摘要:

J. Chem. SOC., Faraday Trans I, 1986,82, 2089-2101 Thermodynamics of Formation of Inclusion Compounds in Water a-Cyclodextrin-Alcohol Adducts at 298.15 K Guido Barone," Giuseppina Castronuovo, Pompea Del Vecchio, Vittorio Elia and Massimo Muscetta Department of Chemistry, University of Naples, Via Mezzocannone 4, 80134 Naples, Italy The interaction in water of ethanol, n-propanol, n-butanol, isobutanol, s-butanol and t-butanol with a-cyclodextrin (hexacycloamylose, crCD) has been studied calorimetrically at 298.15 K. The results indicate that, with the exception of the t-butanol, these alcohols form inclusion complexes with this dextrin. The calorimetric method employed allows the determination of the thermodynamic parameters characterizing the binding process, namely the enthalpy, APB, the apparent equilibrium constant, I&, and then the free energy, AGE, and entropy, TASOB/.The complexes that result are weak, as shown by the low values of the constants. The absolute values of AIPB and K;3 increase with increasing length of the alkyl chain. Branching of the alkyl chain lowers the value of Kb. Evidence is given to show that t-butanol does not form a complex with aCD, probably because of steric hindrance: for this system the calorimetric data are treated according to an approach derived from the McMillan-Mayer treatment of solutions. The most important property of cyclodextrins (cycloamyloses) is their ability to form inclusion complexes with a variety of organic substances, either in solutionl-A or in the solid lo These complexes are promising systems for widespread applications in technological, pharmaceutical and biological fields.1° The crystalline structures of stoichiometric complexes of cycloamyloses with water and/or alcohols have been determined by Saenger and coworkers 11-17 and by Stezowsky and l9 However, there are few hypotheses concerning the forces involved in these interactions, and many problems must be clarified concerning the mechanism of binding, the changes experienced by water in the hydration shells of the cyclodextrins and of the guest molecules, the conformational changes that the cyclodextrins undergo and finally the kinetics of the inclusion process.Our present contribution is part of a program aimed at understanding the forces involved in the interaction of a-cyclodextrin (aCD) with some non-electrolytes in aqueous 21 The interaction between aCD and ethanol, n-propanol, n-butanol, isobutanol, s-butanol and t-butanol has been studied calorimetrically at 298.15 K.Calorimetry was employed by Lewis and HansenZ2 for studying complexes of a- and B-cyclodextrin with a series of essentially aromatic organic molecules and ions and some inorganic electrolytes. Recently Maeda and Tagaki23~ 24 studied calorimetrically the interactions of a- and /3-CD with methanol, propanol and pentanol at 298.15 K. Our microcalorimetric approach, used as both an analytic and a thermodynamic tool, permits one to evaluate the apparent binding constants, K;3, to verify the stoichiometry of the reaction and to determine the enthalpic effect connected with the inclusion process, AWB.The other thermodynamic parameters, AG; and TAP;, can then be calculated. An accurate description of the enthalpic properties of the binary aqueous solutions of aCD 20892090 Formation of Inclusion Compounds in Water is given here for the first time. For systems which do not present the inclusion phenomenon, the experimental data will be treated according to the McMillan-Mayer approach for the solutions, as adapted by several a u t h o r ~ . ~ ~ - ~ ~ Experimental Materials a-Cyclodextrin was a crystalline Sigma product. Alcohols (Fluka products) were of the highest purity commercially available and were used without further purification. The solutions were freshly prepared by weight before each run using doubly distilled water.The concentration of each aCD solution was determined by optical rotation using sodium light. Calorimetry Calorimetric measurements were carried out with an LKB 10700- 1 flow microcalorimeter at 298.15 0.02 K. The following three kinds of experiments were arranged. (i) The determination of the heat of dilution from the initial (m’) to the final (m) molality, AHdil(m’ + m), of binary aqueous solutions of aCD (the heats of dilution of alcohols are known from the literature). The values of the experimental heats were obtained from AHdi1@’ --+ m) = - (dQ/dt)/P, where dQ/dt is the heat flux and Pw is the total mass flow rate of water through the calorimeter; AHdi1 is given in J per kg of solvent in the final solution. (ii) The determination of the heat of mixing AHmix of two binary aqueous solutions of the aCD and of an alcohol, keeping the final molality of the cyclodextrin approximately constant: (2) where (mk) and (m;) indicate the initial molalities of the binary solutions containing each of the solutes x and y, and (m,,rny) are the aquomolalities in the final ternary solutions of each solute. (iii) The determination of the heat of dilution in water of aqueous ternary solutions containing initially both aCD and an alcohol at the aquomolalities mk and mk, respectively : AHdil[(mk, mi) + (mx, my)] = - (dQ/dt)/P,.Details of the calorimetric experiments and of the accurate determination of the concentration changes during the dilution and mixing processes have been reported p r e v i ~ u s l y . ~ ~ - ~ ~ (1) AHmix[(mXm;> -+ (mx, my)] = - (dQ/dt)/P, (3) Treatment of the Data The values of the experimental heats of dilution for the aqueous binary solutions of aCD, can be used to fit the following power-law expansion : (1 / m ) AHdil(m’ + m) = hXx(m - m’) -t hxxx(m2 - mf2) + .. . . (4) The coefficients of eqn (4) have recently been reported in the literature for dilute and moderately concentrated aqueous solutions of 9 29-31 and can be used, together with our data, for determining the contributions to the total enthalpy changes from the dilution of each of the solutes during the mixing or dilution of ternary solutions. When there are indications that in a ternary solution an association process occurs, mixing experiments are useful for gaining information on the thermodynamic parameters characterizing that process. With regard to the present systems the main process assumed to occur is aCD+L = aCD.L ( 5 )G.Barone et al. 209 1 where L is a ligand (in this case an alcohol). The enthalpy of formation of the complex, or in general the enthalpy of interaction between solutes, can be evaluated through an auxiliary function AH*, defined as 2 9 9 32 AH* = AHm’X [(m:)(mS,) -+ (mx, my)] - AHdil(mk + my) - AHdi1@: + m,). (6) The standard molar enthalpy of inclusion (binding) of the ligand, AHO,, can be simply obtained from the relation where mmCD.L is the aquomolality of the adduct formed in the final solution. In the presence of a large excess of the ligand mmCD. + m,cD; hence at saturation A P B = Ap/macD.L (7) AwB = (AH*/maCD)satm (8) The enthalpy of inclusion of one guest molecule in the cycloamylose, APB, has the same significance as the standard molar enthalpy of binding per independent site in the case of the binding of small molecules to proteins or biopolymers.For this reason the symbolism used in these cases,38 was kept in this case. It is possible to relate AH*, normalized for the total molality of the dextrin, maCD, to the actual molality of the free ligand, mi, to the saturation value of eqn (8) and to the apparent association constant K;3 through the following relationship : The inverse of eqn (9) is useful for fitting the data, as it assumes a linear form, granted that the model is correct: m,,,/AH* = 1 /APB KL mfL + 1 /AHB. miA = mL - [AH*/AH*(sat)] macD (10) (1 1) where mL is the total stoichiometric molality of the ligand.ALhloB and Kb are obtained from eqn (10) and (1 1) by an iterative least-squares analysis of the data. The condition of best fit was assumed to be achieved when the change in two successive runs of the values of AWB was < 3%. The values of AG; and TAP,’ are then obtained from For each value of AH* the actual concentration (molality) of the free ligand is given by AG; = - R T In K;3; TAP; = AWB - AG;. (12) The absence of any information about the activity coefficients leads to the evaluation of binding parameters that are thermodynamically not exactly defined. Only an apparent constant, Kk, can be determined, and consequently the standard free energy and entropy suffer of the same limitations. If the simple model discussed previously does not describe the studied system, eqn (7)-(11) cannot be used: in the absence of the process defined by eqn (5), the least-squares iterative procedure does not converge, and saturation is not easily detectable.All this is a clear indication of the inadequacy of the model. In this case it is again possible to obtain information about the weak interactions between solutes by mixing two binary solutions, at varying concentrations of both solutes, or diluting ternary solutions. In the latter case another auxiliary function, AH**, can be introduced, defined as:32 AH** = AHdil[(m’ x, m’) + (mx, my)] - AHdil(rnk -+ m,) - AHdil(m; + my). (1 3) Expressing the dilution enthalpies as function of the self- and cross-interaction coefficients, it results that (14) AH**/my(m, -m;) = 2 4 , + 3h,,,(rn: +m,) +- 3h,,,(m; +my) + .. . .2092 Formation of Inclusion Compounds in Water Table 1. Heats of dilution in water of a-cyclodextrin at 298.15 K 0.1664 0.161 1 0.1440 0.1393 0.1226 0.1194 0.1058 0.0979 0.0896 0.082 1 0.0626 0.0573 0.0559 0.0779 0.0767 0.0689 0.0667 0.0591 0.0575 0.05 13 0.0475 0.0448 0.0400 0.0307 0.0282 0.0270 27.7 27.7 20.1 18.1 14.5 14.1 10.8 9.4 7.6 7.1 3.9 3.3 3.1 355.5 325.1 296.2 271.6 246.0 245.2 210.1 199.1 168.9 178.2 126.8 119.0 116.1 Table 2. Heats of mixing for aqueous solutions of a-cyclodextrin (aCD) and ethanol (EtOH) at 298.15 K 0.0468 0.0469 0.0469 0.0468 0.0468 0.0468 0.0468 0.0469 0.0469 0.0469 0.0469 0.0240 0.0238 0.0236 0.0234 0.0234 0.0233 0.0233 0.0233 0.0232 0.0232 0.0232 I .3898 0.9894 0.7288 0.4769 0.41 14 0.3402 0.2667 0.1897 0.0951 0.046 1 0.0245 0.06778 0.4870 0.3612 0.2380 0.2058 0.1704 0.1339 0.0953 0.0480 0.0233 0.0 124 203.9 115.5 76.3 50.3 43.6 37.5 31.5 24.1 14.0 5.5 1.8 24.2 34.8 36.8 36.0 33.8 31.6 28.8 23.9 15.6 7.5 3.9 1008 1462 1555 I538 1444 1350 1236 1026 672 323 168 The values of the cross-coefficients are obtained from a least-squares fit of the data with eqn (14).The polynomial of highest degree was chosen, whose coefficients still exceed their own 95% confidence limits. Results In table 1 the experimental heats of dilution of aqueous solutions of a-cyclodextrin at 298.15 K are reported, along with the initial and final molalities. According to the criterion quoted before only the second coefficient, h,, = - 3920 f 65 J mol-l (mol kg-l)-l, was found to be significant, within the 95 % confidence limit.Higher-order polynomials did not give other significant coefficients. This is due partly to the small range of concentrations available for the low solubility of aCD. In tables 2-7 AHmix, AH* and AH*/rnacD values are given, together with the initial and final molalities of aCD and of each of the following alcohols: ethanol (EtOH), n-propanol (PrnOH), n-butanol (BunOH), isobutanol (BuiOH), sec-butanol (BuSOH)G. Barone et al. 2093 Table 3. Heats of mixing for aqueous solutions of a-cyclodextrin (aCD) and n-propanol (Pr"0H) at 298.15 K - - p i x -AH* (-AH*/m,,,) mhCD maCD mbrnOH mPrnOH /J kg-l /J kg-l /J mol-1 - ~ _ _ _ _ ~~ 0.0663 0.0663 0.0646 0.0646 0.0646 0.065 1 0.065 I 0.065 1 0.0651 0.065 1 0.065 1 0.065 1 0.065 1 0.065 1 0.065 1 0.065 1 0.065 1 0.0349 0.0343 0.0330 0.0327 0.0325 0.032 1 0.0323 0.0323 0.0321 0.0320 0.0320 0.03 I9 0.03 19 0.03 19 0.03 18 0.03 18 0.03 18 2.1291 1.5504 1.1422 0.8837 0.7003 0.5432 0.4647 0.3856 0.307 1 0.2192 0.1828 0.1539 0.1249 0.1 139 0.0740 0.0489 0.0322 1.0089 0.7490 0.5582 0.4359 0.3475 0.275 1 0.2341 0.1943 0.I556 0.1113 0.093 1 0.0784 0.0637 0.058 1 0.0378 0.0250 0.0 165 1320 691.7 433.2 3 13.4 253.0 213.7 197.0 181.1 167.8 146.3 135.2 123.1 105.4 102.4 76.3 54.0 36.3 131.2 141.9 167.3 167.6 168.3 167.1 165.0 161.1 157.1 143.1 134.3 123.8 107.3 104.6 79.6 57.8 40.3 3759 41 37 5070 5125 5178 5206 5108 4988 4894 4472 4197 3881 3364 3279 2503 1818 1267 Table 4. Heats of mixing for aqueous solutions of a-cyclodextrin (aCD) and n-butanol (BunOH) at 298.15 K 0.0465 0.0468 0.0465 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0469 0.0239 0.0241 0.0234 0.0237 0.0235 0.0234 0.0233 0.0233 0.0232 0.0232 0.023 1 0.0232 0.023 I 0.0230 0.9930 0.9706 0.8373 0.7386 0.5737 0.5009 0.4255 0.3509 0.2824 0.2087 0.1415 0.0803 0.0571 0.0283 0.48 16 0.4706 0.4 150 0.3642 0.2850 0.2489 0.2 I32 0.1762 0.1422 0.1054 0.0717 0.0405 0.0290 0.0144 694.4 657.7 528.3 455.8 350.0 31 1.0 276.0 25 1.4 237.0 217.9 195.4 169.6 138.7 82.3 2 14.9 204.7 2 12.8 224.0 224.0 219.8 214.1 2 12.2 213.7 206.9 19 I .8 170.0 140.0 84.3 8992 8494 9094 945 1 9523 9393 9189 9107 921 1 8918 8303 7328 606 1 3665 and t-butanol (ButOH).The values refer to the mixing of two binary solutions, keeping the final molality of aCD approximately constant.In fig. 1 the AH*/macD (J mol-l) values are reported as a function of the ratio between the final molality of the alcohols and that of aCD. These AH*/macD values are negative and tend toward a saturation plateau, after which they differentiate. The values concerning the two branched alcohols i- and s-butanol seem to increase towards a new, more complicated binding equilibrium, while the normal alkanols seem to show a slightly decreasing trend at high m,l,oho,/m,cn ratios. The AWB and K& values are evaluated by the above-mentioned iterative least-squares analysis of the data, using eqn (10) and (1 1) and considering only that part2094 Formation of Inclusion Compounds in Water Table 5.Heats of mixing for aqueous solutions of a-cyclodextrin (aCD) and s-butanol (BuSOH) at 298.15 K maCD -AH* /J kg-l 0.0468 0.0468 0.0468 0.0469 0.0469 0.0469 0.0469 0.0469 0.0468 0.0469 0.0468 0.0468 0.0238 0.0236 0.0235 0.0238 0.0235 0.0234 0.0233 0.0233 0.0232 0.0232 0.0232 0.0232 0.6779 0.5555 0.3902 0.3318 0.2609 0.1994 0.1500 0.1025 0.0525 0.0477 0.0380 0.0262 0.3329 0.2748 0.1941 0.1635 0.1304 0.1000 0.0753 0.0516 0.0264 0.024 1 0.0 192 0.01 32 323.8 265.7 211.1 194.6 177.3 156.2 137.5 111.1 71.8 64.9 54.6 39.0 220.8 197.3 178.4 171.6 163.9 149.3 134.5 110.8 73.3 66.6 56.5 41.0 9277 8360 759 1 7210 6974 6380 5773 4755 3159 287 1 2435 I767 Table 6. Heats of mixing for aqueous solutions of a-cyclodextrin (olCD) and isobutanol (BdOH) at 298.15 K - A p i X -AH* (-AH*/macD) m:cD maCD mLuiOH mBuiOH /J kg-' /J kg-' /J mo1-l 0.0468 0.0467 0.0467 0.0468 0.0468 0.0468 0.0468 0.0467 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0468 0.0467 0.0240 0.0239 0.0238 0.0239 0.0238 0.0237 0.0237 0.0235 0.0236 0.0235 0.0234 0.0232 0.0232 0.0232 0.023 1 0.023 1 0.9353 0.08589 0.7941 0.7835 0.7424 0.6449 0.5937 0.5117 0.4959 0.4018 0.2926 0.2056 0.1203 0.0849 0.0558 0.0262 0.4555 0.4192 0.3900 0.3839 0.3643 0.3178 0.2928 0.2536 0.2458 0.2001 0.1463 0.1032 0.0605 0.0428 0.0282 0.0132 539.5 472.9 410.6 398.0 374.6 316.0 295.8 253.1 248.2 213.0 178.0 148.5 112.9 91.0 67.1 34.6 323.3 290.9 255.3 246.9 239.2 214.3 210.0 189.9 188.9 174.8 158.7 140.1 11 1.5 91.3 68.5 36.6 1 347 1 12172 10727 1033 1 10050 9042 886 1 808 1 8004 7438 6782 6013 4806 3935 2965 1584 of the curve which goes up to the plateau.Completely different is the behaviour of the system aCD-ButOH, for which were found positive AH*/macD values, and was not possible to detect the presence of a plateau. Finally, in tables 8 and 9 the heats of dilution and the function AH** are reported for the ternary systems aCD-PrnOH and aCD-ButOH along with the initial and final molalities of the species involved. In these cases the cross interaction coefficients were obtained through eqn (14) with a least-squares method. In table 10 the h,, values of the aCD aqueous solutions are compared with those concerning alcohols and other oligosaccharides: only for aCD do we find h,, < 0, while for other sugars the h,, values are positive and increase with increasing molecular 3 9 9 40 Positive values of h,,G.Barone et al. 2095 Table 7. Heats of mixing for aqueous solutions of a-cyclodextrin (aCD) and t-butanol (ButOH) at 298.15 K 0.0558 0.0463 0.0463 0.0463 0.0463 0.0463 0.0463 0.0458 0.0458 0.0458 0.0458 0.0458 0.0236 0.0237 0.0235 0.0232 0.0232 0.023 1 0.0230 0.0228 0.0227 0.0226 0.0226 0.0227 1.1739 0.9840 0.7336 0.5005 0.4249 0.3525 0.2939 0.2376 0.2026 0.0205 0.0604 0.1182 0.5677 0.4802 0.3619 0.2492 0.2121 0.1766 0.1476 0.1193 0.1019 0.0104 0.0306 0.0596 - 384.5 - 237.3 -9.2 - 32.1 - 15.9 - 6.7 - 2.5 1.1 2.6 2.8 3.5 4.2 39.6 37.7 36.2 22.6 21.2 17.0 12.7 10.0 8.4 0.8 2.1 4.7 1678 1591 1540 974 914 736 552 439 370 35.4 92.9 207 Fig. 1. AH*/mcD plotted against the ratio between the final molality of alcohols and the final molality of a-cyclodextrin. The concentration of &D was kept approximately constant (ca.0.023 mol kg-l, except for aCD-Pr"OH, for which it was ca. 0.033 mol kg-l). (a) BuiOH, (b) BunOH, (c) BuSOH, (d) PrnOH, (e) EtOH and ( f ) ButOH. are also shown by other monosaccharides, 41-46 alcohols, 29-319 47-52 and glycols and p ~ l y o l s , ~ ~ - ~ ~ except for sorbitol, perseitol and the cyclitol myo-in~sitol.~~* 55 The h,, values for aCD and alcohols have been used to calculate the AH* and AH** values given in the last columns of tables 2-9, by means of eqn (4), (6) and (13).2096 Formation of Inclusion Compounds in Water Table 8. Heats of dilution for ternary aqueous solutions of a-cyclodextrin (aCD) and n-propanol (PPOH) at 298.15 K 0.1 136 0.09 I6 0.0773 0.0663 0.0654 0.0553 0.0462 0.0398 0.0383 0.0333 0.0328 0.0254 0.0204 0.0131 0.0552 0.0449 0.0379 0.0327 0.0322 0.0273 0.0225 0.0195 0.0191 0.0 163 0.0 163 0.0125 0.0102 0.0065 0.1180 0.095 1 0.1637 0.0689 0.1385 0.1 171 0.1043 0.0899 0.08 1 1 0.0752 0.0694 0.0574 0.0433 0.0295 0.0573 0.0466 0.0803 0.0340 0.0682 0.0579 0.0508 0.0440 0.0404 0.0368 0.0346 0.0282 0.02 17 0.0146 50.6 41.5 30.4 30.2 29.2 25.3 22.2 20.0 19.1 16.8 16.0 12.8 8.9 5.1 40.0 34.6 28.6 26.6 27.8 24.4 21.7 19.6 18.6 16.5 15.7 12.7 8.7 5.0 Table 9.Heats of dilution for ternary aqueous solutions of a-cyclodextrin (aCD) and t-butanol (ButOH) at 298.15 K 0.0649 0.066 1 0.065 1 0.0562 0.0560 0.0550 0.0459 0.038 1 0.0406 0.048 1 0.0414 0.0344 0.0322 0.0329 0.0324 0.028 1 0.0279 0.0275 0.0230 0.0191 0.0206 0.0241 0.0208 0.0 174 0.1785 0.1785 0.1659 0.1517 0.1541 0.1402 0.1263 0.1049 0.1034 0.1012 0.087 1 0.0724 0.0886 0.0888 0.0826 0.0757 0.0768 0.0702 0.0633 0.0527 0.0524 0.0507 0.0438 0.0366 ~ 15.4 14.9 13.3 10.2 9.8 8.1 7.2 5.2 4.6 4.5 2.4 1.9 13.5 13.2 12.4 9.0 8.5 7.4 6.4 4.6 4.3 4.9 2.7 2.2 Discussion The structure of a-cyclodextrin (aCD) in water can be hypothesized to resemble that of aCD hexahydrate in the crystalline state.Namely, the ‘void’ molecule of aCD has two water molecules entrapped in the cavity, hydrogen-bonded to each other and to two OH groups of the glucopyranose rings. The formation of a complex implies, at first, that these water molecules are ‘squeezed out’ on relaxing to the bulk, and then the guest molecule can penetrate the cavity with a modification of the solvent in its hydration cosphere.In the literature there are many data concerning the crystalline structures of the inclusion complexes between aCD and some alcohols. 12-19 The interactions between the ‘host’ and ‘guest’ molecules must be weak, and largely due to dispersion and van der Waals forces. For this reason, the smaller molecules are statistically disordered at the interior of the cavity, even though hydrogen bonds could involve the OH(6) groups. In solution the low values of the association constants are an indication of theG. Barone et al. 2097 Table 10. Selected values of the coefficients of the excess enthalpies of aqueous binary solutions of some hydroxylated compounds at 298.15 K compound hxx /J mol-l(mo1 kg-l)-l hxxx /J mol-l(mo1 kg-1)-2 aCD D-galactosea D-mannosea D-fructosea L-sorboseC lactosea sucroseb ma1 tosed trehalosed cello biased raffinosea melezitosed MeOH EtOH Pr*OH BunOH BuiOH BuSOH ButOH D-glucoseb - 3920 (65) 133 (8) 207 (1 4) 343 (10) 264 (18) 395 (9) 506 (32) 577 (6) 483 (7) . 605 (20) 756 (14) 811 (50) 607 (25) 218e9h 242.9 (9.6)f~ ~ - - 14 (5) - 13 -7 (4) - 16 (4) -33 (8) - - - ~ - - - 64.8 (5.4) 558.9 (14.2)fg 158.4 (7.9) 1003 (15)flh 646 (56) 91699 - 1 ooog, - 656 (33)f7 334.4 (1 8.4) a Ref.(39). Ref. (31). Ref. (42). Ref. (40). Ref. (51). f Ref. (29). 9 Ref. (3 I ) and (49). Data used for the calculations of the present work; other values are reported in the literature [ref. (30), (47), (48) and (50)] even very recently [ref.(52)]. The figures in parentheses are the 95% confidence limits. formation of weak complexes. To understand better which type of interactions exists between the two molecules, it is important to know how the guest molecules penetrate the cavity, if the alkyl chain or the hydroxy group first. The differentiation among the alcohols studied, we think, is a proof that the alcohols enter the cavity through the alkyl chain. In fact, AWB becomes more negative and K;3 increases as the normal alkyl chain lengthens as reported in table 1 1 . Probably the fitting to the cavity improves as the dimensions of the guest molecule increase. However, the branched butanols give weaker complexes with respect to the linear one, as shown by the data of table 11 and as can be deduced also from fig. 1.The small dimension of the aCD cavity is the cause of a steric hindrance that, in the case of ButOH, is such to prevent the formation of a stable inclusion complex. In table 1 1 the values of K& and AWB for the aCD-ButOH system are not given, owing to the difficulty of evaluating saturation values. However, a rough estimate of these quantities gives a Kk value lower than unity and a AWB value positive, thus indicating that, in this case, the above chemical model is devoid of significance. This is in agreement with the results obtained for the interaction between aCD and urea or the alkylureas.21 The interaction becomes stronger with increasing length of the alkyl chain: for monomethylurea it is not possible to obtain a value for KL, probably because the complex is too labile, owing in part to the presence of interactions with urea residue and in part to the very weak interaction between a simple methyl residue and the cavity of aCD.This can also be deduced from fig. 2, where the AWB values are plotted against the number of aliphatic carbon atoms, nc, of the alcohols. It can be extrapolated that2098 Formation of Inclusion Compounds in Water Table 11. Thermodynamic parameters for the binding of guest alcohols to the a-cyclodextrin at 298.15 K -AG; -AWB TAP; G3 system /kJ mol-1 /kg mol-l aCD-EtOH 4.7 2.5 2.2 6.7 &D-PrnOH 8.2 6.1 2.1 27.0 aCD-Bu"OH 11.4 9.9 1.5 99.9 aCD-BuSOH 8.3 9.0 -0.7 28.4 aCD-Bu"OH 7.7 9.4 - 1.7 21.9 I I 1 1 3 5 nC Fig. 2. AIPB plotted against the number of carbon atoms, n,, in a molecule of alcohol. The values given as open circles are from ref.(24). the formation of aCD-MeOH adduct is about athermal, as we found experimentally. The agreement with the AWB values reported by Maeda and Takagi is very g 0 0 d . ~ ~ 1 ~ ~ In table 11 the thermodynamic parameters AG; and TASE are also reported for systems other than aCD-ButOH. As for other systems reported in the literature the AS: values are positive or negative, thus indicating that several forces giving contrasting effects are involved in complex formation. A plot of AWB against AS:, reported in the literature22 for many aCD-guest complexes, shows an almost linear relationship for the complexes with AS; negative or small and positive. No positive values of AWB were observed.Hence, changes in A F B will compensate for changes in A9,l to a large degree. This compensation effect, frequently observed in water, has been related to changes of the solvent in the hydration cospheres of the interacting substances. However, note that some substances fall beyondG . Borone et al. 2099 this rough linearity, e.g. indole and 2-aminobenzoic acid. They are characterized by a very small and negative AffB for the inclusion process, by a high and positive AS: and by very high association constants. Hence, in these cases the inclusion would be an entropy-driven phenomenon, relying on the modifications experienced by the solvent upon the association process. Namely, the gain in entropy for water relaxed from the hydration cospheres to the bulk is only partially balanced by a decrease in entropy due to the formation of the complex.Accordingly, the nature of the hydration cosphere of the guest molecule plays an important and often fundamental role, making possible or not the formation of a stable inclusion complex. The signs of the AWB (negative) and TAP,’ (negative or positive) for the systems investigated are an indication that, as for a- or p-CD complexes, the hydrophobic interactions do not always play the major role in the formation of these inclusion compounds. Dipole-induced dipole, ‘ host-guest ’ interactions (remember that the glycosidic oxygens come out from the inner surface of the aCD cavity) and the decrease of the cavitation energy when a hydrophobic residue fills the cavity must also be considered as important effects.Other processes that can be involved are overall endothermic, as (1) the squeezing out to the bulk of the two water molecules anchored to the ring with consequent relaxation of the aCD ‘tense’ conformation (in these rearrangements the external hydration shell also undergoes some resettlemeiits) and (2) the relaxation to the bulk of the shell of water molecules surrounding the alkyl chain of the guest molecule (the hydrophobic hydration co~phere~~), which is only in a small part compensated by the building up of the hydration shell of the complexes. In considering the TAS; values, it must be reminded that the cratic contribution at 298.15 K is ca. 10 kJ mol-l. This quantity must be added to the values reported in table 11 for evaluating the entropy changes involved primarily in the expulsion of water from the hydration shells of both solutes, that in turn overwhelm the losses of degree of freedom experienced by the alkyl derivatives.As stated previously, the system aCD-ButOH behaves in a different manner from others. Hence, this system has been treated according to the McMillan-Mayer approach for real solutions. Heats of dilution of ternary solutions permit one to evaluate the cross interaction coefficients through eqn (14). In table 12 the experimental cross interaction coefficient h,, for the system aCD-ButOH is reported: it is positive and its value falls outside the range of the homogeneous coefficients for aCD and ButOH [ - 3920 and 656 J mol-1 (mol kg-l)-l, respectively]. For comparison, in table 12 the cross interaction coefficients for aqueous solutions of aCD-PrnOH, as obtained by the heats of dilution of ternary solutions, are also given.The high value of the coefficient h,, (where x is aCD and y is PrnOH) reported in table 12 suggests that interactions, not specific, involving one aCD and two PrnOH molecules would be possible at high alkanol concentrations. On the other hand, the positive h,, value for aCD-ButOH system could be ascribed to a kind of weak interaction, probably prevailingly hydrophobic, that does not lead to the formation of an adduct. This kind of ‘secondary’ interaction probably also occurs at increasing concentrations of the other alkanols: it could account for the relative decrease of the AH*/macD values at high malcohol/maCD ratios.As shown in table 10, the self-interaction coefficient, h,,, for aCD is negative, while for all the other saccharides studied it is positive, roughly increasing along the series mono-, di- and tri-saccharides. The behaviour of binary aqueous solutions of aCD resembles that of some higher members of the polyols, whose coefficients become negative with increasing molecular weight and depend on the stereochemistry of the solute. A suggestive case is represented by myo-inositol, a cyclic hexanol, whose co- efficient is negative and relatively high in absolute value. For the polyols, this trend has been explained in terms of the separation of the hydrophilic and hydrophobic domains in the solute m 0 1 e c u l e . ~ ~ ~ ~ ~ Namely, as the separation becomes more pro- nounced, the negative contribution to the self-interaction coefficients seems to increase.2100 Formation of Inclusion Compounds in Water Table 12.Enthalpic interaction coefficients for aqueous solutions of ol-cyclodextrin, t-butanol and n-propanol from dilution experiments at 298.15 K x = aCD, y = PrnOH - 3920 (65) 558.9 (14.2)' - 18 623 (1690) 18 752 (5839) 30 392 (4759) x = aCD, y = ButOH - 3920 (65) 656 (33)" 2167 (104) - - Units: J mol-l(mo1 kg-l)-l. Units: J mol-l(mo1 kg-1)-2. Ref. (29). The sign and absolute value of the aCD self-coefficient, haCD-aCD, which shows this feature particularly well, leads to the hypothesis that ccCD maintains in solution the con- figuration prevailing in the solid state, where it has been found to be hydrophilic outside the cavity and hydrophobic inside.Work is in progress in our laboratory to clarify this and other aspects of cyclodextrin aqueous solutions as well as of the interactions with alcohols and other families of simple organic molecules. This work was supported by the Italian National Research Council (C.N.R.), Target Project on Fine and Secondary Chemistry and by the Ministry of the Public Education (M.P.I.), Rome. References 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 F. Cramer, W. Saenger and H. Ch. Spatz, J . Am. Chem. Soc., 1967, 89, 14. R. L. van Etten, J. F. Sebastion, G. A. Clowes and M. L. Bender, J . Am. Chem. SOC., 1967,89, 3242. C. van Hooidonk and J. C. A. E. Breebart-Hansen, Recl. Trav.Chim. Pays-Bas, 1971, 90, 680. T. S. Straub and M. L. Bender, J . Am. Chem. SOC., 1972, 94, 8881. D. W. Griffiths and M. L. Bender, Adv. Catal., 1973, 23, 209. R. J. Bergeron, M. A. Channing, G. J. Gibeily and D. M. Pillor, J . Am. Chem. SOC., 1977, 99, 5146. M. Komiyama and M. L. Bender, J. Am. Chem. SOC., 1978,100, 2259. I. Tabushi, Y. Kiyosuka, T. Sugimoto and K. Yamamura, J . Am. Chem. SOC., 1978, 100, 916. R. K. Mcmillan, W. Saenger, J. Fayos and D. Mootz, Carbohydr. Res., 1973, 31, 37. W. Saenger, Angew. Chem., Int. Ed. Engl., 1980, 19, 344. P. C. Manor and W. Saenger, Nature (London), 12, 237, 392. P. C. Manor and W. Saenger, J . Am. Chem. Soc., 1974, 96, 3630. W. Saenger, R. K. McMullan, J. Fayos and D. Mootz, Acta Crystallogr., 1974, 330, 2019. B. Hingerty and W.Saenger, Nature (London), 1975, 255, 396. B. Hingerty and W. Saenger, J . Am. Chem. SOC., 1976, 98, 3357. K. Lindner and W. Saenger, Angew. Chem., Int. Ed. Engl., 1978, 17, 694. K . Lindner and W. Saenger, Biochem. Biophys. Res. Commun., 1980, 92, 933. J. J. Stezowsky, K. H. Jogun, E. Ekkle and IS. Bartels, Nature (London), 1978, 274, 617. J. M. MacLennan, J. J. Stezowski, Biochem. Biophys. Res. Commun., 1980, 92, 926. G. Barone, G. Castronuovo, V. Elia and M. Muscetta, Therrnochirn. Acta, 1985, 85, 443. G. Barone, G. Castronuovo, V. Elia and M. Muscetta, J. Solution Chem., 1986, 15, 129. E. A. Lewis and L. D. Hansen, J . Chem. SOC., Perkin Trans. 2, 1973,2081. M. Maeda and S. Tokagi, Netsusokutei, 1983, 10, 43. M. Maeda and S. Tokagi, Netsusokutei, 1983, 10, 103.W. G. McMillan Jr. and J. E. Mayer, J . Chem. Phys., 1945, 13, 276. J. J. Kozak, W. S. Knight and W1 Kauanann, J. Chem. Phys., 1968, 48, 675. H. L. Friedman and C. V. Krishnan, J . Solution Chem., 1973, 2, 119. T. H. Lilley and R. P. Scott, J. Chem. SOC., Faruday Trans. 1, 1976,72, 184. F. Franks, M. D. Pedley and D. S. Reid, J . Chem. SOC., Faraday Trans. 1, 1976, 72, 359. J. E. Desnoyers, G. Perron, L. Avedikian and J-P. Morel, J . Solution Chem., 1976, 5, 631. J. J. Savage and R. H. Wood, J . Solution Chem., 1976, 6, 733. G. Barone, P. Cacace, G. Castronuovo and V. Elia, J. Chem. Soc., Furuduy Trans. I , 1981, 77, 1569.G. Barone et al. 2101 33 V. Abate, G. Barone, G. Castronuovo, V. Elia and V. Savino, J . Chem. Soc., Faraday Trans. 1 , 1984, 34 G.Barone, P. Cacace, V. Elia and A. Cesaro, J . Chem. Soc., Faraday Trans. 1, 1984,80, 2073. 35 G. Barone, G. Castronuovo, G. Della Gatta, V. Elia and Kh. Stassinopoulou, J . Chem. SOC., Faraday Trans. 1, 1984, 80, 3095. 36 G. Barone, G. Castronuovo, V. Crescenzi, V. Elia and E. Rizzo, J . Solution Chem., 1978, 7, 1979. 37 Water: a comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1975), vol. 4. 38 M. Eftink and R. Biltonen, in Biological Microcalorimetry, ed. A. Beezer (Academic Press, London, 1980). 39 G. Barone, P. Cacace, G. Castronuovo and V. Elia, Carbohydr. Res., 1981, 91, 101. 40 G. Barone, P. Cacace, G. Castronuovo, P. Del Vecchio and V. Elia, to be published. 41 G. Barone, P. Cacace, G. Castronuovo, V. Elia and F. Iappelli, Carbohydr. Res., 1981, 93, 11. 42 G. Barone, P. Cacace, G. Castronuovo and V. Elia, Gazz. Chim. Ztal., 1982, 112, 153. 43 G. Barone, G. Castronuovo, D. Doucas, V. Elia and C. A. Mattia, J . Phys. Chew., 1983, $7, 1931. 44 G. Barone,, P. Cacace, G. Castronuovo, V. Elia and U. Lepore, Carbohydr. Rex, 1983, 115, 15. 45 G. Barone, P. Cacace, G. Castronuovo and V. Elia, J . Solution Chew., 1984, 13, 625. 46 G. Barone, P. Cacace, G. Castronuovo, V. Elia, M. Muscetta and KH. Stassinopoulou, Fluid Phase 47 E. Lange and H. G. Markgraf, Z . Elektrochem., 1950,54, 73. 48 W. Dimmling and E. Lange, Z . Elektrochem., 1951, 55, 322. 49 E. Lange and K. Mohring, 2. Elektrochem., 1953, 57, 660. 50 R. B. Cassel and R. H. Wood, J . Phys. Chem., 1974,78, 2460. 51 G. Perron and J. E. Desnoyers, J . Chem. Thermodyn., 1981, 13, 1105. 52 D. Halltn, S-0. Nilsson, W. Rothschild and I. Wadso, J. Chem. Therrnodyn., in press. 53 G. Barone, B. Bove, G. Castronuovo and V. Elia, J . Solution Chem., 1981, 10, 803. 54 I. R. Tasker and R. H. Wood, J . Phys. Chem., 1982, 86, 4040. 55 G. Barone, P. Cacace, G. Castronuovo and V. Elia, Carbohydr. Res., 1983, 119, 1. 56 F. Franks, M. D. Pedley, J. Chem. Soc., Faraday Trans. 1, 1983, 79, 2249. 80, 759. Equilibria, 1985, 20, 177. Paper 511382; Received 8th August, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202089

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chern. Soc., Faraday Trans. I , 1986,82,2103-2110 Reactions of Hydroxyalkyl Radicals with Uracil Sudhindra N. Bhattacharyya* and Parikshit C. Mandalt Nuclear Chemistry Division, Saha Institute of Nuclear Physics, Sector-1, Block-'AF', Bidhannagar, Calcutta 700 064, India The base degradation yields when aqueous solutions of uracil containing methyl, ethyl or t-butyl alcohol were y-irradiated, were 2.0 0.2, 4.2 f 0.5 and 1.3 & 0.1, respectively. These differences are attributed to differences in reactivity of CH,OH (I), CH,CHOH (11) and CH,(CH,),COH (111) with uracil (U). I11 does not react with U whereas I and TI react with U to form U- and aldehyde. To obtain an insight into the mechanism of the radiation-chemical degradation of DNA, many radiolytic studies with model compounds such as pyrimidine bases have been undertaken.'* Recently3 it has been established that the reducing species, eiq, which is an important product of water radiolysis, inactivates the DNA molecule, and studies with individual pyrimidine bases have indicated that eLq readily forms adducts with them, leading ultimately to their degradation.l? 4-6 Reactions of eLq are usually studied in the presence of OH scavengers such as alcohols.Hence a study of the reactions of the hydroxyalkyl radicals that are formed as a result of the reaction of OH with an alcohol with pyrimidine bases is important in this context. The hydroxyalkyl radical is formed ztia where ROH represents the alcohol and RO the hydroxyalkyl radical. In this paper we show that the methanol and ethanol radicals reduce uracil in deaerated solution, whereas no such reduction was discernible in case of radicals derived from t-butyl alcohol.ROH+OH + RO+H,O (1) Experimental Materials Uracil (Merck) used was recrystallised three times from triply distilled water. [2J4C]uracil (14 mCi dm3 mmol-l) was procured from BARC, Bombay, Other chemicals and solvents were of analytical-reagent grade. Deaeration was carried out by bubbling pure argon gas through each 5 cm3 of experimental solution for 30 min. Pure N20 was used in the investigation. Irradiation The irradiation source comprised a 6oCo gamma cell. The dose rate (1.92 x lo1' eV min-l or 30.7 Gy min-l) was measured with the help o f a Fricke dosimeter, taking G(Fe3*) = 15.6; Fe3+ was determined spectrophotometrically at 305 nm.Analysis Uracil solutions containing various alcohols were irradiated in neutral solution (pH 5-6) without a buffer. No significant shift in pH was observed after irradiation in the dose range (3-40) x 1017 eV cmP3, i.e. 48-640 Gy. The products of radiolysis were separated t Present address : Research Institute for Nuclear Medicine and Biology, Hiroshima University, Hiroshima, Japan. 21032104 Reactions of Hydroxyaikyi Radicals with Uracil by paper chromatography using n-butyl alcohol-water (86: 14) mixtures as solvent and they were identified as described earlier.’-ll The base degradation yield and the yields of the radiolytic products were determined from the relative activity in the area of the respective peaks in the active chromatogram.The degradation of the base was also determined by spectrophotometry from the loss of absorbance at 260 nm. The carbonyl compounds were determined spectrophotometrically by using DNPH reagent. l2 The extinction coefficient was determined with respect to formaldehyde only. Results and Discussion Both the degradation of uracil and the product formation in the case of radiolysis of uracil in the presence of alcohols were determined at different absorbed doses, and the respective G values were ascertained from plots of yield vs. dose, which were found to be linear. The observed G values are shown in tables 1 and 2. Fig. 1 shows the effect of t-butyl alcohol concentration on the observed G(-U) values under different conditions. It is evident from the figure that under argon- saturated conditions G( - U) decreases with increasing [ButOH] until [ButOH] = 5 x lop3 dm3 mol-l, whence there is no further change in G( -U) with increasing alcohol concentration.At this high concentration of alcohol the OH radicals are scavenged,ll leaving only eiq. Hence the observed plateau value of G( - U) is due to elq. only. This is further corroborated by the observation that G(-U) becomes insignificant when radiolysis is carried out in the presence of t-butyl alcohol (> 5 x Fig. 2 shows the variation of G( - U) with MeOH concentration. A consideration of the rate-constant data1, l3 indicates that at the high concentrations of methyl alcohol [(2-100) x mol dm-3] 51-98% of the OH radicals will be scavenged, the result being the formation of CH,OH radicals.The G( -U) value at such high concentrations of MeOH is ca. 1.3. Similar results were also obtained with EtOH. The effect of uracil concentration on G(-U) under conditions where OH radicals are expected to be scavenged by alcohols was also studied, and the results are shown in fig. 3. Note that the range of uracil concentration chosen is sufficient to attain steady-state conditions for its radiolysis. Note also that the steady values of G( -U) (fig. 3) under such conditions were found to be 2.0 0.2 for MeOH, 4.2 & 0.5 for EtOH and 1.3 0.1 for ButOH. Under these conditions eiq reacts with uracil, giving U- which subsequently protonates to mol dm-3) in air- or N,O-saturated solution. UH 11, 4-6, 1 4 U+eiq + U- (2) U-+H++UH. (3) Hence these differences in G( - U) values may be attributed to the differences in reactivity of hydroxyalkyl radicals (RO) with uracil : RO + U -+ products RO + UH + products.(4) ( 5 ) or with the protonated electron adduct of uracil, UH : In the presence of MeOH, however, when the radiolysis was carried out in N,O-saturated solution, the same yield values as in deaerated solutions were also observed (fig. 2). Under these conditions eiq is scavenged by N,O and hence neither U- nor UH is formed from elq. Neglecting the small contribution of [G(H) z 0.51 it may be assumed that the system comprises mostly uracil and RO. Hence it must be concluded that the major pathway for uracil degradation under these conditions is the reaction between RO and uracil [reaction (4)]. Now, depending upon its reactivity, the hydroxyalkyl radical (RO) might undergoS.N . Bhattacharyya and P. C . Mandal 2105 Table 1. Yields of carbonyl compounds and G( - U) in the y-radiolysis of uracil (2 x mol dm-3) in the presence of various alcohols [alcohol] alcohol /mol dm-3 condition G( - U)" G( > C=O) MeOH 0.4 argon-saturated 2.0 1.5 MeOH 0.4 N,O-saturated 2.0 3.6 EtOH 0.4 argon-saturated 4.5 2.4 EtOH 0.4 N,O-saturated 4.5 5.0 ButOH 0.4 argon-saturated 1.3 0.0 ButOH 0.4 N,O-saturated 0.2 0.0 a Measured by loss of absorbance at 260 nm. Table 2. Yields of various radiolytic products in the y-radiolysis of uracil in the presence of alcohols at ca. pH 5.6 ~~ condition G(products) A B G(dimer) 0.2 (20%) 0.2 (9%) G(hydroxydihydrouraci1) 0.2 (20%) 0.3 (13 % ) G(dihydrouraci1) 0.4 (40%) 1.6 (69%) G( - uracil) 1.0 (1.3)a 2.3 (2.0)' ~ ~~~~ (A) 2 x mol dm-3 uracil in the presence of 0.4 mol dm-3 ButOH in argon-saturated solution.(B) 2 x mol dm-3 uracil in the presence of 0.4mol dm-3 MeOH in argon-saturated solution. a Obtained from the loss of absorbance at 260 nm. [ButOH]/mol dm-3 Fig. 1. Effect of t-butyl alcohol concentration on G( - U) : 0, N,O-saturated solution; 0, argon-saturated solution; A, aerated solution; pH 5.6. 70 FAR 12106 Reactions of Hydroxyalkyl Radicals with Uracil 3 2 5 I u v 1 --0 0 A I I I I 1 1 1 1 I I I 1 1 1 1 1 I I I I I I l l 0.1 1.0 10 100 [ MeOH]/mmol dm-3 Fig. 2. Effect of methyl alcohol concentration on the radiolytic decomposition yield of uracil: 0, aerated solution; 0, N,O-saturated solution; A, argon-saturated solution; [U] = 2 x mol dm-3, ca. pH 5.5.0 4 8 12 16 20 [ U]/ 1 OA4 mol dm-3 Fig. 3. Effect of increasing concentration of uracil on G( - U) in the y-radiolysis of uracil in the presence of methyl, ethyl and t-butyl alcohol at pH 5.6: A, in the presence of 0.4 mol dm-3 EtOH in argon-saturated solution; A, in the presence of 0.4 mol dm-3 EtOH in N,O-saturated solution; ., in the presence of 0.4 mol dm-3 MeOH in argon-saturated solution; 0, in the presence of 0.4 mol dmP3 MeOH in N,O-saturated solution; 0, in the presence of 0.4 mol dm-3 ButOH in argon-saturated solution.S. N. Bhattacharyya and P . C. Mandal 2107 reaction with uracil to form an alcohol adduct, as observed by Kamal and Garrison15 and Brown et aZ.16: (6) CH,OH + U + U-CH,OH or undergo oxidation by uracil, giving U- and carbonyl compound:17 CH,OH+U + U-+CH,O+H+.(7) The formation of carbonyl compounds in the presence of MeOH and EtOH (table 1) indicates the occurrence of reaction (7). This is further supported by the fact that G(carbony1) yields increased appreciably in N,O-saturated solution, whereas there was no change in G( - U). However, an insignificant value of G( - U) and the absence of any carbonyl compounds in N,O-saturated solution in the presence of ButOH are indicative of the fact that the t-butanol radical does not reduce uracil to U-. The observation is in agreement with the redox behaviourl* of 'CH,(CH,),COH compared with those of 'CH,OH and CH,CHOH radicals, since the former is known to be oxidizing and the latter two are reducing.lQ9 2o The observed data are therefore consistent with the major path for radiolytic degradation of uracil in the presence of alcohols involving the formation of U- by the reaction of uracil with eiq or with a hydroxyalkyl radical.U- is then protonated to UH [reaction (3)], whch can then undergo the following reactions: 2UH + UH+ + UH- (8) UH+ + UH- + dimer (9) This scheme is consistent with the observation that dihydrouracil is the major product (table 2). However, the increased yield of dihydrouracil(69 % ) in the presence of MeOH compared with that (40%) in the presence of ButOH cannot be explained by means of the above scheme. In addition to the above-mentioned radical reactions there might be some other pathway leading to dihydrouracil from UH in the presence of reducing hydroxyalkyl radicals.Further, it has been observed (table 1) that the G( - U) and G(>C=O) yields are relatively higher in the presence of EtOH compared with those in the presence of MeOH. These differences in yield can only be accounted for if it is assumed that hydroxyethyl radicals are much more reducing than hydroxymethyl radicals. However, this assumption is not improbable. This is supported by the fact that the tendency for electron transfer from the hydroxyethyl radical is much higher than that from the hydroxymethyl radical. 21 Note that in our product analysis we could not detect any alcohol adduct, as was reported earlier.l5? l6 However, from kinetic studies in an aerated solution of uracil there is some indirect evidence for the formation of such products.In the presence of oxygen no product is expected to be formed from eiq,,, and base degradation is achieved mainly through the reactions U+OH+UOH (12) UOH 3 products. (13) 70-22108 Reactions of Hydroxyalkyl Radicals with Uracil However, in the presence of alcohols which compete for OH [reaction (l)] degradation yields would follow the relationship where G(-U), is. the degradation yield in the absence of alcohol, G(-U) is the degradation yield in the presence of alcohol, [U] is the concentration of uracil, [ROH] is the concentration of alcohol and k,, and k, are the reaction rate constants for reactions (12) and (l), respectively. However, the hydroxyalkyl radicals might undergo reaction (4), as discussed, or undergo the reaction Thus, considering reactions (l), (4), RO + 0, -+ products.(15) (12) and (1 5) one finds23 In the limiting case, when the alcohol concentration is sufficiently high to scavenge the OH radicals completely, the remaining uracil degradation depends only on reaction (4). Under these conditions eqn (16) should take the form k15 Q , = I+-- k, [UI' Qco is 12.9 & 1.2 and 16f 1.7 for MeOH and EtOH, respectively, from which k15/k4 is estimated to be 10.9 & 2.4 and 1 1.1 _+ 1.3, respectively. Eqn (16) can be rearranged to give where Qco is independent of [ROH]. If Q is plotted against ([ROH]/[UI)( 1 - Q/Q,), a straight line is expected with slope k,/k,,. The results are plotted in fig. 4. From the plots the values k,/k,, have been determined to be 0.11 for MeOH and 0.4 for EtOH.It follows that the ratio of the rate constants for the reactions of OH with MeOH and EtOH is 0.11/0.4 = 0.28, which is compatible with earlier results. l3 t-Butyl alcohol, however, shows no such effect. Fig. 1 shows the effect of increasing concentration of ButOH on the G( - U) values and the plot of G( - U),/G( - U) against [ButOH]/[U] ratio is shown in fig. 5. The observed results have been shown to yield a straight line, from which kl/k12 = 0.08, close to the literature va1~e.l~ The observed straight-line course of the kinetic plot as shown in fig. 5 points to the conclusion that the radical RO formed in case of ButOH does not undergo any processes of the type envisaged in reaction (4). From the above discussion it is evident that even in the presence of oxygen, the hydroxyethyl or hydroxymethyl radicals react with uracil to give products via reaction (4).In the previous section it was argued that in a deaerated medium the radiolytic products are formed mainly from U- formed in reaction (7) and/or from reaction between eiq and uracil. However, in the presence of oxygen U- is known to revert back to uracil:,, u-+o,+u+o, (19) (20) UH + 0, -, U(0,)H -5+U(02)- + U + 0;.S. N. Bhattacharyya and P . C. Mandal 2109 Q Fig. 4. Plot of (l-Q/Q,)([ROH]/[UI) against Q for methyl (0) and ethyl (A) alcohol. [U] = 2 x lo-* mol dm-3, CQ. pH 5.6. I I 1 I I I I 0 5 10 15 20 25 30 [ BufOHl/ [Ul Fig. 5. Dependence of G( - U),/G( -U) on [ButOH]/[U] in neutral aerated solution. [U] = 2 x loA4 mol dm-3.21 10 Reactions of Hydroxyalkyl Radicals with Uracil Hence the product formed under the condition might be some adduct of alcohol radicals with uracil as was observed earlier by other workers.l59 l6 Failure to detect such products under our experimental conditions might be due to the fact that, owing to the small yield it could not be well separated from the uracil peak when developed in the paper chromatogram.We are grateful to Dr. 0. Yamamoto of Research Institute for Nuclear Medicine and Biology, Hiroshima University, Japan, for useful discussions. References 1 G. A. Infante, E. J. Fendler and J. H. Fendler, Radiat. Res. Rev., 1973, 4, 301. 2 C. V. Sonntag, Int. J. Radiat. Biol., 1984, 46, 507. 3 F. J. Nabben, J. P. Karman and H. Loman, Int. J. Radiat.Biol., 1982, 42, 23. 4 G. A. Infante, P. Jirathana, E. J. Fendler and J. H. Fendler, J. Chem. SOC., Faraday Trans. I , 1974,70, 5 P. C. Shragge, A. J. Varghese, J. W. Hunt and C. L. Greenstock, Radiat. Res., 1974,60, 250. 6 S. Das, D. J. Deeble, M. N. Schuchmann and C. V. Sonntag, Int. J. Radiat. Biol., 1984, 46, 7. 7 S. N. Bhattacharyya and P. C. Mandal, Int. J. Radiat. Biol., 1983, 43, 141. 8 S . N. Bhattacharyya and P. C. Mandal, J. Chem. SOC., Faraday Trans. I , 1983, 79, 2613. 9 S. N. Bhattacharyya and P. C. Mandal, J. Chern. SOC., Faraday Trans. I , 1983,80, 1205. 10 S . N. Bhattacharyya and P. C. Mandal, J . Chem. Soc., Furday Trans. I , 1985,81,2569. 11 S. N. Bhattacharyya and P. C. Mandal, Radiat. E$, 1985, 91, 9. 12 G. R. A. Johnson and G. Scholes, Analyst (London), 1954, 79, 217. 13 M. Anbar and P. Neta, Int. J. Appl. Radiat. Zsot., 1967, 18, 493. 14 M. N. Schuchmann and C. V. Sonntag, J. Chem. SOC., Perkin Trans. 2, 1983, 1525. 15 A. Kamal and W. M. Garrison, Nature (London), 1965, 206, 13 15. 16 P. E. Brown, M. Calvin and J. F. Newmark, Science, 1966, 151, 68. 17 C. L. Greenstock, Trans, Faraday SOC., 1970, 66, 2541. 18 P. Neta, J. Chem. Educ., 1981, 58, 110. 19 M. Kelm, J. Lilie, A. Henglein and E. Janata, J. Phys. Chem., 1974, 78, 882. 20 G. V. Buxton and J. C. Green, J. Chem. Soc., Faraday Trans. I , 1978,74, 697. 21 G. E. Adams and R. L. Willson, J. Chem. SOC., Faraday Trans. I , 1973, 69, 719. 22 M. Simic, in Fast Processes in Radiation Chemistry and Biology, ed. G . E. Adams, E. M. Ficlden and B. D. Michael (The Institute of Physics and John Wiley, London, 1975), p. 162. 23 H. Loman and J. Block, Radiut. Res., 1968, 36, 1. 1162. Paper 511438; Received 19th August, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202103

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J . Chern. Soc., Faraday Trans. I , 1986,82, 2111-2121 Kinetic and Equilibrium Studies associated with the Formation of Inclusion Compounds involving n-Butanol and n-Pentanol in Aqueous Cyclodextrin Solutions Dorothy Hall, Derek Bloor, Khalid Tawaraht and Evan Wyn-Jones" Department of Chemistry and Applied Chemistry, Uniuersity of Salford, Salford M.5 4 WT The equilibrium constants associated with the inclusion compounds of n-butanol and n-pentanol with a-cyclodextrin in aqueous solutions have been measured using the head-space analysis technique involving gas chromatography. These data confirm that a 1 : 1 inclusion complex is formed. The kinetics associated with the formation of these inclusion compounds have been studied using the ultrasonic relaxation method. From a detailed analysis of relaxation data it is clear that the kinetics of the process do not conform to the simple behaviour suggested by the equilibrium experiments. In addition there are also discrepancies between values of the volume change of the reaction determined independently from relaxation and density data. The significance of these results is discussed in terms of a two-step mechanism for the formation of the complex.Inclusion compounds in which the host can admit a guest component into its cavity without any covalent bonds being formed have been used extensively in fundamental studies and have also found a wide variety of applications.' The cyclodextrins are known to form several inclusion compounds with substrates ranging from hydrophobic to ionic in Most fundamental studies on these systems have involved equilibrium measurements on solutions in which the following equilibrium exists : guest + cyclodextrin % inclusion compound.(1) From these data there has been much discussion concerning the forces responsible for the stability of the inclusion complex and also on the nature of the interaction between host and guest? Kinetic studies on these systems have been restricted to few measure- ments using stopped-flow59 and temperature techniques and also ultrasonic relaxation.1° In principle, kinetic studies complemented by equilibrium measurements provide a powerful method to study the properties of solution in which fast equilibria such as (1) exist. Although processes with times ranging from milliseconds to microseconds have been measured in connection with aspects of the formation of several a-cyclodextrin inclusion complexes, several questions still remain unanswered concerning the mechanism of the formation of the inclusion compound.In this work we have adopted a combined equilibrium and kinetic approach to investigate the inclusion complexes of a-cyclodextrin with n-butanol and n-pentanol. The reason for choosing these alcohols is that they are known to form complexes with a-cyclodextrin and we have shown that alcohol monomer activities in solution can be measured conveniently using the technique of head-space analysis, involving gas chromatography.ll9 l2 In addition, the ultrasonic relaxation method can be used to monitor the dynamic exchange processes involving these alcohols and receptors.13 'f Present address: Department of Chemistry, Faculty of Science, University of Yarmouk, Irbid, Jordan.211121 12 Alcohol-Cyclodextrin Inclusion Compounds .72.8 I 0 0.2 0.1 0.6 0.8 1:O r Fig. 1. Scatchard plot of the a-cyclodextrin-butanol system. Equilibrium Measurements Quantitative information concerning the partitioning of the alcohol between the bulk phase and the cavity of the host molecule, a-cyclodextrin, has been obtained using the head-space analysis method involving gas chromatography.ll In this method solutions of alcohol and alcohol-cyclodextrin are placed in a conical flask with fitted septum and placed in a thermostatted bath. After appropriate equilibration times vapour samples are withdrawn from the head space using a gas-tight syringe and are analysed by gas chromatography for the amount of alcohol in the vapour from each solution.The underlying principle of this method is that the amount of alcohol per unit volume of head space from solution in equilibrium with vapour is the same for any two solutions in which the alcohol chemical potential is the same. If we assume, therefore, that a given activity corresponds to a given concentration in the continuous phase we may then, from the total concentration, estimate the amount complexed in the host molecule. The measurements were carried out on a Shimadzu model GC8A fitted with an integrator CR18 and a flame-ionization detector. The sensitivities were adjusted so that the samples in a given set could be run under the same conditions.The handling procedure was similar to that described by Spink and Colgan.12 If we assume that equilibrium (1) between the alcohol and a-cyclodextrin involves a I : 1 complex then, in principle, we can evaluate the equilibrium constants for each solution from the head-space data. In practice, several head-space measurements wereD. Hall, D. Bloor, K. Tawarah and E. Wyn-Jones 21 13 0 0.2 0.4 0.6 0.8 1 . i r Fig. 2. Scatchard plot of the a-cyclodextrin-pentanol system. taken for a range of alcohol and a-cyclodextrin concentrations and in these circumstances the Scatchard equation in the form y. := K-Kr [A1 can be used to analyse the data in graphical form. In this equation concentration of alcohol complexed to a-cyclodextrin total concentration of a-cyclodextrin r = [A] = monomer alcohol concentration in the bulk phase and K is the equilibrium constant.The plots of r / [ A ] vs. r are shown for the n-butanol and n-pentanol data in fig. 1 and 2, respectively. The linearity of these plots confirms the 1 : 1 stoichiometry of the complex and the following equilibrium constants were obtained: n-butanol-a-cyclodextrin K = 72.8(89.1) dm3 mol-I n-pentanol-a-cyclodextrin K = 302( 323) dm3 mol-I. These results compare well with an independent determination using the indirect spectrophotometric method following the inhibitory effect of the alcohol on the association of a-cyclodextrin with azo dyes.14 These values are in parentheses.21 14 Alcohol-Cyclodextrin Inclusion Compounds 120 - 100 - d I E 80 60 10 20 1 2 4 6 8 10 20 frequency/M Hz Fig.3. a-Cyclodextrin-pentanol system : a[f2 us. frequency. 0, 0.05 mol dm-3 a-CD + 0.034 mol d ~ n - ~ pentanol; 0, 0.05 mol dmF3 a-CD; a, 0.1 mol dmP3 pentanol. Kinetic Measurements Measurements of ultrasonic absorption were made in solutions thermostatted at 25 "C. The Eggers resonance technique was used for frequencies between 2 and 20 MHz and a standard pulse technique was used for frequencies up to 95 MHz.15 During the course of these experiments it was found that over a period of time the n-pentanol solutions containing a-cyclodextrin showed a certain degree of instability and became cloudy, sometimes even exhibiting a certain amount of precipitate. This is not an unusual phenomenon for a-cyclodextrin complexes and other slow degradation processes have indeed been observed.1° On the other hand, all the n-butanol solutions were extremely stable and showed no evidence of this type of behaviour.Measurements were made on several solutions of varying concentrations of alcohol and a-cyclodextrin. Aqueous solutions of a-cyclodextrin exhibit a very weak ultrasonic relaxation as is shown in fig. 3. The molecular origin of this process has recently been the subject of much discussion.lG~ l7 On the other hand, in pure solutions of butanol and pentanol up to their maximum solubility there is no ultrasonic relaxation and hardly any excess sound absorption. When alcohol and a-cyclodextrin are present together in an aqueous solution a well defined relaxation occurs, as is shown in fig. 3. This relaxation is clearly due to an interaction between them and is assigned to the exchange process between alcohol molecules in the bulk solution and those which are guests in the a-cyclodextrin cavity.For all the solutions used in this work the relaxation data were analysed using the single relaxation equation : where a is the sound absorption coefficient at frequencyf, A is the amplitude parameter, fc is the relaxation frequency and B is the frequency-independent excess absorption term. In all cases very good fits to the above equation were observed. Strictly speaking the analysis of the data should really take into account the weak relaxation observed in the pure a-cyclodextrin solutions. However, it has recently been shown that once a guest molecule enters the a-cyclodextrin cavity the weak ultrasonic relaxation specific to a-cyclodextrin itself is further reduced in amplitude.16 In these circumstances it is not possible to estimate this contribution, which is very small, and in all cases the data couldD.Hall, D . Bloor, K. Tawarah and E. Wyn-Jones 21 15 9 0 - I E N n VJ 55 El A 3 U 20 I I I I 1 I 1 2 L 6 8 10 20 frequencylMHz Fig. 4. a-Cyclodextrin-butanol system : a / f 2 us. frequency. 0, 0.05 mol dmP3 a-CD + 0.047 mol dm-3 butanol; A, 0.05 mol dm-3 a-CD+0.036 mol dm-3 butanol; V, 0.05 mol dm-3 a- CD + 0.025 rnol dm-3 butanol; @, 0.05 mol dm-3 a-CD + 0.020 mol dm-3 butanol. 120 100 * I 80 v) \ h N 3 S 60 40 20 I f I 1 1 1 2 L 6 8 10 20 frequencylMHz Fig, 5. a-Cyclodextrin-pentanol system: a / f 2 us. frequency. 0, 0.05 rnol dm-3 a-CD +0.034 mol dmA3 pentanol; [7, 0.05 mol dmP3 a-CD+0.025 rnol dm-3 pentanol; a, 0.05 mol dmP3 a-CD+0.014 mol dmA3 pentanol; (>, 0.05 mol dm-3 a-CD+0.010 mol dm-3 pentanol.be described well within the experimental error of 5% by the above single relaxation equation. It must be mentioned at this point, however, that for most of the data the amplitude parameter A was fairly small and usually much less than s2 m-l. As a result, the range of acceptable values of A , B andf, obtained from the fitting procedure, which will give calculated a / f 2 values to within experimental error of the measured values,21 16 Alcohol-Cyclodextrin Inclusion Compounds 0 4 8 [ butanollfree X [a-cy~lodextrin]~,,~/ 1 0-4 mo1’ dmS6 Fig. 6. ,u,,,/z us. [butanolIfree x [a-cy~lodextrin]~,,~.0 0.05 mol dm-3 a-cyclodextrin; a, 0.068 mol dmP3 a-cyclodextrin; v, 0.09 mol dmP3 a-cyclodextrin. increases as the magnitude of A decreases. When A x 2 x s2 m-l a tolerance of f 10% or better is expected for the relaxation parameters. On the other hand, it is very difficult to put a precise value on the accuracies of these parameters, especially when the amplitude is small and the fits to eqn (3) are good. An inspection of the data from the analysis reveals that the relaxation frequency shows very little apparent variation with change in the solution concentration. On the other hand, the relaxation amplitude, as is evident from fig. 4 and 5 , clearly shows a dependence on solution composition. In these circumstances the conventional method of investigating the kinetics of the alcohol inclusion complex involving the concentration dependence of the relaxation time is not necessarily the best way to proceed in analysing the kinetic data, especially when one takes into account the possible errors in this relaxation parameter.In the present work the relaxation amplitude is sensitive to concentration changes ; measurements have been taken at several concentrations and this parameter has been used to probe the kinetics of the processes by taking advantage of a phenomenological treatment which has recently been proposed18 and applied suc~essfully.~~~ 2o The equilibrium measurements have shown that the formation of inclusion com- pounds of a-cyclodextrin with both butanol and pentanol is a well defined process giving a 1 : 1 complex which can be defined by the equilibrium (4) Since a single relaxation has been observed, the phenomenological treatment can be used to evaluate the forward and backward rates of the above equilibrium.As we have shown previously,1s-20 the relaxation amplitude (11) divided by the relaxation time (7) is given by ROH + a-cyclodextrin $ ROH. - -a-cyclodextrin. !! = cr* ( 5 ) z where C is the thermodynamic term given by Z [ A V - ( A H O / C ~ ) ] ~ C = ~ R T K ,D. Hall, D . Bloor, K. Tawarah and E. Wyn-Jones 21 17 / * A 0 4 8 12 16 [pentan~l]~,,, X [a-cyclodextrinlfr,,/ 1 0-5 mol’ dm-6 Fig. 7. pma& us. [pentanol],,,, x [a-cy~lodextrin]~,,~. 0, 0.05 mol dm-3 a-cyclodextrin ; A, 0.08 mol dm-3 a-cyclodextrin. and Y* are the equilibrium forward and backward rates of the process defined by equilibrium (4).In eqn (6) AV is the volume and AH is the enthalpy change associated with equilibrium (4); IC, is the adiabatic compressibility of the solution at infinite frequency, 8 is the coefficient of thermal expansion, and cp is the specific heat per unit volume at constant pressure. For equilibrium (4) the forward and backward rates are given by: forward rate = r* = k , [ROH] [a-cyclodextrin] backward rate = r* = k-, [ROH.-.a-cyclodextrin] (7) (8) where the concentration terms in square brackets refer to equilibrium values which were evaluated from the initial weighed-in amount of alcohol and a-cyclodextrin and the equilibrium constant derived from the Scatchard plots. On this basis a plot of p / x against [ROH] [a-CD] should be a straight line passing through the origin with a slope of k,C.In the same way the plot of ,U/T against [ROH.--a-CD] will exhibit a similar behaviour, but this time with a slope of k-,C. These plots are shown in fig. 6-9 for n-butanol and n-pentanol. The scatter observed in these plots is considerable and reflects the difficulty mentioned earlier in the analysis of the relaxation data. As a result this necessitated the determination of a large number of experimental points. In addition the n-butanol data appear to be better than those of n-pentanol, a factor which may be associated with the slow processes involving the solutions becoming cloudy over a long period of time. The thermodynamic term, C, given by eqn (6) can be evaluated from p = cr (9) in which the term I‘ is given by21 18 Alcohol-Cyclodextrin Inclusion Compounds 0 2 L 6 [ROH~~*a-cyclodextrin]/ mol dm-3 Fig.8. a-Cyclodextrin-butanol system urna ax/^ us. [ROH-..a-cyclodextrin]. 0, 0.05 mol dm-3 a-CD; A, 0.068 mol dm-3 a-CD; V, 0.09 mol dm-3 a-CD. 6 - I v) 9 . h . x 2 -3 nD. Hall, D. Bloor, K , Tawarah and E. Wyn-Jones 21 19 3 - x 3 2 - 2 J 0 L 0 12 I?/ I 0-3 rnol dmm3 a, 0.068 mol dm-3 a-CD; V7 0.09 mol dm-3 a-CD. Fig. 10. a-Cyclodextrin-butanol system ,urn,, us. r, 0, 0.05 mol dm-3 3 5 :2 2 1 0 0 I 1 2 3 r/ 1O-j mol dm3 a-CD; Fig. 11. a-Cyclodextrin-pentanol system pmax us. r. 0, 0.05 mol dm-3 a-CD; A, 0.08 mol dm-3 a-CD. The magnitudes of the forward rate constants are similar to those quoted by Eyring and coworkerslO for the inclusion of some anions into p-cyclodextrin.It is generally regarded that the forward step in equilibrium (4) is a diffusion-controlled process and we can estimate the diffusion-controlled rate constant kD from : k , = 4 4 ~ , + D,) R, N x 10-3 (10)2120 Alcohol-Cyclodextrin Inclusion Compounds where N is the Avogadro number, D, and D, are the diffusion coefficients of free alcohol and free cyclodextrin, respectively, R, is the effective reaction distance and the factor arises because concentrations are expressed in mol dm-3. The diffusion constants of n-butano121 and n-pentanol,, are, respectively, 9.9 x and 9.0 x low6 cm2 s-l. We were unable to find a value for the diffusion constant of a-cyclodextrin; however, a value of 3.2 x cm2 s-l has been quoted for P-cy~lodextrin.~~g 24 If we assume that R, is the external radius of a-cyclodextrin we find that k , is ca.7 x lo9 mol-l dm3 s-l. The measured rate constants are a factor of nearly 20 less than the diffusion-controlled value. We can also carry out further checks on the measurement described in this work. From eqn (6), the slopes of the graphs fig. 10 and 11 and the measured sound velocities we can evaluate AV, the volume change in equilibrium (4). In this calculation the AH term can be neglected for aqueous solutions. The IA VI values from the present ultrasonic data are 1.5 x m3 mol-1 for n-butanol and n-pentanol, respectively. These values can be checked independently by carrying out accurate density measurements in the following manner. In principle, apparent molar volumes can be evaluated from density measurements ; partial molar volumes are then calculated by differentiation of the experimental plot of the apparent molar volume against c~ncentration.~~ The partial molar volumes of the inclusion complex described in equilibrium (4) can be evaluated from the equilibrium constants calculated from the Scatchard treatment.This treatment gives values of AVin the range (15-25) x m3 molt1, which are an order of magnitude higher than those found from the ultrasonic data. A significant feature of this work is the observation that the equilibrium data, which involve measurements of the monomer alcohol activity in various a-cyclodextrin solutions, show very simple behaviour and are consistent with the 1 : 1 complex described in eqn (4).Given the simple form of these equilibrium data, one might have expected the relaxation and also the density data to be equally simple. However, we have shown that this is not so. In an attempt to account for these discrepancies it is necessary to invoke more steps in the mechanism describing the formation of the inclusion compound than that described by equation (4). For the present purposes, we propose that the inclusion compound is formed as a result of the following two-step scheme: and 3.1 x ROH + a-cyclodextrin f (ROH . . . a-cyclodextrin), f (ROH-a-cyclodextrin),. In the above mechanism, the first step is the diffusion-controlled encounter of the alcohol with cyclodextrin, mentioned previously, and the second step is some rate-limiting process.This mechanism would be consistent with the results obtained in this work. A clue to the nature of the rate-determining step may be associated with the observation that the measured forward rate constants for both n-butanol and n-pentanol are very close. This suggests that the nature of the rate-determining step is perhaps not associated with the alcohol, but rather with the a-cyclodextrin structure. On this basis a conforma- tional change may be involved similar to the equilibrium between a collapsed distorted form and the ‘normal’ form of a-cyclodextrin, which has already been proposed in connection with experiments on a-cyclodextrin methanol pentahydrate complex. Indeed, this conformational change may also be the one associated with the weak ultrasonic relaxation that has been observed in pure a-cyclodextrin solutions and the subject of two recent independent studies.lB* l7 The main conclusions from these studies is that the ultrasonic relaxation is associated with a conformational change involving a re- arrangement of the water molecules inside the cavity of a-cyclodextrin.Finally, the mechanism proposed above must not be confused with that reported by Hersey and Robinson6 to account for spectroscopic and stopped-flow measurements. The present work involves the inclusion compound of a-cyclodextrin with alcohol, whereas Hersey and Robinson’s experiments were concerned with inclusion compounds and dyes with a-cyclodextrin. The time scale of the kinetic measurements carried out in this work is in the nanosecond range, whereas the stopped-flow experiments essentially measureD.Hall, D. Bloor, K. Tawarah and E. Wyn-Jones 2121 events in the millisecond time range. It may be that the mechanism of formation of the different inclusion compounds is different. If this is not so, then the two mechanisms can still be reconciled by assuming that the fast step in Hersey and Robinson’s mechanism is described by the two-step process proposed in this work. Clarification can only be obtained by carrying out kinetic experiments on these inclusion compounds covering the complete time range to s. We thank the S.E.R.C. for a research grant to construct the ultrasonic apparatus. D. H. thanks the S.E.R.C. for a research studentship. K.T. thanks Yarmouk University and the British Council for a short-term visit and also the E.E.C.for sponsoring the Sal ford-NEWI-Yarmouk Link. References 1 W. Saenger, Angew. Chem., Int. Ed. Engl., 1980, 19, 344. 2 D. W. Griffiths and M. L. Bender, Adv. Catal., 1973, 23, 209. 3 F. Cramer, W. Saenger and H-Ch. Spatz, J. Am. Chem. SOC., 1967,89, 14. 4 E. A. Lewis and L. H. Hansen, J. Chem. Sac., Perkin Trans. 2, 1973, 2081. 5 R. J. Clarke, J. H. Coates and S. F. Lincoln, Carbohydr. Res., 1984, 127, 181. 6 A. Hersey and B. H. Robinson, J. Chem. Soc., Faraday Trans I , 1984, 80, 2039. 7 R. L. Schiller, J. H. Coates and S. F. Lincoln, J. Chem. SOC., Faraday Trans. I , 1984, 80, 1257. 8 N. Yoshida and M. Fujimoto, Chem. Lett., 1980, 23 1. 9 N. Yoshida and M. Fujimoto, Chem. Lett., 1980, 1377. 10 R. P. Rohrbach, L. J. Rodrguez, E. M. Eyring and J. F. Wojcik, J. Phys. Chem., 1977, 81, 944. 1 1 K. Hayasea and S. Hayene, Bull. Chem. SOC. Jpn, 1977, 50, 83. 12 C. H. Spink and S. Colgan, J. Phys. Chem., 1983, 87, 888. 13 D. G. Hall, P. L. Jobling, J. E. Rassing and E. Wyn-Jones, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 14 Y. Matsu and K. Moshida, Bull. Chem. SOC. Jpn, 1979, 52, 2808. 15 See for example R. A. Pethrick in Techniques and Applications of Fast Reactions in Solution, ed. J. Gettins 16 S. Rauth and W. Knoche, J. Chem. Soc., Faraday Trans. I , 1985,81, 2551. 17 S. Kato, H. Nomura and Y. Miyahara, J. Phys. Chem., 1985,89, 5417. 18 D. G. Hall, J. Gormally and E. Wyn-Jones, J. Chem. SOC., Faraday Trans. 2, 1983, 79, 645. 19 J. Gormally, N. Natarajan, E. Wyn-Jones, D. Attwood, J. Gibson and D. G. Hall, J. Chem. SOC., 20 J. Gormally, B. Sztuba, E. Wyn-Jones and D. G. Hall, J. Chem. SOC., Faraday Trans. 2, 1985,81, 395. 21 P. A. Lyons and C. L. Sandquist, J. Am. Chem. SOC., 1953,75, 3897. 22 S. Glasstone, K. J. Laidler and H. Eyring, The Theory of Rate Processes (McGraw-Hill, New York, 23 L. G. Longsworth, J. Phys. Chem., 1954, 58, 772. 24 H. Uedaira and H. Uedaira, J. Phys. Chem., 1970,74,2211. 25 E. Vikingstad in Aggregation Processes in Solution, ed. E. Wyn-Jones and J. Gormally (Elsevier, 1582. and E. Wyn-Jones (D. Reidel, Dordrecht, 1979), p. 115. Faraday Trans. 2, 1984, 80, 243. 1941), p. 529. Amsterdam, 1983), p. 100. Paper 511453; Received 21st August, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202111

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1986,82,2123-2132 A Kinetic and Equilibrium Study of the Inclusion of Pyronine B by p- and y-Cyclodextrin Robert L. Schiller, Stephen F. Lincoln* and John H. Coates* Department of Physical and Inorganic Chemistry, University of Adelaide, South Australia 5001, Australia Temperature-jump visible spectrophotometric studies show the dye pyronine B (PB) in its monovalent cationic form to be included by y-cyclodextrin (yCD) in a fast step to form a 1 : 1 complex (PB- yCD), which is followed by the slower formation of the 2: 1 complex (PB),. yCD, in which pyronine B dimerises: ki PB + yCD PB - yCD (K,) (fast) (1) PB + PB * yCD (PB), - yCD (K,) (slow) (2) k-I kz k-9 where, in aqueous 1.00 mol dm-3 NaCl at pH 5.70 and 298.2 K, Kl = (4.3 _+O. 1) x 10, dm3 m o P , K, = (1.28 kO.04) x lo5 dm3 mol-l, k, = (8.2_+ 0.2) x lo8 dm3 mo1-1 s-l and k-, = (6.40f0.05) x lo3 s-l.In the presence of PCD only the formation of PB-PCD is detected and is characterised by k , = (1.1 k0.1) x lo8 dm3 mol-1 s-l and k-, = (1.5f0.5) x lo4 s-l. Equi- librium u.v.-visible spectrophotometric measurements are consistent with these kinetic data. No inclusion of PB by aCD was detected. These data are compared with those characterising related systems. The dimerisation of PB in the absence of yCD is fast, an observation which forms the basis for some of the mechanistic deductions concerning the formation of (PB), -yCD, and also for a reassessment of some mechanistic interpretation in earlier studies. The a-, p- and y-cyclodextrins (CD), which are six-, seven- and eight-membered a- 1,4-linked cyclic oligomers of D-glucopyranose with internal annular radii of 5-6, 7-8 and 9-10 A, respectively, form a wide range of inclusion complexes, some of which are in medical and industrial The gradation in size of the cyclodextrins presents an opportunity to study systematically the effect of spatial relationships on both the stability and lability of inclusion complexes.The majority of such studies of the cyclodextrin inclusion complexes have been thermodynamic in nature however, and few kinetic studies have been reported until recently.6-10 In this study the dye pyronine B (PB), whose structure is shown below, is used as a probe to determine the kinetic and related aspects of inclusion-complex formation by a-, B- and y-CD, and represents a systematic extension of our earlier studies in this Experimental Pyronine B was obtained from Sigma as the metal salt (PB),Fe,Cl, and was purified by extraction with ethyl acetate, rotary evaporation of the extract to dryness and recrystal- hation of the residue from ethyl acetate.Calculated for C4,H,,N40,Fe,Cl, (% ) : C, 21232124 Inclusion of Pyronine B by Cyclodextrins 48.40; H, 5.22; N, 5.38; C1, 27.21. Found ?A): C, 48.25; H, 5.29; N, 5.32; C1, 26.95. Analyses were performed by the Canadian Micro-analytical Service. The a-, p- and y-cyclodextrins (Sigma) were stored as the anhydrous material over P,O, in a vacuum desiccator prior to use. A.R.-grade NaCl (B.D.H.) was used as the supporting electrolyte in all solutions investigated, which were prepared by weight in doubly distilled water.PB adsorbs slightly to glass surfaces, but not significantly to silica surfaces. Accordingly, precautions similar to those described for the earlier crystal violet studiess were taken, to ensure that all surfaces had a similar history and that significant variations in the amount of PB adsorbed did not arise. All solutions were prepared immediately prior to spectrophotometric study, and exposure to light was kept to a minimum. Visible spectra were determined in silica cells using a Zeiss DMR 10 double-beam spectsophotometer equipped with a thermostatted ( f 0.1 K) cell block. All spectra were run in duplicate, recorded digitally onto paper tape at 1 nm intervals over the range 450-650 nm, and were analysed using a Cyber 173 computer.Temperature-jump spectrophotometric studies were performed at 555 nm for the BCD system, and at 533 and 553 nm for the yCD system, on equipment constructed in these laboratories to a design similar to that described in the 1iterature.ll In the studies of the PB-CD systems a temperature-jump cell of optical pathlength 1 .OO cm was used. The temperature jump was 8.8 K and the heating time was ca. 2 ps. To study the dye alone, a new cell of optical pathlength 0.23 cm was constructed (which permitted studies of higher PB concentrations than was possible with the 1 cm pathlength cell) for which the temperature jump was 9.7 K and the heating time was ca. 5 p s . In all cases the observation temperature was 298.2 f 0.1 K. Photomultiplier voltages from each transient were collected as 4096 8-bit data points using a Data Lab DL910 transient recorder and stored on magnetic tape.At least six transients were collected for each solution and the stored transients were averaged and subjected to kinetic analysis using a Computer Products Spectrum I1 minicomputer. Results Equilibrium and Spectroscopic Aspects The visible spectrum of pyronine B exhibits no variation in the pH range 2.00-7.00, consistent with its existence as a monovalent cation at pH 5.70 in the aqueous 1.00 mol dm-3 NaCl solutions used in this study. The visible spectra of PB alone, and in the presence of yCD in aqueous 1 .OO mol dm-3 NaCl, are shown in fig. 1, from which it is seen that there is a shift of the absorbance maximum to shorter wavelengths and a pronounced change in the shape of the spectrum as [yCD] increases.(Thirty solutions were studied in which [yCD] was varied in the range 0-1 .OO x mol dmP3 and [PB] was constant at 9.7 x mol dm-3.) These spectral changes are consistent with the dimerisation of PB being enhanced on formation of the yCD inclusion complexes. The near approach to an isosbestic point indicates that, whilst two species dominate the observed spectral variation, a third species is also present consistent with equilibria (1) and (2): kl k-1 k2 k-2 PB + yCD =$ PB - yCD (Kl) (fast) (1) PB + PB - yCD (PB), * yCD (K,) (slow) (2) in which PB - yCD is a 1 : 1 inclusion complex, and (PB), * yCD is a 2 : 1 complex in which PB is included as a dimer. However, it proved impossible to obtain a unique fit of the spectral data to these two equilibria, if the absorption spectra of PB in both inclusion complexes were allowed to be different, and to vary during the fitting procedure usingR.L . Schiller, S. F. Lincoln and J . H. Coates 2125 0 450 550 h/nm 650 Fig. 1. Variation of the pyronine B spectrum in the presence of yCD at pH 5.70 in aqueous 1 .OO mol dm-3 NaCl at 298.2 K. The molar absorbance at 550 nm decreases systematically as the total yCD concentration increases sequentially in the range 0, (2.03 and 5.10) x (1.01, 2.04 and 5.12) x lop4 and (1.028, 2.042 and 4.093) x mol dmp3 with the total pyronine B concentration constant at 9.7 x mol dm-3. These nine spectra exemplify the spectral variation observed for all thirty solutions studied.the program DATAFIT. Accordingly, the approximation was made that the spectrum of PB-yCD was identical to the known spectrum of PB alone and Kl and K , values of (2.64 0.72) x 10, and (2.5 1.4) x lo5 dm3 mol-l, respectively, were then obtained from a fit of the absorbance data to eqn (3) in the ranges 520-537 and 540-567 nm at 1 nm intervals : (3) In this fitting procedure the spectrum of PB monomer was taken as the spectrum of PB in the absence of yCD shown in fig. 1 . (The molar absorbances of this spectrum are substantially greater than those published elsewhere,12 but in view of the high purity of our PB (see Experimental) and the acknowledged low purity of PB used in the earlier study, the molar absorbances shown in fig. 1 are probably the more reliable.) Pyronine B dimerizes in aqueous so1ution,12 but at the low concentrations used in the determination of the PB spectrum shown in fig.1, and used in the PB-CD studies, no free (PB), dimer was detected. Thus, it is concluded that dimerization of PB to free (PB), occurs to a small extent only at these concentrations. (Dimerisation of PB at higher concentrations in the absence of cyclodextrin is discussed below.) Within the approximations employed in the fitting of the spectral variations shown in fig. 1 to equilibria (1) and (2), the spectrum of (PB), * yCD was derived from these data and is shown in fig. 2. The shape of the spectral envelope is similar to that of free (PB),.', An analysis of this spectrum according to exciton theory13 using program EXCITON, indicates that the transition moments of the PB monomers (along the long axis of the xanthene group) are aligned at an angle of 20-25" to each other in (PB);yCD, in contrast to (PB),, in which the analogous angle is zero.This difference may arise from constraints imposed on the dimer by inclusion in the yCD cavity but could also be to some extent an artefact arising from the approximations used in deriving the (PB), - yCD spectrum. Nevertheless, the spectral variations shown in fig. 1 can only be explained in terms of dimer inclusion through equilibria (1) and (2). The variation of the spectrum of PB in the presence of PCD is substantially less than A = &p,[PB] + &pB. yCD[PB ' yCD1 + 2E(PB)s.yCD[(PB)2. yCD1*2126 Inclusion of Pyronine B by Cyclodextrins - ' 12 5 I t I I -450 550 650 A/nm Fig.2. The spectra of PB (-), PB .PCD (---) and (PB)2 - yCD ( .-. - - ) derived as described in the text. 0 450 550 X/nm 650 Fig. 3. Variation of the pyronine B spectrum in the presence of PCD at pH 5.70 in aqueous 1.00 mol dm-3 NaCl at 298.2 K. The molar absorbance at 550 nm decreases systematically as the total PCD concentration increases sequentially in the range 0, 4.00 x lo-*, and 4.02 x mol dmP3. These three spectra exemplify the spectral variation observed for all sixteen solutions studied. mol dm-3, with the total pyronine B concentration constant at 1.02 x in the presence of yCD, as may be seen from fig. 3, and is consistent with the formation of the 1 : 1 complex PB-PCD, as the greatly predominant complex according to equilibrium (4).(Sixteen solutions were studied in which WCD] was varied in the range 0-1.00 x mol dmP3 and [PB] was constant at 1.02 x mol dm-3): Fitting of the spectral data to this equilibrium in the ranges 500-556 and 559-576 nm at 1 nm intervals using the program DATAFIT yields Kl = (4.0 & 2.4) x lo3 dm3 mol-1 and the spectrum of PB-BCD in fig. 2. It is apparent from fig. 2 that the spectral change induced in PB on inclusion by BCD is quite small. It is reasonable to assume that st similarly small change will be induced on inclusion of either PB or (PB), by yCD as a result of the changes in their environment. However, the major cause of the spectralR. L. Schiller, S . F. Lincoln and J . H. Coates 2127 I I I 1 I I I I 1 [ PBItotaI / 1 O-* mol dm-3 Fig.4. The variation of AI/Io with total [PB]. The solid curve represents the best fit of the data to eqn. (6). Experimental data points refer to observations at 553 nm. change induced by yCD is the considerable increase in the total concentration of the (PB), dimer in solution resulting from the formation of (PB), - yCD. Similar observations have been made for the spectral changes accompanying the inclusion of methyl orange,' crystal violetg and tropaeolins by yCD. The spectrum of PB was found to be insignificantly changed in the presence of K D , indicating a low stability for PB - K D . The dimerisation of PB in 1.00 mol dm-3 aqueous sodium chloride was investigated by temperature-jump spectrophotometry. At PB concentrations similar to those used in the cyclodextrin studies (ca.mol dm-3) no relaxation was observed, but at higher PB concentrations relaxations producing a decrease and an increase in absorbance at 533 and 553 nm, respectively, were observed. The larger-amplitude change occurred at 553 nm. Over the PB concentration range studied (2 x mol dmP3) the relaxation occurred within the instrumental heating time (ca. 5 ,us) and therefore no quantitative kinetic data could be derived. However, the variation of the relaxation amplitude ( A I ) with the total PB concentration was consistent with that expected for the Pb dimerisation equilibrium (5): to 1 x which may be shown to be of the form of eqn (5) in which I, is the detected light intensity prior to the temperature jump and C is a constant characterizing the dimerisation eq~i1ibrium.l~ The variation of AI/I, with the total PB concentration is shown in fig.4, as is the best fit of these data to eqn (6) obtained through a non-linear least-squares analysis using the program DATAFIT, which yielded Kd = (1.3 k0.5) x lo3 dm3 mol-l. Ohling15 has obtained k, = 2.3 x log dm3 mo1-l s-l and k-, = 1.7 x lo5 s-l at 298.2 K for the dimerisation of pyronine Y and it is therefore expected that the dimerisation of its homologue, PB, will be similarly fast. In earlier studies of dye-cyclodextrin systems involving crystal violets and tr~paeolin,~ substantially slower relaxations than those observed for PB were thought to arise from dye dimerisation. As a consequence of the more recent observations on PB, and with the advantage of the new and shorter-pathlength temperature-jump cell, the dimerisations of crystal violet and tropaeolin have been reinvestigated.At dye concentrations > 5 x mol dm-3 a relaxation occurring within the instrumental heating time and2128 Inclusion of Pyronine B by Cyclodextrins with an amplitude concentration dependence of the form of eqn (6) was observed for both dyes. This suggests that the dimerisation of crystal violet and tropaeolin occurs much more rapidly than was previously thought, but this conclusion does not change the kinetic parameters derived for the interactions of these dyes with cyclodextrins. Nor does it change the mechanistic deductions derived therefrom, with the single exception that the deduction that the increased stability of the included dimer arose from the enhancement of the forward rate constant for the dimerisation in the presence of the cyclodextrin no longer appears tenable.Kinetic Aspects Temperature-jump spectrophotometric studies at 533 and 553 nm of PB and yCD in 1.00 mol dm-3 aqueous sodium chloride at pH 5.70 detected a single relaxation at both wavelengths. At 533 nm the relaxation was characterised by a decrease in absorbance, whereas at 553 nm the relaxation produced an increase in absorbance consistent with the absorbance changes arising predominantly from the shift of equilibria (1) and (2) to the left for the yCD system. At both wavelengths the relaxation was characterised by relaxation times (z) identical within experimental error, and the l / z values are plotted in fig.5. Similar studies of PB in the presence of QCD at 555 nm also detected a single relaxation which was characterized by a decrease in absorbance. The amplitude of this relaxation was much less than that characterising the yCD system as anticipated from the spectra in fig. 1-3. The l/z data derived for the QCD system are plotted in fig. 6. The dependence of l / z on the total cyclodextrin concentration is quite different for the yCD and PCD systems. The variation of l/z for the yCD system is consistent with the observed relaxation arising from the coupled fast formation of PB - yCD in equilibrium (1) and the slower formation of (PB), - yCD in equilibrium (2) and the slower formation of (PB),-yCD in equilibrium (2), and may be expressed through eqn (7): l / z = k,[PB]([PB * yCD] + [PB] + 4[yCD])/([PB] + [yCD] + 1 / K l ) + k-, (7) which may be derived using the method of either Bernasconi16 or Czerlinski,17 and in which all concentrations are the equilibrium values existing prior to the relaxation.A non-linear least-squares fit of the l/z data to eqn (6) using program DATAFIT produced the best-fit curve shown in fig. 5 and the K,, K,, k , and k-, values shown in table 1. The agreement between the Kl and K2 values derived from the temperature-jump and equilibrium spectrophotometric studies is satisfactory when the highly derived nature of these constants is considered and indicates a reasonable internal consistency. The 1 /z data were also fitted to a relaxation scheme in which a third fast relaxation (8) (PB), - yCD + yCD 2 (PB), * (yCD), (K3) k-3 (analogous to that observed for the inclusion of both methyl orange and tropaeolin, but not crystal violet, by YCD~-~) was added to those arising from equilibria (1) and (2), but this did not improve the data fit.The error in K3 was greater than the magnitude of that constant, and accordingly this third relaxation process was not further considered. The variation of 1/z for the QCD system could only be studied at total QCD concentrations up to mol dm-3, above which the relaxation time became comparable to the instrumental heating time. Nevertheless, the variation of l/z seen in fig. 6 (when considered in conjunction with the equilibrium spectrophotometric data discussed above) is consistent with relaxation arising through equilibrium (4) alone, such that the variation of 1 /z may be expressed through eqn (9) : (9) l/z = k,([PB] + WCD]) + k-l.R.L. Schiller, S. F. Lincoln and J . H. Coates 2129 14 12 -.- c 10 m 2 8 6 4 0 2 4 6 8 10 [rCD]/ 1 0-3 mol dm-3 Fig. 5. The variation of l / z (298.2 K) for the PB-yCD system with variation of the total [yCD]. The total [PB] varied in the range (9.1-9.8) x mol dm-3. The solid curve represents the best fit of the data to eqn (7). 14 2 0 2 4 6 8 1 0 [pCD]/ mol dm-3 Fig. 6. The variation of l / z (298.2 K) for the PB-PCD system with variation of the total WCD]. The total [PB] varied in the range (1.0s1.02) x mol dmP3. The solid curve represents the best fit of the data to eqn (9). The linear least-squares best-fit curve of the data to eqn (9) is shown in fig, 6 and the derived K,, k, and k-, values appear in table 1, from which it is seen that the Kl values derived from the temperature-jump and equilibrium spectrophotometric data are the same within experimental error.No significant relaxations attributable to inclusion processes were observed in temperature-jump spectrophotometric studies of PB in the presence of aCD.2130 Inclusion of Pyronine B by Cyclodextrins Table 1. Equilibrium and kinetic parameters (298.2 K) Kl/102 ~ ~ 1 0 5 K ~ / 103 k,/ 109 system dm3 mol-l dm3 mol-1 dm3 mol-l dm3 mol-' s-l k-,/103 s-l pyronine B-yCDa 4.3f0.1 tropaeolin-yCDC 4.18f 1.47 methyl orange-yCDd 0.45 f 0.07 crystal violet-yCDe 4.63 f 0.06 pyronine B-PCD" 73 f 31 (40 f 24)b tropaeolin-pCDC 7.1 f0.7 (2.64f0.72)b 1.28 f 0.04 - 0.82 0.02 6.40 f 0.05 (2.5 f 1 .4)b 16.8 f0.54 0.177 f 0.154 2.27 k0.61 1.35 & 0.23 20+ 1 1 6.1 k2.5 9.4 f 5.1 4.8 f 0.8 10.3 f 0.9 - 1.73 k 0.08 1.68 f 0.07 - - (k, =) (k-1 =) 0.11 kO.01 15k5 40 f 70 - 5 f 6 1.3 f 1.5 a This study.This study: equilibrium spectrophotometric values. Ref. (9). Ref. (7). Ref. (8). Discussion The differing interactions of a-, p- and y-CD with PB are clearly a consequence of their differing annular radii of 5-6, 7-8 and 9-10 A, respectively. It is therefore appropriate to compare these dimensions with those of PB determined approximately (in A) from CPK (Corey, Pauling and Koltun) models, which indicate the length of PB to be 17, the distance from tip to tip of the ethyl groups in the -NEt, substituent of PB to be 11.5, the smallest and largest distance perpendicular to the long axis of the superimposed xanthene groups in (PB), to be 7 and 8, respectively, and the distance between substituent nitrogens at opposite ends of (PB), to be 8.5, which compares with the depth of the CD annulus of 8.Clearly the -NEt, substituents prevent significant inclusion of PB by aCD, as found in this study, whereas PB fits closely into the larger annulus of DCD to form PB -#KD characterised by a stability constant Kl = 4.0 x lo3 dm3 mol-l determined from kinetic data (table 1). A PB-yCD complex is also formed but is of decreased stability (Kl = 4.3 x 10, dm3 mol-l), probably as a consequence of the relatively loose fit of PB into the larger yCD annulus. The larger annulus yCD does, however, accommodate the dimer (PB), in (BB),.yCD characterised by K , = 1.28 x lo5 dm3 mol-l, whereas the analogous (PB),.PCD species is not formed to a detectable extent. The formation of PB'yCD and (PB),.yCD is shown below, where the truncated cone represents the yCD annulus.PB CD The inclusion process is initiated by a diffusion-controlled encounter between PB and yCD, which then form PB * yCD at a slower rate as PB enters the yCD annulus (probably through its wider end, which is delineated by sixteen secondary hydroxy groups) in the equilibrium characterised by k, and k1. Subsequently a second PB makes a diffusion-R. L. Schiller, S. F. Lincoln and J . H. Coates 2131 controlled contact with PB-yCD and is then included to form (PB);yCD in a slower step, as shown in the equilibrium characterised by k, and k-,.These inclusion processes probably involve a sequence of sterically orienting interactions and solvational changes not specifically detected in this study, but nevertheless the above scheme provides a convenient basis for discussion of the equilibrium and kinetic data of table 1. Pyronine B is the fourth dye for which inclusion by yCD produces a 1l.r dependence of the type seen in fig. 5, and in which the stability of the included dye dimer in (dye),.yCD is substantially increased over that in the free state. No rate parameters are available for the more rapidly formed precursor dye.yCD species, but it is reasonable to assume that its more facilie formation is a consequence of the loose fit of dye monomer in the yCD annulus, whereas CPK models indicate that the entrance of the second dye monomer into the annulus to form (dye), ' yCD is relatively hindered.Nevertheless, the magnitude of k , is large for all four dyes. Of the dye-yCD systems listed in table 1, pyronine B-yCD exhibits the smallest Kz as a consequence of having the smallest k , and the largest k+,. However, k , only varies by a factor of 1 1.5 and k-, varies by a smaller factor of 4.7 with variation of the nature of the dye. This is despite the substantial structural differences between the dyes, and the existence of pyronine B and crystal violet as monovalent cations and tropaeolin and methyl orange as monovalent anions under the conditions of the inclusion studies. These observations are in marked contrast to those made on the inclusion of dyes by aCD to form the 1 : 1 species dye - aCD, where it was found that the formation and dissociation rates were very sensitive to the nature of the dye and varied over several orders of magnitude (whilst the equilibrium constants exhibited a lesser variation consistent with variation in the nature of the dye affecting the formation and dissociation rates to a similar extent).6* lo This suggests that either there is a fortuitious combination of interactions controlling k , and k-, characterising (dye), - yCD to produce the modest variation in these rate constants, or that some interactions are particularly dominant in the systems appearing in table 1.Amongst such interactions dispersion forces between the high n-electron densities of the aromatic substituents of the dyes in (dye), - yCD should be an important factor determining the lability of this species.The approximately 100-fold increase in stability of the included dye dimer over that of the dimer in the free state suggests that a combination of dispersion-force interactions between the interior of the yCD annulus and (PB), and solvational changes and motional restrictions imposed by yCD is important in changing the energetics of dimer formation and therefore in determining the magnitude of k, and k-,. Whilst the dimerisation of PB [reaction (5)] has been shown to be fast, a direct determination of k, and k-, is not available. However, it is reasonable to assume that the values of these rate constants will be similar to k, = 2.3 x lo9 dm3 mol-1 s-l and k-, = 1.7 x lo5 s-l observed for its homologue pyronine Y.15 A comparison of these values with k, and k-, (table 1) characterising the formation of (PB), - yCD in eqn (2) suggest that the increased stability of (PB), included in (PB),-yCD is a consequence of a decreased dimer dissociation rate arising from the dispersion force interactions, solvational changes and motional restrictions experienced by (PB), in the yCD annulus.We thank the Australian Research Grants Scheme for partial support of this research, and Dr Tom Kurucsev for the use of his programs EXCITON and DATAFIT and for advice on some spectroscopic aspects of this study. References 1 M. L. Bender and M. Komiyama, Cyclodextrin Chemistry (Springer, Berlin, 1978). 2 W. Saenger, Angew. Chem., Int. Ed. Engl., 1980, 19, 344. 3 I. Tabushi, Acc. Chem. Res., 1982, 15, 66. 4 J. Szejtli, Cyclodextrins and their Inclusion Complexes (Akadkmiai Kiado, Budapest, 1982). 5 R. Breslow, Chem. Br., 1983, 126.2132 Inclusion of Pyronine B by Cyclodextrins 6 F. Cramer, W. Saenger and H-Ch. Spatz, J. Am. Chem. SOC., 1967,89, 14. 7 R. J. Clarke, J. H. Coates and S. F. Lincoln, Carbohydrate Res., 1984, 127, 181. 8 R. L. Schiller, J. H. Coates and S. F. Lincoln, J. Chem. SOC., Faraday Trans. I , 1984, 80, 1257. 9 R. J. Clarke, J. H. Coates and S. F. Lincoln, J. Chem. SOC., Faraday Trans. I , 1984, 80, 31 19. 10 A. Hersey and B. H. Robinson, J. Chem. SOC., Faraday Trans. I , 1984, 80, 2039. 1 1 G. G. Hammes, in Techniques of Chemistry, ed. G. G. Hammes (Wiley, New York, 1974), vol. vi, part 12 L. P. Gianneschi, A. Cant and T. Kurucsev, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 664. 13 L. P. Gianneschi and T. Kurucsev, J. Chem. SOC., Faraday Trans. 2, 1976, 72, 2095. 14 D. Thusius, J. Am. Chem. SOC., 1972,94, 356. 15 W. Ohling, Ber. Bunsenges. Phys. Chem., 1984, 88, 109. 16 C. F. Bernasconi, Relaxation Kinetics (Academic Press, New York, 1976). 17 G. H. Czerlinski, Chemical Relaxation (Marcel Dekker, New York, 1966). 2. Paper 511461; Received 23rd August, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202123

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. 1, 1986,82, 2133-2140 The Reaction of Ferrimyoglobint with Me thylhydroperoxide Mordechai L. Kremer Department of Physical Chemistry, The Hebrew University of Jerusalem, Jerusalem 91 904, Israel Methylhydroperoxide reacts with ferrimyoglobin in an irreversible reaction. In the course of the reaction two intermediates are formed in two consecutive steps. The experimental data are consistent with a mechanism in which MeOOH is dehydrated catalytically by ferrimyoglobin to yield CH,O. MeOH is not formed in the system, even as a transient. The relative rate constants of the reaction have been evaluated: k, = 1 dm3 mol-1 s-l. k, = 2.23 x s-l and k, = 4.0 x lop6 s-l. The molar absorptivities of the intermediates have been determined at II = 549 nm: = 1.75 x lo4 dm3 mol-l cm-l and e,, = 6.32 x lo3 dm3 mol-1 cm-l.In the systematic study of the reactions of haemoproteins with H202 and its derivatives, the reactions of ferrimyoglobin, Mb(rrr), have drawn considerable interest. Mb(r1r) reacts with these compounds by forming bright-red-coloured intermediates which are stable for a certain period of time but revert ultimately to the parent ferrimyoglobin m01ecule.l-~ The rate of decay of the intermediates depends on the nature of the substrate used. The intermediates formed with alkylhydroperoxides disappear quite quickly, but when H202 is used as a substrate the intermediate is relatively stable.lP2 The mechanism of the reaction, the structure of the intermediates formed and their possible relation to the intermediates formed in the analogous reactions of catalase and peroxidase with the same substrates were the subject of extensive studies.These studies established the essentially irreversible nature of the formation of the intermediates, but the exact course of the reaction remained unknown. Experimental data could be interpreted in a qualitative fashion only. No specific steps were identified and no individual rate constants were determined.lY By using the method of computer simulation, the experimental data obtained in the Mb(rrr)-H,O, system have been examined recently.6 The data were in quantitative agreement with a scheme in which the appearance of the red compound was preceded by the formation of a precursor complex. Rate constants of the individual reaction steps were evaluated and the molar absorptivities of the intermediates at one wavelength were determined .It was the purpose of the present work to extend the scope of the investigations to include the reaction of Mb(rr1) with alkylhydroperoxides. It was expected to find characteristic changes in the mechanism, similar to those found when EtOOH was exchanged as a substrate for H202 in its reaction with catala~e.'-~ Ex per h e n t a1 When solutions of Mb(r1r) and MeOOH or EtOOH are mixed, the mixture attains a temporary bright-red colour. The transient red colour disappears gradually and the spectrum of Mb(m) is regenerated. The maximal change of the optical absorption at a t This compound is commonly indexed under metmyoglobin. 21332134 React ion of Fer r im yoglob in with Me th y lh y droperoxide given wavelength, AAmax, depends on the initial concentration of the respective hydroperoxide (x,): at a constant concentration of Mb(m), to be denoted in the following as e, the value of AA,,, increases as x , is increased.AA,,, itself reaches a maximum, AAMAX, in a large excess of the hydr0peroxide.l George and Irvine found that between pH 8 and 9, a three- to four-fold molar excess of peroxide was necessary to reach this limit. At a higher pH, a larger excess was necessary. The ratio f = AAmax/AAMAx for the reaction between Mb(II1) and MeOOH (interpreted by George and Irvine as the fraction of conversion of Mb(m) to a single intermediate) has been determined as a function of x,/e at different values of e and of the pH.The results are summarized in fig. 4 of ref. (1). Data relevant to the present analysis are reproduced below as points in fig. 1. The experimental results show that there is a marked deviation from 1 : 1 stoichiometry in the formation of the assumed intermediate at a constant e and at all x, studied. This deviation increases as the ratio x,/e is increased. The most remarkable feature of the data is the very slight dependence off on the absolute values of e and x,, i.e. the very slight change off when e is varied ten-fold while x,/e is kept constant. It is this finding which excluded the possibility of formation of the intermediate in an equilibrium reaction from Mb(m) and MeOOH. The deviation from 1 : 1 stoichiometry has been explained by George and Irvine by assuming a chain decomposition of MeOOH.This decomposition was thought to have been initiated by a radical species Y. Y was assumed to have been formed simultaneously with the ‘red complex’. [This complex was denoted as Fekv by George and Irvine. The symbol Mb(1v) is used here to denote the same species.] Mb(m)+MeOOH + Mb(Iv)+Y (1) the destruction of MeOOH. (2) Y + MeOOH + chain reaction leading to Reaction (2) has not been elaborated in greater detail by George and Irvine. It explained, in a qualitative manner, why the entire amount of MeOOH present initially in the reaction mixture was not available for the formation of Mb(1v): by increasing the excess of MeOOH over Mb(Irr), an increasing amount of MeOOH disappeared through induced decomposition. It thus explained ‘the bend toward the abscissa’ of the curve showing ‘the percentage formation of the intermediate compound’ as a function of x,/e.The explanation of George and Irvine nevertheless remained essentially qualitative. In order to find a quantitative explanation of the experimental facts observed by them, a detailed kinetic analysis was carried out as described in the following section. It will be shown that the application of computer analysis will provide important details regarding individual reactions steps occurring in the system and thereby will make a better understanding of the underlying mechanism possible. Derivation of the Mechanism Several patterns of reaction were considered. These included the following schemes. (A) The formation and subsequent decomposition of a single intermediate C:* Mb(II1) + MeOOH -+ C (1) C-+ Mb(1Ir) + CH,O + H,O.(2) (B) The same as (A), but with a release of MeOH simultaneously with the formation of c: Mb(w) + MeOOH -+ C + CH,OH (1) (2) C + CH,OH + Mb(II1) + CH,O + H,O. * The letter C denotes different species in different schemes.M. L. Kremer 2135 0.2 0.6 I .o 1.4 1.8 x0le Fig. 1. AAmax/A&x as a function of x,/e [experimental data of George and Irvine, ref. (l), fig. 41. pH 8.6; e = (a) 5.2 x mol drn-,. Parameters of the calculated curves: and (b) 5.2 x k, = 1 dm3 mol-l s-l; k, = 2.23 x lo-* s-l; k, = 4.02 x s-l; cc = 6.10 x (C) A modification of (A) involving the subsequent oxidation of CH20 to HCOOH by the intermediate C: Mb(rri) + MeOOH --+ C (1) C * Mb(Ir1) + CH20 + H,O (2) C + CH,O -+ Mb(m) + HCOOH + MeOH.(3) (D) The modification of (A), involving two intermediates formed in two consecutive (1) CI --+GI (2) CII --+ Mb(rrr) + CH,O + H,O. (3) steps : Mb(Irr) + MeOOH -+ C, e - p - q x , - p - q - s P Q S [Lower case letters in scheme (D) denote concentrations.] All the above schemes are irreversible, in accordance with George and Irvine’s conclusion regarding the nature of the reactions occurring in the system. The rate equations relevant to each scheme were derived and integrated numerically. The quantity fcalc = AAmax/AAZax was evaluated for different initial conditions on the basis of each of the schemes (A)-(D). AA,,, and AAZ,, are a pair of maximal absorption changes calculated for identical values of e, but for differing x,.The first quantity refers to any arbitrarily chosen value of x,, the second refers to the fixed value of x, = 4e. This2136 React ion of Fer r im y oglo b in with Met h y lh y dr oper oxide Table 1. Results of the calculations mechanism r.s.s. A 1.41 x lop2 B 1.73 x 10-l C 1.34 x 10-3 D 1.65 x 10-4 definition follows closely the experimental determination of this quantity by George and 1rvine.l In schemes (A), (B) and (C),fcalc was put equal to [C],,,/[C]&,x (the asterisk denoting here and also in the following the special initial condition x, = 4e). In the case of scheme (D), where two species, CI and CII, contribute to the change of the optical absorption, fcalc was equated with the expression (p + aq),,,/(p + q)&axa Here a = (cII-cE)/(cI-cE). eI, cII and cE denote the molar absorptivities of C,, CII and Mb(m), respectively.By varying the rate and absorption parameters, relevant to each scheme, agreement between the calculated quantities off,,lc and the observed ones, fob,, was sought. Details of the method of calculation have been described in a previous communi~ation.~ Note that, since the time did not appear explicitly among the variables considered, the values of the calculated rate constants could be determined only up to a common constant multiplication factor. Thus in each scheme considered the rate constants have been determined relative to that of the respective first step, i.e. always relative to k, = 1 dm3 mol-l. Results and Discussion The results of the calculations are summarized in table 1.The second column of the table shows the residual sum of squares of the deviations (r.s.s.) of fcalc from fobs in the case of each mechanism. Scheme (D) offers the most accurate description of the reactions occurring in the system. The optimal values of the parameters in scheme (D) are: k, = 1 dm3 mol-l s-l k, = (2.23 + O . O S ) x k, = (4.0f 1.0) x s-' s-' a = (6.1 k0.4) x lo-,. The calculated values o f f as a function of x,/e at e = 5.2 x mol dm-3 and e = 5.2 x mol dm-3, respectively, are shown as the solid curves (a) and (b) in fig. 1. Fig. 2 shows the dependence of p [curve (a)] and of p + a q [curve (b)] as a function of time at e = 5.2 x mol dm-3 and x, = 6.5 x mol dm-3. (The time is given in arbitrary units because of the undefined scaling of the rate constants.) The quantityp + ccq is proportional to the optical absorption change AA = ( E I - E E ) (p+o~q).Fig. 2 shows that the dominant term in the expression p + aq, near its maximum, is p , i.e. in this range the change of AA reflects mainly the change in the concentration of C,. The effect of CII on the AA us. time curve is essentially secondary. It mainly causes a shift of the maximum of AA to longer times and reduces its rate of decay in the later stages of the reaction. The compensation of the decay of C, by the (temporary) accumulation of CII, in respect of its effect on AA, causes thus 'the (dominant) intermediate' to appear to be more stable than it is in reality. The calculations showed that a four-fold molar excess of MeOOH is not sufficient for a quantitative saturation of Mb(rI1).For x, = 2.8 x mol dm-3 and e = 5.0 x mol dm-3 the sum of concentrations of C, and CII at the maximum ofM. L. Kremer 2137 ‘41 0 2 4 6 8 1 0 1 2 1 4 time (arb. units) Fig. 2. Dependence on time (in arbitrary scale units) of (a) p , (b) p + q and ( c ) x. e = 5.2 x mol drnp3. Parameters as in fig. 1. The concentration scale on the left refers to curves (a) and (b), the concentration scale on the right refers to curve (c). mol dm-3; x, = 6.5 x p + aq has been calculated as 3.14 x rnol dmd3. Thus, under these circumstances, only ca. 60% of Mb(n1) was present in the compounds C, and CII. An inspection of the experimental data of George and Irvine [see the two lowest curves in fig. 4 of ref. (l)] also suggests that at xo/e = 3-4 a 100% formation of ‘ the intermediate compound’ could not have been reached.For this reason, in the definition off the ratio of AA,,, at a given x,/e was taken relative to AA,,, at x,/e = 4 and not relative to its calculated upper limit AAMA,. Because of the existence of two intermediates and because of the incomplete saturation of Mb(rr1) under the conditions of the experiments, the evaluation of the transient spectrum of the system requires a reappraisal. Thus, in table 1 of ref. (l), the absorbance of a 4.15 x lov5 rnol dm-3 solution of Mb(m) at A = 549 nm is given as 0.232 cm-l [experiment (a)]. From here E~ = 5.59 x lo3 dm3 mol-l cm-l can be calculated. In experiment (b) listed in the same table, the maximal absorbance using e = 5.40 x mol dme3 and x, = 1.75 x mol dm-3 has been found to be 0.506 cm-I.A computer simulation of (p+aq),,, for the above initial conditions gives 1.71 x mol dm-3. From here = (0.506-0.302)/1.71 x = 1.19 x lo4 dm3 mol-1 cm-’ and Using a = 6.1 x lop2, E ~ ~ - E ~ can be calculated as 0.73 x lo3 dm3 mot1 cm-l. E~~ thus becomes 6.32 x lo3 dm3 mol-1 cm-l (A = 549 nm). The present value of is considerably higher than that given by George and Irvine [9.52 x lo3 dm3 mo1-I cm-l at A = 549 nm; fig. 1 of ref. (l)]. Note that the value of George and Irvine is based on the assumption of a 100% saturation of Mb(II1) by MeOOH under the conditions of experiment (b), an assumption which is not supported by the present calculations. No previous estimate of cII exists in the literature. The lack of saturation of Mb(1Ir) by MeOOH at x , / e = 4 also affects the interpretation of the results of the ‘ titration of the intermediate’ with K,[Fe(CN),].George and Irvine performed several experiments to determine the ‘state of oxidation of the intermediate’ = 1.75 x lo4 dm3 mol-1 cm-l(L = 549 nm). 71 FAR 12138 Reaction of Ferrimyoglobin with Methylhydroperoxide by adding K,[Fe(CN),] to a solution of the intermediate preformed from its components. After having made a suitable correction for the ‘ spontaneous decomposition’ of the intermediate, the stoichiometric ratio of K,[Fe(CN),] reacting with an excess of the intermediate was determined [see table 2 of ref. (l)]. It was concluded that the intermediate was in a formal oxidation state of +4 carrying thus one oxidation equivalent relative to Mb(rr1). Although the exact value of e used in the experiments was not stated, it must have been ca.4 x 1OP5-5 x mol dmP3. By performing a model simulation, taking e = 5.2 x mol dmP3 and x,/e = 4, the sum of concentrations of C, and CI, [of mechanism (D)] at the maximum of AA has been calculated as 3.13 x loA5 mol dm-3, ca. 60% of e. (The percentage saturatidn decreases when e is decreased, other factors, such as the excess of peroxide, remaining unchanged.) The reduction, roughly by half, of the number of moles of the intermediates reacting with the same amount of K,[Fe(CN),] increases the number of oxidation equivalents which each intermediate carries by a factor of two. This result is consistent with the assumption of an oxidation state of + 5 of both C, and Cll.A similar conclusion was reached in the analysis of the related reaction of Mb(Ir1) with H,O,., Curve (c) in fig. 2 shows the gradual disappearance of MeOOH during the reaction. The decomposition of MeOOH ultimately becomes complete. Thus, after the lapse of a certain time not enough MeOOH will remain in the system to give a perceptible reaction with a new addition of Mb(rrr) to the reaction mixture. It thus explains qualitatively the observations of George and Irvine [table 1, ref. (l)], according to which no further amount of ‘the intermediate compound’ was formed when a fresh portion of Mb(m) was added to a fully reacted reaction mixture. Note that the exact course of the reaction will depend also on the exact time of addition of the second portion of Mb(Ir1).The model of two intermediates appears to characterize a wide variety of reactions of haemoproteins with H202 and its derivatives. Thus there is increasing evidence for the existence of a precursor complex to ‘catalase compound I’ formed both with H,O, and Et00H.9-11 The analysis of the Mb(1r1)-H202 reaction also leads to the postulation of an intermediate preceding the formation of the ‘red The present model implies that there is no transient release of MeOOH during the reaction. To check this point further, a model was constructed [mechanism (E)] in which steps (2) and (3) of mechanism (D) were substituted by the reactions 12* l3 CI --+ C,,+MeOH (2’) (3’) C,, + MeOH --+ E + CH20 + H20. A computer simulation of this mechanism resulted in r.r.s.= 3.93 x lo-,. The assumption of a release of MeOOH during the reaction thus gives a less complete explanation of the experimental data than the alternative mechanism [mechanism (D)] does, in which such a breakup of the molecule is not envisaged. It may be added that the kinetic analysis of the bacterial catalase-Et00H reaction also led to the conclusion that there is no release of EtOH from EtOOH. (The opposite case is argued by Schonbaum and Chance, who quote unpublished observations of Schonbaum, according to which the alcohol component is released during the reaction of catalase with methyl and butyl hydroperoxide. 14) The present evidence supports the hypothesis of a heterolytic breakup of the peroxidic bond.l5?l6 The following scheme embodies the ideas of Bell and McDougall regarding the mechanism of the base-catalysed decomposition of peroxides and reflects current views on the mechanism of the catalase mediated decomposition of H202:17-21 In the first step MeOOH becomes attached to Mb(rrI) both by an 0 bridge to the Fe3+ centre of the haem and by an H bridge to a histidine residue (B) near the haem group. In the second step there is heterolytic fission of the 0-0 bond with a release on an OH- ion.At the same time a proton becomes attached to histidine. In the final step CH,O is freed and Mb(rr1) regains its protolytic equilibrium with the medium.M. L. Kremer 2139 HCH,-0-OH . . B Fe3+ B Fe3$ CH,-0 . B Fe3+ yH+ r+ +OH- (3) I [ +CH,O+H,O CII Mb(IIr) Chemiluminescence Finally, the interesting experiments of George and Irvine involving the hydroperoxide- Mb(1II) system and luminol should be mentioned.l* They observered a bright flash when either H20, or MeOOH was added to a mixture of luminol and Mb(II1).However, if the luminol was added after the complex has been formed, only a dull luminescence was obtained. It was concluded that ‘an oxidizing entity more powerful than the complex itself was present during the formation of the complex’. The species (denoted here as Y) was identified with OH (or either with OH or OR in the case of MeOOH) on the basis of the following criteria: (1) Y is an oxidizing agent, (2) Y is transient and (3) the flash is observed, if luminol is present, only during the formation phase of the complex. No flash is obtained after the complex has been formed.None of these criteria is specific for free radicals. Thus the intermediate C, or a short-lived precursor to it may satisfy the above criteria. The chemiluminescent reaction of luminol itself is a complex process and at present there is no unambiguous evidence supporting any of the mechanisms that have been proposed.22 It may be initiated by free radicals, as suggested on the basis of pulse-radiolytic studies by Baxendale.23 The participation of free radicals, however, is not an absolute requirement for the observation of cheliluminescence. Thus in aprotic solvents at -60 “C and in the presence of 0,, chemiluminescence is observed with no e.s.r. signal detectable in the reaction In aqueous solutions, evidence for free-radical participation in chemically induced luminescence is mainly circumstantial.22 It should also be mentioned that the mode of reaction of Fe(m) of Mb with H,O, or ROOH, as proposed by George and Irvine, is unusual within the framework of free-radical reactions.In general, a reduction of Fe3+ by H,O, (or ROOH) is Fe3+ + H202 + Fe2+ + H+ + HO,. Thus, from a mixture of alkaline Fe(CN),3- with H20,, which actually produces a ‘bright glow of short duration’ with luminol, Fe(CN)*- ions have been recovered quantitatively.26 In summary, the studies of the chemiluminescence of luminol are not able to establish at present with certainty the nature of the active species formed in the initiating mixture. 71-22140 React ion of Fer r im yoglob in with Met h y lh y droper oxide References 1 P. George and D.H. Irvine, Biochem. J., 1953, 55, 230. 2 P. George and D. H. Irvine, Biochem. J., 1952, 52, 51 1. 3 P. George and D. H. Irvine, Nature (London), 1951, 168, 164. 4 P. George and D. H. Irvine, J. Colloid Sci., 1956, 11, 327. 5 P. George and D. H. Irvine, Biochem. J., 1955,60, 596. 6 M. L. Kremer, Zsr. J. Chem., 1981, 21, 72. 7 E. Zidoni and M. L. Kremer, Arch. Biochem. Biophys., 1974, 161, 658. 8 M. L. Kremer and S. Baer, J. Phys. Chem., 1974, 78, 1919. 9 M. L. Kremer, J. Chem. SOC., Faraday Trans. 1, 1985, 81, 91. 10 P. Jones and H. B. Dunford, J. Theor. Biol., 1977,69, 457. 11 M. L. Kremer, J. Chem. SOC., Faraday Trans. I , 1983, 79, 2125. 12 N. K. King and M. E. Winfield, J. Biol. Chem., 1963, 238, 1520. 13 T. Yonetani and H. Schleyer, J. Biol. Chem., 1967, 242, 1974. 14 G. R. Schonbaum and B. Chance, in The Enzymes, ed. P. D. Boyer (Academic Press, New York, 3rd 15 L. Bateman and K. R. Hargrave, Proc. R. SOC. London, Ser. A , 1954, 224, 389. 16 L. Bateman and K. R. Hargrave, Proc. R. SOC. London, Ser. A , 1954, 224, 399. 17 R. P. Bell and A. 0. McDougall, J. Chem. SOC., 1958, 1697. 18 B. Plesnicar, in The Chemistry of Peroxides, ed. S . Patai (John Wiley, Chichester, 1983), p. 573. 19 P. Jones and A. Suggett, Biochem. J., 1968, 110,621. 20 M. L. Kremer, Zsr. J. Chem., 1971, 9, 321. 21 G. R. Schonbaum and B. Chance in The Enzymes, ed. P. D. Bayer (Academic Press, New York, 3rd 22 E. H. White, in Life and Light, ed. W. D. McElroy and B. Glass (The John Hopkins Press, Baltimore, 23 J. B. Baxendale, J. Chem. SOC., Faraday Trans. 1, 1973,69, 1665. 24 E. H. White and D. F. Roswell, Acc. Chem. Res., 1970, 3, 54. 25 W. G. Barb, J. H. Baxendale, P. George and K. R. Hargrave, Trans. Faraday SOC., 1951,47, 591. 26 F. H. Stross and G. E. K. Branch, J. Org. Chem., 1938,3, 385. edn, 1976), vol. XIII, p. 392. edn, 1976), vol. XIII, p. 395. 1961), p. 183. Paper 511466; Received 27th August, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202133

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraduy Trans. I, 1986,82,2141-2149 Characterization of Anion Solvation in N-Methylacetamide Transfer Enthalpies of Anions and the Reaction Rates of Ethyl Iodide with Bromide Ion in N-Methylacetamide-Acetonitrile and N-Me thy lace tamide-N,N- Dimet hy lace tamide Mixtures Yasuhiko Kondo,* Akikazu Nakano and Shigekazu Kusabayashi Department of Applied Chemistry, Faculty of Engineering, Osaka University, Suita, Osaka 565, Japan Single-ion transfer enthalpies have been determined in N-methylacetamide- acetonitrile and N-methylacetamide-N,N-dimethylacetamide mixtures for bromide, perchlorate, and tetra-n-butylborate ions on the basis of the tetra-n-butylammonium/tetra-n-butylborate assumption. From bothempiri- cal and theoretical treatments of these quantities it is concluded that N-methylacetamide interacts with these spherical and pseudo-spherical anions through the N-H dipole directed towards the anion, i.e. it behaves as a protic solvent.Transfer enthalpies for the transition-state anion have also been determined for the ethyl iodide plus bromide ion reaction in the two solvent mixtures. The presence and the magnitude of the specific interaction between the rod-like transition-state anion and the amide linkage, due to dipoledipole association, have been evaluated on various grounds. The relative alignment of N-methylacetamide towards the anion seems to change as the reaction proceeds from reactants to a transition state. N-Methylacetamide is an unusual liquid because of its high dielectric constant (162 at 40 "C), which results in its having high solubilities for inorganic salts compared with general organic solvents; furthermore, it has long been taken as a model compound for proteins having a peptide 1inkage.lY Recently, much attention has been paid to methods of elucidating the physicochemical properties of N-methylacetamide as a solvent ; these include dielectric-constant and dipole-moment various electrochemical measurements1* 2- ' 9 in the bulk liquid and in solutions of N-methylacetamide and thermodynamic studies in which the focus has been on the structural properties in the cybotactic region of a solute in aqueous N-methylacetamide mixt~res.~-ll Interaction of a solute with N-methylacetamide as solvent can involve either the hydrogen-bond donating function (-NHMe) or the hydrogen-bond accepting function (-N-C=O).Such interactions should be revealed by studying the physicochemical properties of solutions containing N-methylacetamide. A combination of thermodynamic and kinetic measurements carried out in mixed solvents instead of single pure solvents has been shown to yield valuable information about the constituent terms of solvation energies for a reactant molecule as well as for an activated l3 An extention of these procedures to a mixed system, in which a multifunctional solvent is included as one of the constituent solvents, might therefore give detailed information as to solute-solvent interactions in such media. In this work, the enthalpy change of solution for various solutes has been measured in N-methylacetamide (NMA)-acetonitrile (AN) and N-methylacetamide (NMA)-N,N- dimethylacetamide (DMA) binary mixtures; the results are dissected into constituent I I 21412142 Anion Solvat ion in N- Methylace tamide Table 1.Enthalpies of solution in NMA-AN mixtures at 25 "C (in kJ mol-l) 0.0 4.4b 16.7b 10.3b 44.0b 0.1 - 8.75 11.0 41.8 0.25 3.6 10.1 13.2 41.5 0.5 2.8 13.1 17.1 38.6 0.75 2.0 18.2 21.3 37.2 0.9 - 20.7 24.3 36.7 1.0 1 .3c 23.2c 26.8c 35.3c " xNMA is the mole fraction of NMA. values, see text. Ref. (13). Extrapolated Table 2. Enthalpies of solution in NMA-DMA mixtures at 25 "C (in kJ mol-l) 0.0 I .6" 16.1" 3.35" 21.4a 0.1 15.3 4.95 22.2 0.25 1.5 16.35 7.8 22.9 0.5 1.4 18.35 12.4 24.3 0.75 1.3 20.4 17.4 27.0 0.9 - 22.2 22.8 31.0 1 .o 1 .3b 23.2b 26.gb 35.3b - " Ref.(13). Extrapolated values, see text. terms for anion solvation in these mixtures. Then, combining the rate data for the reaction between ethyl iodide and bromide ion reaction measured in these mixtures, the characteristics of anion solvation in these media are discussed. Experimental Materials N-Methylacetamide, after storage over calcium hydride, was distilled under reduced pressure and the procedure repeated three times. Other, materials, i.e. acetonitrile, N,N-dimethylacetamide, ethyl iodide, tetra-n-butylammonium bromide, tetra-n-butyl- ammonium perchlorate and tetra-n-butylammonium tetra-n-butylborate, were treated as described previously.l2? l3 Heats of Solution Measurements Heats of solution were measured with a Tokyo Riko twin isoperibol calorimeter (TIC-2D) at 25.0 k0.05 "C;l2? l3 the final concentration ranges were 3.10 x to 1.64 x mol dm-3 for tetra-n-butylammonium bromide and perchlorate, 3.0 x lop3 to 6.1 x mol dm-3 for tetra-n-butylammonium tetra-n-butylborate and 1.3 x 10-1 to 1.7 x 1 OW1 mol dmp3 for ethyl iodide.Since the freezing point is higher than 25.0 "C, i.e. 30.6 "C, in pure N-methylacetamide all the measurements were performed at higher temperatures, 35.0,45.0 and 55.0 "C, and the results were extrapolated to 25.0 "C. Partial molar enthalpy measurements were performed as described previously at 35.0 "C.l29 l3 Experimental errors in the measurements were estimated to be ca. +2%.Y. Kondo, A . Nakano and S. Kusabayashi 2143 Table 3. Single-ion enthalpies of transfer from AN to solvent mixtures, AHtN+miX at 25 "C (in kJ mol-l) XNMA Br- 0.0 0.1 0.25 0.5 0.75 0.9 1 .o 0 - 6.85 - 5.35 - 0.9 4.9 7.65 10.85 CIO, __- 0 1.8 4.15 9.5 14.4 17.65 20.85 B(Bu"), 0 - 1.1 - 1.25 - 2.7 - 3.4 - 3.65 -4.35 ~_____ TS- 0 - 2.65 2.45 6.45 9.8 18.15 - Table 4.Single-ion enthalpies of transfer from DMA to solvent mixtures, AHFMA-+mix at 25 "C (in kJ mol-l) XNMA Br- C10, B(Bun); TS- 0.0 0 0 0 0 0.1 - 1.2 1.2 0.4 - 0.65 0.25 -0.5 3.7 0.75 2.7 0.5 0.8 7.6 1.45 2.8 0.75 1.5 11.25 2.8 9.9 0.9 1.3 14.65 4.8 1 .o 0.15 16.5 6.95 17.35 - ~ - ~ _ _ _ ~ _ _ _ _ _ _ Table 5. Partial molar enthalpies of mixing, hy, at 35 "C for NMA-AN mixtures (in kJ mol-l) 0.99326 5.81 1 .o 0.0 0.892 2.83 0.9002 0.1 12 0.742 1.39 0.751 0.433 0.493 0.473 0.504 0.973 0.245 0.121 0.260 1.55 0.0980 0.0354 0.113 2.00 0.0 0.0 0.0160 2.44 xAN is the mole fraction of AN.Kinetic Measurements Rates were measured by following the increasing concentration of iodide ion produced in the reaction by potentiometric titration using silver nitrate solution. Rate constants were determined at four of the following temperatures, as described previously,12, l3 0.0, 20.0, 30.0, 40.0, 50.0, 60.0 and 70.0 "C, and experimental errors were estimated to be ca. f 2 % . Results Enthalpies of solution in NMA-AN and in NMA-DMA mixtures are summarized in tables 1 and 2, and single-ion enthalpies of transfer calculated on the basis of tetra-n-butylammonium/tetra-n-butylborate assumption are summarized in tables 32144 An ion So lva t ion in N -Met h y lace tamide Table 6. Rate constants and activation parameters for the reaction between ethyl iodide and bromide ion in NMA-AN mixtures at 30 "C 0.0 108.0" 77.4" - 27.2" 0.1 11.0 81.9 -31.1 0.25 4.19 86.0 - 26.0 0.5 2.07 86.35 - 30.7 0.75 1.42 84.7 - 39.1 1 .o 1.01 87.8 -31.8 a Ref.(13). Table 7. Rate constants and activation parameters for the reaction between ethyl iodide and bromide ion in NMA-DMA mixtures at 30 "C kmix AHhx ASLx XNMA /lop4 dm3 mol-1 s-l /kJ mol-l /J K-l mol-l 0.0 6250a 70.3" - 17.1" 0.1 676 70.85 - 34.2 0.25 138 73.6 - 37.9 0.5 23.2 72.5 - 56.2 0.75 5.23 79.0 - 47.2 1 .o 1.01 87.8 -31.8 a Ref. (13). Table 8. Comparison of single-ion enthalpies of transfer (in kJ mol-l) and dipole moments (with AN as standard solvent) solvent p/D B(Bun); ClO; Br- 22.6b { 25.7b*" - 5.0b - 2O.Ob'.-4.35 20.85 {-13.0" 23.85" MeOH 1 .70" NMA 3.85" DMA 3.96f - 1 1.3b 4.35b 10.7b NMePy 4.099 - 12.6b 6.7b 1 6.8b a Ref. (16). Ref. (13). AHtfG!. AHtg+s. " Ref (6). f Ref. (4). 9 Ref. (14). and 4. With respect to perchlorate and tetra-n-butylborate ions, transfer enthalpies indicate rather linear changes with the composition of the solvent, irrespective of the nature of the solvent mixtures. However, for bromide ion the transfer enthalpy shows a minimum in NMA-AN mixtures, but only small changes in NMA-DMA mixtures. These patterns of behaviour are reminiscent of those observed in methanol-acetonitrile and methanol-N,N-dimethylacetamide mixtures. l3Y. Kondo, A . Nakano and S . Kusabayashi 2145 Partial molar enthalpies of mixing, hi-hp = r;f“ for NMA-AN mixtures are summarized in table 5.The mixtures indicate an endothermic heat of mixing and the results satisfy the Gibbs-Duhem relation within an error of +2%. However, NMA-DMA mixtures show rather small enthalpy changes on mixing, i.e. they form a nearly athermal solution; reliable values could not be obtained by our procedure. Rate constants and activation parameters for the reaction between ethyl iodide and bromide ion in the two solvent mixtures are summarized in tables 6 and 7; the last columns of tables 3 and 4 give the transfer enthalpies of the transition-state anion, AH~p’mix(TS-), calculated through the following thermodynamic cycle : AHtP+mix(TS-) = AHki, AHi, + AHpP+mix(Br-) + AH,AP’mix(EtI) (1) where the suffix AP stands for acetonitrile or N,N-dimethylacetamide. Discussion Characterization of Anion Solvation in N-Methylacetamide In N-methylacetamide there exist two functional groups which are expected to exhibit different responses towards anion solvation, i.e.hydrogen-bond donating, -N-H, and accepting, -N-C=O, groups. In order to empirically characterize the mode of inter- action in N-methylacetamide, transfer enthalpies of anions so far determined by us are compared in table 8 (acetonitrile was chosen as a standard solvent). The transfer enthalpies of both perchlorate and tetra-n-butylborate ions give similar values in methanol and in N-methylacetamide, and, as will be discussed below, the transfer enthalpy of bromide ion into N-methylacetamide consists of two terms, a more ‘physical’ interaction enthalpyl2* l3 and a specific interactionl29 l3 enthalpy, similar to the pattern observed in methan01.l~ In contrast, two other amide solvents, N,N-dimethylacetamide and N-methyl-2-pyrrolidone (NMePy) exhibit quite different, although common to each other, patterns of behaviour, i.e.a more exothermic trend than with methanol and N-me thylacetamide. The rather large dipole moments, ca. 3.9 D, found for the three amides have their origin in the presence of a polarizable -N-C=O group in these r n ~ l e ~ ~ l e ~ . ~ ~ 1 4 7 l5 If the amide group in N-methylacetamide participates in anion solvation, a common transfer enthalpy pattern might be observed throughout for the three amide solvents. However, this is not the case. On the other hand, small bond dipoles have been evaluated for 0--H and N-H bonds, i.e.1.58 and 1.3 D, respectively.16 The agreement in the transfer enthalpy pattern and in the bond dipole for methanol and N-methylacetamide leads to the conclusion that N-methylacetamide participates in anion solvation through an N-H bond directed towards the anion, at least for these spherical or pseudo-spherical anions. The association constants for hydrogen-bonding interactions have been determined spectroscopically in carbon tetrachloride for -N-H. * .Br- and -0-H. * .Br- bonds, i.e. 44 and 49 dm3 mol-l, respectively.l79 l8 From relaxation-time measurements, a hydrogen-bonding interaction between the fluoride ion and the relevant solvent has been suggested to occur in methanol and formamide X-ray diffraction studies have also suggested the presence of a hydrogen-bonding interaction between the chloride ion and formamide.20 These results are in agreement with our views on hydrogen bonding given above. Transfer-enthalpy profiles observed for the three anions in NMA-AN and NMA-DMA mixtures are similar to those observed in MeOH-AN and MeOH-DMA mixtures.13 For the perchlorate anion, the transfer enthalpy profiles always indicate linear changes with solvent composition and do not have any minima or significant curvatures which could be taken as the sign of a specific solute-solvent interaction operating in the media.Thus the transfer enthalpy for the perchlorate ion could be taken as mostly consisting of more2146 Anion Solvation in N-Methylacetamide ‘physical’ interaction energies.l27 l3 In contrast, in NMA-AN mixtures the transfer enthalpy for the bromide ion consists of at least two terms, i.e.a term which corresponds to a sharp change at small compositions of N-methylacetamide, and a term which gives the corresponding pattern to that observed for the perchlorate ion. These terms may be identified with the enthalpy due to a specific interaction between the bromide ion and N-methylacetamide, and the more ‘physical’ enthalpy as shown from the similarity to methano113 and from the various features cited above.1s-20 In dissecting the transfer enthalpies into their constituent terms, the same procedures and equations as derived previouslyl29 l3 could be adopted for the bromide ion in NMA-AN mixtures, except for one slight modification, i.e. eqn,(2)-(4) : (4) AHtAN+mix = AHttEh;;SmiX + AHAN+mix r, si where suffixes 1 and 4 stand for NMA and AN, and a, and a4 are the activities of NMA and AN.In eqn (2) the coefficient 0.5 is used in place of the coefficient 1.23 used in acetonitrile-methanol mixtures,13 since it better simulates the transfer enthalpies for the perchlorate ion in the present mixtures. Activity coefficients for AN and NMA have been calculated from eqn (5) on the basis of the assumption that the entropy term does not have a significant contribution in the system: h y = RT In yi. ( 5 ) Calculated values are compared with the.experimenta1 results in fig. 1 . So far, three types of indices are available which are expected to be connected with the hydrogen- bonding donor ability in the respective solvent, i.e.a specific interaction enthalpy for the bromide ion, AHt$-’s(Br-), activation enthalpy differences for the reaction between ethyl iodide and the bromide ion, AH4 - AHI,, and bond dipole moments. Respective values for methanol and N-methylacetamide solvents are - 20.013 and - 13.0 kJ mol-l for AH&+ S(Br-), 17.613 and 10.4 kJ mol-l for AH&-AaN, 1.58 D for the 0-H bond and 1.3 D for the N-H bond? All the results point to a weaker hydrogen-bond donor ability for the N-H group. Specific Interaction at the Transition State A critical evaluation of solute-solvent interactions at the transition state could be made more effective by choosing a suitable model for the transition-state anion as a comparison.l27 l3 In previous work the perchlorate ion was chosen for this purpose.l2, l3 The combination of three transfer enthalpies defined by CAH,AN’S = AH,AN+S(C1O;) -[AH,AN+S(Br-)+AH,AN-’S(EtI)] (6) corresponds to the differential activation enthalpy, A@ - AMAN, to the extent that the perchlorate ion mimics the behaviour of the transition-state anion with respect to solvent variations [see also eqn (l)]. Comparisons are shown in fig.2, including our previous re~u1ts.l~ The correlation line has a slope of 0.98. To some extent, the perchlorate ion is a good model for the transition-state anion as described.129 l3 However, a closer look reveals that the points representing amide solvents systematically deviate downward from the correlation line. A more quantitative thermodynamic discrimination between the perchlorate ion and the transition-state anion can be made in terms of differential transfer enthalpies, AH,AN+S(CIO,) - AHkN’S(TS-).The respective values for three amide solvents, some of which have been calculated from previous data,l29 l3 are 2.7, 3.5 and 9.7 kJ mol-1 for NMA, DMA and NMePy, respectively. The rising trend in the valuesY. Kondo, A. Nakano and S. Kusabayashi 2147 1.0 0.5 0 XNMA Fig. 1. Comparison of the calculated values with the experimental results for single-ion enthalpies of transfer in NMA-AN mixtures at 25 "C. 0, 0, Experimental results for the bromide ion and for the transition-state anion. (-) Calculated from eqn (2)-(4) with K,, = 8.0. AHt:h;2MA = 23.85 kJ mol-l and = - 13.0 kJ mol-l. (-) Calculated from eqn (2)-(4) with K,, = 6.0, AHtFh;FMA = 24.15 kJ mol-1 and = -6.0 kJ mol-l.has, although partly fortuitous, a correspondence in the trend of the dipole moments, 3.85, 3.96 and 4.09 D for NMA,6 DMA4 and NMePy,14 respectively. In the two mixed solvents, perchlorate ion enthalpies indicate a steady increase with increasing NMA content (see tables 3 and 4). However, in NMA-AN mixtures the transition-state anion gives a minimum in transfer enthalpy us. solvent composition profiles ; this is direct evidence for a specific transition-state anion-solvent interaction operating in the mixtures. In NMA-DMA mixtures, the bromide-ion enthalpy shows no substantial variation, and the perchlorate-ion enthalpy shows a linear increase with increasing NMA content (see table 4). In contrast to this behaviour, the transition-state anion gives very characteristic features, i.e. the transfer enthalpies from pure DMA to pure NMA are close to each other for the perchlorate and transition-state ions (16.5 and 17.35 kJ mol-l); however, the difference in the transfer enthalpies, AHf)MA+miX(CIO;) - AH,DMA'mix(TS-), reaches a maximum, 4.8 kJ mol-l, at a mole fraction of 0.5 (see table 4).All these features can be accommodated into our scheme if the assumption is made that for the transition-state anion-amide solvent combinations a specific solute-solvent interaction other than hydrogen bonding makes a significant contribution to the transition-state anion transfer. The transition-state anion has the backbone (Br- * -C. - .I)-, with a negative charge spread2148 Anion Solvat ion in N- Me t h y lace tamide 0 60.0 1 7 -10.0 0 10.0 20.0 XAHPN'S/kJ mol-' Fig.2. Comparison of activation enthalpies with the combined enthalpy change, ZAH,AN'S. Three protic solvents are located on the right-hand side of acetonitrile. (1) MeOH, (2) 2-PrOH; (3) N-methylacetamide, (4) acetonitrile, ( 5 ) propylene carbonate, (6) N,N-dimethylacetamide and (7) N-methyl-2-pyrrolidone. over three atoms. The rather large dipole moment of the amide molecule has its origin in the polarizable N-C=O ba~kbone.~? 1 4 9 l5 The ability of amide molecules to discrimi- nate between the pseudo-spherical perchlorate ion and the rod-like transition-state anion seems to be connected with the geometrical requirements of the pertinent molecules. Thus, as one of the most plausible candidates for such a specific interaction, dipole- dipole association could be invoked between the transition-state anion and an amide molecule with a parallel alignment of the two molecules.Such a supposition agrees with the fact that N-methyl-2-pyrrolidone, for which the parallel alignment should be most favourable because of a fixed geometry around the N-CO bond, shows the largest discrimination between the perchlorate and transition-state anions, i.e. AHtN-+NMePy(CIO;) - AH,AN'NMePy(TS-) = 9.7 kJ mo1-l.12 The same theoretical procedures as applied to the analysis of bromide-ion enthalpies have also been adopted for the analysis of the transition-state anion enthalpies in NMAFAN mixtures. Comparisons are shown in fig. 1. The assumption of a specific interaction between the transition-state anion and the amide, i.e.AH('&+NMA = -6.0 kJ mol-l and Kse = 6.0, leads to a better simulation of the transfer enthalpies. The presence of the specific interaction both for the transition-state anion and for the bromide ion, leads to the necessary consequence that the activation entropy vs. solvent composition profile does not show a sharp maximum over the entire range (see table 6). This is in marked contrast to the same reaction carried out in methanol- acetonitrile mixtures,13 where the combination of no sizeable specific interaction at the transition state with a definite specific interaction for the bromide ion results in a sharp maximum in the activation entropy vs. solvent composition pr0fi1e.l~ Permittivity studies have indicated the formation of a 1 : 1 NMA-DMA complex in carbon tetrachloride,6 and the presence of multimers due to chain association in pure NMA have been suggested on various grounds.21 3 9 6 7 21 The addition of DMA to NMA leads to a variation in solution structures, this effect being most pronounced at the mole fraction 0.5.This should have a direct bearing on the presence of the minimum observedY. Kondo, A . Nakano and S. Kusabayashi 2149 at 0.5 mole fraction on the activation entropy us. solvent composition profile in MNA-DMA mixtures (see table 7). A minimum has also been observed in the profile for MeOH-DMA mixtures at xMeOH = 0.75. Comparative studies of the two systems would be useful in understanding the relations between rate behaviour and the structural properties of the solvents.Conclusion N-Methylacetamide participates in anion solvation through an N-H bond dipole directed towards anions. With respect to a rod-like transition-state anion, a specific interaction due to dipole-dipole association makes a significant contribution to the transfer enthalpies. These results lead to a scheme in which the relative orientation between solute and solvent (in other words the mode of the solute - solvent interaction) changes as the reaction proceeds from the initial state to the transition state. Although the perchlorate ion serves as a useful model for the transition-state anion in some solvents, detailed discrimination between anions can be achieved by a judicious choice of solvent mixtures. We thank Dr M. H.Abraham, University of Surrey, for a critical reading of the original version of this manuscript. References 1 J. W. Vaughn, in The Chemistry of Non-aqueous Solvents, ed. J. J. Lagowski (Academic Press, 2 L. A. Knecht, Pure Appl. Chem., 1971, 27, 283. 3 S. J. Bass, W. I. Nathan, R. M. Meighan and R. H. Cole, J. Phys. Chem., 1964, 68, 509. 4 W. P. Purcell and J. A. Singer, J. Phys. Chem., 1967, 71, 4316. 5 M. M. Omar, J. Chem. SOC., Faraday Trans. I , 1978, 74, 1 15. 6 K. Prafat, J. Jadiyn and S. Balanicka, J. Phys. Chem., 1983, 87, 1385. 7 G. Kortum and C. Hebestreit, Z. Phys. Chem. (Frankfurt am Main), 1974, 93, 235. 8 G. Kortum and C. Hebestreit, Z. Phys. Chem. (Frankfurt am Main), 1975, 95, 65. 9 C. de Visser and G. Somsen, J. Chem. SOC., Faraday Trans. I , 1973,69, 1440. New York, 1967), vol. 11. 10 C. de Visser and G. Somsen, J. Chem. Thermodyn., 1973, 5, 147. 11 J. Manczinger and G. Kortum, 2. Phys. Chem. (Frankfurt am Main), 1975,95, 177. 12 Y. Kondo, K. Yuki, T. Yoshida. and N. Tokura, J. Chem. SOC., Faraday Trans. I , 1980, 76, 812. 13 Y. Kondo, M. Itto and S. Kusabayashi, J. Chem. SOC., Faraday Trans. 1 , 1982, 78, 2793. 14 E. Fischer, J. Chem. SOC., 1955, 1382. 15 S. K. Knudson and J. P. Idoux, J. Org. Chem., 1979, 44, 520. 16 Landolt Bornsrein Tabellen (Springer-Verlag, Berlin, 195 l), band I. 17 R. R. Ryall, H. A. Strobel and M. C. R. Symons, J. Phys. SOC., 1977,81,253. 18 M. C. R. Symons, T. A. Shippey and P. P. Rastogi, J. Chem. SOC., Faraday Trans. I , 1980,76, 2251. 19 I. Birkel and H. G. Herz, J. Chem. SOC., Faraday Trans. I , 1981, 77, 2315. 20 H. Ohtaki and H. Wada, J. Solulion Chem., 1985, 14, 209. 21 W. L. Jorgensen and C. J. Swenson, J. Am. Chem. SOC., 1985, 107, 569. Paper 51 1503 ; 2nd September, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202141

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1986,82,2151-2154 Superacid Sites in Zeolite H-Mordenite Kurt A. Becker Fritz- Haber- Institut der Max- Planck-Gesellschaf t, Ab teilung Grenzjlachenreakt ionen, Faradayweg 4-6, 1000 West Berlin 33, Federal Republic of Germany S tanisYaw Kowalak * Faculty of Chemistry, A . Mickiewicz University, Grunwaldzka 6, 60-780 Poznari, Poland The presence of superacidic sites in fluorinated H-mordenite has been proved by calorimetric measurement of the differential heat of ammonia adsorption, catalytic testing for paraffin conversion at low temperatures and by means of the indicator test. The high proton-donor strength of superacids results generally from an interaction of Lewis and Brarnsted centres present in the system.l The Lewis sites can either form a complex with the conjugate base of the Brnrnsted acid or react with this acid.l The high acidity of zeolite catalysts is also believed to be a result of interaction of strong protonic acid centres (acidic OH groups) and Lewis-type sites.273 Among the many attempts to prepare solid superacidic catalysts, zeolites were also taken into consideration.Tanabe and coworkers4 have presented heterogeneous superacid catalysts prepared on the basis ofmetal oxides and zeolites by treatment with SbF, and other fluorine-containing compounds. The authors found the following order of decrease in catalytic effectiveness : SbF, > FSO,-SbF, > SbCl, $ FS0,H > NH4F. The activity of modified zeolites was usually lower than that of the modified oxides. Mirodatos and Barthomeup have suggested the presence of superacid sites in mordenites modified by steaming.This presumption is based on the results of ammonia thermodesorption measurements. The authors believe that, owing to dealumination occurring during steaming, some aluminium atoms removed from the lattice are deposited in the channels as polymeric oxoaluminium species (AIO);. These species can act as strong Lewis sites and interact with protonic sites. The interaction can result in the creation of superacid sites. Fajula and Gaults have reported the formation of allylic cations during the reaction of 2-methylpropane catalysed by H-mordenite. Such a mechanism involving hydride ion abstraction requires very strong acidic centres, which, according to the authors, can be obtained in H-mordenite by cooperation of Lewis and Brarnsted acidic sites.Hydride-ion abstraction was also suggested by Minachev et aL7 as the mechanism of pentane isomerization catalysed by H-mordenite. Haag and Dessau* have suggested a pentacoordinated carbenium ion as an intermediate product during isomerization of paraffins catalysed by ZSM zeolites. Such a mechanism was proposed for paraffin reactions involving superacid centres. Previous reports do not show clear evidence of the presence of superacid sites in zeolites, although they do suggest such a possibility. Attempts to find or prepare the superacidic centres in zeolites have been carried out mostly with highly silicious zeolites, known to possess very strong acidic sites. 215121 52 ... . . . .._..., ...: , . . ... '. HM-F Ln +-' .- c . . . . . ' ', : IHh . . --I- Superacid Sites in Zeolite H-Mordenite 84 I00 I25 I50 I75 ElkJ mol-' Fig. 1. Ammonia adsorption energy distribution on unfluorinated H-mordenite. The strength of acids is often expressed by means of the Hammett acidity function. The population and strength of acidic sites can be estimated by a titration with butylamine and using indicators of known pK,. The reaction of paraffins at low temperatures is widely performed as a characteristic test, proving the presence of superacid sites in a catalyst. Recently, Taniguchi et aZ. lo have proposed calorimetric measurement of the differential heat of ammonia adsorption as a powerful method to determine the strength of acidic sites. They correlated the values of the Jammett function and the heat of ammonia adsorption.These authorsll have found that for the solid superacid prepared by modification of a silica-alumina catalyst with SbF,, the value of the initial heat of ammonia adsorption was 170 kJ mol-1 and the energies of adsorption were distributed at ca. 137 kJ mol-l. According to the authors, the above energy values correspond to the acidity strength H, = - 14.. In a former study12 we found a significant increase in catalytic activity for cumene cracking of H-mordenite modified with fluorine. The results of catalytic testing were in good agreement with the data of the heat of ammonia adsorption. Fig. 1 shows the ammonia adsorption energy distribution for both fluorinated and unmodified H-mor- denite samples.The microcalorimetric measurement of the heat of adsorption was carried out by means of the twin Calvet-type microcalorimeter. The samples were pretreated at 450 "C under vacuum for 12 h. The heat of adsorption was measured at 26 "C. The initial heat of ammonia adsorption for fluorinated samples is higher than 200 kJ mol, whereas in the case of unmodified H-mordenite it approaches values in the range of 160 kJ mol-l. The maximum of the energy peak for the flourinated form is ca.175 kJ mol-l and ca. 140 kJ mol-1 for the unmodified sample. The values of the heat of adsorption for the fluorine-modified sample are much higher than those measured by Taniguchi et d.l1 for the superacidic catalyst. In the case of untreated H-mordenite the heat values approach those relating to superacids.In addition to microcalorimetric measurements, the indicator tests and catalytic reactions were performed to prove the presence of superacid sites in the samples under study. The following indicators were used : p-nitrotoluene (pK, = 1 1.39, p-nitrochloro- benzene (pK, = - 12.7) and rn-nitrochlorobenzene (pK, = - 13.75).13 The indicators were stored in evacuated, sealed bulbs connected directly to glass tubes where the zeoliteK. A . Becker and S. Kowalak 2153 Table 1. Comparison of n-hexane conversion on fluorinated and unfluorinated H-mordenite for different reaction conditions n-hexane conversion (% ) reation conditions H-mordenite HM-F 25 "C, 240 h traces below 0.005 0.3 60 "C, 120 h 0.01 0.3 90 O C , 50 h 0.5 5 ~____ samples were thermally activated (450 "C) under vacuum.After activation (20 h) the tubes were sealed and the bulbs were broken. Then the tubes were heated over night at 50°C. The colour changes were estimated by visual observation, where samples of Na-mordenite with adsorbed indicator were used as references. The colour change was observed for the p-nitrotoluene and p-nitrochlorobenzene indicators used in the case of fluorinated H-mordenite. This means that acid sites stronger than pK, = - 12.7 but weaker than - 13.75 were present in this catalyst. In the case of unmodified sample only a colour change in p-nitrotoluene was observed, which confirms the results of the calorimetric test and proves the supposition that some superacid sites can also be found in unmodified H-mordenite.The catalytic test was performed using the same glass vessels as for the indicator test. Instead of indicators, 1 cm3 of n-hexane was stored in breakable bulbs. Zeolite samples (200 mg) were activated the same way as for an indicator test. After breaking the wall of the connecting n-hexane containing bulbs, the vessels were kept in thermostats at various temperatures and lengths of time. The product analysis was performed by means of a gas chromatograph. The results are compiled in table 1. Catalytic testing of the investigated samples shows low activity of the catalysts under the conditions used. There is no doubt, however, that some activity can be found even at room temperature. The activity of the fluorinated sample predominates significantly, but some catalytic effectivity of unmodified H-mordenite confirms the calorimetric results, suggesting that the strength of acidic sites in H-mordenite approaches the value designated for superacids.Enhanced acidity of HM-F probably results from the inductive effect of F-atoms on the remaining acidic OH-groups, or is an effect of interaction between strong Brarnsted sites (OH-groups), and strong fluorine-bearing Lewis acid centres. We have also observed the possibility of generation of superacid sites in the case of ZSM-5 zeolites after fluorine treatment.14 The authors are very thankful to Dr M. Marczewski (Technical University, Warsaw) for valuable discussions and helpful suggestions. We are also indebted to Mrs U. Kockeritz and Mr W. Kollmitt for their valuable assistance during the experiments.References 1 B. C. Gates, J. R. Katzer and G. C . A. Schuit, Chemistry of Catalytic Processes (McGraw-Hill, New 2 J. A. Rabo and M. I. Poutsma, Adv. Chem. Ser., 1971, 102, 248. 3 P. A. Jacobs, Catal. Rev. Sci. Eng., 1982, 24, 415. 4 H. Hattori, 0. Takahashi, M. Takagi and K. Tanabe, J . Catal., 1981,68, 132. 5 C. Mirodatos and D. Barthomeuf, J. Chem. Soc., Chem. Commun., 1981, 39. 6 F. Fajula and F. G. Gault, J . Catal., 1981, 68, 291. York, 1979), p. 20.2154 Superacid Sites in Zeolite H-Mordenite 7 K. Minachev, v. Garanin, T. Isakova, U. Kharlamov and V. Bogomolov, Adv. Chem. Ser., 1971, 102, 8 W. 0. Haag and R. M. Dessau, 8th Int. Congr. Catal., Berlin, 1984 (Verlag Chemie, Berlin, 1984), 9 G. Olah, J. Am. Chem. SOC., 1972, 94, 808. 441. vol. 2. 10 H. Taniguchi, T. Masuda, K. Tsutsumi and H. Takahashi, Bull. Chem. SOC. Jpn, 1978,53, 1970. 1 1 H. Taniguchi, T. Masuda, K. Tsutsumi and H. Takahashi, Bull. Chem. Soc. Jpn, 1980,53, 2463. 12 K. A. Becher, S. Kowalak, J. Chem. SOC., Faraday Trans. I , 1985,81, 1161. 13 R. J. Gillespie, T. E. Peel and E. A. Robinson, J. Am. Chem. Soc., 1971, 93, 5083. 14 K. A. Becker, S. Kowalak and K. Fabianska, to be published. Paper 511504; Received 2nd September, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202151

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1986, 82, 2155-2166 Application of Infrared Spectroscopy to the Measurement of Surface and Bulk Oxidation/Reduction States of MOO, Jong S. Chung and Carroll 0. Bennett* Department of Chemical Engineering, University of Connecticut, Storrs, Connecticut 06268, U.S.A. The oxidation/reduttion state of finely divided MOO, catalyst has been studied by measuring the background transmission of infrared radiation, Experimental results have shown that the background transmission at lower wavenumber is mainly sensitive to the surface oxidation/reduction state of MOO,, and the transmission at higher wavenumber is more sensitive to the subsurface layers. Assuming a different penetration depth of infrared radiation through the MOO, particle at a different frequency, based on electromagnetic theory, it is possible to explain the above results by a simple model.A normalized absorbance is defined and used to measure the degree of reduction on the surface and in the bulk of Moo,. When MOO, is reduced by hydrogen, by methanol alone, and by methanol-0, mixtures below 140 "C, the surface is more reduced than the bulk. Above 140 "C exposure of MOO, to methanol-oxygen mixtures produces particles for which the subsurface layers are generally more reduced than the surface. The degree of reduction is dependent on oxygen concentration and is controlled by the way in which the reduction force by methanol is balanced by the oxidation force by oxygen. The existence of a space-charge barrier caused by ionosorbed oxygen is a characteristic of the catalyst during the reaction.The partial oxidation of methanol to formaldehyde over oxide catalysts is an active field for studies aimed at understanding the sequence of steps and identifying the intermediates. We are pursuing such work by means of i.r. spectroscopy and by kinetic studies implemented by mass spectrometry. Although methoxy and hydroxy groups have been proposed as important intermediates,l it has not been possible to observe these or other adsorbed species by i.r. spectroscopy owing to the low transmission of typical catalysts, which have B.E.T. areas of the order of 5 m2 g-l. By using as a model catalyst MOO, made in a flame reactor (an aerosol), we are able to obtain useful i.r. spectra. The stable surface area of this is 27 m2 g-l.In parallel with a number of transient and steady-state experiments performed in the i.r. rea~tor-cell,~:~ we have conducted similar experiments in a differential reactor with gas analysis by mass spectrometry.2, The catalyst in these studies is made of broken discs of the same thickness as those used in the i.r. experiments. In this way i.r. absorbances can be associated with quantities adsorbed. The present work is concerned with certain aspects of the i.r. studies. When methanol-oxygen mixtures are passed over the MOO, catalyst we can see by i.r. various kinds of OH and CH bands. In addition we can measure the oxide structural bands, which become less intense as the oxide is partly reduced by methanol-containing mixtures. These results have been p~blished~-~ and have permitted some progress in understanding the kinetics of methanol oxidation over MOO,.In the present work we are concerned with the observation that the background transmission level at frequencies where there are no absorption bands (we have chosen 4000,2500, and 1320 cm-l for most of the work) is reduced as the catalyst is exposed to methanol. Supposedly the concentration of electrons in the conduction band of the n-type semiconductor MOO, is being increased 215521 56 Measurement of OxidationlReduction States of MOO, so that the electrical conductivity increases. In this paper we propose to show the experimental relation between the i.r. background transmission and the degree of reduction of MOO,. We then use the method to study the state of MOO, as it is exposed to methanol-oxygen mixtures of various proportions at various temperatures. The electrical conductivity is one of the important properties of semiconducting catalysts.Since Hauffe6 reviewed this area, there has been considerable work reported on the change in the electrical conductivities of metal-oxide catalysts during gas adsorption or reaction. So far the conductivity of powder samples has been measured principally by RC bridge methods. Bielanski et aZ.7 applied the variable frequency a s . method in order to discriminate the surface conductance of V,O,-MOO, from the bulk one. However, these conventional methods often induce contacting problems between the probes and the sample, even with an improved four-point probe method.* There have been indications that the background transmission of i.r.is related to the electrical conductivity of catalysts. Hertelg observed that the background transmission was a function of the electrical conductivity of cobalt oxide catalyst under the reaction conditions of CO and hydrocarbon oxidations. Ueno et aZ.l0 observed significant changes in the background transmission of NiO catalyst when 0,, CO or CO, was added to the catalyst. Scholton and van Montfoortll studied thechange in the background transmission of ZnO catalyst caused by hydrogen and based their explanation on variation in the electron concentration in the conduction band. During the i.r. study of the methanol oxidation reaction over MOO,, we have found that the change in the background transmission of MOO, during methanol chemisorption or the CH,OH-0, reaction was a function of the frequency of the i.r.radiation. We shall show how this can be used to measure the oxidation/reduction state of the surface and bulk layers of MOO,. The frequency dependence on the absorption and scattering of electromagnetic waves is described by general electromagnetic theory.12 Wave propagation through the conducting medium is determined by the complex index of refraction and this is dependent on the dielectric constant and the conductivity of the material. The frequency dependence on these values is described according to classical dispersion theory.12 Recent studies with ionic crystals and wide-gap semiconductors show that the absorption index depends on the frequency as exp ( -co).l3g l4 In any event, the penetration depth of i.r.radiation into a uniform semiconducting material must increase with increasing frequency. The literature thus is in general agreement with the results which will be presented: reduction lowers the i.r. transmission. However, to make the method quantitative some samples of a known, uniform degree of reduction are needed, whereas almost all treatments in methanol-oxygen lead to particles with an internal gradient of degree of reduction. Thus data obtained by mass spectrometry, based on the oxygen uptake of reduced sample, give us the global degree of reduction, not that on or near the surface, and this is the region probed by the i.r. experiments. The degree of reduction from studies by a microbalance also refer to the whole parti~le.~ A small degree of reduction leads to non-uniform particles and a large degree of reduction, perhaps corresponding to a uniform particle, leads to zero transmission.Thus the problem of calibration will occupy our attention considerably in what follows. We have found that an extended treatment at 350 "C in helium will lead to pure, uniform MogO,,, as determined by X-ray diffraction. Extended treatment in oxygen at 380 "C leads to pure MOO,. Thus we have these two reference points. It is not possible to use lower oxides, since they involve major changes in the lattice structure. For the MOO, aerosol catalyst the particle size (45 nm) is smaller than the typical i.r. wavelength (I pm) so that Rayleigh scattering should prevail. In any event, scattering affects transmission and is a function of wavenumbers.In our analysis we have used a normalized absorbance A , which is responsive mainly to the oxidation/reduction state,J . S. Chung and C. 0. Bennett 2157 with the scattering effect removed. Thus, for the uniform samples MOO, (completely oxidized) and Mo,O,, we find that the two different values of A , observed are not functions of frequency. For non-uniformly reduced samples, the observed variation of A , with frequency is related to the profile of degree of reduction within the catalyst particles. Experimental Catalyst The catalyst used for the experiment consists of finely divided unsupported MOO, particles made in a flame reactor and obtained from Teichner's laboratory in Lyon.15 The B.E.T.surface area after extensive treatment in oxygen at 380 "C is 27 m2 g-l, and the average particle size measured by X-ray line broadening is 45 nm. The disc for the i.r. cell was made by pressing the catalyst powder onto a stainless-steel mirror of 0.865 cm2 using a hydraulic press. Usually 3-10 mg of catalyst was pressed at ca. 20000 psi? for 2 s. To reduce irregularities in the disc thickness, the surface of the disc was smoothed with additional powder and pressed again with 70% of the original pressure for 2 s. The fully oxidized MOO, was considered to be made after the original sample was heated in the reactor cell with 0, flow at 380 "C for more than 8 h. X-Ray diffraction showed it to be orthorhombic MOO, and its colour was yellow-white. Mo,O,, was made from the original sample by heating with an inert gas flow at 350 "C for more than 2 h.The triclinic structure was also confirmed by X-ray diffraction, and the colour was grey-blue. Apparatus The optics of the i.r. system and the design of the reactor cell were the same as those previously described.16 An additional thermocouple (0.02 in$ 0.d.) was installed to contact the side wall of the stainless-steel mirror which held the catalyst disc. Because the light travelled through the catalyst disc twice with a 24" angle, an effective mass was used for the experimental results : 1 effective mass = m 1 +- = 2. Im ( cos 24") where m is the weight of original catalyst deposited on the mirror. ZnS, of thickness 2 mm, was used for the i.r. window. The same Spex monochromator and Harrick optical system were used as in previous studies.l6 However, a mercury-doped cadmium telluride detector, operated at liquid- nitrogen temperature, gave more than an order of magnitude higher sensitivity than the thermocouple detector used previously. A 4: 1 Ar-He mixture was used as an inert gas, so that its thermal conductivity was about that of pure 0,. This gas was passed through a deoxo unit, packed with Cu,O-Al,O,, and then through a drier filled with zeolite 5A. Other gases of high purity grade were used without further purification. Methanol was further dried by zeolite 5A. A saturator was used to vaporize methanol into the gas streams. It consisted of two consecutive glass columns filled with glass beads. The first column vaporized methanol at 0 "C and the second one, without methanol, was used to remove the entrained liquid methanol in the gas.The flow rate of gas was 35 cm3 min-l, corresponding to a residence time of <2 s in the reactor. 7 1 psi z 6.89 x lo3 Pa. $ 1 in = 2.54 cm.2158 Measurement of Oxidation1 Reduction States of MOO, Experimental Procedure The frequencies for the background transmission measurements were chosen at levels where no specific absorption band from either chemisorbed methanol or the MOO, structure was observed. The lowest frequency which is considered as background transmission is 1320 cm-l, since there exist a series of structural bands of MOO, below 1320 cm-l. Because of the severe reduction in the transmission at frequencies above 4000 cm-l even with oxidized MOO,, 4000 cm-l is chosen as the highest frequency for the background transmission. After the catalyst was fully oxidized, a desired gas was introduced and the transmission at fixed wavenumber was recorded.Usually, balanced beams between the catalyst in the sample cell and the mirror in the reference cell were used in order to increase the sensitivity when the change in the transmission was very small. Gas was allowed to flow only through the sample cell except when methanol vapour was introduced to the reactor below 120 "C. In this case there was a deposition of formaldehyde on the i.r. windows, although the cooling water was turned off. This affected the background transmission at 1320 cm-l. In order to correct this, methanol was allowed to flow through both of the cells below 120 "C and the temperature of both the cells was kept the same.The transmission used for the calculations was obtained by subtracting emission from the observed transmission. When MOO, is pressed onto the mirror of the i.r. cell and then fully oxidized, as already described, a certain level of transmission is observed, for instance at 4000 cm-l. This result is to be compared to that obtained in similar conditions but after the MOO, is washed off the mirror at the end of a series of experiments. Thus we correct for the reflectivity of the mirror, the absorption by the ZnS window and the strength of the Nernst source. The measured decrease in transmission in the presence of MOO, is thus caused by the absorption and scattering of the oxide.The experiment is similar to the usual scattered transmission method. The only difference is that the i.r. beam traverses the sample a second time, after being specularly reflected by the mirror. We are not expressly measuring radiation which is diffusely reflected. Results and Discussion Fig. 1 shows the absorption of oxidized MOO, at 27 "C as a function of the mass of the oxide disc on the mirror. Since it is necessary to change the sample occasionally, and it is not convenient to use exactly the same amount in successive preparations, this type of curve is useful. As will be shown later, knowing the variation of absorption with mass will permit the estimation of certain parameters arising from the Beer-Lambert law. Note that fig. 1 involves the ratio TJT,,, the transmission of the mirror divided by that of the oxidized sample.We now want to consider the effect of reduction on the transmission, so the MOO, will be exposed to various mixtures containing methanol at various temperatures and thereby produce various degrees of reduction. Fig. 2 presents the effect of the reduction caused by the adsorption of methanol at 27 "C from a 3.6% methanol in oxygen mixture. In this figure the results are presented in terms of the ratio of the transmission of the oxidized sample ( qx) to that of the reduced sample (T,) at the same temperature and wavenumber. It is clear that the reduction has a much larger effect on the transmission at 1320 cm-l than on that at 4000 cm-l. At conditions for which the MOO, is slightly, but uniformly reduced to form Mo,0z6, we shall show presently that the transmission at 1320 cm-l is not affected differently from that at 4000 cm-l.However, to show this we need a more quantitative description of the phenomena, and this will be developed presently. For the moment, we propose that at 27 "C the effect of methanol adsorption is to reduce (partially) only the surface layer of the oxide.J. S. Chung and C. 0. Bennett 21 59 0 _ - . - 0 5 10 15 mass of MoOJmg 0 Fig. 1. Effect of the disc mass of fully oxidized MOO, particles on background transmission at 27 "C. 0,4000 cm-l, 0.13 m+O.S; A, 1320 cm-l, 0.09 m i 3.045. mass of Mo03/mg Fig. 2. Effect of the disc mass on the background transmission upon exposure to 3.6% methanol in oxygen at 27 "C. 0,4000 crn-l, 0.62 x m; A, 1320 crn-l, 4.10 x m.This behaviour is in qualitative agreement with the physical principles referred to in the introduction.12 Radiation at 1320 cm-l probes the surface layers, somewhat reduced at 27 "C; radiation at 4000 cm-l probes a number of layers, most of which are not reduced by the adsorption of methanol at 27 "C. We can now explain some of the results of fig. 1 and 2 by a simple application of the Beer-Lambert law, which can be written for the oxidized MOO, as In (%/ Tox) = w I,(v) c o x * + r(v). (2) Thus the slopes of the lines in fig. 1 are determined by the extinction coefficient e(v), the path length Zp(v), which is proportional to the depth of penetration of the light into an individual particle, and Cox, the free-electron concentration for oxidized MOO,.m is the2160 Measurement of Oxidation/Reduction States of MOO, cr f Fig. 3. Simplified concentration profiles of degree of reduction, assumed proportional to free electron concentration for methanol chemisorbed on oxidized MOO, at 25-100 "C. (a) C,(x), (b) C,( 1320), (c) cr(4000), ( d ) I,( 1320), (e) 1,(4000). - mass of sample and v is the frequency. The intercept ~ ( v ) has been added, and as seen in fig. 1 it is a function of v. As the sample size tends to zero, the absorbance should also, so ~ ( v ) must represent a reflection effect from the physical surface of the disc and this is constant as m varies. If MOO, is partially reduced by some treatment by methanol-containing gas, a gradient of free-electron concentration (reduction) will be established in the MOO, so that C,(x) depends on x, the distance from the surface.Then we write (3) T, %(") T, In- = Jb E(v)mC,(x)dx+q(v). When eqn (2) is subtracted from this equation, the result is $(v) In - = me(v) Jo [C,(x) - Cox] dx. (4) GX T, Taking ~,[l,(v)] as the average free-electron concentration up to lp, we rewrite eqn (4) as For a given degree of reduction and Y, this equation gives In ( Gx/ T,) as a linear function of m, with a zero intercept. The data of fig. 2 agree with this equation. Now we can use the slopes of the lines in fig. 1 and 2 to estimate the relative magnitudes of 1,(1320) and Zp(4000) and ~(1320) and ~(4000). To do this we assume that the chemisorption of methanol at 27 "C influences only the surface, and we let I,( 1320) equal the depth representative of the surface, probably one or two layers of atoms.For this particular reduction state, eqn ( 5 ) becomes In (T,,/T,) = me( 1320) I,( 1320) {c,[Zp( 1320)] - Cox} In ( GX/ T,) = me(4000) Zp(4000) {c,[lp(4000)] - Cox}. ( 6 4 (W The free-electron concentration profiles for this case are shown in fig. 3. The integral values Zp(Cr - Cox) must be the same, since we are assuming no reduction beyond I,( 1320). Thus we obtain ~(1320)/~(4000) = 4.10/0.62 = 6.6, from fig. 2. In fig. 1, the slopes are proportional to E(V) Zp(v) Cox. Since the sample is completely oxidized, we obtain Zp(4000)/Zp( 1320) = 6.6(0.18/0.09) = 13.2. Thus a simple application of the Beer-Lambert law to the results of fig. 1 and 2J. S. Chung and C. 0. Bennett 2161 indicates that radiation at 27 "C, for this case, penetrates ca.13 times as deeply at 4000 as at 1320 cm-l. However, E and Zp may be influenced by the state of the solid and the temperature, so we shall try to eliminate most of their effect by dividing eqn (5) by eqn (2). At the same time, this manipulation gives us a relation between an optically measured quantity and the free-electron concentrations. We obtain where Now we define a normalized absorbance A , as follows: A , then represents the relative increase in free electrons in a reduced sample compared to an oxidized one, in the layer probed by light of frequency v. The correction factor F(v) is calculated from the values of ~ ( v ) and E(V) Zp(v) Cox, which correspond to the intercept and slope in fig.1. E.g. F(4000) = 1.34 and F( 1320) = 1.06 for rn = 8 mg. Based on the preceding experiment, we shall assume in what follows that A , at 1320 cm-l is proportional to the surface degree of reduction, whereas A , at 4000 cm-l is related to the degree of reduction in the deeper layers: [1,(4000)/1,(1320)] = 13.2. For the catalyst after chemisorption of methanol at 27 "C, this leads to A,(1320) = 0.046, and A,(4000) = 0.0034. We also can get the following equations from eqn (7a). ACs = C'JO to I,( 1320)] - Cox - - = AN( 1320) c o x COX c o x cox A r b - cr[Zp( 1320) to Zp(4000)] - cox - A,(1320) 13.2 = (A,(4000)- 13.2 )(m)' Thus eqn (9b), for the layers below the surface, leads to an increased free-electron concentration of zero, as assumed for this sample. In order to relate AC/Cox, (AC = Cr-C,,) to the degree of reduction we use a uniformly reduced sample, analogous to the uniformly oxidized sample of fig.1. For this we use M o , ~ , , , as discussed previously. The result is shown in fig. 4. AC/Cox is not a function of v for this sample, uniformly reduced to the extent of 3.4% with respect to MOO,. A,(tO00) is also the same as (Acb/Cox) in eqn (9b) in this case. With the value of A , or AC/C,, for Mo,02, in fig. 4, we deduce in general 3.4 reduction (surface) = )% 3.4 reduction (bulk) = - ( z:) ( A N (M09026) where (AC/Cox) is found from eqn (9a) and (9b). It is important to notice that our definition of A, results in a quantity which is not sensitive to the variation of scattering with frequency. For the uniformly reduced Mo,0z6 of fig. 4, AN is the same for 4000 and for 1320 crn-l, while of course Rayleigh scattering2162 - - - - - Measurement of OxidationlReduction States of MOO, 0.7 0.6 0.5 n - 0 .4 5 G x c, - 0.3 0.2 1.75 2 .oo 2.25 2.50 1 0 3 K/T Fig. 4. The normalized absorbance and the absorbance of the background of M O , O ~ ~ at various wavenumbers. m = 8.2 mg. 0, @, 4000 cm-l; A, A, 1320 cm-l. is much higher at 4000 than at 1320 cm-l. Thus, for non-uniform particles the variation of A , with v is a measure of oxidation/reduction state as outlined above. Having calibrated the method using the known samples of Moo,, Mo,O,,, and MOO, with surface reduction only, we are ready to test the procedure further. To do this we consider first the adsorption of methanol below 100 "C.At these temperatures there is no bulk reduction and no generation of oxidized products by reaction between adsorbed methanol and lattice oxygen. After adsorption, a switch to helium returns the measured AN(1320) to < 10% of its value in the presence of 3.6% methanol in the gas phase. Simultaneous with the measurements of background transmission, we have measured the absorption by the OH and CH bands for the adsorbed methanol, and some results at 27 "C appear in fig. 5. We now compare the surface reduction measured by transmission at 1320 cm-l and the surface concentrations of methanol deduced from the CH bands. These, in turn, can be calibrated by adsorption studies of methanol via mass ~pectrometry.~ Some results are shown in fig. 6. At 27 "C the amount of methanol adsorbed has been measured by mass spectrometry (referred to as 'volumetric'), and the areas of the CH bands for this condition have also been measured.This absorbance is then set equal to the volumetric result at 27 "C. As the temperature is increased, the i.r. and volumetric data agree well, corresponding to smaller equilibrium coverage by methanol. The value of AN( 1320) has also been set equal to the value from the volumetric measurements at 27 "C. Although we can relate AN( 1320) to the degree of reduction on the surface [eqn (8)-(10)], we do not know how much reduction corresponds to the adsorption of methanol. Thus AN(1320) at 27 "C is also made equal to the volumetric results. This curve falls more rapidly than the others, perhaps because of some chemisorption of oxygen at higher temperatures.It is also possible to monitor the degree of reduction on MOO, by measuring the structural bands of the oxide. There is a structural band at 1140 cm-l, It is a combination band and represents bridged oxygen Mo-0-Mo which has its principal bands atJ . S. Chung and C. 0. Bennett 2163 0 5 10 15 20 mass of Mo03/mg Fig. 5. Effect of the disc mass on the absorption bands of methanol at 27 "C. Absorbance for v(0H) (0) is based on the area of all OH bands. Absorbances for v,(CH) (A) and v,(CH) (0) are based on the line intensities at 2965 and 2854 cm-l, respectively. 2 00 150 * I w 3 2 5 3 100 0, Lc 50 0 30 50 70 90 T/*C Fig. 6. The amount of chemisorbed methanol measured by three different methods at 30-100 "C: a, i.r.; 0, volumetric and 0, A , at 1320 cm-l.Absorbance is based on the combination of the v,(CH) and v,(CH) bands.2164 Measurement of OxidationlReduction States of MOO, 8 6 n E: 3 4 0 W E 2 m / * / ' 0 260 340 0 100 180 TIo C Fig. 7. The reduction states of the surface and subsurface of MOO, after exposure to 3.5% methanol-5% oxygen for 30 min: 0, A , at 4000 cm-l; A, A , at 1320 cm-l; ., Mo=O; 0, Mo-0-Mo. frequencies below those accessible through our ZnS i.r. windows. There is also a sharp band at 997 cm-l, arising from double-bonded oxygen, Mo=O. By using MOO, as a fully oxidized reference, and Mo,O,, as a 3.4% reduced reference, the i.r. data on the structural bands can be interpreted as the degree of reduction. Such data are shown in fig.7, along with percentage reduction at the surface and in the subsurface layers based on A,(1320) and AN(4000) for MOO, exposed to 3.6% CH,OH-5% 0,-He for a range of temperatures. Below 100 "C there is no reduction, as measured by the Mo-0-Mo structural bands. This evidence supports our interpretation of the results of fig. 6 and earlier figures as arising solely from surface perturbations. The curve for AN( 1320) [reduction (surface) % 3 below 100 "C is the same as that in fig. 6. There is some loss of the structural band M=O upon methanol adsorption below 100 "C. This can be explained by the weak adsorption of methanol on top of the protruding double-bonded surface oxygens.2 The hydrogen bonding involved weakens the double-bond nature of the M=O band. After adsorption at 100 "C, a switch to helium returns the intensity loss of the M=O band to 5% of its original value in the presence of methanol.Above 100 "C the loss in both Mo=O and Mo-0-Mo bands increases rapidly. It is seen that the percentage reduction (bulk) data have the same trend as those based on the structural bands. This is reasonable, for the structural bands represent the state of the subsurface layers, not just the outermost layer, as does AN( 1320). We have performed gravimetric and mass spectrometric studies of the global degree of reduction, corres- ponding to fig. 7 (and fig. 8),v4t5 and we found that the global reduction by 3.6% methanol in oxygen, even at 370 "C, is <0.5%. Thus the reductions of 2-10% observed by i.r. must represent the subsurface layers and not the whole bulk.In addition, the X-rayJ. S. Chung and C. 0. Bennett 2165 20 I 0.5 1 5 10 0.051 0.05 0.1 reduction (bulk) (%) Fig. 8. The difference in the concentration of free electrons on the surface and in the subsurface of molybdenum oxide under various reaction conditions. Open symbols (see below) measured after 30 min reaction when A , reach steady states at various temperatures: (1) 150, (2) 200, (3) 250, (4) 300 and (5) 360 "C. 0, measured after 30 min reduction at 200 "C and 15 min reduction at 250 "C; @, 10% H, in He, measured after 5, 10 and 23 min reductions; A, Mo,O,,, measured at three different temperatures. symbol MeOH 0 2 0 0.036 0.94 0 0.036 0.18 n 0.036 0.05 0 0.036 0.00 diffraction patterns of catalysts treated as in fig. 7 and 8 remain those of MOO,.The curve based on percentage reduction (surface) in fig. 7 above 140 "C shows that the surface of MOO, is more oxidized than the subsurface layers. This must arise from the development of a space-charge barrier at the surface caused by an increased concen- tration of ionosorbed oxygen above 100 "C. Above 300 "C, with the gas mixture used, the surface loses its adsorbed oxygen and the more normal situation is observed: the surface layers are more reduced than the subsurface layers. Experiments like those of fig. 7 have been performed with a number of other gas mixtures, and the results are given in fig. 8. In general, for temperatures above 140 "C, the subsurface layers are more reduced than the surface, for oxygen-containing mixtures. As the oxygen content is lowered the opposite profile may be attained at high temperature.For CH,OH-He, with no oxygen, the surface is always more reduced than the subsurface. The same holds true for reduction by hydrogen. Note that in the case that percentage reduction (surface) is higher than percentage reduction (bulk) the solid2166 Measurement of OxidationlReduction States of MOO, does not reach a steady state; it is being continuously reduced, so that the percentage reductions are a function of time. In these experiments, the original catalyst state can be recovered by extended oxidation at 380 "C. However, CH,OH-He at temperature above 350 "C leads to a deep reduction, with a formation of black MOO,. The original state can no longer be recovered.Conclusions The oxidation/reduction state of the surface and bulk of a semiconducting catalyst can be studied by using the frequency effect on the background transmission of i.r. radiation. The experimental results are explained by a simple model in accord with general electromagnetic theory. It is also possible to measure the degree of reduction of MOO, quantitatively after the absorbance is normalized. The normalized absorbances at 1320 and 4000 cm-l can be correlated to the concentration change of free electrons on the surface and in the subsurface layers, respectively. Measurements of the reduction state of MOO, under various reaction conditions shows several features. The reduction of MOO, by methanol alone or by hydrogen gives a continuous reduction, with more reduction on the surface than in the subsurface.When methanol-oxygen mixtures are passed over MOO,, the degree of reduction of MOO, during reaction is determined by the way in which the rate of reduction of MOO, by methanol is balanced by the rate of oxidation by oxygen. The existence of a space-charge barrier caused by ionosorbed oxygen is a characteristic of the catalyst during the reaction. The support of the National Science Foundation (grant CPE 79-01018) is gratefully acknowledged. References 1 N. Pernicone, F. Lazzerin, F. Liberti and G. Lanzavecchia, J. Catal., 1969, 14, 293; J. Edwards, J. Nicolaidis, M. B. Cutlip and C. 0. Bennett, J. Catal., 1977, 50, 24. 2 J. S. Chung, R. Miranda and C. 0. Bennett, J . Chem. Soc., Faraday Trans. I , 1985,81, 19. 3 J. S. Chung and C. 0. Bennett, J. Catal., 1985, 92, 173. 4 J. S. Chung and C. 0. Bennett, Adsorption and Catalysis on Oxide Surfaces, ed. M. Che and G. C. Bond 5 R. Miranda, J. S. Chung and C. 0. Bennett, Proc. 8th Znt. Congr. Catal. (Dechema, Frankfurt am Main, 6 K. Hauffe, Adu. Catal., 1955, 1, 213. 7 A. Bielanski, M. Najhar and J. Zientarski, Bull. Acad. Pol. Sci., 1979, 27, 417. 8 H. M. Naguib and R. Kelly, J. Chem. Solids., 1972, 33, 1751. 9 W. Hertel, J. Catal., 1973, 31, 231. (Elsevier, Amsterdam, 1985), Stud. Surf Sci. Catal., 1985, 21, 185. 1984), vol. 3, p. 347. 10 A. Ueno, J. K. Hochmuth and C. 0. Bennett, J. Catal., 1977, 49, 225. 1 1 J. J. F. Scholton and A. van Montfoort, Proc. 5th Int. Congr. Catal., ed. J. W. Hightower (North- 12 J. R. Reitz, F. J. Milford and R. W. Christy, Foundations of Electromagnetic Theory (Addison-Wesley, 13 M. Sparks and L. Sham, J. Phys. Rev. Lett., 1973, 31, 714. 14 B. Bendow, H. G. Lipson and S. P. Yokon, Appl. Opt., 1977, 16,2909. 15 M. Fromenti, F. Juillet, P. MCriaudeau, S. J. Teichner and P. Vergnon, J. Colloid Interface Sci., 1972, 16 C. 0. Bennett, ACS Symp. Ser., 1982, 178, 1 . Holland, Amsterdam, 1972), vol. 1, p. 385. Reading, Mass., 3rd edn, 1979). 39, 79. Paper 5/ 15 13 ; Received 3rd September, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202155

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Furuduy Trans. I, 1986, 82, 2167-2174 Osmotic Coefficients of Aqueous NaCl and KC1 Solutions Temperature-Concentration Behaviour of the Bahe Lattice Model Juan Luis Gdmez-EstGvez Departament de Termologia, Facultat de Fisica, Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain An expression for the osmotic coefficient, 4, has been obtained in the context of the Bahe lattice model for electrolyte solutions. As a first application, the available osmotic coefficient data for aqueous NaCl and KCl solutions in the temperature range 0-100 "C have been analysed. The results show a good agreement with the behaviour predicted by the model up to molalities as high as 3 . 6 4 mol kg-l for the two salts over the whole temperature range. In studying concentrated electrolyte solutions, lattice theories1-* have gained increasing acceptance. The basic idea of these models is to consider a lattice-like structure for the ions in the solution.The Bahe lattice model5 takes into account (apart from the direct Coulombic interactions between two ions) the interaction of the dielectric gradient near ions with the classical Coulombic field. The final result of this treatment gives the following expression for the mean mole fraction (rational) activity coefficient f + , - as a function of the molarity c : ~ , lnf, = Ac-li3+Bc (12) where A is directly related to Coulomb interactions and can be calculated from fundamental values if the structure is known (f.c.c. or b.c.c. for alkali halides) and B, which is related to dielectric gradient interactions, is unknown at present and remains an adjustable parameter.The comparison of the model with experimental data shows good agreement, covering a high range of c~ncentration.~ In a previous paper6 the temperature effect was studied employing tabulated molal activity coefficient data for aqueous NaCl and KCl solutions in the temperature range 0-100 "C. The agreement is excellent up to molalities as high as 4 mol kg-l for all temperatures. However, there are two problems associated with the treatment of tabulated data. First, the tabulated data are obtained after extrapolation to infinite dilution using a limiting square-root behaviour different from the cube-root behaviour typical for lattice models. Secondly, the influence on the final numerical values due to the kind of mathematical treatment can be of importance.Hence, it is necessary to be able to analyse the original experimental data directly. In this paper, an expression for the osmotic coefficient, #, is obtained in the context of the Bahe lattice model. The experimental data (lowering of the solvent vapour pressure, isopiestic, etc.) can be studied without the problems mentioned above. Expression for the Osmotic Coefficient The osmotic coefficient, #, and the solute mean molal activity coefficient, y + , - are related by7 # = l + m jommdlnyl (2) 21672168 The Bahe Lattice Model or taking in account the relation between y+ - andf,,’ - we have rn (3) where v = v+ +v- for the salt C,+ Av-, Ml is the molar mass of the solvent and rn is the molality.(4) A 3 From eqn (1) we get d lnf, = ---~-~/~dc+Bdc. Moreover, we need a relation between c and rn. The expression employed originally by Harned and Owen6 is used in this work c =porn+Brn2+C‘rn3 ( 5 ) where po is the density of the solvent and B’, C‘ are adjustable parameters which depend on solute and temperature. Eqn (5) is used to obtain dc in terms of drn. After substitution of dc in eqn (4) we can write the integral which appears in eqn (3) as A mporndrn A m2B’rn2drn crn3 drn 6 rnd lnf+ - = -7 I. 7 - 3 Jo c2/3 -A 6-e.i;- +B ~om(po+2Brn+3Crn2)drn. (6) In order to evaluate r 2 i 3 we can use a simplified form of eqn (5) where p is an adjustable parameter. The error in the computation of by means of eqn (7), instead of eqn (5), is ca. 0.1% for KCl and NaCl solutions at the greatest molalities considered in this work.Therefore, we have for c 2 l 3 c zporn+B”’rn2 (7) 00 where (9) The expansion of (1+B’/porn)-2/3 as a power series is permissible because [p/pornsatl < 1, where msat is the molality of the satured solution. Performing the integration in the eqn (7) using eqn (8) for we have for 4, after rearranging terms, the following expression where 00 # = l + x Afrnjf1l3+Bx(rn)- j-0 A d = - - - A ”’’ (aj + 2B’aj-l + 3Caj+); j 2 2 3 j + 4 / 3 p 2 B 3 c 2 3 4 x(m> -= A r n + - r n 2 + - r n 3 .J. L. Gtjmez-Esttvez 2169 Table 1. A$ values for the aqueous solutions of alkali-metal halides at 25 "C (f.c.c. structure)a solute 103 A! 105 ~t lo* A t 109 A$ LiCl LiBr LiI NaF NaCl NaBr NaI KF KC1 KBr KI RbF RbCl RbBr RbI CsF CSCl CsBr CSI 2.700 5 2.852 17 4.872 3 2.372 32 3.31077 4.932 58 1.21861 3.705 10 4.683 61 6.273 3 1.8068 4.58345 5.466 8 6.91589 2.703 56 5.63695 6.387 89 7.83964 - 0.102 0 14 - 2.88426 -2.641 18 -9.26238 0.003 37 2.044 13 -2.961 18 - 9.647 29 0.480 79 - 2.337 77 - 8.665 58 - 16.1095 1.05693 - 7.44233 - 11.9183 - 20.91 8 2 2.36646 - 10.9503 - 17.709 1 - 3 1.578 8 - 20.093 6 - 18.4942 - 0.00 1 7 23.1856 - 28.232 5 - 105.852 - - 124.146 2.837 13 -21.398 1 - 121.29 - 268.5 16 9.247 37 -94.638 - 179.064 - 398.964 30.98 1 2 - 150.549 - 31 8.012 -725.561 - 2.033 82 - 1.91903 17.460 2 0.000009 3.075 23 - 4.206 19 - 23.2269 0.193 128 - 3.42206 - 24.996 9 - 64.943 3 0.933 32 - 17.848 5 - 39.266 8 - 109.971 4.678 82 - 30.409 - 82.725 1 - 238.268 a A$ = 0.166158.Table 2. A$ values for aqueous NaCl solutions at different temperatures (0-100 "C; f.c.c. structure) T/"C A$ 103 ~f 105 ~t 107 ~g 109 A$ 0 5 10 15 20 25 30 40 50 60 70 75 80 90 100 -0.161 754 - 0.162 59 -0.163449 - 0.164 340 - 0.165 23 -0.166158 -0.167 114 -0.169 112 -0.171 242 -0.173513 -0.175938 -0.177235 -0.178 531 -0.181 302 - 0.184 259 1.920 55 2.220 17 2.14384 2.228 14 2.31243 2.37232 2.42261 2.493 47 2.53846 2.563 8 2.499 48 2.53428 2.569 08 2.484 19 2.456 33 6.33292 1.624 78 4.087 15 3.309 13 2.531 11 2.044 13 1.636 5 1.123 15 0.81549 0.621 55 1.958 94 1.25143 0.543929 1.90921 2.12063 5.930 59 1.777 77 4.093 18 3.421 52 2.749 86 2.318 56 1.949 69 1.47871 1.189 52 1 .ooo 47 2.30223 1.605 26 0.908 3 1 2.204 12 2.3723 7.12024 2.243 6 1 5.1304 4.35808 3.585 75 3.075 22 2.628 39 2.048 64 1.683 52 1.436 97 3.123 64 2.21 138 1.299 1 1 2.939 27 3.1 1465 As can be seen from eqn (11)-(13), Af are related directly to A , and hence to the Coulomb interactions. Considering only the first three terms in the series, the error is ca.0.04% for aqueous KCl solution with a f.c.c. structure at msat and 25 "C. The behaviour for different solutes and at other temperatures is similar. The numerical values of Af (0 < j < 4) for aqueous alkali-metal halides at 25 "C and for aqueous solutions of NaCl and KCl in the temperature range 0-100 "C are given in tables 1-3. The values 72 FAR 12170 The Bahe Lattice Model Table 3. A$ values for aqueous KC1 solutions at different temperatures (0-100 "C; f.c.c. structure) T/OC 103 ~f 105 A$ 107 A$ 109 ~a 0 10 20 25 30 40 50 60 70 80 90 100 3.15859 3.438 8 3.622 87 3.705 1 3.751 27 3.859 17 3.894 19 3.905 16 3.75925 3.87886 3.68451 3.772 8 1 5.74672 1.14885 - 1.308 -2.33777 - 2.700 06 - 4.089 62 - 4.333 08 -4.15645 - 3.832 83 -3.741 13 - 2.769 32 -2.3442 8.58893 2.597 08 - 0.726 88 - 2.139 81 - 2.640 72 -4.551 24 - 4.886 75 - 4.637 8 1 - 4.461 98 - 4.039 07 -2.86429 -2.126 12 15.633 7 5.20648 - 0.820 88 - 3.42206 -4.359 11 - 7.879 76 - 8.503 31 - 8.039 55 - 8.166 05 - 6.872 -4.932 92 - 3.342 3 of B', C' and & have been obtained using known density data8? following the procedure described in ref.(6). Eqn (1 1) gives the limiting slope for the osmotic coefficient within the Bahe lattice model. In this regime (low molalities) we have and the term within the brackets disappears.On the other side the terms related with B(terms in m, m2 and m3) are negligible and only the term withj = 0 contributes in the series. As a consequence, the limiting behaviour in m1I3 emerges as The lowest molality at which cube-root behaviour is expected to hold is ca. mol kg-l for 1 : 1 aqueous electrolytes at 298.15 K.l0~ l1 A discussion of the transition from cube-root to square-root behaviour can be found in ref. (2), (4) and (1 1). Results and Discussion Aqueous Solutions of NaCl and KCl(0-lo0 "C) In order to apply eqn (lo), we must remember that B is unknown. Following a similar procedure to the treatment of activity coefficient data, we rewrite eqn (10) in the form Dd = Bx(m) (17) where Dd is given by Using #exp, Dd can be computed for each molality.Next, Dd can be plotted against x(m), and should give a straight line of slope B. Such plots for KC1 and NaCl at several temperatures are shown in fig. 1-5. In order to separate clearly the results for different temperatures and references, the vertical scale has been shifted by plotting D, + 0 3 (n = 0, 1,2, . . .). Straight lines are observed up to molalities as high as 4 mol kg-l for NaCl and 3-3.6 mol kg-l for KCl. The slopes, determined by linear regression analysis for theJ. L. Ghez-Estivez 2171 0.4 0.3 ? c! c! tt: I=I" 0.2 0.1 0 0 0.4 0.8 1.2 1.6 x(m)/mol dmT3 Fig. 1. Plot of D, for NaCl at 0 "C. @, Experimental points. Lines are linear least-squares 1.0 0.8 0.6 ? L! L! ru W 13" 0.4 0.2 0 0.4 0.8 1.2 1.6 2.0 2.4 x(rn)/mol dm-3 Fig.2. Plot of D, for NaCl at 25 "C. a, Experimental points. Lines are linear least-squares fits: 1, ref. (12); 2, ref. (16); 3, ref. (17); 4, ref. (13); 5, ref. (13) (smoothed values); 6, ref. (18); 7, ref. (15); 8, ref. (19). 72-221 72 The Bahe Lattice Model 0.7 0.6 0.5 0.3 Q i 0.2 0.1 0 0 0.4 0.8 1.2 1.6 2.0 2.4 x(m)/mol dm-3 Fig. 3. Plot of D, for NaCl at 40, 50 and 60 "C. 0, Experimental points. Lines are linear least-squares fits. 0.7 0.6 0.5 ? c! 0.4 c! r=, '" 0.3 0.2 0.1 0 Fig. 4. Plot of D, for NaCl at 70, 75, 80 and 90 "C. 0, Experimental points. Lines are linear least-squares fits.J. L. Gdmez-Esttvez 2173 0.7 0.6 0.5 h 4 0.4 c! % Q i 0.3 0.2 0.1 0 0 0.4 0.8 1.2 1.6 2.0 x (m)/mol dm-3 Fig. 5. Plot of D, for KC1 in the temperature range 0-100 "C.0, Experimental points. Lines are linear least-squares fits. linear portions of these curves, are shown (averaged values at each temperature) in table 4. For the two salts the correlation coefficients of the linear fit are >0.999 and the standard deviations of the fits, 0, are ca. Also, we have that B(NaC1) > B(KC1) and (aB/aT) > 0 in the complete domain of temperatures. As a test, a comparison between B values for NaCl and KCl at 25 "C obtained using the tabulated yk values of Hamer and Wu31 with the corresponding B values derived from osmotic coefficient data has been made. The results are B(HW) = 0.2389 +0.0007 for NaCl and B(HW) = 0.1598+0.0003 for KCl. These results agree, within experi- mental error, with the corresponding B4 values which are, respectively: 0.237 f 0.004 for NaCl and 0.157 & 0.003 for KCl.The straight lines obtained over a broad concentration range for NaCl and KC1 at all temperatures (fig. 1-5) indicate that the behaviour of these salts can be interpreted on the basis of a model of a loose f.c.c. structure in solution. The importance of the field-dielectric gradient interactions reflected on B, is clear. However, the agreement between the model and the experimental behaviour exceeds the expected domain of validity. In the development of the model5 the change of the solution's dielectric constant with molality, the overlap of the cospheres of the ions and the disturbances from a perfect lattice are not taken in account. It is surprising that the behaviour predicted by a 'rigid' lattice model is not appreciably affected by increasing the temperature.The linearity is good even at temperatures near 100 "C. Further studies on these aspects are necessary employing more sensitive thermodynamic data as, for example, enthalpies of dilution and heat capacities. These results will be presented in the near future.2174 The Bahe Lattice Model Table 4. B values for aqueous NaCl and KCl solutions from 0 to 100 "C T/"C B(NaC1) ref. B(KC1) ref. 0 15 20 25 30 40 50 60 70 75 80 90 1 00 0.1953 0.2285 0.23 16 0.2365 0.2434 0.2535 0.2541 0.2601 0.2655 0.2574 0.2687 0.2702 0.2753 12-14 12 15 12, 13, 15-19 15 12 13 20 20 13 20 20 13, 20, 21 0.1 173 0.1536 0.1566 0.1677 0.1832 0.1813 0.1917 0.1944 0.1969 - - - - 14 22 17-19, 23, 24 25 26 27, 28 26 29 30 - - - - References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 H.P. Bennetto, Annu. Rep. Progr. Chem., Sect. A, 1973, 70, 223. R. M. Pytkowicz and K. S. Johnson in Activity Coefficients in Electrolyte Solutions, ed. R. M. Pytkowicz (CRC Press, Boca Raton, Florida, 1979), pp. 209-264. I. Ruff, G. Palinkas and K. Gombos, J. Chem. SOC., Faraday Trans. 2, 1981, 77, 1189. G. W. Murphy, J. Chem. SOC., Faraday Trans. 2, 1982, 78, 881. L. W. Bahe, J. Phys. Chem., 1972,76, 1062; L. W. Bahe and D. Parker, J. Am. Chem. SOC., 1975,97, 5664. J. L. Gomez Estevez and V. Torra, An. Quim., 1983, 79A, 49. B. R. Staples and R. L. Nuttall, J. Phys. Chem. Ref. Data, 1977, 6, 385. Znternational Critical Tables, ed. E. W. Washbrun (McGraw-Hill, New York, 1928), vol.111. J. L. Fortier, P. A. Leduc and J. E. Desnoyers, J. Solution Chem., 1974, 3, 323. H. S. Frank and P. T. Thompson, in The Structure of Electrolyte Solutions, ed. W. T. Hamer (Wiley, New York, 1959), chap. 8. J. C. Rasaiah, in The Physical Chemistry of Aqueous Systems, ed. R. L. Ray (Plenum Press, New York, Q. D. Craft and W. A. Van Hook, J. Solution Chem., 1975, 4, 923. H. F. Gibbard Jr., G. Scatchard, R. A. Rousseau and J. Creek, J. Chem. Eng. Data, 1974, 19, 281. R. F. Platford, J. Chem. Eng. Data, 1973, 18, 215. P. Olynyk and A. R. Gordon, J. Am. Chem. SOC., 1943,65,224. C. N. Pepela and P. J. Dunlop, J. Chem. Thermodyn., 1972, 4, 255. J. N. Pearce and A. F. Nelson, J. Am. Chem. SOC., 1932,54, 3544. G. Scatchard, W. J. Hamer and S. E. Wood, J. Am. Chem. Soc., 1938,60, 3061. M. C. Petit, J. Chim. Phys., 1965, 62, 11 19. R. P. Smith, J. Am. Chem. SOC., 1939,61, 500; R. P. Smith and D. S. Hirtle, J. Am. Chem. SOC., 1939, 61, 1123. N. Miljevic, G. Dessauges and W. A. Van Hook, J. Solution Chem., 1981, 10, 29. B. F. Lovelace, L. C. W. Frazer and V. B. Lease, J. Am. Chem. SOC., 1921,43, 102. Z. Shibata and K. Niwa, 2. Phys. Chem., 1935, A173, 415. 0. L. I. Brown and C. M. Delanney, J. Phys. Chem., 1954,58, 255. M. A. Smith, R. L. Combs and J. M. Googin, J. Phys. Chem., 1954, 58, 997. T. M. Herrington and R. J. Jackson, J. Chem. SOC., Faraday Trans. I , 1973, 69, 1635. R. A. Robinson, Trans. Faraday SOC., 1939, 35, 1222. W. T. Humphries, C. F. Kohrt and C. S. Patterson, J. Chem. Eng. Data, 1968, 13, 327. J. T. Moore, W. T. Humphries and C. S. Patterson, J. Chem. Eng. Data, 1972, 17, 180. C. S. Patterson, L. 0. Gilpatrick and B. A. Soldano, J. Chem. Soc., 1960, 2730. W. J. Hamer and Y.-C. Wu, J. Phys. Chem. Ref. Data, 1972, 1, 1047. 1973), pp. 203-236. Paper 5/1546; Received 9th September, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202167

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1986,82, 2175-2184 Solution of Hydrogen in Thin Palladium Films Shozo Kishimoto," Minoru Inoue and Noritetsu Yoshida Department of Chemistry, Faculty of Science, Kobe University, Nada-ku, Kobe 657, Japan Ted B. Flanagan Department of Chemistry, University of Vermont, Burlington, Vermont 05405, U.S.A. Solubility enhancements have been observed for hydrogen in thin Pd films (250-820 A) in the aphase. Thermodynamic data for the films are determined from isotherms at temperatures between 25 and -40 "C and are compared with those for bulk Pd. From the values of WH-H (interaction energy of hydrogen in the film) and the solvus behaviour, it is suggested that the hydride phase in the films is destabilized relative to the dilute phase as the film thickness decreases.The dependence of the solubility on annealing has been examined for a thin film. The results indicate that the special characteristics of thin films disappear after annealing at ca. 400 "C. The absorption of hydrogen by Pd has been studied by many workers using samples with different physical forms (wires, foils, powders etc.) and different histories. These studies have revealed that the differences in sorptive properties of the various samples can be significant. For the case of a sample having a large surface-to-volume ratio, a shift of the hydrogen pressure - concentration curve, P us. n (n = H : Pd atom ratio), to higher n values is observed in the a phase owing to chemisorption of hydrogen on the surface of the sample.l, Fig. 1 shows schematically a typical isotherm for this case.The steep branch (I) is known to correspond to solution of hydrogen into the Pd lattice, forming the dilute a phase. Since chemisorption is nearly complete at very low hydrogen pressures n Fig. 1. Schematic isotherm for a sample with a large surface to volume ratio. 21752176 Solution of H, in Thin Pd Films Fig. 2. Isotherms at 0 "C for films (pretreatment at 200 "C) of different thicknesses (A): a, 250; A, 380; 0, 510; 0, 820. where hydrogen solution in the bulk is negligible, branch (11) is attributed mainly to chemisorption. Besides chemisorption, there are other possible reasons for the occurrence of dilute-phase hydrogen solubilities which exceed those formed in bulk well annealed Pd. Flanagan et al. have shown that hydrogen solubility in the a phase is enhanced by cold-working3 or by the a-B phase ~ h a n g e .~ They concluded that the source of this enhanced solubility is the accumulation of hydrogen in the strain fields around dislocations. Furthermore, in the case of heavily cold-worked samples, initial deviations from linear plots of P112 us. n occurred leading to a non-zero intercept on the n axis at P112 = 0. This intercept was suggested to be due to some specially energetic absorption since the surface was too small to account for such an effect. Similar phenomena were also observed by Kirchheim5 for cold-worked Pd samples. In spite of the widespread interest in thin films, there have been few attempts to determine the detailed thermodynamic behaviour of hydrogen in thin films of Pd.It is well known that physical and chemical properties of thin films or small particles differ markedly from those of bulk materials. In the earliest investigation, Suhrmann et a1.6 studied the sorption of hydrogen in Pd films below - 78.5 "C. The anomalously large hydrogen content (n = 1 .8) observed by these workers has not been confirmed by other workers. Frazier and Glosser' studied isotherms of Pd films at 27 "C and found that the P us. n characteristics are highly dependent on film thickness (< 1000 A) and the P-n relationships approach bulk behaviour as the thickness increases. Feenstra et aZ.$ studied films in the range 1000-4000 A and found that the two-phase region of the isotherms is considerably narrower than that for bulk Pd.None of these earlier investigations determined detailed solubility data in the dilute phase. The purpose of this study is to investigate the effects of film thickness on the thermodynamic properties of the Pd-hydrogen system in the dilute phase and the effects of annealing. The dilute phase is of special interest because in this region the effects of lattice defects are most pronounced.S. Kishimoto et al. 2177 2 1 0 0.05 0.1 0 0.1 5 n Fig. 3. Isotherms for a film of 250 8, thickness at different temperatures ("C). Experimental The solubility of hydrogen in films was determined with a conventional u.h.v. apparatus (base pressure, 1 OP9 Torr)i. using a volumetric technique. Films of different thicknesses (200-800 A) were prepared by evaporation of Pd wire (0.2 mm in diameter, 99.97% purity) onto glass surfaces at room temperature.The films were stabilized by heating in vacuo at 200 "C for 1 h. The films were not heated above 200 "C except in the annealing experiments described below. For the determination of solubility data in the a phase the dosing pressures of hydrogen were always less than the plateau pressures. Solubility data were determined at temperatures between 25 and - 40 "C. The equilibrium pressures were measured with an M.K.S. diaphragm gauge. Sorption of hydrogen in the a phase was very fast for all the films used in the present experiments. Hydrogen and helium gases were obtained from Takachiho Co. and were used without further purification. After the experiments were completed, the films were dissolved with aqua regia and the amount of Pd in each film was determined by atomic absorption analysis.Results and Discussion Fig. 2 shows absorption isotherms at 0 "C for films of different thicknesses. It can be seen that the slope of the a+P coexistence region increases and that G,, (maximum solubility of hydrogen in the a phase) shifts to higher values of n with decreasing film thickness. The plateau pressures increase with a decrease in film thickness and the plateau pressures for the thick films were found to be approximately the same as those for bulk Pd. These results are in good agreement with the observations by Frazier and Glosser.' Fig. 3 and 4 show isotherms in the low hydrogen content region for films of 250 and 800 A thickness at different temperatures.It is clear that the isotherms in the very low n region approximately obey the e q ~ a t i o n , ~ n = KO+ K, I'll2, where KO is most likely due t 1 Torr x 133.3 Pa.2178 Solution of H, in Thin Pd Films n 0.05 0.1 0 Fig. 4. Isotherms for a film of 820 8, thickness at different temperatures (“C). l r 4.0 4.5 103 KIT 3.5 Fig. 5. Plots of lnK, us. T1. Film thickness/A: 0, 250; a, 380; 0, 820. to chemisorption and K, is Sieverts’ constant, which is given by K, = exp (- ACH/RT), where AG& is the relative chemical potential at infinite dilution. Frazier and Glosser7 have presented arguments to support the view that KO reflects chemisorption. Some plots of lnK, us. T1 are given in fig. 5. From the slopes and intercepts of these plots we can obtain ATH and ATH, which are summarized in table 1 together with the values for bulk Pd.It can be seen that the relative partial molar enthalpies are more exothermic (ca. 2.1-2.9 kJ mol-l H) than that of bulk Pd and these tend to be moreS. Kishimoto et al. 2179 Table 1. Thermodynamic parameters for Pd films - AG; thickness ATH AFH /kJ mol-1 H /8, /kJ mol-1 H /J K-l mol-1 H (273 K) 250 - 12.9 -49.8 0.75 380 - 12.5 - 50.4 1.25 510 - 12.2 - 50.1 1.48 820 - 12.1 - 50.9 1.82 bulka - 10.0 - 55.0 4.50 1 3.0 z r( 3 I E 2 1 iy I 12.5 12.0 a Taken from data given in ref. (9) and (10). I I I I I I 200 400 600 800 film thickness/A Fig. 6. Dependence of ACH((B), AFH(H(0) and TATH(.) on film thickness at 0 "C. exothermic with decreasing film thickness. It also appears that the entropies are more positive for the films (5 J K-l mol-1 H) than for bulk Pd, but these values are indepen- dent of film thickness.Fig. 6 shows the values of APH, ATH and TAFH as a function of film thickness at 0 "C. This figure suggests that the contribution of entropy to the solubility enhancements in the range of the present thicknesses is insignificant. Fig. 7 shows a comparison of the solubilities of hydrogen in Pd films with well annealed bulk Pd and with two deformed bulk Pd samples at 0 "C. The data for the films have been obtained by subtracting the chemisorption contributions, KO, from each value of n. Solubility enhancements for the films can be expressed as n'/no at a given equilibrium pressure, where no is the solubility for well annealed bulk Pd and n' is the corrected value for the film.These values at P112 = 0.2 (Torr)l12 are: 3.0, 3.7 and 4.8 for films of 820, 510 and 250 A thickness, respectively. It is also clear from fig. 7 that these are larger than for the deformed samples and that n'/n, increases with a decrease in film thickness. It is clear from fig. 3 and 4 that the increased solubilities in the thin films continue to2180 Solution of H, in Thin Pd Films 0.4 0.3 k \ “ZL 0.2 0.1 0 1 2 3 4 5 (n or n‘) x 103 Fig. 7. Isotherms for films and bulk samples at 0 “C. Bulk samples: (a) well annealed sarnple;I1 (b) cold-worked samp1e;ll (c) or-j? cycled sample after undergoing cold-work (Flanagan and Kishimoto, unpublished data). Isotherms for films have their thicknesses (A) shown. -601 I I I 3.0 4.0 5.0 103 KIT Fig.8. Solvus plots for films and a well annealed bulk sample.12 0, Bulk Pd; a, 820 8, film; 0, 250 8, film.S. Kishimoto et al. 2181 the a+ phase boundaries, i.e. it is not a phenomenon due to a limited number of trapping sites. This means that the entire film somehow differs from bulk Pd because solubility in the a phase is a bulk phenomenon. The increased solubility in the dilute phase is due to energetic factors (table 1) and presumably arises from the metastable nature of the film as compared to the bulk. In fig. 8 are shown solvus values for the thin films as compared to bulk Pd,12 where the solvus value, %ax, is the terminal solubility of hydrogen in the dilute phase. Because of the experimental scatter it is not possible to evaluate meaningful solvus enthalpies and entropies for the films but, nonetheless, it can be seen that there is a decrease in the solvus free energy (AGsolv), - RT In (amax/ 1 - amax), as the film thickness decreases, i.e.the hydride phase is destabilized relative to the dilute phase as the film thickness decreases. The solvus free energy represents the free-energy change for the transfer of one mole of hydrogen from the hydride phase to the dilute phase without the partial configurational entropy contribution of the dilute phase. From data shown in fig. 3 it is possible to calculate the parameters for hydride formation for an unannealed thin film (250 A). We obtain AHa+B = - 18.9 kJ mol-1 H and ASa + p = -49.2 J K-l mol-l H, where the determination at 0 "C has been omitted from the calculation of these parameters.These values agree quite well with other measurements of these values for bulk Pd.lo Hence the stability of the hydride phase in the thin film is comparable to that found in bulk Pd, whereas the dilute phase is more stable in the thin film. The thermodynamic parameters for hydride formation and for hydrogen solution in the dilute phase (table 1) are consistent with the solvus behaviour. It has been shown12 that Using the values of these parameters from the present work we find AGSo1,(250 A) x 5.0 kJ mol-1 H as compared to the experimental value of 5.7 kJ mol-1 H (fig. 8). The fact that the experimental and calculated values are in good agreement indicates that the measured values of AIPH and ATH for the thin film are fully compatible with the location of the phase boundaries.This would not be the case if the number of deep trapping sites were small compared to the phase-boundary composition %ax- Using the relationship9 RT In [P1j2( 1 - n')/n'] = AGE = APH - TAPH + WH--H n' where WH-H is the partial interaction energy of hydrogen, values of WH-H can be obtained from plots of ACE us. n'. The values are somewhat scattered; however, it is apparent that the absolute values of WH-H (ca. 6-12 kJ mol-1 H) are much smaller than for well annealed bulk Pd (ca. 50-55 kJ mol-1 H).99 lo Bakker et aZ.13 have found a decrease in the critical temperature for thin films as compared to bulk Pd and this is consistent with the results shown in fig. 9. They attribute the decrease in the value of WH-H in the thin films to the effects of the boundary conditions .as discussed by Alefeld.l* Their films were thicker than those used here so that it is not possible to make a quantitative comparison of WH-H.It can be noted that the values of WH-H for cold-worked Pd are slightly smaller in magnitude than for the well annealed ample.^ Thermodynamic parameters calculated from the extrapolation of the plots of AGE us. n' to n' = 0 give results which are in agreement with the values in table 1 within experimental error. The dependence of the solubility on annealing was examined for a thin film (250 A). The anneals were carried out at different temperatures (100400 "C) for 1 h in uacuo. In fig. 9 we show the effects of annealing on the function ACE at 0 "C. Fig. 10 shows the effects of annealing on values of WH-H and r~'(Pl/~ = 1.0 Torr1/2) as a function of2182 Solution of H, in Thin Pd Films m W W m m W 8-0- -0- A I h A - 6- 0 0.01 0.02 0.03 0.04 0.05 n' Fig.9. Plots of AGE us. n' at 0 "C; annealing temperatures/OC: A, 100; a, 200; @, 250; 0, 300; a, 350; 0,400. Film thickness 250 A. 0.0 5 n -k a03 9 c1 II W : 0.02 E bulk - 0.01 200 300 400 annealing ternperaturel'c 50 40 30 5 A z 3 I c. z 20 4 10 0 Fig. 10. Dependence of solubility, n'( a), and WH-H( 0) on annealing temperature.S. Kishimoto et al. 2183 0 a2 0.4 0.6 Fig. 11. Effect of annealing temperature on isotherms at 0 "C for a film of 250 8, thickness: 0, 200 "C; m, 400 "C. annealing temperature. The values of WH-H have been obtained from the slopes in fig.9. It can be seen that bulk behaviour is approached after annealing at ca. 400 "C. Fig. 11 shows the effect of annealing on the a-p coexistence region at 0 "C for the 250 A film. These results suggest that the special characteristics of thin films disappear after annealing at elevated temperatures. Frazier and Glosser' concluded that the thickness dependence of the P us. n curves can be explained by both substrate surface-film interaction and by crystallite size effects. It is well known that unannealed thin films contain a large number of lattice defects (vacancies, dislocations, stacking faults, etc.) than bulk material and these are annihilated by annealing at elevated temperatures. As stated in the introduction, the role of dislocations on solubility enhancements in Pd has already been elucidated by Flanagan and coworkers3~ and Kir~hheirn.~ It does not seem likely that only hydrogen dislocation interactions can explain the present results because the magnitude of the solubility enhancements observed in the thin films are too large to be explained solely by dislocation-enhanced solubilities.It is interesting that after annealing, the hydrogen solution behaviour of the thin films approaches that of the bulk material (e.g. fig. 9-1 1). This suggests that the thickness of the films per se does not play a major role in their hydrogen solubility characteristics. Provided that the annealing does not lessen the anchoring of the films to the glass surface, the annealing results appear to rule out the influence of the boundary ~0nditions.l~ Anchoring can, in principle, decrease I WH-H_L by-not allowing the film to assume its normal, bulk value of VE, where I WH-H( cc BPH/VPd, where vH and Fpd are the partial molar volumes of hydrogen and Pd, respectively, and B is the bulk modulus.It can be seen from fig. 1 1 that although the shapes of the isotherms are quite different for the 200 and 400°C annealed films (250 A), their plateau pressures for hydride formation are comparable. This suggests that the effective H-H interaction, WHWH, leading to hydride formation must also be similar, therefore the measured values of2184 Solution of H, in Thin Pd Films WH--H in the dilute phase must differ from the values which obtain for hydride formation. The behaviour of the thin, unannealed films therefore appears to be contradictory.In the dilute phase the measured thermodynamic parameters indicate the structure to be uniformly perturbed, whereas the hydride formation behaviour indicates an unperturbed structure. It may be that stresses are present in the thin films which are relieved in the hydride phase, but restored when the hydride is decomposed. T.B.F. thanks the N.S.F. for financial support of his research on metal hydrides. References 1 P. C. Aben, J. Catal., 1968, 10, 224. 2 P. A. Sermon, J. Catal., 1972, 24, 460. 3 T. B. Flanagan, J. F. Lynch, J. D. Clewley, B. V. Turkovich, J. Less-Common Met., 1976, 49, 13. 4 J. F. Lynch, J. D. Clewley, T. Curran and T. B. Flanagan, J. Less-Common Met., 1977,55, 153. 5 R. Kirchheim, Acta Met., 1981, 29, 835. 6 R. Suhrmann, G. Schumicki and G. Wedler, Z. Naturforsch., Teil A , 1964, 19, 1208. 7 G. A. Frazier and G. Glosser, J. Less-Common Met., 1980, 74, 89. 8 R. Feenstra, G. J. de Bruin-Hordijk, H. L. M. Bakker, R. Griessen and D. G. de Groot, J. Phys. F, 9 T. B. Flanagan and W. A. Oates, Adv. Chem. Ser., 1978, 167, 283. 1983, 13, L13. 10 E. Wicke and G. H. Nernst, Ber. Bunsenges. Phys. Chem., 1964, 68, 224. 1 1 T. B. Flanagan and S. Kishimoto, in Electronic Structure and Properties of Hydrogen in Metals, ed. P. Jena and C. B. Satterthweile (Plenum Press, New York, 1983), p. 623. 12 S. Kishimoto, W. A. Oates and T. B. Flanagan, J. Less-Common Met., 1982, 88, 459. 13 H. L. M. Bakker, G. J. de Bruin-Hordijk, R. Feenstra, R. Griessen, D. G. de Groot and J. Rector, in Electronic Structure and Properties of Hydrogen in Metals, ed. P. Jena and C. B. Satterthweile (Plenum Press, New York, 1983), p. 659. 14 G. Alefeld, Ber. Bunsenges. Phys. Chem., 1972, 76, 746. Paper 5/1555; Received 10th September, 1985

ISSN:0300-9599    

DOI:10.1039/F19868202175

出版商:RSC    

年代:1986

数据来源: RSC