 2 4 0 20,26 0 29 032 0 30 0 31 I I -20 -10 0 10 20 lo2 kL0 (obsd) Fig. 1.Comparison of observed and calculated Setschenow coefficients k r for volatile solute U in aqueous salt solutions at 298.15 K: (1) CH, in NH,Br(aq); (2) C,H, and (3) C,H, in NH,Br(aq); (4) CH,, (5) C,H,, (6) C,H, and (7) C,Hlo in Me,NBr(aq); (8) CH,, (9) C,H6, (10) C,H, and (11) C,Hlo in Et,NBr(aq) ; (1 2) CH, (1 3) C2H6 (14) C,H, and (1 5) C,H,, in Pr,NBr(aq) ; (1 6) CH,, (1 7) C,H,, (18) C,H, and (19) C,H,, in Bu,NBr(aq); (20) CH,, (21) C,H,, (22) C4HI0, (23) C,H,, (24) C6H,, (25) toluene, (26) C,H,, (27) ethylbenzene in NaCl(aq); (28) butane in KCl(aq); (29) CH,, (30) C,H,, (31) C,H,, and (32) C2H6 in KI(aq). The straight line signals perfect agreement between theory and experiment. This assumption is valid21 within the domain of the Debye-Huckel limiting law.In these terms we have assumed that the cosphere overlap contribution to the non-ideal properties of a salt solution can be used to calculate these interaction parameters. This extrathermodynamic treatment of solute-solute interactions is based on the realistic assumption that there is an underlying pattern to the deviations from ideal in the thermodynamic properties of salt solutions. For example, we assume that K+ ++ K+ interactions in aqueous salt solution are independent gf counter-anion. The calculated pairwise interactions are broadly in agreement with expectation based in part on an intuitive approach to ion-ion interactions. The interaction parameters involving OH groups and halide ions (table 2) are negative, indicative of hydrogen- bonding interactions, 0-H - - - X-.The small estimate for K+ ++ CH, and Na+ t) CH, interactions are a little surprising. The positive estimate for cation-anion interaction parameters points to a strong repulsion between cospheres, indicating that their structures are incompatible. The strongest repulsion is calculated for interaction between cospheres around K+ and F- ions, pointing to an incompatibility between theA . C. R. Antonini et al. 1895 structure-broken B zone around the K+ ions and electrostrictive hydration shells around the F- ions. The large positive interaction parameter describing overlap between Na+ and Br- points to strong repulsion between structure-broken B-zones around these ions. The large standard errors on some of the derived parameters is a little disappointing, although not out of line with some previously reported estimates of pairwise interaction parameters.These estimates can be in part supported by their application to other data. We report in fig. 1 an updated version of the figure in ref. (8). Included in the new version are observed and calculated Setschenow coefficient^^^-^^ for a range of volatile solutes including saturated and unsaturated hydrocarbons, together with aromatic derivatives in aqueous solutions containing sodium chloride, potassium chloride and potassium iodide.25 The basis of the calculation is the equation (10) kl'(U; MX) = (2/2.303RT)[G(M+ * U) + G(X- t* U)] in which substance U is a volatile solute in the presence of M+ cations and X- anions; k'O(U; MX) is the Setschenow coefficient calculated to logarithm base 10.The calculated pattern is in fair agreement with that observed. At the very least, the outcome points to a convenient method for predicting whether a given salt will salt-in or salt-out a particular solute. In these terms the derived interaction parameters have practical significance rather than being the outcome of some purely arithmetic exercise. We will discuss later applications of these parameters to kinetic data. At this stage, the derived parameters are further tested by examining differences in Setschenow coefficients as a function of salt and solute. An interesting example concerns amide-salt interactions in aqueous solutions.26 This is an important subject in view of its links with salt-peptide interaction^.^'-^^ In the present context we represent N-methylacetamide (NMA) by (3 x CH, + CONH) and N-methylpropionamide (NPA) by (4 x CH, + CONH).Hence in aqueous solutions containing M+X- [cf. eqn (1 O)] Ak"" = k'"'(NPA; M+X-)-k'l')(NMA;M+X-) = (2/2.303RT) [G(CH, c) M+) + G(CH, c) X-)]. (1 1) Hence Ak'lO) can be calculated knowing the relevant pairwise interaction parameters. For NaCl the observed Ak'l') is 0.065 at 298 K, both amides being salted out. Using the parameters given in table 5, Ak(l0) is estimated at 0.027. Similarly for NaBr, the calculated Ak'l') is 0.022, compared to the observed coefficient, 0.053. Thus the pairwise interaction parameters predict the correct sign and order of magnitude for the salting- out of these two amides, together with the increase in Ak'") on going from NaCl to NaBr.The agreement is poor between observed and calculated Ak"') for NaI solutions, 0.038, (observed) and - 0.041 (calculated). We commented on the unexpectedly large negative estimate for G(CH, c) I-) reported earlier.' This estimate was calculated using practical osmotic coefficients for aqueous solutions containing alkylammonium iodides. We are considering alternative methods of calculating this important quantity. Nevertheless we have observed recently3' that added iodides have a striking effect on the kinetics of reaction between hydroxide ions and bromophenol blue in aqueous We defer comment at this stage except to note that the pattern does not seem out of line with our estimate of pairwise interaction parameters involving iodide ions.We thank the S.E.R.C. for an award (to A. W. H). References 1 R. W. Gurney, Ionic Processes in Solution (McGraw-Hill, New York, 1953). 2 M. J. Blandamer, Q. Rev. Chem. SOC., 1970, 24, 169. 3 R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworths, London, 2nd edn (revised) 1963), table 1.3.1896 Pairwise Interaction Parameters 4 M. J. Blandamer, Adv. Phys. Org. Chem., 1977, 14, 203. 5 H. L. Friedman and C. V. Krishnan, in Water: A Comprehensive Treatise, ed. F. Franks (Plenum Press, 6 J. B. F. N. Engberts, M. J. Blandamer, J. Burgess, B. Clark and A. W. Hakin, J . Chem. Soc., Faraday 7 M. J. Blandamer, J. Burgess and J. C. McGowan, J. Chem. Soc., Dalton Trans., 1980, 616. 8 M. J. Blandamer, J. Burgess, M. R. Cottrell and A. W. Hakin, J . Chem.Soc., Faruday Trans. I , 9 K. S. Pitzer, Activity Coeficients in Electrolyte Solutions, ed. R. M. Pythkowicz (CRC Press, 10 J. J. Savage and R. H. Wood, J . Solution Chem., 1976, 5, 733. 11 J. J. Spitzer, S. K. Suri and R. H. Wood, J. Solution Chem., 1985, 14, 571. 12 S. K. Suri and R. H. Wood, J. Solution Chem., 1986, 15, 705. 13 M. J. Blandamer, J. Burgess and A. W. Hakin, Thermochim. Acta, in press. 14 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J . Chem. Soc., Faraday Trans. I , 1985, 81, 2191. 15 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J. Solution Chem., 1986, 15, 99. 16 H. E. Kent, T. H. Lilley, P. J. Milburn, M. Bloemendal and G. Somsen, J . Solution Chem., 1985, 14, 17 W-Y. Wen and S. Saito, J . Phys. Chem., 1965, 69, 3569. 18 G. E. Boyd, A. Schwarz and S. Lindenbaum, J . Phys. Chem., 1966, 70, 821. 19 D. Rosenzweig, J. Padova and Y. Marcus, J. Phys. Chem., 1976, 80, 601. 20 W-Y. Wen, K. Miyajima and A. Otsuka, J. Phys. Chem., 1971, 75, 2148. 21 G. N. Lewis and M. Randall, Thermodynamics, rev. K. Pitzer and L. Brewer (McGraw-Hill, 22 R. Battino and H. L. Clever, Chem. Rev., 1966, 66, 395. 23 E. Wilhelm, R. Battino and R. J. Wilcock, Chem. Reu., 1977, 77, 219. 24 H. L. Clever, J. Chem. Eng. Data, 1983, 28, 340. 25 T. J. Morrison and F. Billett, J. Chem. Suc., 1952, 3819. 26 E. E. Schrier and E. B. Schrier, J. Phys. Chem., 1967, 71, 1851. 27 E. R. Stimson and E. E. Schrier, J. Chem Eng. Data, 1974, 19, 354. 28 E. E. Schrier and R. A. Robinson, J . Solution Chem., 1974, 3, 493. 29 E. Eagland, ref. (5), 1975, vol. 4, chap. 5. 30 M. J. Blandamer, J. Burgess, A. W. Hakin and F. Sanchez, unpublished work. 31 E. S. Amis and V. K. Lamer, J . Am. Chem. Soc., 1939, 61, 905. New York, 1973), vol. 3, chap. 1. Trans. 1, 1987, 83, 865. 1987, 83, 3039. Boca Raton, Florida, 1979), vol. 1, chap.7. 101. New York, 1961), chap. 34. Paper 7/998; Received 8th June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401889

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . (7hern. Soc., Faraday Trans. I , 1988, 84(6), 1897-1910 Activity Measurements and Spectroscopic Studies of the Catalytic Oxidation of Toluene over Silica-supported Vanadium Oxides Bo Jonson," Bernd Rebenstorf and Ragnar Larsson Research Group on Catalysis, Inorganic Chemistry 1, Chemical Center, University of Lund, P.O. Box 124, S-221 00 Lund, Sweden S. Lars T. Andersson Research Group on Catalysis, Department of Chemical Technology, Chemical Center, Lund Institute of Technology, P.O. Box 124, S-221 00 Lund, Sweden Vanadium oxides on silica gel (Davison 952) with varying vanadium concentrations have been investigated for the activities and selectivities in the oxidation of toluene. These results were correlated with the features shown by X.r.d., ESCA, u.v.-visible and i.r.studies of adsorbed CO. The catalysts show a low activity in comparison with V/alumina. The support alone shows 60% selectivity for benzaldehyde formation. From i.r. data it is suggested that this benzaldehyde is formed on sites in association with sodium impurities. Introduction of vanadium increases the rate of reaction but the selectivity is shifted towards carbon oxides. The increased activity at low vanadium content is due to isolated four-coordinated vanadium species. U.v.-visible data show that the isolated species agglomerate at further loading to polyvanadium chains. These polymers possess an increased activity as a further route of reduction of Vv to VIV is present. Five- coordinated vanadium of the same activity as the chains and producing selective oxidation products other than benzaldehyde is found with a loading of 2 YO V. Larger agglomerates and V,O, crystallites are present on the 10% V catalyst.These species only give a somewhat increased activity, but with a product pattern different from crystalline V,O, and 10% V/alumina. Supported vanadium oxide catalysts have been subjected to a large number of studies and considerable efforts have been made to elucidate the structure of the surface species and the catalytic properties. The detailed nature of the surface species, and in particular of the surface vanadium compounds, is, however, still unknown. To further the knowledge on the structure of vanadium oxide catalysts, we are presently investigating catalysts based on different supports and with varying loading of the active phase.Recently we reported results on vanadium oxides supported on alumina,' and in the present work results on silica as support are reported. Experimental Catalyst Preparation The catalysts were prepared by adding aqueous solutions of [NH,VO,] to silica gel. The silica used was Davison 952, for which the manufacturer reports a surface area of ca. 350 m2 8-l. The samples were dried at 393 K in air and calcined at 773 K for 6 h in air (which we will call fresh catalysts). For the i.r. and u.v.-visible studies the catalysts were evacuated and heated in oxygen at 773 K for 1 h (oxidized form) or reduced at 773 K 18971898 Catalytic Oxidation of Toluene with CO for 1 h (reduced form). The concentration of vanadium is given in wt % defined as (g V/g SiO,) x 100.Electron Microscopic Studies Scanning electron microscopic investigations were performed with an ISI- 1 OOA instrument equipped with EDAX (energy dispersive analysis of X-rays). Gas Adsorption Measurements A gravimetric B.E.T. apparatus was used for measuring the adsorption of N, at 77 K. Prior to adsorption the samples were outgassed overnight at a temperature of 573 K and a pressure of 6 x lop6 Torr.? The B.E.T. method was used for calculation of the surface area. The pore-size distribution was calculated from the adsorption and desorption branch by the method of Dollimore and Heal., ESCA Measurements ESCA measurements were performed on an AEI ES200B electron spectrometer equipped with an A1 anode (1486.7 eV). The full width at half maximum of the Au 4f,/, line was 1.8 eV.Charging effects were corrected for by giving the Si 2p peak from the support a binding energy (E,,) of 103.9 eV, as determined previously by the Au calibration techniq~e.~ For the quantitative estimations, peak areas above a linear baseline were measured with a planimeter. The V2p intensity was estimated by measuring the area of the partly overlapping peaks of V 2p,,, + V 2p3,: + 0 1s K,,,,. The 0 1s K,,,, intensity was estimated from the 0 1s K,,., intensity by multiplying with the 0 1s Ka3,,/0 1s K., 'L ratio measured on the pure support. The V 2p3,, intensity was finally obtained by subtract- ing the estimated 0 1s K,,,, intensity from the V 2p,/, + V 2p3,! + 0 1s K,,,, intensity, followed by multiplication by 0.67.The calculation of theoretical I(V 2p3,,)/1(Si 2p) intensity ratios was performed assuming monolayer formation. The equations used earlier for V/Al,03 catalysts were applied here also.' The intensity ratio for pure samplcs [P(V 2~1,,,)/1~(Si 2p)l was calculated to be 2.05. The monolayer thickness, a = 2.34 A, was calculated from thc density of V,O,. The electron mean-freeopaths for V 2p3/, in V,O,, 3Lv(V205) = 25.3 A, and for Si 2p in V,O,, Lsi(V2O5) = 29.1 A, were calculated by the equation for inorganic samples given by Seah and D e n ~ h . ~ I.R. Measurements Infrared measurements of adsorbed CO were performed with self-supporting discs placed in an i.r. cell which has been described previ~usly.~ Spectra were taken (128 scans) on a Nicolet 20 SXC Fourier-transform i.r.spectrophotometer. The resolution was 2 cm-'. Surface area, X-ray diffraction and diffuse reflectance spectroscopy (d.r.s.) measure- ments were performed in the same way as in the previous study on alumina-sup- ported vanadium surface compounds. ' Activity Measurements A conventional flow apparatus operated at atmospheric pressure was used as described previous1y.l The gas flow was 15-46 dm3 h-l. Toluene was introduced with saturators, giving a toluene concentration of 1.1-1.3 vol % in an oxygen-nitrogen mixture (20/80). The catalyst bed contained 0.1-3 g of catalyst of 100-200 pm particle size and was t 1 Torr = 101 325/760 Pa.J. Chern. Soc., Faraday Trans. I , Vol. 84, part 6 ( a ) Plate 1 Plate 1. Electron micrographs of V/silica catalysts. Magnification of 5000.(a) Silica, D952, 10&200 pm, tilt 25"; (b) 2 % V/SiO,, D952, 10&200 pm, tilt 25"; (c) 10% V/SiO,, tilt 25". B. Jonson ct a/. (Fucing p . 1899)B. Jonson et al. 1899 operated isothermally at 673 1 IS. The 10 % V/SiO, catalysts were diluted with quartz to avoid adverse thermal effects. The analytical methods have been described in detail elsewhere.6 The change in the catalytic performance with time on-stream was small for all catalysts, and no colour change indicating coke formation was observed. The reactor was operated with conversions between 0.06 and 4.6 Yo and the rates were obtained from the differential data by division with W/F (g h mmol-l). Mass-transport effects were of no importance for the catalysts with low loading of vanadium because of their very low activity.The possible influence of diffusion for 10% V/silica was checked by taking measurements at different gas velocities. No effect on the selectivities was observed, but the rate for toluene oxidation varied by 20%. Results and Discussion ESCA, X.R.D. and Morphology Studies The surface composition of the 100-200 pm catalyst particles was studied by ESCA. No pure V 2p3/2 core-line spectra could be obtained owing to the overlap with the K,,,, satellite of the 0 1s line. The V 2p3/2 peak could be distinguished clearly only in the case of 10 % V/silica (fig, 1). The V 2~~~~ binding energy was estimated to be ca. 517.0 eV for this catalyst, indicating the predominance of the pentavalent state. This is slightly higher than the 516.6 eV observed for pure V205,7 but similar binding energy shifts have been observed earlier1~s~9 and may be due to the presence of aggregates which undergo considerable interaction with the support.The V 2p3/JSi 2p intensity ratio as a function of the vanadium concentration in the V/silica catalysts is shown in fig. 2. The intensity ratio increases with loading. For comparison, a calculated curve for the hypothetical case of VO, units in a monolayer uriiformly distributed in all pores of the particle is also shown. Full coverage occurs at 24.6% V for a silica surface area of 314.5 m2 8-l. The difference between these two curves indicates that on increasing the vanadium loading, a greater fraction of the vanadium is present on the surface of the silica particles.There is thus not a homogeneous distribution of vanadium throughout the silica particles. The X.r.d. measurements on the catalysts showed that crystalline materials were only present at the highest loading. Thus eight peaks between (28 of 15 and 35") due to V,O, were detected for the 10% V sample, but not at any other concentrations. The same observations were made in the electron microscopic work, where long, flat crystals and thin, needle-like crystallites were found on the surface on 10% V/SiO, (see plate 1). EDAX confirmed the vanadium content of the needles, whereas the surrounding support gives both vanadium and silicon peaks. The presence of a considerable amount of the vanadium in 10% V/SiO, in the form of crystallites is of importance for the interpretation of the ESCA intensities.For the abovementioned sample a large portion of vanadium inside the crystallites is not detected in the ESCA analysis due to the low escape depth of the photoelectrons. The measured intensity ratio Z(V 2p3/&/1(Si 2p) (see fig. 2) is, on the contrary, higher than that calculated for a uniform distribution. It is therefore suggested that a considerable amount of the vanadium in 10% V/SiO, is present on and near the surface of the 100 pm catalyst particles. A gross estimation indicates that 90 YO of the intensity originates from a 50 A thick surface layer. The ESCA analysis thus gives informatiof;l from the external surface of the particles and pore surfaces down to a depth of 50 A (deeper in the case of free line-of-sight). The decrease of the surface areas of the V/Si02 catalysts with increasing vanadium content is shown in fig.3. The surface areas per gram of catalyst and per gram of silica are both shown. The latter would not be affected by fractions of a monolayer of vanadium oxide on the silica support. If the vanadium oxide were present as a separate1900 Catalytic Oxidation of Toluene (U 1 ( b l ( C l i L 5 30 525 5 20 515 510 binding energy/eV Fig. 1. Photoelectron spectra of the V 2p binding energy region for V/silica catalysts: (a) 10 YO, (6) 2 %, ( c ) 0.5 %, ( d ) 0.2 YO, (e) 0.1 YO, (f) silica. crystalline phase without pore blocking, one would expect it to contribute some additional area per gram of silica. A slight increase in surface area per gram of support was found for V/alumina.' Neither of these cases occurs for the V/silica catalysts, for which the surface area decreases from 314.5 m2 g-' for silica itself to 192.6 m2 g-' of the 10% V catalyst.This large effect indicates blocking of pores by deposited vanadium oxide and eventually sintering or agglomeration of silica primary particles. It is interesting that the first 0.2 % V gives a large reduction in surface area from 3 14.5 toB. Jonson et al. 1901 5 6 7 8 9 10 0 1 2 3 4 V concentration (wt 7%) Fig. 2. I(V 2 ~ ~ , ~ ) / 1 ( S i 2p) intensity ratios from ESCA measurements versus vanadium con- centration in V/silica catalysts: x , experimental data; 0, calculated data. 0 2 4 6 0 10 V concentration (wt %) Fig. 3. Surface areas of V/silica catalyst : x , m2 per g of catalyst; 0, m2 per g of silica.262.4 m2 g-l. Vanadium loadings between 2 and 10 YO also result in a large reduction in surface area. However, introduction of additional vanadium from 0.2 to 2 % results in a slightly increased area per gram of silica, indicating formation of vanadium oxide particles without pore blocking. A tentative interpretation is that the first vanadium species are highly dispersed surface- bonded species, some of which interconnect silica primary particles to give a big reduction of the surface area. Intermediate loading gives in addition small oxide particles or agglomerates, whereas high loadings also give blocking of pores in addition to the effects just mentioned. The above results are in qualitative agreement with the pore structures of the silica support. The pore distribution is bimodal, with one wide maximum at a mean pore diamFter of ca.290 A. The contributions to the maximum arose mainly from 450 to 100 A pores, and these constitute ca. 70 O h of the total surtace area. The main part of the remaining area arose from pores of between 25 and 10 A.1902 Catalytic Oxidation of Toluene x absorption ;, I \ I \ I wavenum ber/cm-' Fig. 4.1.r. spectra of adsorbed CO (10 kPa) at room temperature. The absorbance scales have been varied in order to illustrate better the different V/silica samples. The vanadium concentrations and scales are: (a) 10 %, x = 0.02; (b) 2 %, x = 0.03 ; (c) 0.5 %, x = 0.01 ; ( d ) 0.2 %, x = 0.01 ; (e) 0.1 %, x = 0.001. I 2220 2200 21 80 2160 2140 wavenumberlcm-' Fig.5. 1.r. spectra of adsorbed CO (10 kPa) at 153 K. Scales in spectra: cf> 0% V, x = 0.02; (b) 2% V, x = 0.05. The other spectra as in fig. 4 with x = 0.02.B. Jonson et al. 1903 I.R. Spectra of Adsorbed CO Fig. 4 and 5 illustrate i.r. spectra from CO adsorbed on the reduced catalysts. On the other hand, no adsorption on vanadium surface compounds was detectable for the oxidized samples. Adsorption at room temperature (fig. 4) on the 0.1 % V sample results in a weak band around 2183 cm-l [fig. 4(e)]. Absorption bands at these wavenumbers have been assigned to CO adsorbed on V"' surface compounds on silica gel.'' The 0.2 YO V catalyst has a more distinct, but still asymmetrical band arising from adsorbed CO [fig. 4(d)]. This band is found at 2185.3 cm-'.The CO band shifts with increasing vanadium loading to 2186.3 cm-l for the 0.5 % V sample [fig. 4(c)] and to 2187.7 cm-' for 2 % V [fig. 4(b)]. Reduction of the 10 YO V catalyst resulted in a disc of very low transmittance. The spectrum of adsorbed CO [fig. 4(a)] suggests a band of about the same position as that above. No adsorption was found for the support under these conditions. One observes a trend of increasing wavenumber with increasing vanadium concentration for the: room-temperature spectra. Spectra of CO adsorption at 153 K are shown in fig. 5. The support alone gives rise to three bands [fig. 5 0 1 . The bands at 2156.1 and 2136.8 cm-' are assigned to CO adsorbed on surface silanol groups'l and to CO clusters, respectively. The 2171.9 cm-l absorption most likely arises from CO adsorbed on sodium impurities of the silica gel.This is concluded from a previous study on CO adsorbed on metals on silica, where CO adsorption on sodium gave rise to a band at 2173 cm-1.*2 This conclusion is also supported by the fact that NaOH is used by the manufacturer for the preparation of the silica gel. Introduction of vanadium (0.1 O/O) causes an additional shoulder at 2185.7 cm-l [fig. 5(u)]. One also observes that the absorption around 2172 cm-' now appears as a shoulder. An increase of the vanadium content to 0.2 YO gives rise to bands at 2186.5 and 2170.9 cm-l besides those from CO adsorbed on silanol groups and CO clusters [fig. 5 (41. The former two bands might arise from V'I'-(CO), surface complexes analogous to those previously found." More likely, however, is the 2170.9 cm-' band due to CO adsorbed on sodium ions which are still present on the surface.The shift of this band should then indicate that some of the original surface sites of sodium are now occupied by vanadium ions and the sodium ions have migrated to other sites. A further increase of the vanadium concentration to 0.5 YO shifts these two bands to 2187.8 and 2173.0 cm-I and reverses the intensities [fig. 5(c)]. One also notices a decreased intensity and a shift to 2157.2 cm-l of the CO band arising from the silanol groups. The spectrum of the 2 */6 V sample has an asymmetrical band at 2 188.9 cm-l and an absorption at 2 1 57.4 cm-l (20 adsorption on 10 YO V/SiO, gives rise to one pronounced band at 2189.4 cm-l [fig.5(a)]. A weak band can be found at 2212 cm-l. This band can be assigned to CO adsorbed on VrV surface compounds. This is concluded from the known i.r. bands of CO adsorbed on V1ll/SiOzlo and the general fact that these i.r. bands are shifted to higher wavenumbers with increasing oxidation state of the metal. In another paper on vanadium surface we assigned a CO i.r. band at 2150 cm-l to VII surface species, i.e. a shift of 35 cm-l towards lower wavenumbers. In the present work a shift of 28 cm-l is observed for VIV. The absence of CO bands around 2157 cm-l (due to adsorption on silanol groups) must mean that the chemically active silica surface is covered completely by vanadium compounds at this loading. Also in low-temperature spectra there is a shift towards higher wavenumbers of the peak due to V"' surface compounds.This is indicative of decreasing electron density of the vanadium surface compounds with increasing vanadium concentration. On V/aluminal bands from CO adsorbed on two types of vanadium surface sites at room temperature (2200 and 2 178-21 85 cm-') and low temperature (2 190 and 21 78-2 18 1 [fig. 5(b)i.1904 Catalytic Oxidation of Toluene 10 20 30 40 50 wavenum ber/ 1 O3 cm-' Fig, 6. Diffuse reflectance spectra of oxidized samples. The scale of F(R',) has been varied to show samples with different vanadium concentrations. (a) 10 YO V, x = 10; (b) 2 YO, x = 5 ; (c) 0.5 YO, x = 5 ; ( d ) 0.2%, x = 5 ; (e) 0.1 YO, x = 5. cm-') were observed. From these results one can conclude that sites of different nature are formed on silica and alumina.Diffuse Reflectance Spectra The diffuse reflectance spectra of the oxidized samples are shown in fig. 6. A charge transfer (c.t.) band at 42000 cm-l and a shoulder at 47500 cm-l are found for the 0.1 YO V sample [fig. 6(e)]. The c.t. band shifts towards lower wavenumbers and the shoulder vanishes with increasing vanadium concentration. The c.t. band is found at 41 000, 38500 and 35000 cm-l for the 0.2 [fig. 6(d)], 0.5 [fig. 6(c)] and 2% [fig. 6(b)] V catalysts, respectively. The shift towards lower wavenumbers is interpreted as an increasing agglomeration of vanadium species with increasing loading. The 10 YO V sample has an absorption at 29 500 cm-l and a shoulder around 21 000 cm-l [fig. 6(g)]. This shoulder indicates the presence of V,O,.One also observes a weak d-d band around 9500 cm-' in spectrum (a), i.e. formation of agglomerates containing vanadium with oxidation number less than five occurs at a loading of 10 YO. Absorptions around the wavenumbers of the shoulder have been claimed to be due to VV-Vrv intervalence charge transfer.14. l5 Reduction of the catalysts with CO results in the appearance of d-d bands in addition to c.t. bands (fig. 7). The samples with low vanadium content (0.1 and 0.2%) have c.t. bands at 42000 cm-l [fig. 7(d) and (e)]. There is a weak absorption at 16500 cm-' in the spectrum of 0.1 Yo V/SiO,, which increases in intensity and shifts to 17500 cm-' at 0.2 YO V [fig. 7 ( d ) ] . New spectral features appear with a vanadium loading of 0.5 YO [fig.7(c)]. A d-d band is found at 16500 cm-l as well as a shoulder at 9000 cm-l. The c.t. band appears at 40000 cm-l. The 2 YO V catalyst [fig. 7(6)] has d-d bands at 8500 andB. Jonson et al. 1905 0.8 0.6 h - 8 s k 0.L 0.2 (c) . . . . . . 10 20 I I I I 1 1 l I I I I 30 LO 51 wavenumber/ lo3 cm-' Fig. 7. Diffuse reflectance spectra of reduced samples. Spectra illustrate : (b) 2 YO V/SiO,, x = 6, right scale only. (cHe) As in fig. 6 with x = 4. 17000 cm-l. A shoulder at 20000 cm-l is also noticed. Two poorly resolved c.t. bands appear at 35000 and 40 500 cm-l. C.t. bands around 35000 cm-l have been assigned to five-coordinated VIV.l4* l5 Reduction of the 10 YO V catalyst resulted in a black sample which could not be studied by diffuse reflectance spectroscopy. Diffuse reflectance spectra of 0.5% V/SiO, under probe molecules (N2, CO and 0,) were recorded in order to examine the reactivity of this sample.Only minor spectral changes were observed in the case of N, and CO adsorption in contrast to earlier investigations on analogous compounds.16* l7 The difference is probably due to a higher degree of agglomeration in our case. Reduced 0.5 Oh V/SiO, under 0, forms the known V'v-O, surface complex,16?17 as indicated by the broad single band in the d-d region (found at 16500 cm-l in this case). The c.t. bands were found at 35500 (shoulder) and 41 000 cm-'. Activity Measurements Fig. 8 shows the reaction rate for the oxidation of toluene on V/silica catalysts as a function of the concentration of vanadium in the samples. The reaction rate increases with the vanadium concentration, which is what one would expect.However, the interesting point is that this effect is not only due to the larger amount of vanadium ions eventually available at higher loadings. Fig. 9 shows the rate of oxidation of toluene per gram of vanadium as a function of the vanadium concentration. The activity per gram of vanadium increases with the loading, and this effect is largest up to 0.5 YO V. Similar effects were earlier observed for vanadium supported on alumina,l which showed a more continuous effect. It was suggested that different vanadium species of different nature and activity exist in these catalysts. In fig. 10 the selectivity for benzaldehyde as a function of the vanadium loading is1906 Catalytic Oxidation of Toluene V concentration (wt %) Fig.8. Reaction rate per gram of catalyst for oxidation of toluene as a function of vanadium loading on silica. 0 2 4 6 8 V concentration (wt %) Fig. 9. Reaction rate per gram of vanadium as a function of vanadium loading on silica. 0 ' I I I I 0 2 L 6 8 Fig. 10. Selectivity for benzaldehyde, CO and CO, as a function of vanadium loading. 0, benzaldehyde, A, CO; x , CO,, for V/silica. V concentration (wt 5%)B. Jonson et al. 1907 shown. The selectivity decreases with increasing loading, whereas the carbon oxide formation increases. In contrast to V/A1,0,' no benzene formation is obtained on any of the catalysts. For the 0.5, 2 and 10% V/silica catalysts some other minor products belonging to the general product scheme as described elsewherel* were observed.The 0.5 YO V/silica gives 0.6 O/O methyldiphenylmethane (MDPM) and the 2 YO V/silica gives 0.9 YO MDPM and 1.5 % benzoquinone. The 10 % V/silica gives 3.4 O h benzoquinone, 3.1 YO maleic anhydride, 1.1 YO MDPM, 0.6% citraconic anhydride, 0.5 O h of phenol, benzoic acid and anthraquinone and 0.2% phthalic anhydride. In addition to these, a few small (ca. 0.5 YO) unidentified peaks were observed. The decreased selectivity for benzaldehyde with increased loading of vanadium suggests that these very selective sites are successively covered by vanadium species. Note that no coking effects could be observed on these catalysts, in contrast to V/alumina catalysts.' Obviously, very acidic sites are not present on V/silica.Catalytic Activity correlated to Surface Structure The investigated catalysts possess a considerably lower activity per gram catalyst in the oxidation reaction than the previously studied alumina-supported vanadium catalysts.' Also the activity per gram vanadium is less than 10% of that of V/alumina, at least for the higher loadings. The selectivity for benzaldehyde formation shows, however, reversed characteristics with ca. 60 % for silica and 7 YO for 10 YO V/silica. The corresponding values for V/alumina were 0 and 29 YO, respectively. Van Hengstum et al., on the other hand, found no selectivity for toluene oxidation on SO,, and only CO, when performing this reaction on silica-supported vanadium catalysts. l9 Note that these authors performed the experiments at higher conversions and at a different toluene/ oxygen ratio.The recorded i.r. spectra of adsorbed CO indicated sodium impurities on the silica surface. These sodium ions are assumed to be associated with the sites possessing benzaldehyde selectivity. This oxidation probably occurs via a radical mechanism which could be furnished by sparsely occurring 0; or 0-. It has been suggested that surface Na+ on y-Al,O, promotes the adsorption of 0- via Na+---O--.-Na+ species.,' Sodium and other alkali metals in connection with vanadium are known to have drastic effects on catalytic activity and selectivity.,l Introduction of vanadium decreases the selectivity for benzaldehyde. Apparently, the sodium sites are covered by or replaced with vanadium species. 1.r. spectra indicated a replacement of sodium ions with vanadium accompanied by a migration of sodium ions to new surface sites.At higher loadings sodium could be solvated in the vanadium oxide (according to phase diagrams,,) to form a P-type sodium-vanadium bronze Na,V,O,. It is reasonable to assume that a similar effect occurs with small polyvanadium aggregates. E.s.r. indicate a slow tumbling motion of V02+ upon Na+ addition to V/SiO,, which suggests the formation of sodium polyvanadate structures. These could function as a supply of loosely bound oxide ions (cf. silica) for reaction with intermediates initially formed at sodium sites. Introduction of Na+ to V/SiO, catalysts increases the selectivity for mild oxidation of but-1-ene to butadiene, but the reaction rate decreases.This is in agreement with our results, since in our case low vanadium addition gives a high Na/V ratio and the highest selectivity and lowest rate. This would be valid at low vanadium loadings, but not at higher loadings (2 and 10%) since the effects are very small at low Na/V ratios. The build-up of vanadium species with increasing loading is accompanied by an iicrease of catalytic activity, both per gram catalyst and per gram vanadium. Species of a different nature must thus be present at different vanadium concentrations. Some suggestions concerning the structure of vanadium species on silica gel can be found in the literature. Vorob'ev et al. suggest isolated or associated (VO,),- units sharing one corner with the (SiO,) chain.,, These authors also suggested the presence of (VO,),- in 63 FAR 11908 Catalytic Oxidation of Toluene ( b ) increasing V concentration association, oligomerization of vanadium species.Fig. 11. Surface compounds on V/silica catalysts. (SiO,) chains. With increasing vanadium concentration the partition between various species are shifted towards associated species and agglomerated finally to form V,O,. The partition between these species is also dependent on the initial OH-group concentration on the silica surface and the calcination temperature and 25-27 Yoshida et aL2' deduced (VO,) units accompanied by distorted square pyramidally coordinated VIV and V20, at higher loadings from e.s.r. measurements. Ohlman and coworkers26* 27 propose various different vanadia phases on silica. Narayana et al.suggest four-coordinated VIV on reduced V/Si0,.29 Horvath et al.' and Kazansky and Our results can be explained by proposing four-coordinated (tetrahedral) complexes. Hanke et aL31 suggest that tetrahedral Vv is formed on silica with a high concentration of OH groups, and otherwise octahedral Vv. The silanol groups of the silica surface [fig. 11 (a)] react with oxovanadium anions present in aqueous solutions of NH,V0332 and give vanadium surface compounds bonded through Si-0-V bridges. The calcined catalysts containing Vv are suggested to have the structure shown in fig. 11 (b), i.e. the same as the one proposed by Horvath et al.8, l7 At low loadings, when mostly isolated species are assumed to be present, the compound is characterized by u.v.-visible bands at 42000 and 47500 cm-l.This compound has the ability to coordinate water molecules [fig. ll(c)] in accordance with the results of Narayana et aLZ9 and Pak.33 This is confirmed by the observation of the sample's change of colour from white to more or less yellow (depending on the vanadium concentration) when kept in air. The colour turns back to white again after moderate heating. Reduction (with CO) gives coordinatively unsaturated V"' surface compounds [fig. 1 1 (41, characterized by the i.r. spectra of adsorbed CO. These V"' species are reactive towards oxygen and they readily form VIV-O, complexes [fig. 11 (e)]. With increasing vanadium concentration, the surface compounds associate as indicated by the shift in d.r.s. data. Chains (or aggregates) of oxygen-bridged vanadium are formed.From i.r. spectra it is evident that these chains do not contain hydroxyl have proposed tetrahedral vanadium surface complexes.B. Jonson et al. 1909 groups as found for the vanadium species on V/alumina.l Hanke et al. have suggested that chains with every second oxygen bridge between pentavalent vanadium doubled exist on silica-supported ~anadium.~* Every Vv ion, except those linked to silica, has one double-bonded oxygen. The V=O groups provide the oxygen that is active in the oxidation process. Chains of this type would increase the rate of reaction, but the selectivity is directed mainly to carbon oxides. The formation of chains provides easier one-electron transfer by the use of the VV-VIv redox couple compared to the case of the VV-V"* system of the isolated species.The former redox process would be more favourable for the oxidation of toluene. This explains the increase in activity per gram vanadium for the first 0.5 YO V. The 2 YO V/silica catalyst has roughly the same activity per gram V as the 0.5% V sample. The enhanced activity due to chain formation has levelled off. From u.v.-visible spectra one can deduce the presence of aggregates with five-coordinate vanadium on the 2% V catalyst. These new vanadium sites can be formed upon linking polyvanadium chains together. They also constitute the sites producing selective oxidation products besides benzaldehyde. Evidently the five- coordinated vanadium produces the same activity as the proposed chains, in contrast to the enhancing effect of agglomeration found for V/alumina.' S.e.m.and u.v.-visible studies revealed V,O, crystallites on the 10% V sample. The presence of agglomerates containing VIV in both the oxidized and the reduced form of this sample was shown by i.r. and u.v.-visible measurements. These crystallites and/or agglomerates only give a slightly increased activity per gram vanadium on going from 2 to 10% V. The needle-like crystallites appear to expose the (010) plane. In a series of papers this crystal plane with the V=O groups has been suggested to determine the selectivity and rate of r e a ~ t i o n . ~ , - ~ ~ However, the selective oxidation products produced by the 10% V catalyst are not typical for V,O, obtained by thermal decomposition of NH,V03.1s A product pattern more typical of V,O, was found for 10% V/alumina.' The difference might be explained in terms of variation of degree of crystallinity, number of structural defects etc.The V,O, crystallites on alumina were not large enough to be visible by s.e.m. in contrast to silica, although the diffractograms were similar. The crystallites are apparently much larger on silica and therefore expose a much lower area. Consequently, the contribution to the overall catalytic activity is most likely negligible in the case of silica, whereas in the case of alumina it is more likely. References 1 B. Jonson, B. Rebenstorf, R. Larsson, S. L. T. Andersson and S. T. Lundin, J . Chem. SOC., Faraday 2 C. D. Dollimore and G. R. Heal, J . Appl. Chem., 1964, 14, 109. 3 S. L. T. Andersson and M.S. Scurrell, J . Catal., 1981, 71, 233. 4 M. P. Seah and W. A. Dench, Surf. Int. Anal., 1979, 1, 2. 5 B. Rebenstorf and R. Larsson, Z . Anorg. Allg. Chem., 1979, 453, 127. 6 S. L. T. Andersson, J . Chromatogr. Sci., 1985, 23, 17. 7 S. L. T. Andersson, J . Chem. SOC., Faraday Trans. 1, 1979, 75, 1356. 8 B. Horvath, J. Strutz, J. Geyer-Lippmann and E. G. Horvath, Z . Anorg. Allg. Chem., 1981, 483, 9 B. Horvath, J. Strutz, J. Geyer-Lippmann and E. G. Horvath, 2. Anorg. Allg. Chem., 1981, 483, Trans. 1, 1986, 82, 767. 181. 193. 10 B. Rebenstorf, M. Berglund, R. Lykvist and R. Larsson, Z . Phys. Chem. N.F., 1981, 126, 47. 11 G. Ghiotti, E. Garrone, C. Morterra and F. Boccuzzi, J. Phys. Chem., 1979, 83, 2863. 12 B. Rebenstorf and R. Larsson, Actu Chem. Scand., 1980, AM, 239.13 B. Jonson, B. Rebenstorf and R. Larsson, Actu Chem. Scand, Part A, in press. 14 J. Hanuza, W. Oganowski and B. Jezowska-Trzebiatowska, Bull. Pol. Acad. Sci., 1983, 31, 153. 15 J. Hanuza, B. Jezowska-Trzebiatowska and W. Oganowski, J. Mol. Catal., 1985, 29, 109. I6 J. Geyer-Lippman, Dissertation (Freie Universitat Berlin, 198 1). 17 B. Horvath, J. Geyer and H. L. Krauss, 2. Anorg. Allg. Chem., 1976, 426, 141. 18 S. L. T. Andersson, J. Catal., 1986, 98, 138. 63-21910 Catalytic Oxidation of Toluene 19 A. J. van Hengstum, J. G. van Ommen, H. Bosch and P. J. Gellings, Appl. Catal., 1983, 8, 369. 20 C. Kordulis, L. Vordonis, A. Lycourghiotis and P. Pomonis, J. Chem. SOC., Faraday Trans. I , 1987,83, 21 R. Fricke, W. Hanke, H-G. Jerschkewitz, B. Parlitz and G. Ohlman, Appl. Catal., 1984, 9, 235. 22 A. A. Fotiev, Zh. Neorg. Khim., 1977, 22, 2531. 23 R. Fricke, H-G. Jerschkewitz and G. Ohlman, React. Kinet. Catal. Lett., 1981, 18, 497. 24 L. N. Vorob’ev, I. K. Badalova and K. Kh. Razikov, Kinet. Catal. (Engl. Transl.), 1982, 23, 119. 25 J. C. W. Chien, J. Am. Chem. SOC., 1971,93,4675. 26 K. Heise, G. Heise, R. Fricke and G. Ohlman, Muter. Sci. Monogr., 1982, 10, 918. 27 M. Richter, K. Heise and G. Ohlman, React. Kinet. Catal. Lett., 1985, 27, 109. 28 S. Yoshida, T. Iguchi, S. Ishida and K. Tarama, Bull. Chem. SOC. Jpn, 1972, 45, 376. 29 M. Narayana, C. S. Narasimhan and L. Kevan, J. Cataf., 1983, 79, 237. 30 A. M. Gritscov, V. A. Shvets and V. B. Kazansky, Chem. Phys. Lett., 1975, 35, 511. 31 W. Hanke, K. Heise, H-G. Jerschkewitz, G. Lischke, G. Ohlman and B. Parlitz, Z. Anorg. Allg. 32 M. T. Pope and B. W. Dale, Q . Rev. Chem. SOC., 1968, 22, 527. 33 V. N. Pak, J. Gen. Chem. USSR, 1975, 45, 920. 34 W. Hanke, R. Bienert and H-G. Jerschkewitz, Z. Anorg. Allg. Chem., 1975, 414, 109. 35 K. Mori, M. Inomata, A. Miyamoto and Y. Murakami, J. Chem. SOC., Faraday Trans. I , 1984, 80, 36 A. Miyamoto, K. Mori, M. Inomata and Y. Murakami, 8th Int. Congr. Catal., Berlin, July, 1984, 37 K. Mori, M. Inomata, A. Miyamoto and Y. Murakami, J. Phys. Chem., 1983, 87,4560. 627. Chem., 1978, 438, 176. 2655. vol. iv, p. 285. Paper 7/1089; Received 19th June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401897

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. Soc., Furuday Trans. I , 1988, 84(6), 1911-1917 A General Calculation of Molecular Solvation Energies Raymond J. Abraham" and Brian D. Hudson? The School of Chemistry, University of Liverpool, Liverpool L69 3BX Mark W. Kermode and J. Roger Mines Department of Applied Mathematics and Theoretical Physics, University of Liverpool, Liverpool L69 3BX A method for the calculation of molecular solvation energies that takes into account both the molecular geometry and the partial atomic charges is presented. A van der Waals sphere is placed at each atomic centre and the molecular surface is taken to be the exterior boundary of the interacting spheres. The energy is expressed as an integral over this surface involving the electrostatic field generated by the atomic charges in U ~ C U O and the electrostatic potential in the presence of the solvent treated as a dielectric medium.Illustrative examples of calculations for some simple ions are presented. Good agreement with known (experimental) values of the solvation energies is found. With the increasing sophistication and accuracy of modern theoretical calculations of molecular structures and energies in the gas phase,' the calculation of solvation energies, by which the theoretical results can be related to experimental studies of molecules in solution, becomes increasingly important. For small molecules and solvents, Monte Carlo methods3 have found some success in reproducing solvation energies and giving detailed insight into the solvent structure, and recently this technique has been applied to the estimation of the solvation energies of the acetate and methylammonium ions and the Ala dipeptide in aqueous solution.* However, these techniques become prohibitively time-consuming for larger molecules and the results also rely on the accuracy of the molecular intermolecular potentials used. (The Lennard-Jones potential is most often used, yet this has known deficiencies in molecular mechanics ~tudies.~) Quantum-mechanical methods such as the supermolecule approach of Pullman' are also of limited applicability, partly due to computational restrictions.The other major theoretical approach relies on the continuum model, in which the solvent is regarded as a continuum of given dielectric constant, and this model has been colntinuously extended since the original Born equation7 for the solvation energy of a spherical ion.Recently Beveridge and SchnuelleR solved the electrostatic problem of a multipole in the centre of a spherical cavity surrounded by a shell of solvated solvent molecules, and Abraham and Liszi have successfully applied this model to the calculation of the energies of solvation of a number of simple ions.' Gersten and Sapsel' have allowed for the fact that most molecules are not spherical and have calculated the solvation energy of a multipole in an ellipsoidal cavity. However, neither of these methods can be applied to the general problem of the calculation of molecular solvation energies. Most molecules are neither spherical nor ellipsoidal and the approximation that one can replace the molecular electrostatic field of, e.g.dopamine (d ihydroxyphenylethylamine hydrochloride) by a series of multipoles placed at the centre of the molecule, assuming it to be spherical, is clearly a very crude approximation. In the present approach the molecular volume is taken to be the interior of a set of it Present address : The Wellcome Research Laboratories, Langley Court, Beckenham, Kent BR3 3BS. 141 11912 Molecular Solvation Energies intersecting van der Waals spheres. The volume integral over the region occupied by the solvent (treated as a dielectric medium) which gives the required energy is transformed into an integral over the molecular boundary. This integral involves the electrostatic field generated by the partial atomic charges in vacuo and also the electrostatic potential in the presence of the dielectric.While the former is treated exactly, it is necessary to make an approximation to the latter. Theory The energy of interaction of the molecular charge distribution with the environment is the solvation energy of the molecule. This is a Helmholtz free energy and is given by the reversible work involved in charging the distribution in the presence of the dielectric." Taking the solvent to be a dielectric medium (with dielectric constant K ) the molecular solvation energy is given by'' - , - 1 w = - J E,.E,(K- 1)dV 8n v where E, is the electrostatic field of the partial atomic charges in uacuo and Es is the modified field in the presence of the dielectric. The region V is that occupied by the dielectric (i.e.the exterior of the van der Waals spheres). Numerical integration over V presents a difficult problem which is resolved by an application of the Gauss divergence theorem, which transforms the volume integral into a surface integral over the boundary. Setting Es = -grad mS and noting that div (as E,) = grad 0, - Eo + OS div E, = grad@,,.E,, (since divE, = 0 in V ) the divergence theorem tells us that E,.E,dV= grad@,.E,dV Jv Jv = ldiv(Q,E,)dV = -Js @,E,.ndS where S is the boundary surface and n the unit outward normal vector on S. Thus, In the simple case of a single point charge e at the centre of a spherical cavity of radius a in a dielectric medium of dielectric K , the above surface integral is 1 - K e 2 K 2a - the well known Born charging en erg^.^ We require to evaluate eqn (1) for molecules of arbitrary complexity.The determination of Eo.n at points on the surface is straightforward and an elegant procedure for performing a numerical integration over S will be presented here. TheR. J . Abraham et al. 1913 Fig. 1. Region of integration (shaded area) in the 6, Q plane for one of the hydrogen atoms in NH;. exact determination of QS involves satisfying boundary conditions on S and is impractical. We adopt the approximation where Eo = -grad (Do. This gives the exact result in the case of the Born charging energy above. In the Appendix, we show that the integral part of eqn (1) may be written where i represents the ith sphere, j(k) labels the contribution of the jth (kth) charge to the potential (field) on the surface.The function&&, q5i) is given in the Appendix [eqn (A 6)] and is easily calculated numerically. The angles (Oi,q5i) are the usual spherical polar coordinates (for sphere i). The sum over m arises from the fact that for a fixed qbt, only certain values of Bi occur for points on the 'exposed ' surface. In fig. 1, we show an example of the region of integration (shaded) for one of the hydrogen atoms in the NHT molecule. We see that for some values of $ (e.g. 30") m = 1, but other values of q5 (e.g. 120") may have m = 2. The boundaries of the curves [B = pm(q5), a,(#)] like those shown in fig. 1 are determined from eqn (A 5 ) in the Appendix. The numerical evaluation of the integrals in eqn (2) were carried out using the NAG library subroutines The integration procedure has been extensively tested on systems where exact solutions of the equations are available, for example the surface area (equal to a: sine, for points lying on the surface) of isolated, overlapping and non-overlapping spheres, and for some test cases of the solvation energy calculation for both isolated spheres and some special cases of two overlapping spheres (where exact solutions exist).Table 1 gives the results of these calculations for the case of two overlapping spheres of equal and opposite charge, for which the exact solutions are given by eqn (3) and (4): exposed surface area = 2n(b2[1 +(d2+b2-a2)/2db]+a2[1 +(d2+a2-b2)/2da]) (3) DOlBAF, DOlBAZ for 8 and DOlGAF for #. solvation energy (a = b, el = e = -e2) kcal? mol-1 = 332.1 4a [ - 2 + g + ( 1 +:)'+:In( 1 +:)I (4) It can be seen that the numerical integration gives excellent approximations to the (valid for 2a 3 d 2 0).7 1 cal = 4.184 J.1914 Molecular Solvat ion Energies Table 1. Exposed surface area and solvation energy of two charged spheres surface area/A2 solvation energy/ kcal mo1-1 Nq5" NB" S(a> S(b) N#a NO" exactb 15.8652 3 5.9769 exactc 10.1356 10 10 15.8652 35.9769 10 10 10.1356 8 8 15.8652 35.9769 4 8 10.1356 4 4 15.8652 3 5.9768 2 8 10.1356 2 2 15.8652 35.9768 8 4 10.1344 2 2 15.6968 35.4069 4 4 10.1344 a Number of increment? on the potential surface. a = 1.19 A, b = 1.734 A, d = 2.5 A. a = b = 1.5 A, d = 1.0 A. exact solution even when the number of steps on the potential surface is very small. (Note that 0 is 0-180" and $ is 0-360"; however, the above example is axially symmetric, i.e.the potential of any given 8 is independent of $.) Application The application of the above scheme to the calculation of the solvation energies of ions in aqueous solution has been performed on the simple ammonium and alkylammonium ions, and also on the formate and acetate ions. There are considerable uncertainties in such applications which are not due to numerical calculations but simply to our lack of knowledge of the charge distribution and effective volumes of even these simple ions. The question of the distribution of the integral charges on the atoms of the ions has been discussed for many years without any generally agreed consensus. We shall avoid this problem by using the charge scheme recently developedx2 in which the integral charges on ammonium and carboxylate ions can be handled in a straightforward manner by modifying the atom's electronegativities, resulting in some dispersion of the charge to the neighbouring atoms.The effective volumes of ions in aqueous solution is also another long-standing problerny6 and this is relevant to the calculation of the van der Waals surface of the molecules under consideration. We have taken the van der Waals radii of the atcms from a recevt theoretical derivation by Franc1 et 0aZ.13 which gave C(sp3) 1.893" A, N(sp3) 1.734 A, O(sp2) 1.549 A and H(C-H) 1.190 A. The van der Waals radius of hydrogen, which is usually on the outside of the molecule, is particularly uncertain in these calculations, as the firstosolvent layer may be included in this radius;9 thus we have used values of 1.19 and 1.5 A in the calculations.The total solvation energy of these ions is equal to the sum of the electrostatic term which is calculated here, and a non-polar term due to the energy involved in the formation of a cavity in the aqueous medium. This second term was estimated following the procedure given in ref. (9), in which the non-polar contribution (AGZ) is given by the equation where r is the radjus of the ion in question. The values used were r(NH,+) = 2.00 A and r(NMe1) = 2.58 A, taken from ref. (9), and the remaining alkylammonium ions were AG; = -2.0514r+9.7426 ( 5 )R. J. Abraham et al. 1915 Table 2. Calculated and observeda solvation energies (kcal mol-l) of ions in aqueous solution ~ ion .,/A AG," AGE AG," AG, (obsd) NH; 1.19 1.5 CH,NHi 1.5 I(CH,),NH; 1.5 I(CH,),NH+ 1.5 (CH,),N+ 1.19 1.5 H * CO; 1.19 1.5 CH, * CO; 1.15 1.19 - 89.7 - 77.5 -71.2 - 64.2 - 58.5 - 57.5 - 52.7 - 83.2 - 82.9 - 82.0 - 82.8 5.64 5.64 5.34 5.04 4.75 4.45 4.45 6.4 6.4 5.9 5.9 - 84.1 - 72.4' - -71.8 -65.9 - 65.4b - 59.2 - 58.5' - 53.7 -51.5b - 53.1 - 48.0b.-48.3 - 76.8 - 76.5 - 99d - 76.1 - 76.9 -lOld, -90e - - - a Standard free energies of solvation of ions (1 atmt gas to unit mole fraction solution). ' Ref. (14); converted to the atm/mole fraction scale by subtracting 4.25 kcal mol-l. Ref. (15). Ref. (4). Ref. (9). interpolated from these. The values for the formate and acetate ions were obtained from the incremental contributions for estimating van der Waals volumes given by Ford and Scribner .l4 The results of the calculations together with observed values of AGZ are given ic table 2. It can be seen that the use of the hydrogen effective van der Waals radius of 1.5 A gives virtually quantitative agreement with the experimental solvation energies of the alkylammonium ions, which is very encouraging. Note that only one hydrogen van der Waals radius was used, although in principal the NH proton should have a different radius from the CH proton. The agreement for the formate and acetate ions is reasonable, although in view of the considerable uncertainties in the experimental values of the solvation energies of these ions it was not considered worthwhile to optimise the van der Waals radii in this case; however, this could in principle easily be performed.Conclusions In conclusion, therefore, a scheme has been developed by which the reaction field theory can be applied to molecules which cannot be approximated by a set of point multipole moments. Furthermore, the integration of the electrostatic field can now be performed for an irregular-shaped cavity by transformation of the volume integral to a surface integral. Thus, the shape of the cavity can be one which more closely approximates the shape of the molecule. The use of charges which accurately represent the dipole moments means that the dipole and quadrupole terms of the reaction field theory are implicitly included in the calculation. The method can therefore be applied to neutral molecules as well as ionic species under study here.We thank Dr M. H. Abraham for helpful assistance and comments on this work, and the University of Liverpool computing laboratory for the computing facilities. We also acknowledge an S.E.R.C. CASE studentship (B.D.H.) with Dr W. A. Thomas (Roche Products Ltd). 1- 1 atm = 101 325 Pa.1916 Molecular Solvation Energies 0 Fig. 2. Vector diagram for two overlapping spheres. Appendix In this Appendix, we show how the repeated integral form [eqn (2)] of the surface integral [eqn (l)] is obtained. In fig. 2, we show the relevant position vectors for the case of the overlapping spheres. The superscript i refers to a point on the exposed surface. We see immediately that the equation I = Is qboE0.ndS may be written where qj is the charge at the centre of thejth sphere. coordinates with respect to its centre by A point on the surface of the ith sphere may be expressed in terms of spherical polar ri = ri + ai(sin Oi cos q5i i + sin 8i sin q5i j + cos Oi k).(A 2) For the overlap of the ith and jth spheres, we require lri -rjl 2 aj (A 3) (A 4) for a given q5i. Thus the effect of all spheres on the surface of the ith sphere is to produce the boundary (intersecting) curves for points exterior to sphere Sj. The equality will hold if Bi satisfies the equation [(xi -xi) cos #i + ( y i - y j ) sin #,I sin 8, + (zi - z j ) cos Oi = [a; -at - (ri - rJ2]/2ai Iri-rjl = a j ; j = 1, ..., N ( j # i) (A 5 ) some of which are shown in fig. 1.R. J. Abraham et al. 1917 Substituting for ri [eqn (A 2)] into eqn (A 1) and using dSi = a: sin Oi doid& we obtain eqn (2), which has in its integrand Ljk(Oi7 # i ) = kik(Oi7 $ i ) + hTk?,3/2(0i $ i ) hG1'2(0i, # i ) (A 6) where gik = (xi - xk) sin Oi cos qbi + (yi --yk) sin Oi sin #i + (zi - 2,) cos Oi hij = Atj + 2aigij Atj = Iri -rjI2 + a:.and References 1 W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab initio Molecular Orbital Theory (J. 2 W. G. Richards, Quantum Pharmacology (Butterworths, London, 1979). 3 J. M. Hammerslay and D, C . Handscomb, Monte Carlo Methods (Methuen, London, 1964). 4 G. Alagona, C. Ghio and P. Kollman, J. Am. Chem. SOC., 1986, 108, 185. 5 U. Burkert and N. L. Allinger, Molecular Mechanics: ACS Monograph 177 (A.C.S. Washington, 6 A. Pullman and B. Pullman, Q. Rev. Biophys., 1975, 7 , 505. 7 M. Born, Phys. Z., 1920, 1,45. 8 D. L. Beveridge and G. W. Schnuelle, J. Phys. Chem., 1975, 79, 2562. 9 M. H. Abraham and J. Liszi, J. Chem. SOC., Faraday Trans. I , 1978, 74, 1604. Wiley, New York, 1986). 1982). 10 J. I. Gersten and A. M. Sapse, J. Am. Chem. SOC., 1985, 107, 3786. 11 C. J. F. Botcher, Theory of Electric Polarisation (Elsevier, Amsterdam, 1952). 12 R. J. Abraham and B. D. Hudson, J. Comput. Chem., 1985, 6, 173. 13 M. M. Francl, R. F. Hout and W. J. Hehre, J. Am. Chem. SOC., 1984, 106, 563. 14 G. B. Ford and J. D. Scribner, J . Org. Chem., 1983, 48, 2226. 15 E. S . Rudakou, Dokl. Akad. S.S.S.R., 1981, 160, 676. Paper 7/ 1106; Received 22nd June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888401911

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1988, 84(6), 1919-1925 Excess Enthalpies of Ternary Aqueous Solutions of Amides and Ureas at 298.15 K Guido Barone,* Giuseppina Castronuovo, Pompea Del Vecchio and Vittorio Elia Department of Chemistry, Universita ' Federico 11, via Mezzocannone 4, 80134 Naples, Italy Aqueous solutions of amides have been studied for their importance as uncharged model molecules of structural units of polypeptide chains and proteins. Urea and its alkyl derivatives are known as denaturing agents against proteins. In this paper some new experimental calorimetric data from our laboratory concerning ternary aqueous solutions, each containing one amide and one urea species, are presented, discussed and compared with literature data. The relevance of the weak interactions occurring between these types of cosolutes in water is discussed from the point of view of the biochemical consequences on the stability of proteins. It is confirmed that both polar and hydrophobic intermolecular interactions perturb the native conformation of polypeptides and proteins, as they compete with intramolecular interactions.Urea is either ineffective or it enhances the hydrophobic interactions, contrary to what apparently is suggested by the increase of solubility of lighter hydrocarbons in urea-water mixtures. The weak interactions in water between model molecules, bearing peptide groups, have received increasing attention in recent years, since they are related to the conformational stability of proteins and naturally occurring oligo- and poly-peptides. '-16 Also of interest, but less studied, are the interactions between urea (and its derivatives) and amides,6,8,17-25 whose knowledge permits an insight into the denaturation processes.In fact, the nature and strength of the most important driving forces ruling the conformational transitions of proteins and polypeptides are still controversial. For instance, differing opinions are currently held on whether the interactions between a pair of urea or peptide molecules are due to direct hydrogen bonds or to solvent-mediated effects. In a series of papers from our laboratory we have shown that the first hypothesis is unacceptable, 26-29 and that the excess thermodynamic properties of these solutions are rather characterized by the perturbation induced in the structure of the solvent by the solute molecules.Although hydrophilic solutes interact well with the solvent, they can also promote aggregates whose geometry is very different from the tetrahedral structure of the ice. Then the solute-solute interaction between a pair of urea or urea-like molecules has been hypothesized to consist of a partial coalescence of the two distorted cospheres.lg~ 29 The analysis of the excess thermodynamic properties of a solution, in absence of direct molecular information, is an important starting point for formulating hypotheses on the nature of the interactions in solution. For unsymmetrical mixtures of non-electrolytes an excess property JE is usually defined as ' ' 9 28-30 n J E = J-J:- m,Jz-J'D x= 2 where J E and J refer to an amount of solution containing 1 kg of solvent and m, ...m, moles of each solute species, JF is the standard property of 1 kg of pure solvent 1, and 19191920 Excess Enthalpies of Ternary Aqueous Solutions the J: are the standard partial molal limiting quantities for each solute; JID is the part of J due to the ideal mixing (on the molality scale). For the enthalpies HID = 0 and HE represents the deviation from the behaviour of athermal or ideal solutions. The excess enthalpy per kg of water can be represented, for binary and ternary solutions, as a power series of the solute molalities :6 (2) and (3) According to the McMillan-Mayer approa~h,~'-~~ the coefficients h of eqn (2) and (3) can be considered as the enthalpic contribution to the coefficients g of the excess Gibbs free energy, GE, that are a measure of the interactions between pairs, triplets etc.of molecules of the corresponding species in the considered solvent. For each of the h coefficients, as for the overall properties HE(m,) = h,, mz + h,,, mz + . . . HE(m,, my) = h,, m: -+ 2hXy m, my + h,, mi + h,,, mz + 3h,,, mz my+. . . The statistical-mechanical significance of both the free energy and enthalpy coefficients has been discussed recently.34 The point was also stressed that they implicitly account for changes occurring in the solute-solvent33 and solvent-solvent interaction~~~' 36 with respect to the standard state (i.e. an infinitely dilute solution) at increasing concentrations. Experimental Materials Urea (a Carlo Erba product, Milan), and monomethylurea (MMU), monoethylurea (MEU), N-methylacetamide (NMA) (Fluka products) were twice crystallized from ethanol-water solutions and dried in vacuo at room temperature.N-ethylformamide (NEF) and N-t-butylformamide (NButF) were Fluka products of the highest purity commercially available and were twice distilled at reduced pressure, avoiding exposure to the light. Finally, they were stored in a desiccator over zeolites. The water used for all experiments was deionized, twice distilled, filtered on a millipore membrane and degassed. Solutions were freshly prepared by weight and used within a few hours. Calorimetry The heats of dilution of binary and ternary solutions in water were determined at 298.15 f 0.02 K with a LKB 10700- 1 standard flow microcalorimeter, by the procedure reported in previous papers.6*8*19~28~29 The heats of dilution Adil H (J kg-l of water in the final solution) for the binary and ternary solutions, respectively, are given by: Adil H(mi --+ mf) = - (dQ/dt)/Pw (6) (7) H[(mix, mi,) -+ (mfx7 mfy)l = - (dQ/dt)/Pw where (dQ/dt) (W) is the heat evolved per time unit under steady-state conditions, Pw is the total mass flow rate of water and mi and m, are the initial and final molalities, respectively.G.Barone et al. 1921 Table 1. Enthalpic interaction coefficients for amides and ureas in water at 298.15 K NEF-urea -2 (1)" -7 (1) 5 (1) 529(7) - -350 (1 22d NButF-urea 37 (7) -28 (2) 7 (4) 1355 (22) - -350(ll) 22 NMA-MMU 189(ll) - - 236 (ll)d 95d -85 ( l l ) e 21 ( l r NMA-MEU 334(13) - - 236(ll) 95 160 (7)f 37 (3)j a Units: J kg mol-2. * Units: J kg2 rn0F3.limits. The uncertainties reported are the 95 YO confidence Ref. (17). The authors give also h,,,, = - 18 J kg3 m01-~. Ref. (19). Ref. (18). Treatment of the Data The enthalpic homotactic interaction coefficients h,,, h,,,, . . . were evaluated for the two formamide derivatives by fitting the heats of dilution of the binary solutions with the polynomial expansion :16 Adil H(mi -+ m,) = h,, rn,(m, -mi) + h,,, m,(m," - m:) -k . . . . (8) For the other solutes in water (urea, MMU, MEU, NMA) the homotactic coefficients are given in the literat~re."-'~ The etherotactic interaction coefficients h hXxy, hxyy etc. were obtained through the fitting of the following power expansion :6p '"' AH** = 'hxymfy(mfx-miJ +3hx,ymfy(m,x-mix) (mfx +mix) The values of the auxiliary function AH** for each pair of initial and final molalities were calculated by the definition subtracting from the experimental values of Adil H(x, y ) those evaluated for the same change of concentration for the two binary solutions.In the fitting of eqn (8) and (9), a least-squares method was employed. The fitting was tried with polynomials of increasing degree, choosing that of highest degree, for which all the coefficients are significant with respect to their own 95% confidence limit. Results The experimental values of the heats of dilution of the binary solutions of the two N-alkylformamides, along with the initial and final molalities for each experiment, as well as the heats of dilution of the ternary solutions containing urea or an alkylurea and an amide (together with the corresponding auxiliary function AH** and the initial and final aquamolalities of both cosolutes for each experiment), are reported in deposited tables [SUP 56 709 (5pp)l.t In table 1 the experimental interaction coefficients are summarized for all the systems studied.In some cases, owing to the limited concentration range explored, only the pairwise parameters were needed for representing the system. These positive values of the h,, coefficients can be compared with that found by Savage and Wood for N- methylformamide (272 J kg m~l-~),'' and they are in agreement with the hypothesis that the homotactic interactions of the alkylamides are prevailingly hydrophobic in nature. From this point of view the behaviour of t-butyl derivatives deserves comment.It was t See Notice to Authors, J . Chem. Soc., Faraday Trans. 1, 1988, 84, January issue.1922 Excess Enthalpies of Ternary Aqueous Solutions found that the h,, values for ButOH and BunOH are, respectively, h,, = 656+ 33 and 1003 f 15 J kg In contrast, in a previous paper we found the opposite behaviour for t-butylurea and n-butylurea : h,, = 3632 f 373 and h,, = 1039 f 31 J kg A rationalization was suggested, assuming that the hydrophobic interaction is not the only one contributing to the positive value of h,,, but in these cases the water-mediated etherotactic interaction between the urea framework and the t-butyl side-chains of two different solute molecules is very important, giving a positive large contribution to h,,.24 The behaviour of NButF and of N-butylacetamide (in the absence of data for N-buthylformamide) is intermediate (h,, = 1355 k22 and h,, = 1477f24 J kg m01-2,17 respectively), probably because of a less positive contribution owing to the interaction between the substituted amide group and the t-butyl chain with respect to that given by the urea framework.Discussion The additivity principle has been fully applied in chemical thermodynamics for obtaining the contributions of groups of atoms to molecular properties. Wood and coworkers suggested17, 20, 2 5 9 37* that this approach can work approximately also for the excess thermodynamic properties of solutions. Among other analogous expressions they obtained for the second coefficients of the excess enthalpy the following relation: h,, = C n: nj" {Hij} i j where h,, is given by the sum over all the contributions {Hij} obtained by coupling each group of atoms i of the solute x (n: being the total number of these groups) with each group j of the solute y (even in the trivial case x = y ) .That approach has been criticized as theoretically ill-founded, especially for the assumption that the potential of mean forces between two solute molecules in a solvent is of an exponential form, vanishing at short-range, and then adequately approximated by a linearization p r ~ c e d u r e . ~ ~ ~ ~ ~ However, Lilley and coworkers have shown that in the limits of validity of the McMillan-Mayer approach and using the Lewis and Randall formalism, eqn (1 1) (regarding only the enthalpic coefficients) can actually work;34 however, it must be expected that eqn (1 1) will fail when cooperative or concerted effects of several groups lead to highly specific interactions.For statistical reasons eqn (1 1) should work better with larger sets of data for systems characterized by weak and scarcely specific interactions. In a previous paper6 we selected a basis of ca. 50 systems containing ureas, amides and uncharged peptides. This choice allowed us to consider only three groups: the urea framework (U), and the peptide (-CONH-) and methylene groups (CH2-). Following Friedman and Savage and Wood, we assumed CH = 0.5 CH, and CH, = 1.5 CH,, all the contributions involving -CONH- and U having the same value, independently of their degree of N- substitution.In table 2 are the results obtained for the evaluation of these group contributions in the last few years and in the present work. The first column of values has been used to recalculate the first set of h,, reported in table 3. Here are also given some recent data, kindly provided by Dr. T. H. Li1ley.t For the evaluation of the group contributions we also take into account recently published data of Lilley's group7' 13714334 and Wood's group.25 However, as Lilley and coworkers showed, a marked difference exists for the interaction between a wholly substituted peptide group and a -CH,- group, with respect to the corresponding t Dr. Lilley also drew our attention to the fact that his experimental h,, values, when plotted uersus the number of equivalent CH, groups, produce two straight lines, one grouping together the data for the U-amides and U-monoalkylamides, and the other the Uklialkylamides.Given its value, the U-NButF system resembles much more the ULdialkylamides, probably because the t-butyl group screens the -NH- group.G. Barone et al. 1923 Table 2. Group contributions {H,,}/J kg mo1F2 to the second virial coefficients of the excess enthalpies for amides, ureas and peptidoamides at 298.15 K functional groups this ref. (7), this worka (1 l), (14) ref. (6) workb ref. (38) ref. (17) i j CONH CONH - 266 (20) - 292 (59) - 307 (22) - 225 (1 33) - 225 (74) - 25 1 (103) - 377' U U -350(exp) - - 350 (exp) - 350 (exp) - - 565" CONH U - 185 (45) - - 196 (75) -256 (88) - - 74 (10) 66 (21) - CONH CH, 65 (9) 47 (8) U CH, CH, CH, 32 (4) rJ & 136 - 129 - * 354 - + 220 81 (29) 93 (11) 82 (39) 71 (29) 42(33) 25 (13) 17 ( 5 ) 21 (11) 36 (11) 42 (8) 61" a (>n the basis of data reported in ref.(6), (7), (lo), (1 1); (13)-(15), (29) and (34) (see the text, more than 80 values are taken into account). On the basis of a restricted set of ca. 35 values concerning only amides and interactions of amides with ureas. " Estimated on the basis 1 U = 1.5 CONH. Table 3. Experimental and calculated values of the h,, coefficients for interactions involving amides and ureas at 298.15 K urea urea urea urea urea urea urea urea MMU MEU urea urea urea urea urea urea NEF NButF FA AA NMF NMF NEF NMA NMA NMP NMA NMA NBA DMF DMA NButF DEF DEA NEF NButF 0.5 1.5 2 2 3 3 3 4 (1.5+ 3) (2.5 + 3) 6 3.5 4.5 5 5.5 6.5 -261 (6)" - 142 (4)e - 132 (2)" - 109 (9y -2 (1)g 0 (9)f 15 (6)" 180 (13)f 181 (ll)g 334 (1 3)g 264 (36)f - 155 (9)" -70 (8)" 37 (7)9 36 (4)" 135 (10)" 529 (7)g 1355 (22)g - 161 -115 -91 -91 -44.5 -44.5 -44.5 2 199 36 1 96 -21 26 49 72 119 416 1196 - 223 - 157 - 124 - 124 - 57 - 57 - 57 9 159 303 141 - 24 42 75 107 174 454 1113 ~ ~~ ~ a Number of CH, equivalent groups (see text) on amides or alkylureas.Units: J kg rnol-,. According to the group contribution of the second column of parameters of table 2. P. J. Cheek and T. H. Lilley, J. Chem. Soc., Faraday Trans. I, 1988, 84, in press. f Ref. (17). Data from this work. FA, formamide ; AA, acetamide ; NMF, N-methylformamide ; NMP, N-methylpropionamide ; NBA, N-butylacetamide ; DMF, N,N-dimethylformamide ; DMA, N,N-dimethylacetamide ; DEF, N,N-diethylformamide ; DEA, N,N-diethylacetamide. Other abbreviations as in the text.According to the group contribution of first column of parameters of table 2.1924 Excess Enthalpies of Ternary Aqueous Solutions unsubstituted peptide-methylene interaction^.^^ However, we do not consider in our calculations solutions of molecules with the former groups. For similar reasons we treated separately, in a preceding paper, the problem of the cis-(-CONH-) group.' It is possible to see that the prediction of the experimental values for two binary and four ternary systems studied is relatively good, within the limits of this method. Moreover, it must be considered that 80 experimental values are still a restricted basis for a statistical analysis of a relation with six interaction parameters, as in the present case.The group contributions evaluated here seem to be in better agreement with those calculated more recently by Wood3' than with those reported by Lilley and coworkers3~ 7,11,13,14+ 34 and by ourselves.' The actual improvement in correlation, as regards the set of values of Wood, is the reduced scattering of the data (see the value of the standard deviation, 0, in table 2) owing to the selection of a more restricted, but more homogeneous basic set of aqueous systems. Following this theme we also tried an evaluation of group contributions, choosing ca. 35 values concerning only amides and their interactions with themselves and ureas. The data reported in the fourth column of table 2 show that CT is almost the same as that obtained for our other calculations.However, because of the much more restricted basic set of data, the uncertainty of each of the parameters is worsened. This is in part compensated by an improvement in the prediction of the single data of the set, especially for the terminal ones. The agreement between the calculated and experimental data is less good for the intermediate values, because of the quoted trends along two lines. It could be significant that the worst agreement is shown by the more hydrophobic normal alkylamides (N-methyl- propionamide and N-butylacetamide) when reacting with urea. The group-contribution analysis suggests that both the polar and the hydrophobic interactions between the denaturing agents and the polypeptide chains make the compact tertiary structure of the globular proteins more labile, probably because they compete with the analogous chain+hain intramolecular interactions. For the CH,-CH, and U-CH, interactions, it is found that both the entropic and enthalpic contributions are positive.This can be attributed to a release of water from the hydrophobic cospheres to the bulk. However, in the case of the CH,-CH, interaction, the entropic term is such as to make negative the (GCHO-CH1) = -318 or -20 J kg m01-2.34 On the contrary, the failure of a favourable free energy term (GCH,-CH,) = 2 f 12 J kg mo1P characterizes the U-CH, interaction in a very different manner. Sup- posing that the positive sign will be confirmed, this could lead to the conclusion that a salting-out effect on the CH, groups exists owing to the release of water molecules from the hydration cospheres of the alkyl chain to the bulk, induced by urea.This model is not in contrast with the experimentally determined increase of solubility of lighter hydrocarbons in the presence of urea. In fact, the reduction of the 'hydrophobic hydration ' promoted by urea gives the same result obtained by the overlap of hydrated apolar molecules. The increase in solubility should arise from a further retrogradation of the 'hydrophobic hydration', due partly to the formation of a greater number of ' associated ' molecules, via hydrophobic interactions, partly due to the direct dehydration induced by urea. This model is in agreement with that already discussed by Ben- Naim,41 who found that urea reinforces hydrophobic interactions.Finally, the statistical disordering effect (chaotropism) that urea exerts on all the ordered s t r ~ c t u r e s , ~ ~ , ~ ~ and the competition (for the hydration water) with all the hydrophilic and hydrophobic groups of the side chains on the surface of the macromolecule and the substitution of the hydration water with urea at high con- centrations, are all to be considered. All these effects cannot be taken into account in the preceding analysis, but they are important factors in determining the destabilizing action of urea on the conformation of polypeptides and proteins.G. Barone et al. 1925 This work was financially supported by the Italian Ministry of Public Education and by the Italian National Research Council.References 1 G. M. Blackburn, T. H. Lilley and E. Walmsley, J. Chem. Soc., Furaday Trans. I , 1980, 76,915. 2 G. M. Blackburn, T. H. Lilley and E. Walmsley, J . Chem. SOC., Chem. Commun., 1980, 1091. 3 G. M. Blackburn, T. H. Lilley and E. Walmsley, J . Chem. Soc., Faraday Trans. I , 1982, 78, 1641. 4 A. Cesaro, E. Russo and G. Barone, Int. J . Peptide Protein Res., 1982, 20, 8. 5 K. Nelander, G. Olofsson, G. M. Blackburn, H. E. Kent and T. H. Lilley, Thermochim. Acta, 1984,78, 6 G. Barone, P. Cacace, V. Elia and A. Cesaro, J . Chem. SOC., Faruday Trans. 1, 1984, 80, 2073. 7 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J. Solution Chem., 1984, 13, 789. 8 G. Barone, P. Cacace, G. Castronuovo, A.Cesaro and V. Elia, Fluid Phase Equilibria, 1985, 20, 9 H. E. Kent, T. H. Lilley, P. J. Milburn, M. Bloemendal and G. Somsen, J . Solution Chem., 1985, 14, 303. 169. 101. 10 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J . Chem. Soc., Chem. Commun., 1985, 299. 11 G. M. Blackburn, T. H. Lilley and P. J. Milburn, Thermochim. Actu, 1985, 83, 289. 12 T. E. Leslie and T. H. Lilley, Biopolymers, 1985, 24, 695. 13 G. M. Blackburn, H. E. Kent and T. H. Lilley, J . Chem. SOC., Furaday Truns. I , 1985, 81, 2207. 14 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J . Chem. SOC., Furadaql Trans. I , 1985, 81, 2191. 15 G. Barone, G. Castronuovo, V. Elia and C. Giancola, J . Thermal Anal., 1985, 30, 1367. 16 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J.Solution Chem., 1986, 15, 99. 17 J. J. Savage and R. H. Wood, J. Solution Chem., 1976, 5, 733. 18 G. Barone, G. Castronuovo, V. Crescenzi, V. Elia and E. Rizzo, J. Solution Chem., 1978, 7, 179. 19 G. Barone, G. Castronuovo, V. Elia and A. Menna, J . Solution Chem., 1979, 8, 157. 20 R. H. Wood and L. H. Hiltzik, J . Solution Chem., 1980, 9, 45. 21 T. H. Lilley and R. H. Wood, J . Chem. Soc., Faraday Trans. 1, 1980, 76, 901. 22 G. Barone, G. Castronuovo, A. Cesaro and V. Elia, J . Solution Chem., 1980, 9, 867. 23 V. Abate, G. Barone, G. Castronuovo, V. Elia and P. Masturzo, Gazz. Chim. Ital., 1981, 111, 85. 24 V. Abate, G. Barone, P. Cacace, G. Castronuovo and V. Elia, J . Mol. Liquids (Adu. Mol. Relaxation 25 J. P. Grolier, J. J. Spitzer, R. H. Wood and I. R. Tasker, J . Solution Chem., 1985, 14, 393. 26 G. Barone, E. Rizzo and V. Vitagliano, J. Phys. Chem., 1970, 74, 2230. 27 G. Barone, G. Castronuovo, C. Della Volpe and V. Elia, J . Solution Chem., 1977, 6, 117. 28 G. Barone, P. Cacace, G. Castronuovo and V. Elia, Gazz. Chim. Ital., 1980, 110, 215. 29 G. Barone, P. Cacace, G . Castronuovo and V. Elia, J . Chem. SOC., Faraday Trans. I , 1981, 77, 30 H. L. Friedman and C . V. Krishnan, J . Solution Chem., 1973, 2, 119. 31 W. G. McMillan and J. E. Mayer Jr, J . Chem. Phys., 1945, 13, 276. 32 J. J. Kozak, V. S . Knight and W. Kauzmann, J. Chem. Phys., 1968, 48, 675. 33 F. Franks, M. D. Pedley and D. S. Reid, J . Chem. SOC., Faraday Trans. 1, 1976, 72, 359. 34 G. M. Blackburn, T. H. Lilley and P. J. Milburn, J. Chem. Soc., Faraday Trans. I , 1986, 82, 2965. 35 V. Abate, G. Barone, G. Castronuovo, V. Elia and V. Savino, J. Chem. SOC., Faraday Trans. I , 1984, 36 G. Barone, G. Castronuovo, V. Elia, Kh. Stassinopoulou and G. Della Gatta, J. Chem. Soc., Faraday 37 Y. B. Okamoto, R. H. Wood and P. T. Thompson, J . Chem. SOC., Faraday Trans. I , 1978, 74, 1990. 38 S. K. Suri, J. J. Spitzer, R. H. Wood, E. G. Abel and P. T. Thompson, J . Solution Chem., 1985, 14, 39 L. P. Pratt and D. Chandler, J . Solution Chem., 1980, 9, 1. 40 F. Franks, and M. D. Pedley, J . Chem. SOC., Faraday Trans. I , 1981, 77, 1341. 41 A. Ben-Naim and M. Yaacobi, J . Phys. Chem., 1974, 78, 170. 42 F. Franks, Philos. Trans. R. SOC. London, Ser. B, 1977, 278, 33. Interaction Process.), 1983, 27, 59. 1569. 80, 759. Trans. I , 1984, 80, 3095. 781. Paper 711 108; Received 22nd June, I987

ISSN:0300-9599    

DOI:10.1039/F19888401919

出版商:RSC    

年代:1988

数据来源: RSC