年代:1977 |
|
|
Volume 73 issue 1
|
|
21. |
Thermodynamics of solution of two forms ofDL-α-amino-n-butyric acid in water |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 181-185
Michael H. Abraham,
Preview
|
PDF (424KB)
|
|
摘要:
Thermodynamics of Solution of two forms of DL-a-Amino-n-butyric Acid in WaterBY MICHAEL HI. ABMHAM,* ERIC AH-SING, ROBERT E. MARKS AND RONALD A. SCHULZDepartment of Chemistry, University of Surrey, Guildford, SurreyANDBIUAN C. STACEDepartment of Chemical Physics, University of Surrey, Guildford, SurreyReceiced 18th August, 1976Two crystalline polymorphic forms of DL-a-amino-n-butyric acid have been obtained and havebeen shown to be identical to the A-form and B-form previously described by Iitaka and coworkers.The two forms differ in heat of solution and in their solubility in water. We find that for the A-format 298 K, AH: = 374 cal mol-l, AG; = 1772 cal mol-I and AS; = -4.7 cal K-' mol-' and for theB-form AH," = 1653 cal mol-l, AG; = 1883 cal mol-' and AS," = -0.8 cal K-l mol-l.Althoughthe B-form is the stable form at 298 K, it is predicted from the above measurements that the A-formshould be the stable form above - 326 K.Themlodynamic quantities for the solution of aminoacids, and for their transferfrom water to other solvents, have been the subject of numerous studies.1-6 In acontinuation of our own work7 on the water+ethanol solvent system, we haveinvestigated the solution of DL-a-amino-n-butyric acid in water, and report hereresults of this study.We observed that slow recrystallisation of the aminoacid from water or aqueousethanol yielded needle-like crystals, whereas recrystallisation from hot water or fromaqueous acetone resulted in plate-like crystals. The influence of the solvent seemsat least partially to be that of determining the temperature at which the initially hotsolutions crystallise.Usually, crystallisation from water or aqueous ethanol tookplace only when the solution was cold, whereas from aqueous acetone the crystalswere invariably deposited whilst the solution was still hot. The two crystalline formseach had the correct microanalysis for the aminoacid, and identical n.m.r. spectra(in D20). The forms are interconvertible by crystallisation, as above, and are clearlypolymorphic forms of DL-a-amino-n-butyric acid. It was recognised in 1951 thatcertain aminoacids (including DL-a-amino-n-butyric acid) were capable of existingin polymorphic forms,8 and a detailed study by Iitaka and coworkers 9* l o has revealedthe existence of a tetragonal form obtained as fibrous needles or square pyramidalcrystals from aqueous ethanol and a monoclinic form crystallising in plates fromwarm water.The tetragonal needles (designated the B form) are of space groupP4Jn and the monoclinic plates (designated the A-form) are of group P21/a.10*Iitaka recorded the infrared (i.r.) spzctra of the A- and B-forms. The i.r. spectra ofour own needle-like crystals and plate-like crystals are identical with the spectra givenby Iitaka and show that our two forms are the B- and A-polymorphic forms ofaIso discovered a low temperature form (the C-form) of space group 12/a, but this isstable only below 200 K.* lltaka18182 THERMODYNAMICS OF AMINOBUTYRIC ACID I N WATERDL-a-amino-n-butyric acid respectively. We have dealt with this in some detail,because previous workers 5 * 12* l3 h ave reported thermodynamic properties forsolution of DL-a-amino-n-butyric acid in water, apparently unaware of the existenceof two distinct crystalline forms.We, therefore, studied the solution process for thetwo separate A- and B-forms.Heats of solution of DL-a-amino-n-butyric acid in water have been determinedcalorimetrically by three previous sets of workers, (table 1) but there is considerableTABLE 1 .-CALORIMETRICALLY DETERMINED HEATS OF SOLUTION OFDL-a-AMINO-n-BUTYRIC ACID IN WATER AT 298 KA ~ * / C ~ I mo1-11528" ref. (12)1350+30 ref. (5)1587+7 ref. (13)374f3 A-form, this work1653+ 18 B-form, this worka One determination only.12disagreement between their various values.We find that the heat of solution of theplatelike A-form is only 374 cal mol-l whereas the heat of solution of the needle-likeB-form is 1653 cal mol-I. However, we found it very difficult to obtain the pureB-form uncontaminated with A-form, and other preparations of the B-form that (asjudged from their i.r. spectra) were slightly contaminated had AHo values of 1601and 1422 cal mol-l. It is known lo that the B-form is nearly always associated withquantities of the A-form, and it is, therefore, probable that the variable AH; valuesobtained by previous workers 9 2, result from use of specimens that were mixturesof the two forms." As far as we can tell from inspection of the i.r. spectra, our samplewith AH; of 1653 cal mol-1 is the pure B-form, although even here we cannot ruleout the possibility of trace contamination.The A-form seems rather easier to obtainpure, and various preparations yielded essentially the same value of AH:.The crystalline form in equilibrium with water at 298 K is the B-form, but no matterwhether the starting material is the A- or B-form, the same equilibrium solubilityshould be reached. Our value of the mol fraction solubility, Ns, of the B-form(0.0351) is in reasonable agreement with previously determined values (table2).12* 15* l 6 Interestingly, we observed that the solubility of the A-form at 298 Kappeared to be higher than that of the B-form.? If the A-form is vigorously shakenTABLE 2.-sOLUBiLITY OF DL-Q-AMINO-ll-BUTYRIC ACiD IN WATER AT 298 KNB/mol fraction0.0363 ref.(15)0.0387 ref. (16)0.0349 ref. (12)0.0410 A-form, this work0.0351 B-form, this work* Crystallisation of DL-a-amino-n-butyric acid from water or water + ethanol (the usual solventsfor recrystallisation) can yield the A-form, the B-form or mixtures of the two, depending on the rateof crystallisation from the saturated solution and on the temperature at which crystallisation takesplace. High temperatures and rapid crystallisation favour formation of the A-form. It is fortuitousthat the three previous sets of xorkers '9 l2> l3 in calorimetry all apparently obtained mainly theneedle-like B-form, since it is the plate-like A-form that is usually regarded as the common form ofthe amin~acid.~~ l4that different samples of aminoacids sometimes lead to slightlydifferent solubilities in water, but no specific reference to DL-a-amino-n-butyric acid has been made.t It has been noted before 4ABRAHAM, AH-SING, MARKS, SCHULZ AND STACE 183with water, it dissolves to a concentration of 0.0410 mol fraction ; subsequently thesolution precipitates out the B-form and reverts to a mol fraction solubility of 0.0351,the equilibrium value.This phenomenon could merely be due to supersaturation,but we find the same effect in aqueous ethanol where the ratio NfJN; is the same as itis in water. We do not think that the constant ratio can be due to exactly the samedegree of supersaturation in each case, and hence regard the value of N i as the actualsolubility of the A-form.The secondary medium molal activity coefficient of DL-U-amino-n-butyric acid in water at 298 K is known up to almost the saturated solution,and we have used the expression for ym given by Ellerton et aZ.18 to calculate ym*s theactivity coefficient in the saturated solution. Conversion to y N p S yields the values intable 3 for yz*s and y : p s . Then since AGg = - RTln (Ns..yNyS) where AG; is the standardfree energy of solution, we can calculate AG; and hence AS," (table 3).TABLE 3 .-THERMODYNAMICS OF SOLUTION OF DL-a-AMINO-n-BUTYRICACID IN WATER AT 298 KaA-form E-formAH," /cal mol-I 374 1653NS 0.0410 0.035 1yN,S 1 . 2 2 6 1 . 1 8 7AG,O/caI mol-l 1 7 7 2 1 8 8 3AS,"/cal K-' mol-I - 4.69 - 0.77uThis work, except for the mol fraction activity coefficients, y." All processes refer to thehypothetical solution of unit mol fraction solute and unit activity.From the data in table 3 we can also deduce the standard thermodynamic para-meters for process (l), at 298 K.DL-a-amino-n-butyric acid (B-form, cryst) +DL-a-amino-n-butyric acid (A-form, cryst.).(1)We find that AGT = 11 1 cal mol-l, AH," = 1279 cal mol-l, and AS; = 3.9 cal K-'mol-1 at 298 K. The increased entropy of the A-form is compatible with the knowncrystal structures. In the A-form the y-carbon atom can occupy three possible posi-tions with respect to the nitrogen atom (trans, gauche I, and gauche 11) whereas in theB-form the y-carbon atom is restricted to the trans position only.1°TABLE 4.-sOLUBILITY (MOL-FRACTION) OF THE A- AND B-FORMS OFDL-a-AMINO-n-BUTRYIC ACID IN WATERT / K A-form B-form298 0 .0 4 1 0 0 . 0 3 5 13 0 8 0 . 0 4 1 2 0 . 0 3 9 03 1 8 0 . 0 4 3 4 0 . 0 4 3 0If AH; and AS; are constant over a rather small temperature range, it can becalculated that AGT = 0 when the temperature is 326K (i.e., only 53°C). Fromsolubility measurements at different temperatures (table 4) it may be calculated thatthe solubility in water of the A- and B-forms becomes identical at 319 K, i.e., thatAGF = 0 at this temperature, in fair agreement with the value of 326 K found above.Iitaka and coworker^,^ however, found that the crystalline B-form was converted intothe A-form on heating only at a temperature as high as 468 K. They suggested thatthe A-form was metastable below 468 K and stable above 468 K.This is in markedcontrast to our results, which indicate that the A-form is stable above - 320 K. W184 THERMODYNAMICS OF AMINOBUTYRIC ACID I N WATERsuggest that the experiments of Iitaka and coworkers on the conversion of onecrystalline form to the other reflect a kinetic stability, and hence are not directlycomparable to our results on the thermodynamic stability.Finally, we consider the implications of our results for the determination ofthermodynamic parameters for the transfer of DL-a-amino-n-butyric acid from waterto other solvents. Values of AGF; are usually obtained from AG: values in water andthe other solvents. Since the equilibrium value of Ns at 298 K is the same whetherthe A-form or B-form is used in solubility experiments, there should be no difficultyover calculations of AGF values, at least for the water+ethanol system.However,values of AH:, as we have shown, depend on the composition of the crystallinespecimen used. Now AH: for the pure A-form is expected to be identical to AH: forthe pure B-form, so that, again, correct values can be obtained for the transfer para-meter. Since it is not easy to obtain absolutely pure forms, it is, however, essentialthat the same sample be used in the determination of AH: in water and the givensolvent.EXPERIMENTALSamples of the A-form of m-a-amino-n-butyric acid were obtained by crystallisaiionfrom water-acetone ; the plate-like crystals were always deposited while the solution wasstill hot.The B-form was best obtained by crystallisation from water-ethanol. After thehot solution had been filtered, it was set aside to stand at - 278 K, whereupon the coldsolution very gradually deposited the needle-like B-form. Crystallisation from water by asimilar procedure also yielded the B-form, but crystallisation from either water or water-ethanol gave the B-form contaminated with A-form whenever the solution crystallised outwhile still hot.Heats of solution of the two forms were determined using an LKB 8700 calorimetrysystem with an 8721-1 solution calorimeter ; thermochemical functions are expressed interm of the defined calorie (4.1840 J) and refer to the isothermal process at 298 I<. Thefinal solutions were about 3 x mol dm-3, and the observed heats of solution here,therefore, taken as those at infinite dilution.Various preparations of the A-form all gaveessentially identical values of AH;, and a final series of experiments on the same samplegave values of 374, 376, 370 and 378 cal mo1-* ; we take AH; for the A-form 2s 374+3 cal mol-l. Different preparations of the B-form, however, gave different AH; 1 dues.Close examination of their i.r. spectra showed that the samples with low AH: values wereslightly contaminated with the A-form. Thus a preparation obtained by crystallisationfrom water at 298 K, had AH; values of 1414, 1422, 1433, 141 8 and 1424 cal rnol-1 (averageI422+ 7) and another preparation obtained by crystallisation from water+ ethanol at 278 K,had A H ; values of 1586, 1594, 1610, 1599 and 1617 cal mol-' (average 1601+ 12). Bothof these preparations were slightly contaminated with the A-form.A sample of the B-formthat, as judged by its i.r. spectrum, was free from the A-form had AH: 1666, 1638, 1632,1654 and 1676 cal mol-1 (average 1653 & 18).The equilibrium solubility of the B-form was determined by shaking an excess of thesolid with water at 295 K for several days. Aliquot portions of the supernatant liquid wereperiodically removed, weighed in stoppered bottles, and then evaporated to constant weight.The solubility of the A-form was determined by adding portions of the solid to a knownweight of water at 298 K. The mixture was vigorously shaken and the weigh? of solidrequired to achieve saturation within about 30 min was noted.The resulting saturatedsolutions ofthe A-form, on contact with a slight excess of the solid A-form, always precipltntedout the B-form and the solution reverted to a concentration, equivalent to the solubility ofthe B-form. This precipitation sometinies took place within a few minutes of (A-form)saturation being reached ; on other occasions precipitation occurred only after a consitlerLihleinterval of time had elapsed. SoluSilities at a number of other temperatures were determinedas described aboveABRAHAM, AH-SING, MARKS, SCHULZ AND STACE 185Proteins, Amino Acids andpeptides, ed. E. J . Cohn and J. T. Edsall (ReinhoId, New York, 1943).P. L. Whitney and C . Tanford, J. Biol. Chem., 1962,237, PC 1735.Y . Nozaki and C . Tanford, J. Biol. Chem., 1963, 238,4074 ; 1965, 240, 3568.G. Tanford and Y . Nozaki, J. Biol. Chem., 1971, 246,2211. ' C . H. Spink and M. Auker, J. Phys. Chem., 1970,74, 1742.G . Conio, L. Curletto and E. Patrone, J. Biol. Chem., 1973, 248, 5448.M. H. Abraham, D. H. Buisson and R. A. Schulz, J.C.S. Chem. Comm., 1975, 693.M. Tsuboi, Y . Iitaka, S. Suzuki and S-i. Mizushima, Bull. Chem. SOC. Japan, 1959, 32, 529.l o T. Ichikawa and Y . Iitaka, Acta Cryst., 1968, B24, 1488; T. Ichikawa, Y. Iitaka and M.Tsuboi, Bull. Chem. SOC. Japan, 1968, 41, 1027.T. Akimoto and Y . Iitaka, Acta Cryst., 1972, B28, 3106.C . H. Spink and I. Wadso, J. Chem. Thermodynamics, 1975, 7, 561.York Acad. Sci., 1957, 69, 94.* B. Dawson and A. McL. Mathieson, Acta Cryst., 1951, 4, 475.l 2 6. C . Kresheck, H. Schneider and H. A. Scheraga, J. Phys. Chem., 1965, 69, 3132.*' R. J. Koegel, R. A. McCallum, J. P. Greenstein, M. Winitz and S. M. Birnbaum, Ann. NewI s E. J. Cohn, T. L. McMeekin, J. T. Edsall and J. H. Weare, J. Amer. Chem. SOC., 1934,56,2270.l 6 P. K. Smith and E. R. B. Smith, J. Biol. Chem., 1937, 121, 607.l 7 M. S. Dunn and F. J. Ross, J. Biol. Chem., 1938, 125, 309.'* H. D. Ellerton, G. Reinfelds, D. E. Mulcahy and P. J. Dunlop, J. Phys. Chem., 1964, 68, 398.(PAPER 6/1610
ISSN:0300-9599
DOI:10.1039/F19777300181
出版商:RSC
年代:1977
数据来源: RSC
|
22. |
Termination errors in Fourier analysis of diffraction data of aqueous electrolyte solutions |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 186-189
Roberto Triolo,
Preview
|
PDF (257KB)
|
|
摘要:
Termination Errors in Fourier Analysis of Diffraction Dataof Aqueous Electrolyte SolutionsB Y ROBERTO TRIOLO" AND IRENE RUFFOIstituto di Chimica Fisica, Universith, Palermo, ItalyReceived 2nd June, 1976Termination errors have been assessed for X-ray and neutron diffraction experiments using aqueoussolutions. Some discrepancies and apparent anomalies of recently published diffraction experimentsare interpreted in the light of this analysis.The Fourier analysis of scattering data involves many kinds of errors. Amongthem, the so-called '' termination errors ", due to the replacement of the infinite limitin the Fourier integral by a finite one [k,,, = (4n/A) sin Om,,], will be discussed inthis note. They cause a kind of modulation of the radial distribution function.As an example we consider the X-ray diffraction pattern of HC1+ H 2 0 solutions.Of at least five papers on this two l * are in good agreement with eachother but not with the other three,2-4 which, however, do agree at least qualitativelywith each other.The main discrepancy is a peak at ~ 2 . 5 A in G(r) that in ref. (1)and ( 5 ) is ascribed to 0-0 interactions due to strong H-bonding. This feature isclosely correlated with the increasing prominence, in the correlation function obtainedfrom neutron ~cattering,~ of the peak around 1.6 A.We think that the above discrepancies in the two categories of papers, are basicallydue to a termination error. In fact from the X-ray data of ref. (5) we calculatedG(r) using five different values of k,,, (16.00, 14.80, 12.60, 10.00 and 7.60 A-'> ; the&st value corresponds to MoK, 1 * radiation, the last one to CuK, ; the values12.60 and 10.00 A-' correspond to k,,, reported in ref.(4) and (2) respectively. Infig. 1 is sketched the effect of such truncations on the same set of data.As can be seen even for the lowest k,,, (7.60 A-l corresponding to CuK, radiation)the peak at % 2.5 Acan be detected for the most concentrated solution (HCl-3.99 HzO) ;the same peak is barely resolved for k,,, = 12.60 A-l in the case of HC1.8.20 H20,while a value of k,,, = 14.8OA-1 is necessary to resolve the same peak in the caseof HCl. 16.0 H20.In the case of HC1.31.5 H20, no peak at 2.5 A is detected even with kmax =16.0 A-'. With these results in mind we can explain the apparent discrepanciesbetween the different sets of data.As far as data of ref. (2) are concerned the mostone could get is a shoulder for the most concentrated solution (NHCl.10 H20).For the other two solutions the weight of short 0-0 interactions is not strongenough to be seen with k,,, = 10 A-l. On the other hand, data of ref. ( 3 ) show noshort 0-0 interactions for the low concentrations and for the low upper integrationlimit. Owing to the relatively highk,,, and concentrations, the 0-0 interactions via strong H-bonding can be barelydetected.In conclusion, the effect of " termination of data " is not only manifested in theappearance of spurious peaks, but also as a general loss of resolution (basicallyThe same is not true for the data of ref.(4).18R. TRIOLO AND I. RUFF0 187broadening of peaks). It follows that if two peaks [either in D(r) or in G(r)] are closeenough, they collapse into a single one in the limit of low k,,, and/or lowconcentration.We now consider other cases where the effect of termination errors is not asdramatic as the one we have just examined. As an example we can take recent X-raydiffraction data from alkali halide solutions. With sodium halide solutions welit I5432I15432IFIG. 1.-Termination effect on HCI + H20 solutions investigated with X-rays. Curves (a), (b), (c) and(d)refer to the four more concentrated solutions of ref. (5). (a) = HC1.3.99 H20, (b) = HC1.8 H20,(c) = HC1.16.0 H20, (d) = HC1.31.5 HzO.Different curves .are shifted vertically for clarity;const. is a scale constant.expect at least three different peaks for 2 A < r < 4 A, in both D(r) and G(r), thoughin practice we find a more resolved Na+-H20 peak in sodium bromide solution thanin a sodium chloride solution at approximately the same concentration. In additionthe peak of Na+-H20 interactions is less resolved than that of Li+-H20, again ofapproximately the same concentration, in spite of the lower scattering cross section a188 TERMINATION ERRORSLi+. We think that these findings are a direct consequence of termination errors,affecting the manifestation of Na+-H20 interactions in sodium chloride solutionmore seriously than in sodium bromide solution, because in the former case, sodium-water, water-water and chloride-water distances (2.5, 2.8, 3.2 A respectively) arecloser than sodium-water and bromide-water distances (2.5 and 3.4 A).That is,with NaCl+H,O solution, the situation is similar to the one found for HCl+ H2Q1*already discussed.On the other hand, the peak ascribed to Li+-H20 interactions is far enoughfrom the water-water and chloride-water peaks ; consequently better resolution isobtained in this last case. The situation is different with neutron diffraction. Owingto the dominance of interactions involving deuterium atonis over other interactions, aI !J1 1 11.0 2.0 3.0.-, , I1.0 2.0 3.0(9)i iFIG. 2.-Termination effects on DCl+ D 2 0 solutions investigated by neutron diffraction. Curves(e), cf), (9) and (h) refer to the solutions of ref.(5). (e) = DC1.3.08 D20, (f) = DC1-9.00D20,(g) DC1.15.6 D20, (h) = DC1.30.9 D20. Different curves are shifted vertically for clarity ;again const. is a scale constantR. TRIOLO AND I. RUFF0 189lower k,,, can be used without seriously affecting the results. To show this, we havetaken the neutron diffraction data of ref. ( 5 ) and calculated G(r) with five differentvalues of k,,, (15.7, 13.7, 11.7, 9.7 and 7.7 .$-I). The results of these calculationsare shown in fig. 2. As can be seen, going from k,,, ~ 1 6 A - l to k,,, -12A-lthe calculated correlation filnctions hardly undergo any appreciable changes. Onlyin the case of the lowest k,,, value do we have a noticeable change in the curve.In practice, values of k,,, of about 10-12 A-' can be used quite confidently whenneutron data are collected. As a conclusion we wish to point out that the con-centration of the solution also plays a fundamental role in this kind of experiment, sothat suitably chosen systems should be studied by both X-ray and neutron diffractionmethods over wide ranges of concentration and with appropriate k,,, values. Suchextensive studies are necessary in order to reduce greatly the non-uniqueness ofinterpretation usually associated with diffraction data from liquids and the consequentquestions concerning the reliability of diffraction experiments.S. C. Lee and R. Kaplow, Science, 1970, 169,477.D. S. Terekhova, J. Struct. Chem., 1970, 11, 483, translated from Zhur. strukt. Khim., 1970,11, 530.G. Licheri, G. Piccaluga and G . Pinna, Chem. Phys. Letters, 1971, 12, 425.R. Triolo and A. H. Narten, J. Chem. Phys., 1975, 63, 3624.G. Licheri, G. Piccaluga and G . Pinna, J . Appl. Cryst., 1973, 6, 392.' D. L. Wertz, J. Solution Chem., 1972, 1, 489.(PAPER 6/1030
ISSN:0300-9599
DOI:10.1039/F19777300186
出版商:RSC
年代:1977
数据来源: RSC
|
23. |
Reviews of books |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 190-191
Mino Green,
Preview
|
PDF (138KB)
|
|
摘要:
Reviews of BooksSurface Physics of Phosphors and Semiconductors. Ed. C. G. SCOTT and C. E. REED. (Academic Press,New York, 1975). Pp. xiv+644. Price 218.80.To be properly equipped for research into the behaviour of semiconductor surfaces, a scientistneeds to know a fair amount of physics, physical chemistry and, if an experimentalist, vacuumtechniques and aspects of electronic instrumentation. This is truly an interdisciplinary field anda highly challenging one at that. The basic challenge at the moment is to obtain and interpret data,devoid of spurious effects, on a restricted number of semiconductors ; silicon, germanium, galliumarsenide and zinc oxide are the substances which head the list. The technological incentive is broad,ranging from control of surfaces involved in electronic devices to the discovery of new chemicalcatalysts. The editors see the task of reviewing this subject as too great for an individual and havetherefore decided upon “ a collection of inter-related articles by authorities on various topics ”.This approach contains the danger of a varying level and quality of presentation, and the occurrenceof overlap as well as gaps : all these faults are, to some extent, present and the editors must acceptsome of the blame.That having been said, this useful book can be recommended to research workersin, or starting in, semiconductor surface research. Contrary to the title there is nothing but a passingremark or two about phosphors.The nine chapters of this book are as follows :(1) Atomic Structure of Surfaces (D.Haneman) ; not a particularly balanced review, it comescloser to being the author’s personal record of two decades of research on the atomic structure ofGe, Si and the group 111-V compounds. The main message of this work is that deviations from theideal surface structure are largely surface relaxations and not rearrangements involving atom migra-tion. When the author discusses experimental matters his obvious experience warrants respect andattention.(2) Aspects of Surface State Theory (R. 0. Jones) ; an excellent survey of the various methods,and their shortcomings, of calculating surface states.(3) The Surface Space Charge Layer (F. Bertz) ; contains the essential mathematical relationsbetween charge density and electric potential for a moderate range of semiconductor conditions.Equilibrium, forced and natural response are considered.A first class introduction to the subject.(4) Theory of Adsorption (S. R. Morrison) ; mostly about adsorption involving charge transfer,with a low level introduction.(5) Techniques and Measurements (C. E. Reed and C. G. Scott) ; principally concerned withdetermining surface state density and energy through measurement of space charge parameters.(6) The Chemical and Physical Properties of Clean Germanium and Silicon Surfaces (F. Meyerand M. J. Sparnaay) ; the pikce de rCsistance of this collection and should be read first by non-specialists. The dominant role oflocal interactions in chemisorption is emphasised.(7‘) Group 11-VI and 111-V Compound Materials (C.G. Scott and C. E. Reed) ; the unsatisfactoryposition concerning our knowledge of these compounds is moderately well summarised. Disappoint-ingly, the thermodynamics and kinetics of defect formation are not discussed.(8) Gas Effects on group IV-VI Semiconductor Films (J. N. Zeniel) ; concerned with the inter-action of hydrogen and oxygen with the surface and bulk of thin films of Pb- and Sn- (S, Se and Te)compounds, and the attendant electrical changes. The status of the subject makes its inclusion inthis book singularly doubtful.(9) The Effect of Surface States on Semiconductor Devices (K. H. Zaininger) ; the effect of surfacecharge trapping on semi-conductor device behaviour is very well discussed, together with the necessarydevice theory.Surface structure is reviewed and shown to be very complex.Up to date surface state data are provided.MINO GREENReceived 18th August, 197619REVIEWS OF BOOKS 191MacromoIecular Chemistry 10.(IUPAC Macromolecular Microsymposium XIV & IVth DiscussionConference on Macromolecules). Ed. B. SEDLACEK. (Butterworths, 1975). Pp. 310. Price E17.50,$35.This book contains papers presented at two IUPAC meetings held in Czechoslovakia in thesummer of 1974. The first part on " Crosslinking and Networks " contains six papers describingboth theoretical and experimental studies of network formation, structure and properties. Theeight papers in the second section deal with " Heterogeneities in Polymers " ranging from thoseoccurring within single chains to larger scale structural heterogeneities in melts, rubbers and glasses,including multicomponent systems. A useful, if somewhat expensive, library investment although,as always, the contents also appear elsewhere : Pure Appl. Chem., 1975, 43 (1-2).G. C. MAITLANDReceived 12th August, 197
ISSN:0300-9599
DOI:10.1039/F19777300190
出版商:RSC
年代:1977
数据来源: RSC
|
24. |
Similarity in the reactivity of O–2and double bond type lattice oxygen as revealed by vapour-phase catalytic oxidation of furan to maleic anhydride |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 193-202
Masamichi Akimoto,
Preview
|
PDF (780KB)
|
|
摘要:
Similarity in the Reactivity of 0; and Double Bond TypeLattice Oxygen as Revealed by Vapour-Phase CatalyticOxidation of Furan to Maleic AnhydrideBY MASAMICHI AKIMOTO" AND ETSURO ECHIGOYADepartment of Chemical Engineering, Tokyo Institute of Technology,00 ka yama, Megur o-ku, To ky 0, JapanReceived 27th October, 1975Vapour-phase oxidation of furan o v a titania and molybdena catalysts has been investigated withreference to the oxygen species incorporated into maleic anhydride. The sequence of the amountof 0; formed on titania and molybdena catalysts modified with VA group oxides was found to beP205 > As203 $- Sbz03 2 Bi2O3, which is in agreement with the electronegativity of the addedVA group elements. The rate of maleic anhydride formation over the modified titania showed thesame sequence, but the rate over the modified molybdena increased in the following order ; Bi203 >Sb2O3 > Pz05 > AszO3, which is in line with the reactivity of the double bond type lattice oxygen,MFO, in these modified molybdena catalysts.Therefore, it is proposed that 0; or double bondoxygen Mo=O plays a very important role in the oxidation of furan to maleic anhydride over a titaniaor molybdena catalyst. This has been supported further by kinetic studies and discussion of thesereaction mechanisms, and these oxygen species are shown to be incorporated into the maleic an-hydride formed. Thus, the results suggest a similarity in the reactivity of 0; and that of the doublebond oxygen in the oxidation of furan.111 vapour-phase catalytic oxidation of olefinic hydrocarbons, selectivity is stronglyaffected by the nature of the oxygen species formed.With the exception of epoxideformation and oxidative dehydrogenation, selective oxidation of olefins proceedsthrough a repetition of hydrogen abstraction and oxygen addition. The latticeoxygen in a metal oxide has an ability to effect hydrogen abstracti0n.l Accordingly,this selectivity is strongly dependent upon the presence of oxygen species which havethe ability to add to the hydrocarbon species, i.e., " selective oxygen species ". Thespecial activity of V205 and MOO, catalysts has been attributed to the presence of alattice oxygen with double bond character 2-5 and 0; causes epoxide formation inthe oxidation of ethylene over a silver catalyst.6-10 The type of the metal-oxygenbond rather than its strength thus appears to strongly affect the reactivity of theoxygen species. However, addition of oxygen to hydrocarbon species is a chemicalreaction. Hence, it is likely that a similarity exists in the nature and the reactivityof " selective oxygen species " in spite of the differences in the oxygen spzcies andoxidation reactions involved.In this paper, the effect of VA group oxide additionon the formation of 0; over titania and molybdena catalysts was studied in relationto the reactivities of 0; and the double bond type lattice oxygen Mo=O in theoxidation of furan to maleic anhydride, and incorporation of these oxygen speciesinto maleic anhydride was proposed.EXPERIMENTALVapour-phase catalytic oxidation of furan was carried out at 563-663 K using a con-ventional flow microreactor under atmospheric pressure.The reactor was a 13 mm4.d.Pyrex tube, 300 mm long, with a concentric thermowell. The catalysts (16-32 mesh) werediluted with quartz (30-50 mesh) and Raschig rings (2.5 x 2.5 mm) were added above the1-7 19194 CATALYTIC OXIDATION OF FURANcatalyst bed as a preheating medium. A cylindrical fluidized thermal bath was used forheating.Reagent grade furan was used in the experiment. Air was purified by passing througha silica-gel and a soda-lime tower to remove the water and carbon dioxide. Nitrogen,hydrogen and helium were supplied from commercial cylinders. By bubbling air throughliquid furan cooled at 273.2K with ice-water, furan was evaporated, and the resultinggaseous mixture was introduced into the reactor after dilution further with air and nitrogen(furan 1.0-2.0 vol %, oxygen 7-21 vol %).Anatase titania was prepared by calcination of titania gel in air current at 773 K for 2 h.Various supported molybdena catalysts were prepared by mixing either silica sol or titaniagel with an aqueous solution of ammonium p-molybdate followed by evaporation andcalcination of the dried mass in air current at 813 K for 4 h.Titania and molybdenacatalysts modified with VA group oxide were similarly prepared. As a source of P2OS,As203, Sb203 or Bi203, 85 wt% phosphoric acid, extra pure As203, Sb203 or bismuthnitrate were used, respectively.The silver catalyst was obtained by reacting silver nitratewith sodium hydroxide in an aqueous medium followed by reduction of the silver oxideformed with 36 % formalin at pH = 11 0.5.11PEG 20M20 wt % on Neopak 1A (4 mm-i.d., 3 m, 353 K, furan), propylene carbonate 40 wt % onCelite 545 (4 mm-i.d., 4 m, room temp., C 0 2 ) and molecular sieve 13X (4 mm-i.d., 2 m,room temp., CO, N2, 02) were used as separating columns. The maleic anhydride formedwas absorbed in water and boiled to remove the dissolved carbon dioxide from the solutionand then titrated with an aqueous solution of sodium hydroxide using a pH meter. Thestructure of the catalysts was investigated by X-ray diffraction using nickel-filtered Cu-K,radiation. The surface area was measured by the B.E.T.method using nitrogen as anadsorbate. The amount of paramagnetic species formed on the catalysts was measured bye.s.r. using Mn2f or l,l-diphenyl-2-picrylhydrazyl (DPPH) as an internal standard. Theeffect of temperature on the amount of adsorbed oxygen species was also studied by thesame method, but a temperature-programmed cavity was used for these studies.The gaseous effluent from the reactor was analysed by gas chromatography.RESULTSFORMATION OF 0, ON ANATASE-TYPE TITANIAAnatase-type titania (50 mg) was weighed into a quartz tube (3 mm-i.d.) and thenevacuated at 813 K and 6.65 x kN m-2 for 1 h. The formation of Paramagneticoxygen species was observed by e.s.r. on introduction of gaseous oxygen (0.665 kNm-2at 173 K) and the amount of this oxygen species, based on the height of e.s.r.signaldue to Mn2+, increased with rise in temperature in contrast to that of a broad signalwith g = 1.97 (Ti3+). Ti3+ disappeared at 233 K while the oxygen species showed amaximum at 323 K and then disappeared at 503 K. Fig. 1 shows the e.s.r. spectrumof the oxygen species adsorbed on the evacuated titania at 433 K, which is a tripletsignal with g = 2.003, 2.010 and 2.020 and is similar to that observed by Slivetsand Kazansky ’ and by Naccache, Meriaudeau, Che and Tench.’ Hence, this e.s.r.signal can be attributed to the 0, formed by the adsorption of gaseous oxygen onTi3+.FIG. 1.-E,s.r. spectrum of 0; formed on anatase titania at 433 KM. AKIMOTO AND E . ECHIGOYA 195The formation of adsorbed oxygen species on titania catalysts modified with asmall amount of VA group oxides was similarly investigated.In the case of P20,+Ti02 (1 : 18 mol ratio), a singlet signal with g = 2.010 and AHmax + 87 G wasobserved at 173 K, but the AH,,, decreased with rise in temperature. At 273 and293 K, g = 2.010 and 2.020 for the oxygen species. However, at higher temperatures,it showed a triplet signal with g = 2.003, 2.010 and 2.020. Hence, this oxygen200 300 400 500 600 700temp./KTi02(1 : IS), (D : As203+Ti02(l : 18).FIG. 2.-Effect of temperature on the amount of 0; over T i 0 2 + v ~ group oxide. 0 : P205+TABLE CORRELATION BETWEEN THE MAXIMUM AMOUNT OF 0; AND THE RATE OF MALEICANHYDRIDE FORMATION OVER T i 0 2 + v ~ GROUP OXIDE afuran conversion rate (M.A.) maximum amount ofcat.(mol : mol) cat./g I % M.A. yield/% / g mol g catalyst-1 h-1 0 / g catalyst-1Pz05+TiOz(l : 18) 0.306 25.9 3.30 1.327 x 102.6As203+TiOz(l : 18) 0.305 5.3 1.21 0 . 4 7 4 ~ 13.2Sb203+Ti02(l : 18) 0.312 2.3 0.21 0.083 x -2Bi203+ TiOz(l : 18) 0.304 6.1 0.17 0.068 x 10-3 -2a feed : furan 1.53 vol %, oxygen 10.8 vol % ; diluent-nitrogen ; total feed, 300 NTP cm3 min-I ;reaction temp., 635 K.species is 0;. The 0; formed on As203 +TiO, (1 : 18) showed a change similarto that on P205 +Ti02. Fig. 2 shows the change in the amount of 0; formed perunit surface area on the titania modified with P205 or As203. The amount reachesa maximum value at 359 K on these catalysts. On the other hand, the formationof 0; was scarcely observed in the case of Sb203+Ti02 (1 : 18) and Bi203 +Ti02(1 : IS), and the amount was in the order of 0.07 and 0.04 m2 catalyst-'. Thesequence of the maximum amount of 0; was0.448 (P205 + TiO,) > 0.379 (As203 +Ti02) & 0.07 (Sb203 + Ti02) >Apparently, this agrees with the order of electronegativity of VA group elementsadded to titania.0.04 (Bi203 +Ti02).OXIDATION OF FURAN OVER TITANIA CATALYSTSOxidation of furan over these modified titania catalysts was performed withreference to the formation of maleic anhydride.In table 1, the maximum amoun196 CATALYTIC OXIDATION OF FURANof 0;: per gram catalyst and the rate of maleic anhydride formation per gram catalystmeasured at 635 K (furan, 1.53 vol % and oxygen, 10.8 ~01%) by means of a differentialreactor are summarized.The rate increased with increase in the amount of 0;.The formation of maleic anhydride over P205 + TiO, was kinetically investigatedusing the differential method. Thus,= k l [furan]*. 39[02]0*8E,,, = 6.29 kJ g mol-Iat 606 K, furan : 1 .O-2.0 vol%, O2 : 7-1 5 ~ 0 1 % .100 1290 310 330 350 370 390reaction temp./"(=FIG. 3.-EfTect of reaction temperature on the oxidation of furan over silver catalyst. Catalyst :silver reduced with hydrogen at 573 K for 1 h. Feed : furaii 1.54 vol %, oxygen 10.7 vol %, diluent-nitrogen. Contact time : W/F = 0.185 g catalyst h g mol-'. 0 : conversion of furan, (3 : yieldof COz, A : yield of maleic anhydride, 0 : selectivity to maleic anhydride.The formation of maleic anhydride was also observed in the oxidation over asilver catalyst (fig. 3).In this case, the catalyst was previously reduced with hydrogen(50 NTP cm3 min-') at 573 K for 1 h, and then a reactant mixture (furan, 1.54 vol% ;02, 10.7 vol% ; NZ, 87.8 ~01%) was introduced into the reactor at - 523 K, followedby a rise in the reaction temperature. The yield of maleic anhydride increased from0.25 % to 4.0 % with a rise in the reaction temperature from 568 to 655 K, with aselectivity of N 5.0 %. Though 30.7 % furan was converted at 656 K, only anegligible amount of maleic anhydride (selectivity 0.75 %) was formed over the silvercatalyst under reaction conditions similar to those of fig. 3 except when using N 2 0(20. I ~01%) instead of gaseous oxygen.OXIDATION OF FURAN OVER SUPPORTED MOLYBDENA CATALYSTSOxidation of furan was similarly carried out using various supported molybdeiiacatalysts.Fig. 4 shows the results of oxidation in air of 1.50 vol% furan over thevarious silica-supported molybdena catalysts at 633 K. The catalytic activity andthe yield of maleic anhydride increased with increase in the content of molybdena toreach a niaximuin value (conversion 40.3 %, yield 13.5 % and selectivity 33.6 %) ata molybdena content of 10 mol%. Various MOO, + SOz catalysts (0.50 g> wereevacuated in the quartz tube at 638 K under 1.33 x kNrn-, for 30 min and werethen investigated by e.s.r. at room temperature. The formation of Mo5+ withg = 1.93 was observed and the amount produced was comparable to that formed iM.AKIMOTO AND E . ECHIGOYA 197the catalyst quenched during the oxidation of furan. The relative amount of Mo5+formed on the catalyst surface by evacuation, i.e., the relative value of Mo5+ con-centration formed per gram catalyst multiplied by its surface area, is also shown infig. 4. Apparently, the catalytic activity and yield of maleic anhydride are enhanced1 h0 10 20 30 40 50MOO, content/mol %FIG. 4.-Correlation between relative amount of Mo5+ and activity of MOO, + SiOz catalyst.Reaction temp. : 633 K, feed : furan 1.50 vol % in air. Contact time : W/F = 0.373 g catalysth g mol-'. 0 : conversion of furan, c) : yield of CO,+CO A : yield of maleic anhydride, 0 :selectivity to maleic anhydride.by an increase in the amount of Mo5+ formed on the surface.A similar correlationwas also observed in the oxidation of 1.50 vol% furan in air over a MoO,+TiO,catalyst at 633 IS. This catalyst was of the anatase-type solid solution within therange of molybdena content of - 25 mol%. The catalytic activity and yield ofmaleic anhydride also increased with increase in the amount of Mo5+ formed on theY1fii43210- 100 0 loo 200temp./"CFIG. 5.-Variation in the amount of 0; over Moo3 + TiOz + VA group oxide with rise in temperature.Catalyst : MOO, +TiO2+vA group oxide, Mo03/Ti02 = 1/99 mol/mol,O : Pz05, P/Mo =5/100 atomlatom, 0 : AsZ03, As/Mo = 5/100 atomlatom, a : Sbz03 or Bi203, Sb/Mo orBi/Mo = 5/100 atomlatom198 CATALYTIC OXIDATION OF FURANsurface and reached a maximum value (conversion 87.3 %, yield 38.8 % andselectivity 44.5 %) at a molybdena content of 25.0 mol%.When gaseous oxygen (0.665kNm-2) was introduced, the formation of 0;with g = 2.003, 2.009 and 2.019 or g = 2.005, 2.010 and 2.015 was observed onMOO, +TiO, (1 : 99) or MOO, + Si02 (1 : 9) catalyst previously evacuated at 813 Kfor 1 h.14 Then, the formation of 0; over MOO, +TiO, (1 : 99) modified with VAgroup oxides (x/Mo = 5/100 atom/atom, x = P, As, Sb or Bi) was similarly investi-gated as in the case of the modified anatase titania (fig. 5).0; over the molybdena40 -20 -l o :8 -6 -?4 1 I I I I I103 KITFIG. 6.-Rate of maleic anhydride formation in the oxidation of furan over Mo03+TiOz+v.4group oxide. Feed : furan 1.50 vol % in air.Rate (M.A.) : g mol m2 catalyst-l h-l. Catalyst :Mo03+Ti02+v~ group oxide, Mo03/Ti02 = 1/3 molelmole X/Mo = lS/lOO atomlatom, X =Bi, Sb, As or P. 0 : Bi203, (D : Sb203, A : As203, 0 : P205.catalyst modified with P205 or As203 showed a maximum value at - 318 K anddisappeared at 413 K. Interestingly, the sequence of the amount of 0; per unitsurface area is P205 > As203 9 Sb,03 2 Bi,03 with respect to the VA group oxidesadded, in good agreement with the results obtained over the modified titania catalysts.TABLE 2.-ABSOLUTE AMOUNT OF PARAMAGNETIC SPECIES FORMEDspecies a [(MoS+)surface/ (Mo5f)surfacecat. (mol : mol) /spin g catalyst-1 (Mos+/total Mo)/% total Mas+] b/% /spin g catalyst-1Pz05+TiOz(l : 18) 053.71 x lO"(359 IS) - - -Moo3 + TiOz(l : 9) Mo5+1 .32 x lo'' 1.91 37.1 5 .0 9 ~ 10"Mo03+TiQ2(18 : 82) Mo5+3.59x 10'' 3.08 20.0 7.18 x 10''MoO3+TiOZ(l : 3) Mo5+4.37 x lo1' 2.85 18.5 8.08 x 10''Mo03+SiOz(l : 9) Mo5+5.56 x lo1* 0.640;- described in this paper ; b ref. (30).Mo03+Ti02(l : 1) M0~~2.93 x 10'' 1.14 14.8 4.34x 10" - -a formation of Mo5+-by evacuation of the catalyst (0.50 g) at 638 K for 30 min ; formation ofHowever, the sequence of the rate of maleic anhydride formation from 1.5 vol %furan per unit surface area determined by the differential method in air over theMoO,+TiO, (1 : 3) catalyst modified by VA group oxides is Bi203 > Sb203 >P205 > As203 with respect to the additives (fig. 6). In contrast to the case of themodified titania catalysts, this sequence does not agree with that presumed from theamount of 0; formed on these catalystsM.AKIMOTO AND E. ECHIGOYA 199The absolute amount of these paramagnetic species formed on the P205+Ti02and the molybdena catalysts was measured by e.s.r. spectroscopy using DPPH as aninternal standard (table 2). Here, the preparation of Mo5+ was carried out byevacuation of the catalysts (0.50 g) at 638 K for 30 min, and the preparation of 0;was described elsewhere in this paper. The amount of 0; and Mo5+ formed on thesurface was 1017-1018 spin g catalyst-l.DISCUSSIONThe good correlation existing between the amount of MoS+ formed on the surfaceand the catalytic activity (fig. 4) gives strong evidence supporting the conclusion thatMo5+ is the active site for oxidation of furan.Maleic anhydride is formed by thereaction of furan with the oxygen species formed on the molybdenum ion in accor-dance with the oxidation of butadiene to maleic anhydride.14 Adsorption of gaseousoxygen on Ti3+ or Mo5+ is accompanied by the transfer of an electron from thesemetal ions to the adsorbed oxygen molecule followed by the formation of Ti4+-O;or Mo6+-O;. Generally, the strength of a metal-oxygen bond ('in this case,Me"+-0;) and heat of oxygen adsorption on a metal oxide are enhanced by theincrease in electronegativity of the metal ion on to which oxygen is ad~orbed.'~~''Hence, the sequence of the amount of 0; per unit surface area of the titania andmolybdena catalysts modified with VA group oxides agrees with that of the electro-negativity of the Ti3+ or Mo5+ species existing on the surface.Thus, it is concludedthat addition of VA group oxides changes the electronegativities of these metal ionsin proportion to those of VA group elements. This addition effect is indicated wellby the thermal stability of O,, that is, 0; showed a maximum at - 359 K anddisappeared above 523 K on the P205+Ti02 and As203+Ti0, catalysts (fig. 2).These temperatures were higher than those observed for unmodified titania (323 and503 K). Similarly 0; on Moo3 +Ti02 (1 : 99) disappeared at 373 K,I4 while 0, onthe molybdena catalyst modified with P205 or As20, did so above 413 K (fig. 5).At higher temperatures, 0; is decomposed by scission of the 0-0 bond followedby formation of lattice oxygen 02-.In the case of molybdena, this lattice oxygenis a doubIe bond type Mo6+=02- . While, in the case of titania, which has nodouble bond type, a-bond type -Ti4+-O-Ti4+- is formed.I4 Hence, redox oftitania (or molybdena) during the oxidation of furan is likely to occur through theformation of Ti4+-O; (or Mo6+-0;) by oxygen adsorption on Ti3+ (or No5+),followed by decomposition into the a-bond lattice oxygen (or double bond one) andreduction of the lattice oxygen with furan forming Ti3+ (or MoS+). Here, the amountof 0; formed on these catalysts during the oxidation reaction and its life are expectedto increase in the following order : P205 > As20, > Sb203 > Bi203, with respectto the additives. The life of 0, is very short during the oxidation of furan.However,if the life is long enough to react with the furan species, participation of this oxygenspecies in the oxidation reaction can be expected. This suggests the possibility of0; participation in the oxidation of furan over a titania catalyst, especially over thatmodified with P205 or As203, where the 0; formed is very stable as compared withthat formed on the molybdena catalysts (fig. 2 and 5). The formation of maleicanhydride was actually observed, and the rate was found to increase by increasingthe amount of 0; over these titania catalysts (table 1). The correlation betweenthe maximum amount and the yield of maleic anhydride is not linear. This is dueto consumption of 0; in other reactions, as seen in the higher furan conversion andthe lower selectivity over the P205 +Ti02 catalyst.In contrast, the rate increasedby decreasing the amount of 0; and electronegativity of Mo5+ existing on th200 CATALYTIC OXIDATION OF FURANmolybdena catalysts (fig. 6). Reducibility and reactivity of the double bond typelattice oxygen in molybdena modified with VA group oxides was already found toincrease with decreasing electronegativity of the VA group elements. At thesehigher temperatures, the formation of adsorbed oxygen species without any chargetransfer and bonding to the metal ions is unlikely. Assuming the presence of suchan oxygen species and its participation in maleic anhydride formation, the samesequence in the rate should be obtained over these modified titania and molybdenacatalysts.However, this is not in line with the results. Hence, the above resultsstrongly suggest participation of 0; over the titania and that of double bond oxygenMo=O over the molybdena in maleic anhydride formation. The formation of 0;when gaseous oxygen is adsorbed on a reduced silver surface has been reported bymany 9* lge2' Observation of maleic anhydride formation by oxidationover a silver catalyst (fig. 3) and the negligible yield of the anhydride obtained withN20 also suggest the participation of 0;. The effect of participation of 0; inmaleic anhydride formation was also reflected in the reaction order with respect tooxygen. In our previous studies on the formation of maleic anhydride from furan,addition of oxygen species to furan was shown to be the rate-determining step.22Butadiene is adsorbed on Mo5+ and then converted to maleic anhydride by oxida-t i ~ n .* ~ Analogously, furan is adsorbed on Mo5+ and Ti3+ competitively withgaseous oxygen. Thus, the initial rate of maleic anhydride formation based on therate-determining step is proportional to CFC6 and is, at the same time, inverselyproportional to (I + KFCF + KcfCG)2 ; where KF or KO represents the equilibriumadsorption constant for furan or oxygen and C, or C, represents the concentrationof furan or oxygen in the feed. rz is 1 or 9. When the adsorbed molecular oxygenspecies (n = I) or dissociatively adsorbed oxygen or lattice oxygen (n = +), reactswith the adsorbed furan, the reaction order with respect to oxygen (fl) would probablybe between 1 and - 1 or 3 and -3.The reaction order with respect to furan liesbetween 1 and - 1. Thus, f l = 0.83 over the P205 + TiO, [eqn (I)] is consistent with theparticipation of the adsorbed molecular oxygen species rather than the dissociatedone in the rate-determining step. This molecular oxygen species seems to bs 0;.On the other hand, the double bond oxygen Mo=O is thought to play an importantrole in the rate-determining step over the molybdena catalysts. The amount of 0;formed during catalysis was not determined. However, the comparable amount of0; and surface Mo5+ (table 2) cannot always deny the participation of this oxygenspecies in the oxidation of furan.Incorporation of these oxygen species into maleic anhydride was discussed fromthe viewpoint of reaction mechanism.Mo6+ on the surface is present in octahedralc~ordination.~~ Hence, the double bonded oxygen on the surface projects out ofthe catalyst plane. Our previous studies on oxidation of butadiene over titania ormolybdena catalyst revealed that the butadiene is adsorbed by donation of its n-electrons to the surface.23 Analogously, the adsorbed furan species is positivelycharged. Hence, addition of the double bonded oxygen 02- to the positively chargedfuran species possibly proceeds through a nucleophilic mechanism. On the otherhand, the electroaffinity of 0; in the homogeneous phase is -4.8 eV.25 0; formedon the surface is not always similar to that in the homogeneous phase.However, itis also very likely that 0; on the surface is not an electrophile. Hence, interactionof 0 ; with the positively charged furan species is nucleophilic rather than electrophilic.In the absence of any contradictions in these reaction mechanisms, participation ofthese oxygen species in the oxidation of furan to maleic anhydride could not bedenied.The intermediate compound formed by addition of 0; to furan may be a kinM. AKIMOTO AND E . ECHIGOYA 201of peroxide (A), which is in agreement with ethylene oxide formation over silver 26* 27and acrolein formation in the oxidation of propylene over Zn0.28CH-CH CH--CH I !I*-0-0-CH CH\/0I il\/Mo"+-O-CH CH0*Ti4+, Ag+ n may be 5On the other hand, addition of the double bonded oxygen to a furan species isexpected to proceed via a surface alcholate (B) as in the case of o-xylene oxidation tophthalic anhydride over a vanadia catalyst.2g The intermediate compound formedby scission of the 0-0 or Mo-0 bond leads to maleic anhydride by further hydrogenabstraction and oxygen addition.HC=CH HC=CH0-0 or Mo-0 scission I !I +o I I4 C CH-+ C C/ \ / -H /\/\ -H( 4 9 (B)0 0 0 0 0In the oxygen addition step, 0; and double bond oxygen Mo=O play the role ofthe " selective oxygen species ", which then are incorporated into the maleic anhydrideproduced.This reactivity of O,, which is similar to that of the double bonded oxygen, alsoprovides the possibility that the double bond oxygen behaves like a radical oxygen 0-during the oxidation of hydrocarbons.H.E. Swift, J. E. Bozik and J. A. Ondrey, J. Catalysis, 1971, 21, 212.K. Tarama and S. Teranishi, Proc. 3rd Int. Congr. Catalysis (Amsterdam, 1964), vol. 1, p. 282.K. Hirota, Y. Kera and S. Teratani, J. Phys. Chem., 1968, 72, 3133.Y. Kera and K. Hirota, J. Phys. Chem., 1969,73, 3973.F. Trifiro and I. Pasquon, J. Catalysis, 1968, 12, 412.L. Ya. Margolis, Advances in Catalysis (Academic Press, New York, 1963), vol. 14, p. 429.S. Shikagawa, H. Tokunaga and T. Seiyama, K6gy6 Kagaku Zasshi, 1968,71,775.a L. Imre, Ber. Bunsenges Phys. Chem., 1968,72, 863.L. Imre, Ber. Bunsenges Phys. Chem., 1970,74,220.lo S. Shikagawa, K. Kiino, H. Nida and T. Seiyama, Kdgy8 Kagaku Zasshi, 1971, 74, 819.l 1 T. Gotii, S. Shimatake, I. Hara and Y . Aono, Japan Pat., 1963, No. 9259.l 2 V. A. Shvets and V. B. Kazansky, J. Catalysis, 1972,25,123.l 3 C. Naccache, P. Meriaudeau, M. Che and A. J. Tench, Trans. Faruday Soc., 1971, 67, 506.l4 M. Akimoto and E. Echigoya, J. Catalysis, 1973, 29, 191.l 5 A. F. Clifford, J. Amer. Chem. Soc., 1957,79, 5404.l6 A. F. Clifford, J. Phys. Chem., 1959, 63, 1227.l7 Y. Moro-oka, Y. Morikawa and A. Ozaki, J. Catalysis, 1967, 7, 23.l a M. Akimoto and E. Echigoya, J. Catalysis, 1974, 35, 278.l 9 R. B. Clarkson and A. C. Cirillo, Jr., J. Yac. Sci. Tech., 1972, 9, 1073.' O R. B. Clarkson and A. C. CiriIlo, Jr., J. Catalysis, 1974,33, 392.21 N. Shimizu, K. Shimokoshi and I. Yasumori, Buli. Chem. Soc. Japan, 1973, 46, 2929.22 M. Akimoto and E. Echigoya, Bull. Chem. SOC. Japan, 1975, 48, 3518.23 M. Akimoto and E. Echigoya, J. Catalysis, 1973, 31, 278.24 G. N. Asmolov and 0. V. Krylov, Kinetika i Katuliz, 1970, 11, 1028.2s H. 0. Pritchard, Chem. Rev., 1953,52, 529202 CATALYTIC OXIDATION OF FURAN26 S. V. Gerey, K. M. Kholyavenko and M. Ya. Rubanik, Ukrain khim. Zhur., 1965, 31, 449.27 I?. A. Kilty, N. C. Rol. and W. M. H. Sachtler, Pruc. 5th Int. Cungr. Catalysis (Miami Beach,28 B. L. Kugler and R. J. Kokes, J. Catalysis, 1974, 32, 170.29 D. Vanhove and M. Blanchard, J. Catalysis, 1975, 36, 6.30 M. Akimoto and E. Echigoya, Bull. Chem. Suc. Japan, 1973, 46, 1909.1972), vol. 2, p. 929.(PAPER 5/2101
ISSN:0300-9599
DOI:10.1039/F19777300193
出版商:RSC
年代:1977
数据来源: RSC
|
25. |
Preferential binding to poly(α-L-lysine HBr) from polar solvent mixtures |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 203-212
Jiro Komiyama,
Preview
|
PDF (742KB)
|
|
摘要:
Preferential Binding to Poly(a-L-lysine HBr) fromPolar Solvent MixturesBY JIRO KOMIYAMA, TOSHIO MORI, KURANOSUKE YAMAMOTO AND TOSHIRO IIJIMA"Department of Polymer Science, Tokyo Institute of Technology,Ookayama, Meguro-ku, Tokyo 152, JapanReceived 19th January, 1976The preferential binding behaviour of one of the components of aqueous organic solvents ororganic solvent binary mixtures to poly(cc-L-lysine HBr)(PLLHBr) has been studied by differentialrefractometry. The aqueous mixtures have dimethyl sulphoxide (DMSO), N,N-dimethylformamide(DMF), N-methylformamide (NMF), 1 , 2-ethanediol (EG), 1-methyl-2-pyrrolidinone (NMP) and2-propanol(2PrOH) as their second component, while nonaqueous mixtures comprise DMF + DMSO,EG + DMSO and NMP+ DMSO. Aqueous mixtures show different behaviour ; large positivebinding of DMSO at lower concentrations is followed by a decrease and eventually by an inversion athigher concentrations, while the binding of 2PrOH is increasingly negative with concentration. Thebehaviour pattern of DMSO binding from nonaqueous mixtures is the same as that of aqueousmixtures, though the extent is different.In DMFS-DMSO, the helix formation of PLLHBr isobserved. The data obtained is interpreted invoking an operational model of PLLHBr, in whichpolar and less polar portions of the polymer residue undergo solvation separately. PLLHBr isexclusively hydrated in aqueous 2PrOH, if the polymer assumes no helix. It is suggested that thehelix formation in DMFfDMSO mixtures is accompanied by the release of solvent molecules fromaround the less polar portion of the polymer.The preferential binding of organic solvents to biological polymers has beenstudied in relation to the conformational transformation.Measurements of certainphysical properties of a system composed of a polymer and binary solvent mixture,give the excess of one solvent component in the polymer containing phase over theamount found in the bulk solvent.2 Thus, Timasheff and coworkers 3 * 4 havemeasured the preferential binding of various alcohols to proteins and discussed thepossible reason of the denaturation in terms of the actual binding of the alcohols.Morcellet and Loucheux have discussed the preferential binding to, and the confor-mational transformation of, poly-(a-L-glutamic acid) in water + dioxan mixtures alongthe same lines.Difficulty in interpreting such preferential binding data comes from two sources ;one is the fact that the excess quantity cannot be immediately associated with theactual binding of the relevant solvent, and the other is that the protein moleculemanifests heterogeneity in affinity to solvents.In the present investigation, we havestudied the preferential binding of polar organic solvents, DMSO, DMF, NMF, EG,NMP and 2PrOH to PLLHBr from aqueous mixtures. Since this polymer dissolvesin DMSO, the study was extended to DMSO mixtures of NMP, EG and DMF. Inthe last mixture, the polymer assumes helical structure with increasing DMF contenL6The wider spectra of the solvent combination and composition, as well as the simplicityof the polymer composition, were thought to help the interpretation of the preferentialbinding behaviour of the poly(amino acid) as a model for biological polymers.20204 PREFERENTIAL BINDiNG FROM POLAR SOLVENTSEXPERIMENTALMATERIALSPLLHBr was obtained by hydrobromination of poly(N-carbobenzoxy-L-lysine) whichwas prepared according to Oya et aL7 The polymer, was washed with acetone and dried,and was then dissolved in water and recovered by precipitation in 2PrOH.Impuritiesand the low molecular weight fraction of the polymer were removed by electrodialysis ofthe solution for 50 h, followed by ultrafiltration through a Diafilter G 10T (BioengineeringCo, Japan) which is impermeable to polymers with molecular weight > lo4.The polymerwas recovered by freeze drying and stored in a desiccator until use. The molecular weightwas estimated by the viscosity in methanol as 1 .Ox lo5. Solvents (reagent grade) obtainedfrom Wako Pure Chemicals were dried and fractionally distilled using an efficient columnunder the appropriate pressures. Anhydrous Na2S04 for EG and 2PrOl3, and CaHz forthe others were used as drying agents. The solvents were stored in desiccators. Densitiesof these solvents agreed within g ~ m - ~ with the literature values in table 1, whichincludes some other physical constants of the solvents. Deionized and distilled water wasused throughout the experiments.5 xTABLE 1 .-PROPERTIES OF SOLVENTS aH20 DMSO DMF NMF NMP EG 2PrORdielectric constant, (25') 78.4 46.7 36.7 182.4 32.0 37.7 19.9dipole moment (25") 1.8Sb 3.89c 3.86 3.86 4.09 2.28 1.66(30") (20") (30")refractive index (25") 1.3325 1.4773 1.4282 1.4300 14.680 1.4306 1.3752density/g ~ m - ~ (25") 0.9970 1.0958 0.9440 0.9988 1.0279 1.1100 0.7813(1 Taken from ref.(9) unless specified. b ref. (10). C ref. (1 I).APPARATUS AND METHODSThe densities of the solvents and solvent mixtures were measured by using an AntonPaar precision density meter DMA 2C, at 25°C. Partial specific volumes were calculatedgraphically from the specific volume against concentration plots. The refractive indexdifference was measured at 25.0"C by a Carl Zeiss Jena laboratory interferometer. The cellswere stoppered with Teflon plugs to prevent evaporation or condensation.The opticalrotatory dispersion in the wavelength 300 to 600 nm was recorded by a Jasco J 5 automaticspectropolarimeter at ambient temperature. The bo parameter was obtained as the slopeof the Moffitt-Yang plot. The helix content, xH, of PLLHBr was estimated according toxH = -601630. (1)Equilibrium dialysis was performed by dialysing 5 cm3 of PLLHBr solution sealed in acellulose casing (Visking) against 40 cm3 of the solvent mixture in a 55 cm3 vial sealedtightly with a ground glass stopper and parafilm (Marathon Division, American Can Co).To assist the dialysis, the vial was tumbled at a rate of 8 r.p.m. in a thermostat at 25.00k0.01"C. After 20 h, the solvent mixture was replaced by a fresh one and the dialysis wascontinued for an additional 20 h.The cellulose tubing, which had been treated with boilingwater prior to use, was pre-equilibrated with the solvent mixture by soaking it 5 times in thefresh mixtures. In organic solvent mixtures, especially with low DMSO content, the tubingbecame brittle. However, dialysis was successfully attained without polymer leakage (seebelow for dialysate titration) for about half of the attempted experiments. The polymerconcentration in the dialysis equilibrium was determined by colloidal titration with a1 /400 residue mol dm-3 aqueous potassium poly(vinylsu1phonate) solution, using ToluidineBlue as indicator. Titration of the dialysate showed that polymer permeation through thecasing was < 1 % and gave an error below that from other sources.eqn (WJ .KOMIYAMA, T . MORI, K . YAMAMOTO A N D T . IIJIMA 205PREFERENTIAL BINDING PARAMETERThe preferential binding parameter of a component from the solvent mixture to a polymeris given by the following relation~hip,~where subscripts 1, 2 and 3 refer to water or another principal solvent, polymer and addedsolvent, respectively.l49 l 5 For organic solvent mixtures, DMSO was taken as component 3.ti is the refractive index, rn is the molality of respective component in solvent 1 and C theconcentration in g C M - ~ of solution. V is the partial specific volume, p is the chemicalpotential, T is the temperature, p is the pressure, M is the molecular weight (per repeatingunit for PLLHBr), and the superscript O indicates infinite dilution of the polymer.(an/dC2)~,p,p3 was approximated by (dn/dC2)T,p1,p3 obtained by taking the difference betweenthe refractive index of the polymer solution and the dialysate in eq~ilibrium.~ Linearrelations betwen the increment in refractive index and C2 were found in this and the followingmeasurements. (an/aC,),,,, was measured on the polymer solution and the referencesolvent mixture which had the same molality as component 3. (an/aC3),,.,, was obtainedfrom the plot of the increment in refractive index against dC3. In these measurements, nopolymer was contained in the solutions. The preferential binding parameter calculatedfrom these quantities was denoted by (8rn3/amz),7,,,,.-RESULTSThe differential quantities in eqn (2) for aqueous DMSO, DMF, NMF, NMP,EG and 2PrOH systems are listed in table 2.Table 3 summarizes the results forNMP +DMSO, EG+DMSO and DMF+DMSO mixtures. The last column of tables2 and 3 gives the helix content of PLLHBr in the relevant mixtures. The preferentialbinding parameters for the water + organic solvent systems are plotted against x3,the mole fraction of component 3 in the solvent mixtures, in fig. 1 and 2. Thesefigures show that, for aqueous DMSO, DMF and NMF, the preferential binding of6.01-80o 0.1 a2 0.3 04 0.5 06x3FIG. 1 .-Dependence on solvent composition of preferential binding of organic solvent to PLLHBr :(e) in HzO+ DMSO ; (A) in HzO + DMF ; (+) in H20 + NMF. Curves were drawn by employingthe exchange constants listed in table 4206 PREFERENTIAL BINDING FROM POLAR SOLVENTSTABLE 2.-PREFERE"MAL BINDING OF ORGANIC SOLVENTS TOPLLHBr FROM AQUEOUS MIXTURESsolvent <3>(A) H20(1)+DMS0(3)vol.% x3 C3lgcm-310 0.028 0.110515 0.044 0.168320 0.060 0.222135 0.122 0.394140 0.143 0.443950 0.198 0.553560 0.272 0.670370 0.361 0.776980 0.530 0.911890 0.645 0.974910 0.032 0.116520 0.058 0.199930 0.091 0.288640 0.135 0.386850 0.190 0.485160 0.249 0.570070 0.349 0.676980 0.469 0.7630(B) H20( 1) + DMF(3)(C) HzO(l)+ NMF(3)10 0.033 0.100720 0.071 0.202140 0.170 0.406560 0.314 0.609680 0.550 0.8086(D) HzO(l)+ NMP(3)10 0.028 0.136120 0.053 0.236930 0.087 0.348040 0.126 0.455060 0.228 0.646080 0.433 0.844990 0.693 0.9620(E) H20(1)+ EW3)10 0.034 0.109520 0.074 0.222340 0.176 0.446160 0.326 0.673380 0.561 0.8934(F) HzO(1)-ZPrOH(3)10 0.025 0.078920 0.055 0.158530 0.091 0.240740 0.135 0.322450 0.190 0.403060 0.260 0.476470 0.354 0.561180 0.484 0.638085 0.571 0.6755-V3/cm3 g-10.85920.85800.85910.86100.86340.87490.88590.89730.90850.91 521.00351.00351.00351.00641.01171.02141.03801.04530.96440.96440.96690.98560.99920.92720.92820.92760.92490.94020.95950.97170.86400.86820.87870.89060.89821.11401.13401.15971.19201.20981.23701.25221.27651.27800.2600.3150.3120.3000.3000.2640.2220.1330.0450.0620.1980.2040.1950.1810.1690.1550.1420.1330.1840.1820.1620.1430.1360.1320.1410.1210.1110.1020.0930.0910.1730.1600.1450.1240.1120.1620.1460.1300.1280.1280.1260.1270.1420.1500.1710.1620.1520.1380.1350.1230.1230.1170.1190.1130.1800.1730.1660.1530.1460.1440.1420.1410.1800.1730.1620.1430.1420.1650.1580.1400.1310.1220.1110.1040.1710.1620.1540.1420.1310.1800.1760.1710.1700.1670.1650.1600.1580.1570.1180.1240.1330.1400.1480.1520.1570.1610.1700.1830.1240.1200.1200.1150.1120.1000.0900.0770.1170.1080.0970.0900.0860.1480.1440.1430.1400.1290.1 110.0960.0900.0900.0880.0840.0820.0970.08 10.0730.0620.0460.0360.0320.0250.0182.233.863.984.694.834.814.140.91- 6.79- 6.920.470.920.971.141.150.270.0- 1.470.140.370.00.0- 1.29- 0.54- 0.32- 0.42- 0.54- 0.85- 1.88-4.190.08- 0.09- 0.57- 1.81-4.15- 0.71- 1.57- 2.77- 3.83- 5.76-9.18- 12.1 7-11.7 49- 11.3 85* The content is 0 unless indicateJ.KOMIYAMA, T. MORI, K. YAMAMOTO A N D T. IIJIMA 207the organic component to PLLHBr increases with their contents, then after passinga maximum at around x3 = 0.2, it is inversed by a change to preferential hydrationbetween x3 = 0.3 and 0.5. NMP, EG and 2PrOH are not preferentially bound atany composition but the hydration progresses with their contents. No helix forma-tion of PLLHBr was detected in these mixtures except in aqueous 2PrOH at x3 > 0.3.TABLE 3 .-PREFERENTIAL BINDING OF ORGANIC SOLVENTSTO PLLHBr FROM BINARY MIXTURES(G) NMP(l)+DMSQ(3)20 0.241 0.208130 0.314 0.277140 0.490 0.455160 0.630 0.612580 0.871 0.9073(H) EG( 1) + DMSO(3)10 0.081 0.110220 0.165 0.220540 0.345 0.441860 0.543 0.662780 0.760 0.8805(I) DMF(l)+DMSQ(3)10 0.107 0.109120 0.213 0.218130 0.317 0.327840 0.419 0.437360 0.616 0.652970 0.714 0.763280 0.799 0.874085 0.859 0.936590 0.906 0.98430.91260.91260.91260.91260.91260.90350.90440.90790.91010.80290.89890.90000.90120.90420.91 120.91 140.91 160.91 190.91250.0970.0890.0830.0810.0760.1540.1500.1350.1010.0800.1480.1380.1270.1160.1100.1080.1070.1050.1030.0760.0770.0840.1070.1270.1230.1070.1000.0980.0950.1430.1380.1360.1360.1360.1360.1350.1350.1350.0160.0170.01 80.0230.0250.0450.0480.0560.0590.0600.0440.0440.0450.0510.0530.0570.0720.0850.1084.342.53- 1.74- 6.71- 33.62.053.002.790.34- 5.660.34 1000.0 93- 0.76- 1.74 71-3.24 46- 4.32 -- 5.12- 6.47- 7.79-FIG.2.-Dependence on solvent composition of preferential binding of organic solvent to PLLHBr :(4) in H20+EG ; (0) in H20+NMP ; (A) in H20+2PrOH. Curves were drawn by employingthe exchange constants listed in table 4208 PREFERENTIAL BINDING FROM POLAR SOLVENTSFig. 3 shows the binding parameters for NMP+DMSO and EG+DMSOmixtures while fig.4 shows them for DMF+DMSO system. The helix content ofPLLHBr in the last mixture is also plotted in fig. 4. In these mixtures, the initialpreferential binding of DMSO to the polymer is followed by an inversion at higherx3FIG. 3.-Dependence on solvent composition of preferential binding of organic solvent to PLLHBr :( 0 ) in NMPS-DMSO; (0) in EG+DMSO. Full lines were drawn according to eqn (10) byemploying the exchange constants listed in table 4. Dashed lines show the curves obtained with" calculated " exchange constants given parenthesized in table 4.0? -O Z n03 E .10080_------_ 60402000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 a0 0.9x3FIG. 4.-Dependence on solvent composition of preferential binding of DMSO to PLLHBr and ofhelix content of the polymer: (0) preferential binding parameters in DMF+DMSO; (0) helixcontent. Dashed line (a) represents the curve obtained by employing the " calculated " exchangeconstants given parenthesized in table 4.Dotted lines were drawn by reducing the value of d from12 (a) to 10 (b), 8 (c) and 6 (d), otherwise with the same constants as for curve (a). Full line (e)represents the best fit curve drawn with the exchange constants listed in table 4.DMSO content. The maximum value of the binding parameter depends greatly onthe first solvent. However, if we compare the DMSO binding parameters at x3 = 0.1in fig. 3 and 4, the positive values indicate that the relative affinity of the four organiJ . KOMIYAMA, T.MORI, K . YAMAMOTO AND T . IIJIMA 209solvents to the polymer is in the order, DMSO > DMF > EG > NMP. The sameorder is also found in fig. 1 and 2 when the organic solvent is taken up by the polymerfrom aqueous mixtures at x3 = 0.1. PLLHBr assumes a helix in DMF+DMSOmixtures over 20 vol % DMF and in aqueous 2PrOH over 70 vol % 2PrOH. Wewill discuss the preferential binding behaviour in the former mixture in terms of thischange in next section. Helix formation in aqueous 2PrOH will be discussed in aseparate paper. The drastic decrease in the bulk dielectric constant of this mixturemay eventually provoke the association of bromide ion with the respective chargeson the polymer, probably leading to the helix formation. We note that desolvationmay be accompanied by such association, giving rise to an additional complicationin the preferential binding behaviour.DISCUSSIONThe various portions of the repeating unit of PLLHBr dissolved in a solventmixture, namely the amide group, the methylene groups of the side chain and thecharged polar groups including the amino group and bromide, have different affinitiesfor solvent molecules.Let the polymer be composed of j kinds of portions and thesolvation number per mole of thejth portion be nj, irrespective of solvent species.Then,x i = ni/nj, (3) c x;=1,i= 1,3and(4)where x i and ni are the mole fraction and the moles of the ith solvent in the j t hsolvation layer, respectively. np represents the total solvation number of a repeatingunit.We assume the following exchange equilibrium with the stoichiometricconstant, K-’, between thejth solvation layer and the bulk,K’3.isolvent (3) + solvent (1) - j + solventi.e.,From eqn (3)- (6), it follows :andKj3J = X5X1/X(X3.ni = nhl/(x, +K{,JI -XJ)(1) + solvent (3) - j(6)(7)ni = K i , l n % 3 / { l +(K$,l - l)x3]. (8)Tanford l6 and Inoue and Timasheff l7 have related the preferential binding para-meter to the actual solvation number asFor practical reasons, we confine the number of j to 2, i.e., the polymer residue isreplaced by a model composed of two different portions. The charge conveyin210 PREFERENTIAL BINDING FROM POLAR SOLVENTSgroup of the polymer and bromide (called the polar portion and numbered 2, here-after) and the rest (called the less polar portion and numbered 1) may thus beclassified.A similar postulate has been successfully applied by D'Acray and Wattto the water sorption isotherms of many proteins and synthetic polymers.18Four parameters, two solvation numbers and two exchange constants, aredetermined according to the following procedure. Fig. 5, which replots the pre-ferential binding parameter for the water +2PrOH system against x3/x1, shows thata linear relationship holds between these two quantities so long as the polymerassumes no helical conformation. In terms of eqn (9) and (lo), this fact indicatessimply that En! = 0 or a11 Ki,l = 0, and that the slope of the plot, calculated as 27,is the total solvation (in this case, hydration) number.Then, the largest variationof the binding curve found for the water + DMSO system is approximated iterativelyby assuming an n1 value and calculating the other parameters through the relation(limiting slope of the binding curve at x3 = 0) = (Ki,l - l)nl +(K:,, - l)n2 (11)and eqn (10) with the x3 value at 0 preferential binding. 12 and 15 ( f 0.5) are obtainedfor the solvation numbers of the 1 st and 2nd portion, respectively. The latter numberis associated with the polar portion because of the high affinity to water as judgedfrom the accompanying exchange constant. These numbers are used in calculatingthe exchange constants for all of the other mixtures. Table 4 summarizesthe exchangeconstants for water + organic solvent and two organic solvent mixtures.Fig. 1-4show that curves drawn by eqn (10) with these constants reproduce the preferentialjiml , , , \ , e ( *0 Q1 0.2 0.3 0.4 0.5 0.9 1.3x3 1x1FIG. 5.-Plot of the preferential binding parameter against x3/x1 for the H20+2PrOH system.binding behaviours of all the systems investigated. However, the relative ease offitting a curve with an equation containing two or three adjustable parameters shouldbe noted here. Nevertheless, the discussion below seems to indicate that thepreferential binding behaviour found in this study is interpreted within the frameworkof the present simplified model. The numbers in parentheses in table 4 given for thethree sets of organic solvent mixtures are calculated according to the principle thatthe free energy difference of solvent exchange is determined by the interactionsbetween the solvents and the polymer portion.Then,Kj3.3' = KS,,IKS~,l, (12J . KOMIYAMA, T. MORI, K . YAMAMOTO AND T. IIJIMA 21 1i.e., the constants for a pair of organic solvents can be calculated from those of thecorresponding water + organic solvent mixtures. These “ calculated ” values agreewithin 10 % deviation with the observed values except in the DMF+DMSO case,in which the discrepancy is significantly large. Curves drawn in full and dashedlines show the agreements in fig. 3 and the disagreement in fig. 4 between the com-position ranges.TABLE 4.-EXCHANGE CONSTANTS FOR PLLHBr SOLVATION amixture K1 K2H20+DMSO 11 0.10HzO+DMF 3.4 0.23HzO+NMF 2.0 0.48HzO+EG 1.8 0.33H28+ NMP 1.2 0.60H20+2PrOH 0 0NMP+DMSO 9.0 (9.2) 0.15 (0.17)EGfDMSO 5.5 (6.0) 0.33 (0.30)DMF+DMSO 1.8 (3.2) 0.48 (0.43)a Numbers in parentheses are calculated according to eqn (12)If nothing is observed with the polymer conformation, the preferential bindingbehaviour in the last mixture should be reproduced in terms of the ‘‘ calculated ”constants.In fact, we find in fig. 4 that the three points which lie on x3 2 0.8, whereno helix is formed, locate on the calculated curve, while the points in the helix regiondeviate downward from it, indicating an apparent decrease in the relative affinity ofDMSO over DMF to the polymer on helix formation. Before discussing this, wemention the trends found in the K values. Table 4 shows that K1 of DMSO overwater has by far the largest value ; this is an indication of the high affinity of DMSOto the less polar portion of PLLHBr.Further, the larger K 1 value is concomitantwith the smaller K2, with ths exception of alcohols. These affinity orders apparentlyhave no relation to the protic nature or the electrostatic properties of the solventsshown in table 1.have reported that uncharged PLL also forms a helixin DMF + DMSO essentially to the same extent as found here for the charged polymer.This and the fact that no helix is formed in NMP (dielectric constant is lower thanDMF) + DMSO, indicate that the observed conformational transformation is notcaused by some electrostatic factor. Table 4 shows that among the three solventsmixed with DMSO, DMF is most accessible to the less polar portion and least tothe polar portion.Thus, the helix is formed according to the extent of the replace-ment of DMSO molecules from the vicinity of the polymer main chain. Now, theeffect of the helix formation on the solvations of respective portions of the polymermay be discussed primarily in terms of the exclusion of solvent molecules fromcontact with the less polar portion. Solvent release accompanied by helix formationhas been suggested in a theoretical calculation of the free energy of a coil to helixtransition of polypeptides. On the basis of the exchange constants calculated fromwater + organic solvent mixtures, the preferential binding parameters are calculatedfor the solvation numbers of the less polar portion being reduced stepwise by 2.Fig.4 gives the results, indicating that the observed deviation is explainable if - 6solvent molecules are gradually released on the helix formation. This figure alsoHatano and coworker212 PREFERENTIAL BINDING FROM POLAR SOLVENTSshows that the solvent exclusion reaches its largest value at N 70 % helix content inpreference of the completion of the helix formation. At the moment, these inter-pretations should be taken only qualitatively because of the assumption included,however, they give rise to an interesting point to be investigated further.In summary, the preferential binding parameters measured for PLLHBr inwater +polar organic solvent mixtures are explained in terms of two kinds of solvationlayers around the polymer.This interpretation is extended to the systems com-prising organic solvent binary mixtures. The deviation of the parameters from thepredicted ones in DMFfDMSO is related to the helix formation of the polymerthrough the reIease of several solvent molecules from around the less polar portionof the polymer.We thank Dr. H. Uedaira of Research Institute for Polymers and Textiles forthe use of a Carl Zeiss interferometer.S. N. Timasheff, Accounts Chem. Res., 1970, 3, 62.E. F. Cassasa and €3. Eisenberg, Ado. Protein Chem., 1964, 19, 287.H. Inoue and S. N. Timasheff, J. Amer. Chem. SOC., 1968, 90, 1890.S. N. Timasheff and H. Inoue, Biochemistry, 1965,7, 2501.hi. Morcellet and C. Loucheux, Polymer, 1975, 16, 401.M. Yoneyama, Y. Hanaoka and M. Hatano, 20th Annual Meeting of the Society of PolymerScience (Tokyo, 197 1) and personal communication.M. Oya, K. Uno and Y. Iwakura, Kogyo Kagakic Zasshi, 1966, 69, 741.A. Yaron and A. Berger, Biochim. Biophys. Acta, 1963,69, 397.J. A. Riddick and W. Bunger, Techniques of Chemistry, Vol. II, Organic Solvenfs, ed. A.Weissberger (Wiley-Interscience, New York, 1970).lo Handbook of Chemistry and Physics, ed. R. C . Weast (Chemical Rubber Co., 48th edn.,Cleveland, Ohio, 1968).C. W. N. Cumper, J. F. Read and A. I. Vogel, J. Chcnz. Soc., 1965, 5323.K. Imahori and N. A. Nicola, in Physical Principles and Techniques of Protein Chemistry,Pnrt C, ed. S . J. Leach (Academic Press, New York-London, 1973).l 3 H. Terayama, J. Polymer Sci., 1952, 8, 243.l4 G. Scatchard, J. Amer. Chem. Soc., 1946, 68, 2315.l 5 W. H. Stockmayer, J. Chern. Phys., 1950, 18, 58.l6 C. Tanford, J. Mol. Biol., 1969, 39, 539.l 7 H. Inoue and S. N. Timasheff, Biopolymers, 1972, 11, 737.l 8 R. L. D’Acray and I. C. Watt, Trans. Faraday Soc., 1970, 66, 1236.19 G . Nemethy and H. A. Sheraga, J. Chem. Phys., 1962,36, 3401.(PAPER 6/113
ISSN:0300-9599
DOI:10.1039/F19777300203
出版商:RSC
年代:1977
数据来源: RSC
|
26. |
Binding-induced conformational transition of sodium poly-L-glutamate by iron(III) complex ions in aqueous solution |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 213-229
Mario Branca,
Preview
|
PDF (1281KB)
|
|
摘要:
Binding-induced Conformational Transition of SodiumPoly-L-Glutamate by Iron(r11) Complex Ionsin Aqueous SolutionBY MARIO BRANCA AND BASILIO PISPISA*Istituto di Chimica Fisica, Universitk di Roina, Rome, ItalyReceived 15th March, 1976The binding of pseudo-octahedral trans- and cis-FeIII complex ions by sodium poly-L-glutamate(PLG) and dextransulphate (DS) in aqueous solution at about pH 7 has been studied. Equilibriumdialysis and “ phase-separation ” data show that the affinity for the complex counterions by bothpolyelectrolytes follows the order trans-lFe(tetpy) (OH),]+ > cis-[Febmen) (OH),]+. Evidence ofa specific site binding, leading to a marked “ renaturation ” effect on the charged polypeptide, isproduced in the case of the trans-FeIII-quaterpyridine compound.The binding isotherm of the trans-complex+PLG system and the circular dichroism patterns of the polypeptide as a function of thebound-trans-complex to polymer-residue ratio were successfully treated by a two-state model forthe polyelectrolyte and a preferential association of the complex ions to the helical conformationof the polymeric matrix. All these features are examined in the light of the structural characteristicsof the interacting species. The implications of the different stereochemistry of the other complexions studied on the binding process with PLG are also discussed.Association complexes between transition metal compounds and polyelectrolytesin solution have been the subject of a number of studies either to elucidate the natureof the binding process in a system where forces other than electrostatic ones can playa significant role or to investigate model systems of enzymatic materials or ofna turally-occurring compounds.4-6We present the results of a study on the binding of some iron(rI1) complex ionsby sodium poly-L-glutamate (PLG) and dextransulphate (DS) in aqueous solution.The Fe”’ derivatives were [Fe(pmen) (OH),]+ and [Fe(tetpy) (OH),]+, where pmen =N,N’-bis-(2-pyridylmethyl)-ethylenediamine and tetpy = 2,2‘,2“,2”’-tetrapyridyl.Owing to the marked difference in flexibility between the two quadridentate nitrogenligands, the complexes are expected to assume a different configuration. The formerexhibits a cis-type geometry whilst the latter a trans topology,8 as schematicallyshown in fig.1. Furthermore, since the chiral cis arrangement of the pmen ligandaround the central metal ion gives rise to two enantiomers (A and A),9* lo [Fe(pmen)(OH),]+ ions are a racemic mi~ture.~The aim of the work was first, to study the association process of complex counter-ions with different configurations and a relatively different hydrophobic characterusing dissymmetric substrates with partially hydrophobic properties ; and second,to investigate the correlation between the structural features and the catalase-likeactivity of the resulting association complexes.While a preliminary report on the latter topic has recently appeared,l the bindingdata, obtained by equilibrium dialysis and “ phase-separation ” measurements, aswell as absorption and extrinsic circular dichroism data of the iron(rI1) complex ionsin the polyelectrolyte solutions are presented here.Evidence of specific site binding,leading to a marked “renaturation” effect on the charged polypeptide and to a21 214 BINDING INDUCED TRANSITIONstabilization of the a-helical structure, is also shown in the case of the trans-Fexxl-quaterpyridine derivative. The binding isotherm of the trans-complex + PLG systemand the circular dichroism patterns of the polypeptide in the presence of this complexwere treated by a model which predicts a conformational transition from the coilHH IH T C nHn OHOHok(b)[Fe(pmen) (OH),]+ (A) ; (b) trans-[Fe(tetpy) (OH),]+.FIG. 1.-Schematic representation of the structural features of the complex ions used : (a) cis-or-form to the a-helical structure of the substrate upon progressive binding of complexmolecules.All these features are discussed in the light of a few general considerationsconcerning the structural characteristics of the interacting species.EXPERIMENTALMATERIALSPoly-L-glutamic acid (M = 30 000) was purchased from Miles-Yeda. It was convertedto sodium salt by 0.1 mol dm-3 NaOH. The stock solution was then exhaustively dialysedagainst water to eliminate excess sodium ions and the material recovered by freeze-drying.Dextran ‘‘ sulfate ” (sodium salt, M = 500 000) was purchased from the Sigma Chemical Co.,(U.S.A.) and used without further purification. Concentration of random coil polypeptidesolutions was determined by U.V.absorption at 205 (E = 3200) or 200 nm (E = 5500) anM. BRANCA AND B . PTSPTSA 21 5that of a-helix at 200 nm (E = 3250). Concentration of DS was determined by weight afterprolonged drying under vacuum since this material does not absorb within the accessibleregion of the ultraviolet. Final concentrations of all polymers were in the range of 5~10-4-1 xThe cis compound may beformulated as a mononuclear species both in the solid state and in aqueous solution, whereit undergoes the following protolytic equilibriummol drr3, referred to the monomeric unit.The complexes were prepared as already de~cribed.~~[Fe(pmen) (H,O) OH],+ + [Fe(pmen) (OH),]++ Hf (pKa = 4.4, 2OoC).7The trans FelI1-quaterpyridine derivative is an 0x0-bridged dimeric compound in the solidstate.s In solution, at the pH's and concentrations with which we are primarily concerned,it is almost entirely in the mononuclear form : [Fe(tetpy) This is confirmed bythe results of binding experiments shown later.Analytical reagent grade sodium chloride (C.Erba, Milan) was employed without furtherpurification. Tris(hydroxymethy1) aminomethane (Sigma Chemical Co.) was used asbuffer in the chloride form (tris buffer) at pH 7-8 and in a concentration of 0.01 mol dm--3.Within experimental error, interactions between buffer and complexes or polymers was ruledout on the basis of preliminary optical measurements. In the acid region, the pH of polymersand complex solutions was adjusted each time by adding the appropriate amounts ofstandard acid or base.All measurements were performed on freshly prepared solutions, using double distilledwater with a conductivity < 2.5 x ohm-l cm-l (20°C).METHODSDIALYSIS EQUILIBRIADialysis equilibrium measurements were carried out at 8+0.2"C in 0.01 rnol dm-3 trisbuffer at pH 7.2.The ionic strength was sufficient to depress any Donnan effect. Beforeuse, dialysis tubings (Thomas, U.S.A.) were carefully purified in the conventional way."Dialysis equilibrium was attained within 48 h, with magnetic stirring of the external solutionand only occasional shaking of the dialysis bag containing the polymer solution. At theend of each set of experiments, the concentrations of complex in the external (polymer-free)solution were determined by U.V.absorption at Amax 252 nm (Emax = 10 100) and 210 nm(hax = 42 000) for [Fe(pmen) (OH),]+ and [Fe(tetpy) (OH),]+, respectively, against areference containing equimolar amounts of buffer. The small concentrations of the complexsused (2 x 10-5-5 x mol dm-3) and their low molar extinction coefficients in the visibleregion prevented use of the spectral bands in that region. The concentrations of complexesin the internal (polymer-containing) solution were determined at around 255 and 300 nm(for the cis and trans compounds, respectively), using previously plotted calibration curves.Comparison between total concentration of complex ions before and after dialysis showederrors to be at most $-6 %, not ascribable however to retention of molecules by membranesnor to polymer leakage, according to blank experiments.The concentration of bound molecules, Cb, was calculated as the difference betweencomplex concentrations inside and outside the dialysis bag, the concentration outside beingthat of the free complex ions, Cf.Dialysis equilibrium measurements on the cis complex in PLG solutions containing0.01 mol dm-3 NaCl were also carried out at pH - 5.7, though the experimental data showeda much larger scattering that that found in all other systems investigated.This was probablydue to some decomposition of the complex, not observed in the other cases. On the otherhand, dialysis measurements on [Fe(tetpy) (OH),]++ polymers solutions at pH c 7 couldnot be performed since precipitation phenomena were observed even at very low complexconcentrations.Finally, a set of back-dialyses were also carried out to ascertain whether any selectivebinding occurs in [Fe(pmen) (OH),]++ polymers systems, since this complex is a racemicmixture of two enantiomers (cis-a-A and ~is-a-A).~.' 9 lo For this purpose, suitable samplesat varied complex to polymer ratios and at fixed ionic strength and pH were prepared216 BINDING INDUCED TRANSITIONAfter 24 h they were dialysed at 8°C against known volumes of water at the same ionicstrength and pH. Circular dichroism measurements on the external solutions were thenrecorded at different times, but without detecting any optical activity, within experimentalerror.PRECIPITATION MEASUREMENTSThe critical amount of complex ions (CS) sufficient to induce a phase-separation in themacroionic solutions at different pH and polymer concentration was determined spectro-photometrically, by measuring the turbidity of the solutions at around 10°C.Small amountsof - 0.02 rnol dm-3 solution of complexes were added via a microsyringe directly into the1 cm optical cell containing 2.5 cm3 of polymer at the given pH, all steps being monitoredat 450 or 500nm. At these wavelengths the absorption of the complexes is low and notappreciably perturbed by the presence of polyions. The values of CS and Cp at precipitationwere corrected for the variation in concentration, assuming volume additivity.PONTENTIOMETRIC MEASUREMENTSPotentiometric titration experiments were carried out at 8+0.2"C on 10 cm3 of solutionof poly-L-glutamic acid containing 0.01 rnol dm-3 NaCl, with 0.01 mol dm-3 NaOH.Polyacid samples were prepared by batch-deionization of PLG ~ 0 1 u t i o n s .~ ~ ~ Pure nitrogen,saturated with the solvent, was passed over the sampIe during the titration to eliminatecarbon dioxide. Blank titrations essentially required no base. The reading accuracy was0.01. A G 2222C glass electrode and a K 41 12 calomel electrode were used and preliminarilystandardized with four buffers over the pH range 3-9.APPARATUSAbsorption measurements were performed using a Beckman DK-2A spectrophotometerwith appropriate quartz cells. Circular dichroism (c.d.) spectra were recorded on a Cary 61instrument with quartz cells of 2, 1 or 0.1 cm path-length.The molar absorption coefficient,E, and the molar ellipticity, EL-ER, were obtained on the basis of the total iron concentration.pH values were determined on a Radiometer 26 pH meter using standard semiinicroelectrodes.RESULTSBINDING DATADIALYSIS EQUILIBRIUM DATATypical results of dialysis equilibrium measurements carried out at a fixed polymerconcentration of 7.4 x mol dm-1 in 0.01 mol dm-3 tris buffer (pH = 7.2) areillustrated in fig. 2, where the bound-complex to polymer-residue molar ratio,/3 = (Cb/Cp), is plotted against the concentration of free counterions (Cf). /? valueslarger than those reported could be reached in all cases but the corresponding solutionsshowed some opalescence which affected reproducibility of the data.In the presenceof both PLG and DS, the curves are concave to the free-complex concentrationordinate, indicating that the binding is not cooperative, but the lower part of thebinding isotherm of trans-[Fe(tetpy) (OH),]+ ions on PLG [curve (c)] also exhibitsa trend which suggests a cooperative process, whereby bound molecules facilitatethe association of other molecules. Furthermore, under the same experiinentalconditions, the affinity for the pseudo-octahedral trans compound by PLG is seen tobe higher either than that shown by DS for the same complex ions or than thatobserved for the cis-iron(rrr) derivative by both polyelectrolytes.For a non-cooperative process, e.g., when a localized binding of independenlparticles takes place without influencing the conformational statistics of the polymera Langmuir isotherm may be applied.It holds true if the binding constant does noM . BRANCA A N D B . PISPISA 217change with the charge of the macroion and if both the interactions between boundmolecules are negligible and multilayer absorption does not occur. In such a case,approximating activity with concentration, the association curves are governed bythe relation 6 = B/Bmax = K C,/(l+ K Cf), which can be also expressed as :where K is the intrinsic association constant of each binding site :14BIG = K P m a x - K B (1)[occupied sites][unoccupied sites][free complex ions]K =and Pmax represents the saturation value.Cf x 105/mol dmd30 0.5 1.0 1.5 2.0C 5 10Cf x 1OS/mol dm-3FIG. 2.-Binding isotherm (SOC) of [Fe(pmen) (OH),]+ on PLG (curve a) and DS (b) and of [Fe(tetpy)(OH),]+ on PLG (c) and DS (d), as bound-complex to polymer-residue molar ratio (/3 = c b / c p )as a function of free complex concentration, Cf.The larger symbols refer to the data obtained fromphase-separation experiments (see text and fig. 6). Polymer concentrations 7.4 x mol dm-3 intris buffer 0.01 mol dmL3 (pH = 7.2).TABLE 1 .-INTRINSIC ASSOCIATION CONSTANT, K[eqn (2)], GOVERNING THE BINDING OF THEIRON(III) COMPLEXES TO THE POLYELECTROLYTES AT 8°C (pH = 7.2) AND SATURATION VALUESOF BOUND-COMPLEX TO POLYMER-RESIDUE MOLAR RATIO [eqn (l)]complex polymer K/dm3 mol-1 Pmaxcis-[Fe(pmen) (OH) DS 6 . 5 ~ 103 0.95cis-[Fe(pmen) (OH),]+ PLG s.1~103 0.95trans-[Fe(tetpy) (OH),]+ PLG 1 .4 ~ lo4 * 1 atrans-[ Fe(tetpy) (OH),]+ DS 2.3 x 104 1.1a evaluated from eqn (3) (see fig. 5 and text)The relations between (PICf) and p based on eqn (1) are linear for all the systemsinvestigated (fig. 3), except for the one formed by the trans complex in PLG. Thevalues of the association constant (which indicate a free energy of binding of som21 8 BINDING INDUCED TRANSITION5 kcal mol-' of complex, corresponding to a strong-binding process)13* l4 and thoseof Pmax are summarized in table 1.Under the experimental conditions used, PLG is fully ionised [fig. 4(a)]. It isnot surprising, therefore, that all saturation values of bound-complex to polymer-residue ratio are around unity (table 1).This implies that each charged group inboth polylectrolytes behaves as an effective binding site.0 1 I 1 I I 1 IP = c b / c pFIG. 3.-The relations based on eqn (1) between (@/Cf) and /3 at pH = 7.2 (tris buffer 0.01 mol dm-3)for [Fe(pmen) (OH),]+ in PLG (a) and DS (b) solutions and for [Fe(tetpy) (OH),]+ in DS solutions (c).4.8 '14.61 I I I4 5 6 7 0 0.2 04 0.6 0.8 1.0FIG. 4.-(a) Titration curve of poly-L-glutamic acid in 0.01 mol dm-3 NaCl plotted as the degreeof dissociation ac (for the reaction R-COOH 4 R-COO-+ H+) against pH (S'C) ; (b) plot ofthe same data as pH-log[cr/(l -or)] as a function of a.On the other hand, the peculiar trend of the binding isotherm of the [Fe(tetpy)(OH)2]++PLG system [fig. 2(c)] can be ascribed, in principle, to different effects.For example, it may be thought that a dimerization process between the planarmolecules takes place upon association to poly-L-glutamate.However, when it isfound that the curve cannot be linearized by plotting @/C:) as a function of fl,according to the relation (PIC?) = K'P,,,-K'P which refers to the equilibriuM. BRANCA AND B . PISPISA 2192 C+P + (C,P), such an hypothesis is ruled out. This also agrees with the lack ofany cooperative behaviour in the association between the trans complex ions and DS,at the same pH (fig. 2 and 3).The binding isotherm of the [Fe(tetpy) (OH),]+ + PLG system was then treatedby a two-state model for the polyelectrolyte, taking into account a change in theassociation constant with the conformational state of the polymer.Poly-L-glutamicacid has been demonstrated to undergo a sharp helix-coil transition in aqueoussolution as the charge distribution on the polymer varies.15* l6 This is illustratedin fig. 4(b) where the data of fig. 4(a) are replotted as pK, = pH-log [a/(l -a)]against a (degree of dissociation for the reaction R-COOH + R-COO-+H+), sothat the pH region of the cooperative transformation is shown.Since the conformational equilibrium between, say, state a and b is a functionof a, the average fraction xio) of polymer molecules present in state a may be derivedfrom potentiometric titration data. Furthermore, it has been shown that the averagenumber of monomers S in the helical (or coil) region of the chain in the middle ofthe transition is around 20.15b In a first approximation, an aqueous solution ofpoly-L-glumatic acid may therefore be represented as a system formed by Mindependent and identical units per unit volume, consisting of S monomeric unitswhich can be either in the a or b state depending upon the experimental conditions.On this basis and on the assumption that each it4 unit may bind s < S complex ions,the binding isotherm of the trans complex in PLG solutions was treated accordingto the following expression, which describes the binding process in terms of theconformational statistics of the polymer :179 l8z 1 + L'O'hy"- 1l + z l+L'o'y" *e = - (3)Superscript (0) indicates that the quantity refers to a system where no complex ispresent and z = K, x a,, where K, is the association constant between complex ionsand the units in the a state while a, is the activity of free complex ions in solution.0 0.5 1.0 1.5Cfx 105/mol dm-3FIG.5.-Experimental data (0) and calculated curves according to eqn (3) (solid line) and to theLangmuir isotherm (broken line) for [Fe(tetpy) (OH),]++ PLG solutions at pH = 7.2 (tris buffer0.01 mol dm-3) : see text.Furthermore, L(O) is the equilibrium constant between the two conformational statesof the polymer and hence depends on the state of charge of the Munits (ie., L(O) =q$,o)/qAo), where qio) and qio) are molecular partition functions for any M units in theb or a state, respectively), h = KJK, and y = [(l +hz)/(l +z)].A detailed description of the statistical approach used to derive eqn (3) isknown,17* l8 though it is worth stressing that the lower ac is, the more it deviates fro220 BINDING INDUCED TRANSITIONa Langmuir isotherm.For instance, if h c 1, y is smaller than unity but increasesmonotonically as z decreases. For very low values of a,, z << 1 and y - 1, so that8 becomes a linear function of z with slope [(I +L(O)h)/(l +L(O))].Using a value of L(O) = 30, which is compatible with the results of titration experi-ments (fig. 4) and assuming S = 20, the other parameters were adjusted by a trialand error procedure to fit the experimental data, as shown in fig. 5. Approximatingactivity with concentration, owing to the very low concentration of the complex insolution, both the Langmuir isotherm and that based on eqn (3) (with same K, =7 x lo4 mol dm-3) are plotted, h and Pmax being 0.2 and 1 , respectively.Despite the crudeness of the fitting procedure, the results are satisfactory andindicate that the affinity for the trans complex by helical polypeptide is much higherthan that by the polymer in coil, the latter being of the same order of magnitude ofthat found for the same complex by DS (table 1).PHASE-SEPARATION DATAA typical plot of the results of " phase-separation " measurements is illustratedin fig. 6.The concentration of trans complex ions (Cs) sufficient to induce yrecipita-tion in the polyelectrolytes solutions, at different ionic strength and pH, is reportedas a function of the polymer concentration (Cp).The results in the figure may beexpressed as Cs = Cf+p' x Cp, where Cf is the concentration of free complex ionsand /?' is the amount of bound counterions per polymer-residue at precipitation,that is, it represents a measure of the association between macroions and complex0 5 toCp x 104/mol dm-3FIG. 6.-Critical amount of trans-complex ions (Cs) necessary to induce a phase-separation in PLGor DS solutions at different concentrations, Cp. Curve 1: PLG in tris buffer 0.01 mol dm-3(pH = 7.2); 2: PLG in tris buffer 0.01 mol dm-3 and NaCl 0.01 rnol dm-3 (pH = 7.2); 3 : DSin tris buffer 0.01 mol dm-3 (pH = 7.2) ; 4 : PLG in NaCl 0.01 mol dm-3 (pH = 6.1 k0.1).ions at that point.'". l9 In agreement with a number of e ~ a r n p l e s , ~ ~ ~ 2o these datashow that the binding decreases as the ionic strength increases.The lowest degreeof association, i.e., the highest Cf value, is however observed for the cis compound inPLG solutions at pH around 7.2 (data not reported here). Since the charge densityof both complex ions should be similar, this finding suggests that forces other thanelectrostatic ones play a significant role in the binding process between [Fe(tetpy)(OH),]+ ions and poly-L-glutamateM . BRANCA AND B . PISPISA 221The results of these experiments are in good agreement with those of dialysisequilibria in the sense that, within experimental error, the value of fl' falls on theisotherm curves. For instance, the values of cb and Cf at the precipitation for[Fe(tetpy) (OH),]+ ions in a PLG buffered solution at 7.4 x in01 dm-3 (pH = 7.2)are : C,, = 39.2 x mol dm-3 and those for the same complexin DS solution are : cb = 26.0 x mol dm-3.All these features indicate that the precipitation data are complementary to thedialysis data.They lead to the conclusion that, independent of polymer concentra-tion, (i) the affinity of both charged polymers for the complex counterions is [Fe(tetpy)(OH),]+ > [Fe(pmen) (OH),]+ and (ii) the affinity for the trans complex follows theorder: a-helical-PLG > coil-PLG N DS. Finally, it is interesting to note thatthe observed values of p' are generally lower than those of Pmax derived from equili-brium dialysis measurements (fig. 3 and table 1). For example, at pH = 7.2,p' = 0.35 and 0.53 whilst Pmax is about unity for both the trans-complex+DS and + PLG systems, respectively.This finding suggests that interchain interactions areoccurring, which determine the aggregation, even before a complete '' neutralization "of the fixed charges in the polymers by the bound molecules is reached.16and C, = 1.8 xand Cf = 2.0 xSPECTRAL AND OPTICAL DATAVISIBLE AND NEAR-ULTRAVIOLET C.D. SPECTRAFor brevity, we confine the discussion to the most significant results which arethose obtained with the complex+PLG systems. Typical c.d. spectra in the visibleand near-u.v. region are presented in fig. 7 and 8, where the dichroic bands originatesolely from the electronic transitions of the bound molecules, since poly-L-glutamateabsorbs only below 250 nm.Note that, under the same experimental conditions,the values of the rotational strength displayed by the bound cis counterions aremarkedly lower than those of the trans complex ions. This finding reflects the factthat the former compound (at variance with the latter which is not dissymmetric)is a racemic mixture of two enantiomers (A and A).'. ' 9 lo On the assumption thatno selective binding occurs,1* as confirmed by back-dialysis data (see experimental),the c.d. spectral patterns of the cis complex in PLG solutions are therefore differencespxtra, e.g. spectra of a mixture of diastereoisomers, owing to the fact that theinteractions of the enantiomers of an asymmetric molecule with a dissymmetricenvironment are different.22 It follows that the optical activity in the trans complexis induced by a dissymmetrical perturbation of the chromophoric electrons of themolecules bound to the asymmetric substrate whereas the optical activity in the ciscompound arises from the same type of perturbation which affects (differently) theelectronic transitions of the D and L forms of the bound counterions.When small molecules are ligated to dissymmetric macromolecules, it has beenshown that induction of optical activity may be achieved by different rnechanisni~,'~not mutually exclusive.The most significant ones are those associated with a chiralarrangement of molecules bound to the structured polymer. In such a case, adegenerate or a non-degenerate exciton interaction between adjacent molecules mayThe former is characterized by the coulombic coupling of an electronicexcitation in one molecule with that of the same energy in another, the latter by thecoupling between two electronic excitations of different energy of two adjacentmolecules.In terms of c.d. spectra, the degenerate exciton effect gives rise to twocod. bands of opposite sign associated with the same absorption band, whereas thenon-degenerate exciton interaction gives rise to a single c.d. band for each absorptionband, the c.d. bands of the two coupled transitions still being opposite in sign.2 222 BINDING INDUCED TRANSITIONObviously, in both cases the rotational strengths decrease as the bound-molecule topolymer-residue ratio decreases,l e. g. as the average distance between the ligatedmolecules increases.By inspection of fig.8, it appears that : (i) the two major c.d. bands of the transZomplex fall at frequencies where the maximum of the absorption bands is found;[ii) the sign of the rotational strength of these bands is opposite in sign; (iii) therotational strength of both bands markedly increases as the a-helical fraction (fh) inthe polypeptide increases, i.e. as the pH is lowered.A/nmFIG. 7.-Circular dichroism (a) and absorption (b) spectra of cis-complex ions in poly-L-glutamate(curves 1-3) or pOly-D-glUtamate (curve 4) solutions, at different pH values. Curve 1 : pH = 7.4 ;2 : pH = 6.4; 3 and 4 : pH = 5.7. Complex to polymer molar ratio (C/P) = 0.20; polymerconc. 7.44 x mol dmA3.Unfortunately, we could not investigate the dependence of the rotational strengthon p at fixedf,, since it was necessary to go to pH values as low as - 5, where verysmall amounts of complex ions are sufficient to determine ‘‘ salting-out ” phenomena.In all other cases, an increase in p brings about an enhancement of the helicalfraction in the polyelectrolyte (see below), which is reflected in an increase of therotational strength of the c.d.bands of the bound complex ions. Nevertheless,these results strongly suggest that optical activity in the bound trans counterions isprimarily induced by a non-degenerate exciton mechanism, though the complexityof the spectra would indicate the occurrence of additional effects. For example, acomplementary mechanism, which sees the y-carboxylate groups of the side-chainMa BRANCA AND B .PISPISA 223acting as unidentate ligands, could also be taken into account. A coordination ofthese groups in the apical position of the complex is sterically allowed and may accountfor the relatively high ellipticity observed even at about pH 7.6, where the binding-induced ‘‘ renaturation ” effect on poly-L-glutamate is not relevant (see below).250 300 3 50 L 00 L 50h/nmFIG. 8.-Circular dichroism (a) and absorption (b) spectra of trans-complex ions in PLG solutions,at different pH values. Curve 1 : pH = 7.6; 2 : pH = 6.7; 3 : pH = 6.0. (C/P) molar ratio =The same hypothesis probably does not apply to the cis complex, however, sincemolecular models suggest that in this case the steric requirements for the formationof such a coordination bonding are not matched.On the other hand, the very lowvalues of the rotational strength of the c.d. bands, together with the occurrence of achange in both the spin-state and configuration of this complex as pH increases,’make it difficult at present to interpret the c.d. data of fig. 7.0.10 ; polymer conc. 7.4 x mol dm-3.ULTRAVIOLET C.D. SPECTRATypical c.d. spectral patterns of [Fe(tetpy) (OH),]+ + PLG solutions at about pH 7and at varied bound-complex to polymer-residue ratios are presented in fig. 9224 BINDING INDUCED TRANSITIONtogether with the c.d. spectrum of poly-L-glutamate at the same pH without addedcomplex ions. The main features of these spectra are the variation in the U.V. regionobserved upon addition of the complex.The 215-217 (R > 0) and 196 nm (R < 0)c.d. bands of PLG, typical of a random coil, disappear while the 220 and 208-210 nrn( R < 0) c.d. bands, typical of an a-helix, appear and strongly increase depending on thevalues of p, which have been derived from equilibrium dialysis measurements atpH = 7.2.In fig. 10 the change in ellipticity at 220nm of poly-L-glutamate in the absenceand in the presence of both complexes is plotted against pH. In this case too thetrans complex determines a stabilization of the a-helical structure in the polypeptidein terms of an upward shift in the pH region of the helix-coil transition, while additionof the cis compound even at a higher complex-to-polymer (C/P) ratios does not- 7apparentlyo aio 0.20 - 1 1I 1I I p ; y P1 ' 1- I$I I I I 1affect200 300 GOOA/nmFIG.9.-Ellipticity of [Fe(tetpy) (OH2)]+ + PLG solutions (pH = 7) at a bound-complex to polymer-residue ratio #3 of 0.095 (curve 2) and 0.15 (3). Curve (1) : ellipticity of poly-L-glutamate withoutadded complex ions at the same pH. All measurements were normalized for an optical path lengthof 1 cm. Complex concentration : (2) 7.55 x lo-' mol dm-3 ; (3) 1.17 x mol dm-3 ; PLGconc. 7.55 x mol dm-3. Insert : variation of ellipticity at 220 nm as a function of fi (pH = 7) :trans-complex-PLG (A) and cis-complex-PLG (0) systems. The filled symbol refers to PLGwithout added complex ions.The relevance of these results is that only the trans complex has the ability toinduce a conformational change in the poly(amino-acid), which in turn suggests theoccurrence of a specific site binding.The data of fig.10 indicate that in the presence of trans counterions the curvelevels off at a value of ellipticity lower than that of helical polypeptide-free complexsolutions. The same effect is observed when the ellipticity at 220 nm is reported asa function of the bound-complex to polymer-residue ratio, at fixed pH (insert offig. 9), although in both cases the trend of the curves suggests that the polymer haM. BRANCA AND B. PISPISA 225almost attained the maximum helical content, under the reported expximentalconditions. A perturbation in the corresponding transition of the polypeptide, dueto the association, or the presence of an induced c.d.band of opposite sign in thebound complex within the same frequency region, or both effects, may be responsiblefor such a result. Although the former hypothesis was invoked to explain similarbehaviour in other complex-poly (amino-acid) system^,^ the latter effect is predominanthere. At fixed pH values < 6, when the binding-induced renaturation effect is ofminor importance, an increase in the complex to polymer molar ratio (C/P) deter-mines a decrease in the rotational strength of the c.d. bands of the polypeptide. Thisimplies that the higher the amount of bound complex, the more the induced dichroicband of opposite sign subtracts rotational strength from the negative ones at 220and 208 nm. For example, the estimated value of ( E ~ - E ~ ) at 220 nm of helicalpolypeptide at pH - 5, in the presence of trans complex ions, is about - 11 at a(C/P) ratio of 0.05 and -9 at a (C/P) ratio of 0.10.-,51 I I I 1 1PHFIG.10.-Variation of ellipticity at 220nm of PLG solutions (0) and of PLG in the presence ofcis-(.) and trans-complex ions (0) as a function of pH. Complex to polymer molar ratios: 0.154 5 6 7 8 9(0) and 0.10 (0) ; polymer conc. - 7.4 x mol dm-3.These values of ellipticity were employed to evaluate the a-helical fraction in thecharged polymer at bound-trans-complex to polymer-residue ratios correspondingapproximately with the (C/P) ratios given above. For this purpose, the conventionalexpression : = fh hh + (1 -fh) hc was used, where subscripts h and c designatethe helical and coil form, respectively, whilst the values of are obtained fromthe c.d. data reported in the insert of fig.9. The results are shown in fig. 1 1 togetherwith the relation between the a-helical fraction and the amount of bound trans-complex ions, calculated by eqn (4) : *x, = l/(l+L‘O’Y”). (4)I-226 BINDING INDUCED TRANSITIONAccording to the statistical treatment previously seen, x, represents the fraction ofall units in the a state, under the reported experimental conditions of bindingequilibrium. The curve of x,, based on eqn (4), was calculated using the same para-meters adopted for the binding data of fig. 5, i.e. L(*) = 30, h = 0.2, Pmax = 1,S = 20 and K, = 7 x lo4. Despite the approximation in evaluating the a-helicalfraction by c.d.measurements, owing to the uncertain values of A E ~ and Act, thefitting in the figure appears surprisingly good. Furthermore, under the experimentalconditions used, the coil-helix transformation occurs at a relatively low bound-complex to polymer-residue ratio, i.e. about 0.10 (x, = 0.50). This implies a highefficiency of the added compound in determining the conformational change in PLGand suggests strongly that a multibonding interaction characterizes the mode ofassociation in the system considered. In contrast, at the same pH a much largeramount of bound cis-complex ions is probably needed to perturb the conforniationalequilibrium of poly-L-glutamate (insert of fig. 9).0 010 0.20 030 0,LOfi = c b / c pFIG.11.-Conformational change in the charged polypeptide (pH = 7) as a function of bound-trans-complex to polymer-residue ratio, p. The solid line is calculated according to eqn (4) whilstthe experimental points (A) are evaluated from c.d. measurements at 220 nm (see text).The results of binding and circular dichroism measurements indicate, therefore,a different degree of participation of the complex ions studied in the associationprocess with PLG. Whether a strict comparison can be made between the complexcounterjons used is, however, debatable since, under the same experimental conditions,the fraction of bound molecules is different and this may well depend not only ontheir different physical characteristics but also on their different stereochemicalfeatures, all of these being responsible for a diverse degree of binding.DISCUSSIONFew examples of the behaviour of complex ions in aqueous solutions of poly-electrolytes have been reported.Furthermore, despite the large number of studieson the interaction between small molecules and macromolecules in solution, in fewinstances have bound molecules been shown to affect the conformation of the poly-meric sub~trate.l-~. 17* 24-26 Recently, it has been reported that acridine orange (AO)exerts a stabilization effect on the helical structure of poly-L-glutamic acid in termsof an upward shift in the pH region of the helix-coil tran~ition.~~ Moreover, on thebasis of a previously proposed model of PLGA-A0 complex in acid solution,28 asuper-helical arrangement around the core of the a-helical polypeptide was thoughtto occur even in the alkaline pH region, at a glutamyl-residue to dye molar ratioof about unity.2M.BRANCA AND B . PISPISA 227As far as we know, however, no other example of binding-induced conformationalchange in PLG has been reported, though many efforts were made to elucidate thestructural features of the association complex between a-helical PLGA and planardyes in acid s o l ~ t i o n . ~ ~ - ~ ~The relevance of the results described here is that the two pseudo-octahedraliron(@ complexes investigated show quite different behaviour in poly-L-glutamatesolutions, probably depending on their different configuration (fig. 1). The observa-tion that only the trans-[Fe(tetpy) (OH),]+ ions determined a marked " renaturation "effect on the charged polypeptide may be phenomenologically explained in terms ofa preferential binding to helical conformation (table 1 and fig.5), which brings abouta shift of the conformational equilibrium towards the orderly form of the polymer.This would in turn facilitate the binding of additional molecules, with a resultingincrease in the a-helical fraction in the polyelectrolyte (fig. 2, 5 , 9 and 11).Molecular models support this conclusion in that they suggest that the transmolecules can approach the polymeric substrate closely enough to make it possiblethat an apical site of the complex matches a y-carboxylate group of the side-chainswithout any severe steric hindrance. Therefore, the formation of a coordinationbonding between the two species may be reasonably considered.As a result, onemay expect depression of the electrostatic repulsion between carboxylate anions and,at the same time, the occurrence of extensive interactions between the tetrapyridylligand and neighbouring side-groups of the polymer, probably also involving hydro-phobic forces. The overall effect is such that a relatively large portion of substrateis engaged by each bound molecule. Note that such intimate contact would bebetter achieved with the a-helical structure than with the coil form since the overallflexibility of the former is lower than that of the latter, the stereochemical requirementsfor a more extensive and possibly strong interaction being thus optimized.Incontrast, the cis topology of [Febmen) (OH),]+ makes a close approach of theseions to the polymeric substrate difficult, an intimate association complex being inthis case overcrowded.In fact : (i) itjustifies the much higher affinity of the a-helical- than the coil-PLG for the transcomplex, as indicated by the values of the binding constants (table 1 and fig. 5 ) ; (ii)it explains both the stabilization effect exerted by the same complex counterionstowards the a-helical structure and the efficient determination of the conformationalchange in the charged polypeptide (fig. 9, 10 and 11) ; (iii) it explains to some extentwhy even the affinity of coil-PLG for the cis compound is lower than that for thetrans derivative ; (iv) it justifies the relatively high values of induced optical activityobserved when the trans complex is ligated to the macroion at pH values where thebinding-induced " renaturation " effect is not relevant (fig.8) ; (v) it accounts forthe " phase-separation " phenomena already observed at low concentrations of thetrans compound (fig. 6).Considering point (v), the equatorial plane of the bound trans molecules must lieparallel to the helical axis of the polypeptide. According to molecular models, anyother arrangement is sterically hindered. A similar stereochemistry has been proposedfor the association complex between haemin and helical poly-~-lysine,~ whose side-chains are even longer than those of poly-L-glutamate. It is reasonable to suppose,therefore, that the trans complex molecules ligated to the macroion can easily interactwith the carboxylate anions of other chains, through the " external " apical site.This should give rise to an insoluble coacervate, as experimentally observed.Lineardichroism measurements on streaming solutions of trans-complex + PLG systems willbe carried out to confirm such an hypothesis.This hypothesis is fully consistent with the experimental results228 BINDING INDUCED TRANSITIONIn conclusion, the results show the occurrence of specific site binding betwentrans-[Fe(tetpy) (OH),]+ ions and poly-L-glutamate, which probably involves acoordination bonding between the y-carboxylate groups of the side-chains and themetal ion of the complex, besides electrostatic and hydrophobic interact ions.Specificity in such a process arises from the topological characteristics of this complexwhich are suitable for an intimate contact with the substrate. This produces in turna conformational transition in the polypeptide in terms of a marked increase ofx-helical structure at those pH values where the coil form normally predominates.The trans-[Fe(tetpy) (OH),]++PLG system may be thus considered as a simple modelfor the allosteric effects attributed to enzymes.The different stereochemical featuresof the cis-[Fe(pmen) (OH),]+ molecules are responsible for the different degree ofparticipation of these counterions in the binding process with the charged polypeptide,which leaves the secondary structure of the macroion unperturbed.We thank Prof.E. Torracca and Dr. M. E. Marini for helpful discussions.(a) F. Ascoli, M. Branca, C. Mancini and B. Pispisa, J.C.S. Faraday I, 1972, 68, 1213 ; (6)Biopolymers, 1973, 12, 2431.M. Barteri, M. Branca and B. Pispisa, Biopolymers, 1974, 13,2161.G . Blauer, Nature, 1961, 189, 396 ; Biochim. Biophys. Acta, 1964, 79, 547 ; G. Blauer and 2.B. Alfassi, Biochinz. Biophys. Acta, 1967, 133, 206 ; G. Blauer and A. Ehrenberg, Acta Chem.Scand., 1963, 17, 8.J. H. Wang and W. S. Brinigar, Proc. Nut. Acad. Sci. U.S.A., 1960, 46, 958 ; J. H. Wang,Accounts Chem. Res., 1970,3, 90.L. Stryer, Biochim. Biophys. Acta, 1961, 54, 397 ; M. Tohjo and K. Shibata, Arch. Biochern.Biophys., 1963, 103, 401 ; T. King, F. C. Young and S. Takemori, B ochem.Biophys. Res.Comm., 1966,22,658.M. Branca, B. Pispisa and C. Aurisicchio, J.C.S. Dalton, 1976, 1543.Inorg. Chem., 1970,9, 1.New York, 1971).’ M. Barteri, M. Branca and B. Pispisa, J.C.S. Dalton, 1974, 543.’ M. Branca, P. Checconi and B. Pispisa, J.C.S. Dalton, 1976, 481.l o See, for example, C. J. Hawkins, Absolute Confisuration of Metal Complexes (Wiley-Interscience,l 1 M. Barteri and B. Pispisa, Gazzetta, 1976, 106, 499.l2 P. McPhie, in Methods in Enzymology (Academic Press, New York, 1971), vol. 22, p. 25.l3 A. Blake and A. R. Peacocks, Biopolymers, 1968, 6, 1225.K. Jackson and S. F. Mason, Trans. Faraday SOC., 1971, 67,966.Is See, for example : (a) D. S . Olander and A. Holtzer, J. Amer. Chem. Soc., 1968, 90, 4549 ;Y . M. Mayer, Macromolecules, 1969, 2, 624; (b) V. E. Bichkova, 0. B. Ptitsyn and T. V.Barskaya, Biopolymers, 1971, 10, 2161, and references cited therein.J. Monod, J. Wyman and J. P. Changeux, J. Mol. Biol,, 1965, 12, 88.M. Mandel and W. H. J. Stork, Biophys. Chem., 1974,2, 137.I 6 M. Barteri and B. Pispisa, Biopolymers, 1973, 12, 2309.l 9 U. P. Strauss and A. Siegel, J. Phys. Chem., 1963, 67, 2683.’O H. R. Mahler and G. Green, Ann. N. Y. Acad. Sci., 1970, 171 (3), 783.’l F. P. Dwyer, M. F. O’Dwyer and E. C. Gyarfas, Nature, 1951, 167, 1036.22 C. Mavroyanis and M. J. Stephen, Mol. Phys., 1962,5,629 ; B. Bosnich, J. Amer. Chem. Soc.,1966, 88,2606 ; 1967, 89, 6143.23 (a) H. Eyring, H.-C. Liu and D. Caldwell, Chem. Rev., 1968, 68, 525; D. J. Caldwell andH. Eyring, The Theory of Optical Activity (Wiley-Interscience, New York, 1971) ; (b) I. TinocoJr., Adv. Cliem. Phys., 1962, 4, 113 ; R. W. Woody and I. Tinoco Jr., J. Chem. Plrys., 1967,46,4927.24 J. M. Rifkind and G. L. Eichhorn, Biochemistry, 1970, 9, 1753.25 I. M. Klotz, G. P. Royer and A. R. SIoniewsky, Biochemistry, 1968, 8, 4752; I. M. Kilobz26 K. Wuiff, H. Wolf and K. G. Wagner, Biochem. Biophys. Res. Comm., 1970, 39, 870; K. G.and J. U. Harris, Biochemistry, 1971, 10, 923.Wagner and K. Wulff, Biochem. Biophys. Res. Comm., 1970, 41, 813M . BRANCA AND B . PISPISA 22927 Y. Sat0 and M. Hatano, Bull. Chem. SOC. Japan, 1973, 46, 3339 ; M. Hatano, M. Yoneyania28 R. E. Baliard, A. J. McCaffery and S. F. Mason, Biopolymers, 1966, 4, 97.29 L. Stryer and E. R. Blout, J. Amer. Chem. Soc., 1961, 83, 1411.30 B. C. Myhr and I. G. Foss, Biopolymers, 1971, 10, 425.and Y . Sato, Biopolymers, 1973, 12, 895, and references cited therein.(PAPER 6/499
ISSN:0300-9599
DOI:10.1039/F19777300213
出版商:RSC
年代:1977
数据来源: RSC
|
27. |
Solutions of lithium salts in liquid lithium. The electrical resistivity of solutions of nitride, hydride and deuteride |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 230-235
Paul F. Adams,
Preview
|
PDF (440KB)
|
|
摘要:
Solutions of Lithium Salts in Liquid LithiumThe Electrical Resistivity of Solutions of Nitride, Hydride and DeuterideBY PAUL F. ADAMS, MICHAEL G. DOWN, PETER HUBBERSTEY ANDRICHARD J. PULHAM*Department of Inorganic Chemistry, The University, Nottingham NG7 2RDReceived 14th April, 1976The electrical resistivities, p , of solutions of lithium nitride, lithium hydride and lithium deuteridein liquid lithium have been determined for concentrations, x, up to 2.77, 5.68 and 2.22 mol % non-metal over the temperature ranges 200-450, 257-551 and 276-50O0C, respectively. For each solute,resistivity increases linearly with increasing concentration, except for very dilute solutions, and thecoefficient, dpldx increases with increasing temperature. Nitride causes the greatest increase inresisitivity [dp/dx = 7 .0 ~ lo-* a m (mol % N)-l at 400°C], and hydride and deuteride show no de-tectable isotope effect [dpldx = 4.9 x Qm (mol % H or D)-l at 400°C]. The resistivities ofmixtures of nitride and hydride in lithium are additive, thereby showing lack of association betweenthese solutes. Ammonia vapour reacts with the metal to form hydride and nitride which dissolve toincrease the resistivity by their characteristic amounts.We noted previously that the electrical resistivity of liquid lithium was appreci-ably dependent on the degree of purification ; metal which had been scavenged of thecommonly occurring oxide, hydride and nitride salts had the lowest resistivity. Wenow describe the specific effects of nitride, hydride and also deuteride on the resistivityof the liquid metal as a function of both concentration and temperature.The resultsshow that changes in the resistivity of the liquid provide an elegant method of follow-ing changes in the concentration of non-metal solutes, particularly if there is no chemi-cal interaction between them. This is the case for mixtures of nitride and hydride,whether added separately or together in the form of ammonia. With other solutecombinations, however, we believe that resisitivity will prove as valuable with lithiumas with sodium at unravelling the nature of solvation and chemical reactions in theliquid metal. Since lithium is a potential candidate as coolant and/or tritiumbreeder in the fusion reactor, such reactions might influence the efficiency of purifica-tion of the metal for this purpose.EXPERIMENTALThe apparatus and method for the determination of resistivities of solutions of salts inmetals have been described previ~usly.~ Nitrogen(Air Products, 99.98 %) and deuterium (Matheson Gas Products, 99.9 %) were purified asbefore and exposed in small volumes (- lo4 m3 at S.T.P.) to liquid lithium (30 g, Koch-Light, 99.98 %) which had been purified in situ by gettering with yttrium sponge at 400°Cfor 72 h.The gases reacted rapidly with the metal to form the corresponding salts whichsubsequently dissolved. Resistivities were determined under equilibrium conditions as afunction of solute concentration at selected, constant, temperatures.(Air Products, 99.99 %), hydrogenRESULTS AND DISCUSSIONCOMPOSITION AND TEMPERATURE DEPENDENCEResistivity isotherms were determined at fourteen, eleven and six temperaturesin the ranges 200-450, 257-551 and 276-500°C for nitride, hydride and deuteride23P . F.ADAMS, M. G . DOWN, P. HUBBERSTEY AND R . J . PULHAM 231solutions, respectively. Typical isotherms, illustrated in fig. 1, show that the resis-tivity increased slightly at first, then more steeply, and linearly, up to saturation,beyond which it remained constant. The change in resistivity at saturation is excep-tionally clear and provides, therefore, a precise method of measuring the solubilities403836 E s 0.m2 343230concentration, x/mol% N, H or DFIG. 1 .-(Resistivity, concentration) isotherms (at 400°C) for solutions of nitride (O), hydride (0)and deuteride (+) in liquid lithium.of the salts.These have been reported earliera3* The dominant feature is theextensive linear increase in resistivity with solute concentration which indicates thateach solute ion scatters electrons to the same extent as its predecessor. DissolvedTABLE 1 .-COEFFICIENTS IN THE (RESISTIVITY, COMPOSITION) EQN (1)eJ"c1 Osm/Qmnitride solutions108 dp/dr/Qm(mol % N)-1108 pc/8mxl/mol % Nxz/mol % Nhydrideldeuteride solutions108 dp/dx/Qm (mol % Hor D)-1108 pc/Qmxl/mol % H or Dxz/mol %Hx2lmol % D200 250 300 35024.16 25.80 27.32 28.725.7 6.2 6.5 6.824.06 25.67 27.17 28.470.03 0.03 0.05 0.150.07 0.19 0.41 0.80- 4.4 4.5 4.7- 25.80 27.32 28.67 - 0.02 0.02 0.02- 0.13 0.31 0.66 - - 0.20 0.5140030.017.029.610.301.454.929.920.021.241-13450 500 55031.19 32.25 33.197.2 - -30.39 - -0.60 - -2.77 - -5.0 5.2 5.330.89 31.75 32.410.30 0.60 0.902.14 3.48 5.682.22 - -nitride has a greater effect on the resistivity of lithium than has hydride.Deuterideand hydride effects are indistinguishable although their solubilities are slightlydifferent.6 This linearity is unlikely to persist over the entire lithium+ lithium saltphase diagram, however, and, at temperatures high enough to maintain completemiscibility, the resistivity is expected to rise steeply and smoothly, as observed inpotassium + potassium bromide mixtures,' to the high values characteristic of thepure molten salts.The resistivity, p/Rm, over the linear region is given as a functio232 SOLUTIONS IN LIQUID LITHIUMof concentration, x/mol % N, H or D, by eqn (1). Values of dpldx, pc and therelevant concentration rangesp = $ x + p ,are given in table 1 at selected temperatures. The resistivity of pure lithium, pLi,which, due to the initial curvature, differs slightly from pc, is also included in table 1 .The coefficient, dpldx, increases with increasing temperature for each solute andmay be represented by eqn (2) and (3). The nitride solutions are more sensitive to(dpldx), = 3.04 x lo-'+ I .66 x 10-108, - 1.66 x 10-138,2(dp/dx)H or D = 3.60 x lo-' + 3.13 x 10-1 '8c(2)(3)[200 < 8, < 450°C; Q = 0.41 x lo-' Rm (mol % N)-l-J[257 f 8, < 551°C ; cr = 0.32 x lo-* Rm (mol % H or D)-l]temperature than are hydride (or deuteride) solutions, which again show identicalbehavi our.COMPARISON OF RESULTSThe relative effects of dissolved non-metals on the resistivities of alkali metals areshown in table 2.The values are for the overall resistivity increase, irrespective ofwhether it is wholly or partially linear or non-linear, found or extrapolated for aconcentration of one mole percent non-metal in the metal solvent. Thus, in somecases, these values will differ slightly from the coefficients, dp/dx.TABLE 2.-INCREASE IN THE RESISTIVITY (108p/&l) OF ALKALI METAL SOLVENTS CAUSED BYONE MOLE PERCENT OF SOLUTE AT 400°C (OTHER TEMPERATURES~"~ IN PARENTHESES)-solute lithiumsolventsodium potassium caesiumnitride 6 .6 a ; 7.4(300)* ; 8.8(300)*t c ; - -hydride 4.5 a ; 3.8(300)* "; 4.6 d ; I -deu t eri deoxide 2.1(300)* ; 2.0* d ; - 3.2(30) f ;fluoride - - 3.9(700) ; -bromide - - 6.9(740) ; I iodide - - 9.1(700) ; -- - - 4.5 a ;* extrapolated ; 1- as Bad?.a present results ; b ref. (8) ; C ref. (2) ; dunpublished results ; e ref. (7) ; f G . Brauer, Z. anorg.Chem., 1947,255,101.We find lower and higher unit resistivity increases for nitride and hydride, respec-tively, in liquid lithium than have been reported previously.* The nitride resistivitieswere reported for a single temperature (300°C) and up to concentrations no greaterthan 0.05 mol % N. On the other hand, hydride resistivities were reported up toconcentrations as large as 0.55 mol % H at 300"C, but the solution saturates at themuch lower concentration of 0.31 mol % H at this ternperat~re.~ For these reasons,we believe our values for nitride and hydride in lithium to be more reliable.Included in table 2 are tentative values for oxygen (as sodium oxide) and hydrogen(as sodium hydride) in sodium which we have obtained from preliminary experiments.The table shows that nitrogen gives a high value in lithium and also in sodium, whenrendered soluble by the addition of barium.Hydrogen and oxygen also give similaP. F. ADAMS, M. G. DOWN, P. HUBBERSTEY AND R. J . PULHAM 233values, whether dissolved in lithium or sodium. Oxygen does not appear to increasethe resistivity of all metal solvents by the same amount, however, since a much highervalue is exhibited in caesium than in lithium and sodium.Chloride ion also givesthe same value in at least two solvents, potassium and bismuth, as observed byBronstein et aL7 Their measurements on the resistivities of solutions of the salts,KF, KBr and KI, in liquid potassium show increasing values of dp/dx with increasingcrystal radius (rlnm) of the anion, approximately according to eqn (4). The inter-polated value for chloride is 5.7 x a m (mol % KCI)-I compared with 17.4 xa m (mol % BiC13)-l for three chloride ions in liquid bi~muth.~ In lithium andsodium also, the increase9 = 1.6 x 101'r2+1.2dx (4)in resistivity appears to be related to the ionic size where, despite taking solutes fromdifferent groups of the Periodic Table, the increase in resistivity caused by a mole ofsolute increases with crystal anionic radius, provided that radii of 0.171,'' 0.154 l 1and 0.145 nm are chosen for N3-, H- and 02-, respectively.I 1 I I I 174I n'6 8 I /*;o-o-*-*;44040.0 0.2 0.4 0.6 0.8 1.0concentration, x/mol% N or HFIG.2.-Concentration dependence of dpldx for small values of x : upper diagram, nitride solutions ;lower diagram, hydride solutions.VERY DILUTE SOLUTIONSThe first few additions of solute caused a smaller increase in resistivity than didsubsequent ones as is clearly shown in fig. 2. The effect was most marked for solu-tions of nitride and was enhanced by an increase in temperature. Whereas the effectcould be detected for nitride solutions at 300°C and greater, for hydride solutions itwas only detectable at 440°C and greater.Hydride and deuteride exhibited similareffects.There is no clear explanation why the resistivity should increase less for smallconcentrations. There were no droplets of lithium on the walls of the vessel to abstractpart of the gas intended for the bulk of the liquid, and removal of nitride and hydridefrom lithium by steel at these temperatures is negligible. Chemical reaction betweensolutes provides no answer ; only dissolved ca-rbon might abstract nitride as cyanid234 SOLUTIONS IN LIQUID LITHIUMor cyanamide but carbon should have no effect on hydride. As described later,hydride and nitride do not react with one another. The effect appears to be funda-mental in nature, therefore, but no simple explanation, based on changing com-plexity of solute with concentration is acceptable.MIXTURES OF NITRIDE AND HYDRIDEAssociation between two solutes in a ternary solution is reflected in a deviationfrom additivity in the resistivity, e.g., N and Ba in Na.2 A polyatomic solute stillincreases the resistivity but not to the extent of the sum of the effects of its components.Where there is no association, the resistivity of the ternary solution is additive asshown by the solute pairs Ge and Sn, and Sn and Pb, in liquid sodium.12Resistivity changes for mixed solutions of nitride and hydride at 420°C are shownin fig.3. Line OL represents the resistivity of solutions of nitride alone in lithium,and line OM that for hydride alone.Addition of 1.5 mol % H, at point A, to asolution already containing 0.9 mol % N increased the resistivity to point B. Theincrease from A to B is that expected for hydride alone. Similarly, addition of nitride(0.9 mol % N) to a solution at point C, which contained 1.5 mol % H, also increased4442403432concentration, x/mol% N, H or NH3FIG. 3.-Resistivities of mixtures of nitride and hydride in liquid lithium at 420°C; OA and CBdenote addition of nitride, OC and AB denote addition of hydride, ON denotes addition of hydride+nitride as ammonia.the resistivity up to point B by the expected amount for nitride alone. Thus, irrespec-tive of the order in which the two solutes are dissolved, they each exert their charac-teristic electron scattering for this solvent and the resistivity is therefore additive,implying that nitride and hydride do not react chemically under these conditions.Association to form amide, for example, would destroy the parallelogram OABCsince the scattering by an NH; ion is unlikely to equal that of the sum of nitrideand hydride.Association is not expected on thermodynamic grounds, since thP. F. ADAMS, M. G . DOWN, P. HUBBERSTEY AND R. J . PULHAM 235enthalpy (in the absence of free energy data for every compound) change for thecorresponding solid state reaction [eqn (5)] is a large negativevalue (- 165.4 kJ mol-l).We have also added the elements together in the form of ammonia vapour. Theresistivity increases almost linearly as shown by the line ON in fig.3, which has slopeca. 22 x a m (mol % NH3)-l. This is to be expected if each mole of ammoniadissociates into one of nitride and three of hydride, and they each dissolve to exerttheir characteristic scattering of 7.1 x Rm (mol% H)-’. Again, this is to be expected from the reaction of excess lithium with am-monia [eqn (6)] for whichAH0 = -392.3 kJ mol-l. (Enthalpies : LiNH2,13 Li3N,14 LiH,” NH3.13)4Li + LiNHz 3 Li3N + 2LiH (5)SZm (mol % N)-I and 4.9 x6Li + NH3 -+ Li3N+ 3LiH (6)The authors thank the S.R.C. and U.K.A.E.A. (Culham) for the award of main-tenance grants (to M. G. D. and P. F. A., respectively).G. K. Creffield, M. G. Down and R. J. Pulham, J.C.S. Dalton, 1974, 2325.C. C. Addison, G. K. Creffield, P. Hubberstey and R. J. Pulham, J.C.S. Dalton, 1976, 1105.P. F. Adams, M. G. Down, P. Hubberstey and R. J. Pulham, J. Less-Common Metals, 1975,42, 325.C. C. Addison, R. J. Pulham and E. A. Trevillion, J.C.S. Dalton, 1975, 2082.R. J. Pulham, J. Chem. Suc. A, 1971,1389.P. F. Adams, P. Hubberstey, R. J. Pulham and A. E. Thunder, J. Less-Common Metals, 1976,46,285.H. R. Bronstein, A. S. Dworkin and M. A. Bredig, J. Chem. Phys., 1962, 37, 677.M. N. Arnol’dov, M. N. Ivanovskii, V. I. Subbotin and B. A. Shmatko, Teplofiz. Vys. Temp.,1967, 5, 812.A. H. W. Aten, 2. phys. Chern., 1909,66,641.lo W. E. Dasent, Inorganic Energetics (Penguin, London, 1970).l 1 F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry (Interscience, New York,l2 P. Hubberstey and R. J. Pulham, J.C.S. Faraday I, 1974,70,1631.l3 F. D. Rossini, D. D. Wagman, W. H. Evans, S. Levine and I. Jaffe, Selected Values of ChemicalThermodynamic Properties (Nat. Bureau Stand., 1952, circular 500).l4 P. A. G. O’Hare and G. K. Johnson, J. Chem. Thermodynamics, 1975, 7, 13.l5 H. R. Ihle and C. H. Wu, J. Inorg. Nuclear Chem., 1974, 36, 2167.1967).(PAPER 6/733
ISSN:0300-9599
DOI:10.1039/F19777300230
出版商:RSC
年代:1977
数据来源: RSC
|
28. |
Photoelectrocatalysis by metal phthalocyanine evaporated films in the oxidation of oxalate ion |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 236-242
Shunsuke Meshitsuka,
Preview
|
PDF (443KB)
|
|
摘要:
Photoelectrocatalysis by Metal Phthalocyanine EvaporatedFilms in the Oxidation of Qxalate IonBY SHUNSUKE MESHITSUKA*Sagami Chemical Research Center, Nishi-Ohnuma 4-4-1,Sagamihara, Kanagawa 229, JapanANDKENZI TAMARUDepartment of Chemistry, Faculty of Science,University of Tokyo, Hongo, Bunkyoku, Tokyo, JapanReceived 5th May, 1976The photoelectrode process of metal phthalocyanine electrodes was studied in the reaction ofhole capture by oxaiate ions. The photo-response was explained by assuming that the charge carriersdecay in a bimolecular recombination and that the electrochemical reaction rate is proportional tothe number of charge carriers. The dependence of the photo-action upon the wavelength of theradiating light was related to the absorption spectra of the metal phthalocyanine.Electrode reactions on semiconductors have been studied by many investigator^,^'^in particular, the photoelectrode processes of semiconductors, iiicluding sensitizedprocesses by several Organic or metal complex semiconductors have alsobeen studied with regard to the photovoltaic effect, and the mechanisms of formationof charge carriers and electric conduction have been discussed.*-lMetal phthalocyanines are generally stable compounds, even in electrochemicalredox reactions, and have structures similar to porphyrins, which have characteristicfunctions in many biological systems such as chlorophyll, haemoglobin, cytochromesand several oxidases. Several types of studies have been carried out using metalphthal~cyanines.~~-" They have been studied as catalysts for the reduction ofoxygen in fuel cells18-20 and in the electrocatalytic reduction of carbon dioxide.21The photovoltaic effect of metal phthalocyanines in an electrolyte has been studiedby several investigator^,^^-^^ but the detailed mechanisms of the phenomena havenot been elucidated.The energy conversion in going from photon to chemicalenergy is an important problem for which the electrochemical method might providea straightforward method of study. Accordingly, in order to clarify the mechanismof photochemistry on metal phthalocyanines, the photo-response of metal phthalo-cyanine evaporated film electrodes has been studied.EXPERIMENTALMetal phthalocyanines were synthesized from phthalonitrile and metal chloride inquinoline and purified by repeated sublimation.All the samples utilized in this investigationwere thin evaporated films, the vacuum evaporation being carried out under a pressure of10-3-10-2 N m-2 onto Pt, Au, Ag and Sn02. The films prepared in this way were alwaysa phase, as confirmed by their absorption The lead wire was connected by Ag23S . MESHITSUKA AND K. TAMARU 237paste and fixed by epoxy resin, and ohmic contact was ascertained. The back face and sideedges of the electrode were covered with epoxy resin. The thickness of the films on atransparent electrode of SnOz was estimated by the interference method and their absorbancefrom electronic absorption spectra, The effect of thickness was revealed only in themagnitude of photo-current. The most effective photo-response was observed for films - ]I p i thick.The electrochemical cell and the apparatus are shown in fig.1. To measure the photo-response a 1 kW xenon arc lamp (Ushio Electric) with a focusing lense was used. Mono-clzromatic light was obtained through a set of interference filters (Koshin Kogaku), eachwith a transmitting band of 8-10nm half width. Ultraviolet, near ultraviolet and nearinfrared wavelengths were cut off by glass filters. The cell was made of Pyrex glass andthe electrode was illuminated through a quartz window. The whole cell was immersed ina water bath, whose temperature was controlled by an electronic thermopile regulator(I'amato Coolnics). The steady state photo-current generated in the sample by the illumina-tion was measured on an ammeter of resistance zero ohm at a certain electrode potential E,which was fixed in the range +0.5 to -0.5 V by a potentiostat (Hokuto HA 110).Spectraof the photo-current were obtained by means of a lock-in amplifier (Ritsu ST 502) with anoscilloscope and pen recorder.The anode reaction on the metal phthalocyanine electrode was the oxidation of oxalateion-+ 2C02+2eand the cathode reaction on the platinized platinum wasFe(CN)g-+ e +. Fe(CN)%-.Na2(C00)2 (Wako Chem., extremely pure) was used without further purification. K,Fe(CN)6was recrystallized from water before use. Both parts of the cell were purged of oxygen bybubbling with purified N2 gas.I.IPC,'PtFILTERXENON LAMPPt (PI)En;.1 .-Schematic diagram of electrochemical photo-cell. The metal phthalocyanine evaporatedfilm ejectrode on Pt (surface area 1 x m2) is connected to a platinized platinum eIectrode. Theanolyte and the cathoIyte are separated by a cation exchange membrane.RESULTS AND DISCUSSIONKINETICS OF PHOTOELECTROCHEMICAL PROCEDUREIn the simple case where the absorption of a photon leads to the appearance ofan electron in the conduction band and a hole in the valence band, and where thedisappearance of conductivity is due to the recombination of the hole and electron,the rate of change in the charge carrier density n is expressed bydnldt = bI-an2 (1238 PHOTOELECTROCATALYSISwhere I is the rate at which quanta are absorbed and a and b are constants.In thesteady state dn/dt = 0Upon integrating, the rate of decay in the dark from the steady state in light is given byn, = J(bIJa). (2)(n, -vz)/n = an,t = J(abI) t = Kt (3)where K is a constant.rate in the dark, under constant illumination isLikewise, the rate of increase of charge carriers, from theIn ((n, +n)J(n, -n)> = 2an,t = 2,/(abI) t = 2Kt.i = kn f(ca, c,)(4)If we assume the photoelectrochemical current to be proportional to the chargecarrier density( 5 )where k is constant andf(c,, c,) is a function dependent upon the concentration ofthe reactants, c, and c, in anolyte and catholyte. By incorporating eqn (3, eqn(2), (3) and (4) may be rewritten, when the reactants have the same concentrations,asi , = k J ( b W f k 9 CC) (6)(7)(8)(i, -i>/i = J(abI) t = KtIn ((i, + i>/(im - i)> = 2 J(abI) t = 2Kt.; ,../ A-0 10 '30 50FIG. 2.-Dependence of steady state photo-current im on the concentration Ca of Na2(COO), for acopper phthalocyanine electrode on Pt at 25.0"C at several applied voltages ; (A) 0.0 V, (a) 0.1 V,( x ) 0.2 V, (0) 0.3 V, (0) 0.4 V.(A > 420 nm).STEADY STATE PHOTO-CURRENTThe photo-current reached a steady state in tens of seconds after turning on thelight. The relation between the observed steady state photo-current and the oxalateconcentration is given in fig. 2. In the counter electrode reaction the current observedwas proportional to the concentration of Fe(CN)%-. For sufficiently dilute solutioneqn (5) approximates toThe steady state photo-current i, is plotted against light intensity in fig.3. Thelight intensity was varied by means of a series of neutral filters which exhibitedi = k' nc, c,. (9S . MESHITSUKA AND K . TAMARU 239constant transmittance in the visible region. The order of its dependence on I was0.58, which suggests that the steady state photo-current is proportional to 14. Thisis in reasonable agreement with eqn (6).I/arbitrary unitsFIG. 3.-Log steady state photo-current i, against log light intensity I for a copper phthalocyanineelectrode at 25.0°C, ca = cc = 10.0 mol m-3, (A) 0.40 V, (0) 0.45 V. (1 > 420 nm).PHOTO-CURRENT DECAY FROM THE STEADY STATEThe current difference j from the steady state is given byj = i, - i (10)tlj = t/i,+i,/K.(1 1)and upon substitution in eqn (7), one obtainsPlots of t / j against t are shown in fig. 4. The value of K is estimated in each casefrom the intercept on the ordinate and the slope of the line.0 2 4 6 8 1 0tlsFIG. 4,-Analysis of the photo-current decay process of a copper phthalocyanine electrode on Ptat 25.0"C and at E = 0.45 V, at various light intensities (arbitrary units) ; (0) 1.00, (A) 0.89, (0)0.64, (0) 0.40 ; Ca = cc = 10.0 mol m-3. (A > 420 nm).The light intensity dependence of K values is shown in fig. 5. The order of thedependence on I is 0.45, which agrees well with eqn (3). The electrode potential Eis a parameter in fig. 3 and 5 , the values of i, in fig. 3 increasing with E, whereasthose of Kin fig. 5 decrease with E.Eqn (3) and (6) show that i, is inversely pro-portional to a* while K is directly proportional to a*. Thus the more E increases,the more the bimolecular decay constant a decreases. It would appear that th240 PHOTOELECTROCATALYSISpolarization of the electrode prohibits the coupling of charge carriers, separatingthem into opposite directions.0.4 0.6 08 1.0Ilarbitrary unitsFIG. 5.-Log K against log light intensity of a copper phthalocyanine electrode on Pt at 25.0’C,The behaviour of cobalt phthalocyanine was similar to that of copper phthalo-cyanine. The effect of the materials on which the metal phthalocyanine wasevaporated, such as Pt and SnO,, on the features of the photo-response was small,but the magnitude of the photo-current was changed. Therefore the characteristicsof nietal phthalocyanines were only revealed in the photo-current.In fig. 6 thephoto-response curve was calculated by using the K value obtained from the analysisof the dark process for cobalt phthalocyanine on Sn02. Although the observedvalue deviated slightly from the calculated immediately after turning on the light,the photo-response of the metal phthalocyanine electrode could be representedreasonably well.Ca = cc = 10.0 mol m-3. (A) E = 0.45 V, (0) E = 0.40 V, (0) E = 0.30 V. (A 3 420 nm).00.12 c 203 . 50 150 200t i sFIG. 6.-The variation of photo-current with time for a cobalt phthalocyanine electrode on SnQ2at 250°C and at E = 0.40 V ; Ca = cc = 10.0 mol m-3. (1 > 420 nm). -, observed ; - - -,calculated.ACTION SPECTRUM OF PHOTOELECTROCHEMICAL CURRENTThe action spectrum of the photoelectrochemical current in an evaporated copperphthalocyanine film on Pt is given in fig.7, together with the absorption spectruS . MESHITSURA AND K . TAMARU 241of the thin film. The current was measured by means of lock-in amplifier as shownin fig. 1. The photo-current component was obtained by alternating the illuminationat 11 Hz, and was corrected by the filter transmittance to obtain the photo-actionwith the same photon flux.1,jnnlFIG. 7.-The action spectrum of photoelectrochemical current for a copper phthalocyanine electrodeon Pt at 2S.O"C and at E = 0.3 V, c, = 1.00 x lo2 mol m-3 and cc = 0.40 x 10' mol m-3. Theupper curve is the absorption spectrum of copper phthalocyanine evaporated on SnOz.When the light was turned on, the photo-current showed a rapid negative response,and when it was turned off, it showed a positive response.When a positive potentialwas applied to the metal phthalocyanaine electrode, no rapid response was observed.Therefore the action spectrum was observed in anodic polarization.The maximum photo-action was observed at a shorter wavelength than themaximum absorption. The same, as yet unexplained, phenomenon has been reportedby Calvin et al. for the photo-conductivity of metal free phthal~cyanine.~ It wouldappear that the efficiency of conversion of the excited electronic state into chargecarriers might not be the same throughout the range of the absorption spectrum.Thephoto-conductivity spectrum of copper phthalocyanine was observed by Harrison butthe features were quite different from our spectrum.1o His spectrum differs fromthe absorption spectrum, and has several maxima, of which the most intense is at800 nm. In our case the spectrum is similar to the visible absorption spectrum. Wecould measure the photo-current only in cases where the charge carriers participatedin the electrochemical reaction, e.g. hole capture. The difference between the twoaction spectra suggests that there are different types of photo-carriers formed in themetal phthalocyanine semiconductor. The photo-action spectrum of the cobaltphthalocyanine electrode was similar to that of copper phthalocyanine, as wasexpected from the similarity of the absorption spectra.H.Gerischer, Adeances in Electrochemistry and Electrochemical Engineering, ed. P, Delahay(Interscience, New York, 1964), vol. 1, p. 784.P. J. Holmes, The Electrochemistry of Semiconductors (Academic Press, London, 1962).T. Freund and W. P. Gomes, Catalysis Rev., 1969, 3, 1242 PHOTOELECTROCATALY SISA. Fujishima and K. Honda, Nature, 1972, 238, 37 ; Bull. Chem. SOC. Japan, 1971, 44, 1148.A. Fujishima, T. Iwase, T. Watanabe and K. Honda, J. Amer. Chem. SOC., 1975, 97, 4134.R. Memming and H. Tributsch, J. Phys. Chem., 1971, 75, 562.'I H. Yoneyama, H. Sakamoto and H. Tamura, Electrochim. Acta, 1975, 20, 341.R. C. Nelson, J. Chem. Phys., 1954,22,885.G. Tollin, D. R. Kearns and M. Calvin, J. Chem. Phys., 1960, 32, 1013, 1020.S. E. Harrison, J. Chem. Phys., 1969, 50,4739.P. Day and R. J. P. Williams, J. Chem. Phys., 1962, 37, 567.H. Baba, K. Chitoku and K. Nitta, Nature, 1956, 177, 672.l 3 S. Meshitsuka, M. Ichikawa and K. Tamaru, J.C.S. Chem. Comm., 1975, 361.l4 M. Ichikawa, R. Sonoda and S . Meshitsuka, Chem. Letters, 1973, 709.l6 J. Manassen and A. Bar-Ilan, J. CataZysis, 1970, 17, 86.Is F. Beck, J. Heiss, H. Hiller and R. Polster in Katalyse an Phthalocyanine, ed. H. Kropf andI9 H. G. Jahnke, M. F. Schoenborn and G. Zimmerman in Proc. Symp. Electrocatalysis, ed.'O A. Kozawa, V. E. Zilionis and R. J. Brodd, J. Electrochem. SOC., 1970, 117, 1470.'I S. Meshitsuka, M. Ichikawa and K. Tamaru, J.C.S. Chem. Comm., 1975, 361.22 K. Hauffe, D. Rein, in Katalyse an Phthalocyanine, ed. H. Kroph and F. Steinbach (Georg23 G. A. Alferov, V. I. Sevastyanov, V. A. Ilatovskii, Yu. S. Shumov and G. G. Koniissarov,24 V. B. Evstigneev and A. N. Terenin, Doklady Akad. Nauk S.S.S.R., 1951, 81, 223.2 5 S. E. Harrison and K. H. Ludewig, J. Chem. Phys., 1966,45,343.R. Taube, Pure Appl. Chem., 1974,38,427.J. H. Weber and D. H. Busch, Inorg. Chem., 1965, 4,469.F. Steinbach (Georg Thieme, Stuttgart, 1973).M. W. Breiter (Electrochem. SOC., 1974).Thieme, Stuttgart, 1973).Doklady Akad, Nauk S.S.S.R., 1972,203, 628.(PAPER 6/859
ISSN:0300-9599
DOI:10.1039/F19777300236
出版商:RSC
年代:1977
数据来源: RSC
|
29. |
Viscosity dependence of diffusion-controlled triplet energy transfer in 2-methylpentan 2-,4-diol |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 243-249
Frederick S. Dainton,
Preview
|
PDF (476KB)
|
|
摘要:
Viscosity Dependence of Diffusion-controlled TripletEnergy Transfer in 2-Methylpentan 2-,4-diolBY FREDERICK s. DAINTON,? MARIAN s. HENRY,$ MICHAEL J. PILLING* AND PHILIPC . SPENCERPhysical Chemistry Laboratory, South Parks Road, Oxford OX1 342Received 2 1 st May, 1976The rate constant for triplet energy transfer from phenanthrene to stilbene has been studiedin 2-methylpentan 2-,4-diol over the temperature range 190-350 K. It shows a non-Arrheniusdependence which can be described by the VTF equation k = ATexp { -B/(T- To)>, where B =(850k 100) K and To = 158 K. The B-value is smaller than that derived from viscosity data.Similar results are obtained from less extensive investigations of naphthalene + chromium acetyl-acetonate and anthracene + perylene. The data for phenanthrene + stilbene, which cover ninedecades in rate constant, are inadequately described by the empirical equation (kq/T) = a+67pYwhere a, 6 and x are constants.The viscosity dependence of diffusion-controlled reactions has received muchexperimental attention.' Several techniques have been employed to vary theviscosity : (a) change in the composition of a mixed solvent, (b) addition of a thick-ening agent, (c) the use of several similar solvents and ( d ) change in temperature.Of these the second is the least satisfactory, since a change in the macroscopicviscosity is not reflected at a molecular level.A change in solvent or solvent com-position suffers from the disadvantage that the solute diffusion coefficient is not onlyaffected by the change in viscosity but also by specific changes in solute-solventinteractions. In the temperature variation method most parameters other thanviscosity are kept reasonably constant ; glass forming liquids are particularly goodcandidates as solvents, since they display a wide and continuous viscosity range.The choice of reaction is also important, since the mechanism can depend in acomplex manner on solvent-solute interactions.Triplet energy transfer is eminentlysuitable, since the interaction leading to transfer is fairly well understood.2 It hasbeen suggested that at very low viscosities such reactions fall below the diffusion-controlled limit,3 whilst at very high viscosities energy transfer takes place at distancesgreater than the normal encounter di~tance.~ At these extremes the rate constantceases to reflect faithfully the dependence of the diffusion coefficient on viscosity.Inthe wide intermediate range, however, triplet energy transfer provides a usefulvehicle for the examination of viscosity effects in transport controlled reactions.EXPERIMENTALTriplet energy transfer was investigated using conventional and laser flash photolysis ;the experimental technique has been previously de~cribed.~The following compounds were used as supplied : naphthalene and anthracene (B.D.H.microanalytical grade), phenanthrene (Oxford Organic Chemicals), perylene (Koch-Light,puriss), t-stilbene (B.D.H. laboratory reagent). Chromium acetylacetonate was prepared$ Department of Chemistry, Boston University, Boston, Massachusetts 0221 5.present addresses : p University Grants Committee, 14 Park Crescent, London W1N 4DH.24244 VISCOSITY DEPENDENCE OF TRIPLET ENERGY TRANSFERby the method of Fernelius and Blanck.6 2-methylpentan 2-4-diol (MPD) (B.D.H., laboratoryreagent) was distilled at 8 Tom pressure through a Vigreux column; the middle fraction,boiling at 90°C was collected and stored under nitrogen.RESULTSThe decay of a triplet (TI) molecule may be expressed as- d[TJ /d t = k1 [T 11 -I- 2k2 [ TJki = ks+L+k~[Qlwhere k3 is the viscosity-independent and k4 the viscosity-dep,3iideiIt first order rateconstant for triplet decay.k2 describes triplet-triplet annihilation and k , thedeactivation of the triplet by an added quencher Q.The methods of analysis ofmixed first- and second-order and of pure first-order kinetics have been described.’In the absence of added quencher, the triplet states of napthalene, antliraceneand phenanthrene showed first order decay constants which depended on tempxaturein the usual sigmoid manner,5 tending to constant low temperature values (k3) of0.45, 20 and 0.27 s-’ respectively, cf. the literature values for the reciprocal naturallifetimes (l/e) of 0.42, 20 and 0.29 s-l.’‘Fhe rate constants (k5) for energy transfer to an added quencher, Q, were examinedover a wide range of quencher concentration. The phenanthrene + stilbene systemwas the most thoroughly studied, afid [stilbene] was varied froin 5 x mol dm-3to 4 x mol d ~ n - ~ .The kinetics were accurately described by a quenchingcomponent first order in Q over this range. Fig. 1 shows an Arrhenius plot for thissystem ; the other two systems showed similar non-Arrhenius bchaviour.DISCUSSIONExergonic triplet energy transfer is generally diffusion-controlled with a rateconstant given bywhere R is the encounter distance and D the relative diffusionbe generally described by the relationwhere k is the Boltzmann constant and ( the friction coefficient.* 5 is defined internis of the force autocorrelation function of the particle ; realistic calculations ofthis parameter can now be made for simple liquids. For more complex systems ainore phenomenological approach has to suffice.The Stokes-Einstein theory of molecular diffusion suggests that D can be expressedin terms of the macroscopic viscosity, q, with 5 lying between 471qr and Gnqr, depxdingon the relative sizes of the solute and solvent molecules; r is the hydrodynamicradius of the solute.’ The viscosity of a glass formifig liquid can usually be rclatedto tcmpsrature via the empirical VTF equation :’* loIn q = In A+B,/(T-To).For MPD, over the temperature range 220-350 K, log (A/kg m-l s-l) = -5.37,B = 1220 K and To = 158 K.1-14 The VTF equation has been rationalized by Adamand Gibbs and Goldstein l 5 has recently presented a heuristic picture of viscousflow applicable for He used a model in which flow is dorninatcdby the presence of potentid energy barriers which are high compared to thermalenergies.Irreversibility arises from the large number of cooperative rearrangementsk5 = 4nRD, ( 1 )D mayD = kT/[ (2)2 1 kg in-’ s-lF. s. DAINTON, M. s. HENRY, M. J . PILLINGANDP. c . SPENCER 245which accompany deformation, thus essentially erasing any memory of the initialniolecular distribution. As the temperature is lowered the size of the coopzrativeregion increases and, at To, includes the whole solvent system. The parameter Brelates the number of coopxative configurational excitations to the temperature inthe entropy theory of Gibbs and co-w~rkers.~Combining eqn (1) and (2) with the Stokes-Einstein and VTF equations gives :where r1 and r2 are the hydrodynamic radii of T1 and Q. Fig. 2 shows a plot oflog(k,/T) against (T- To)-, for phenanthrene + stilbene.There is, perhaps, somehint of positive curvature at low temperatures, but the experimental scatter in thisregion, arising from the extreme tempzrature depzndence of the viscosity, precludesa clear assignment of such an effect and a straight line fit seems justified. Similarbehaviour was found for napthalene +chromium acetylacetonate and anthracene +perylene. The slopes of plots such as that shown in fig. 2 were used to obtain Bvalues ref. eqn (3)] for the diffusion-controlled quenching processcs and these areshown in table 1 (B,). The B, values obtained directly from viscosity measurementsare also shown. In all cases B , < B,, i.e., eqn (3) is not directly applicable and thediffusion-controlled reactions show a weaker dependence on temperature than does(11 IT).This could reflect microscopic differences in the cooperative rearrangementsrequired for diffusion of the solute molecules and for viscous flow. Molecularrearrangements occur constantly and may be biased by an external stress to produceviscous flow, or utilized by a solute molecule to effect diffusion. With the exceptionof naphthalene + Cr(acac),, which shows a B-value quite close to that pxtaining toviscosity, the solute pairs are all planar aromatic molecules and there seems no apriori reason why their diffusive motion should require exactly the same degree ofcoopxative rearrangement of the surrounding solvent molecules as does viscous flow.The suggestion of curvature at low temperatures might be due to a variation in To,which has been found in some solvent systems, or to the onset of non-VTF behaviour,possibly of the type discussed by Dainton et a l l 6TABLE 1 .-EXPERIMENTAL B-VALUES FOR A FIT OF QUENCHlNG DATATO THE V.T.F. EQUATIONdonor acceptor BqIK BlK B2K anaphthalene Cr(acac), 1220+70 1180+ 160 1240+ 160ant hracene perylene 1220+70 6 920f 160 920f 160phenanthrene t-stilbene 1220+ 70 850+ 100 940+ 100a Calculated after the data were modified to recognise the increase in encounter distance.withviscosity, using the model of Pilling and Rice,4 with A = lo8 s-I and L = 0.14 nrn. The anthracene;perylene data were obtained over a small viscosity range (2 decades) and at high temperatures,where the effect of an expanding encounter distance is minimal ; b errors quoted correspond totwice the standard deviation; C errors are nominal and assume no uncertainty in the parametersA and L.Fig.2 shows a plot of In (/<,IT) where k, is the rate constant calculated from theStokes-Einstein equation :kT R(r, + r2)k , = -vl l'1r2The hydrodynamic radii were set equal to the van der Waals radii (0.34 nm for bothphenanthrene and stilbene) and the encounter distance to the sum of the van de246 VISCOSITY DEPENDENCE OF TRIPLET ENERGY TRANSFERWaals radii; the rate constant thus reduces to 4kT/q. The experimental rate constantlies above the theoretical value except at very low viscosities (q - kg m-' s-I).Skipp and Tyrrell l 8 showed that ground state aromatic molecules have diffusioncoefficients compatible with hydrodynamic radii equal to half their van der Waals+ t+PQ3 4 5lo3 KITFm.1 .-Arrhenius plot for the rate constant (k) for deactivation of triplet phenanthrene by t-stilbenein 2-methylpentan 2-,4-diol.radii in propan-1,2-diol at - 300 K. Other evidence 19* 2o suggests that excitedaromatic molecules show more normal behaviour in this viscosity regime, and wemight expect k, to be a lower limit for a diffusion-controlled reaction. There islimited evidence that the experimental rate constant is beginning to fall below thediffusion-controlled value at high temperatures, suggesting the onset of transfer-contr01.~. sIt is well known that in rigid media triplet energy transfer can occur over distanceslarger than the encounter distance. Pilling and Rice developed a model wherebythe continuous expansion of the effective encounter distance with increasing viscositycan be calculated. They showed thaF.S . DAINTON, M. S . HENRY, M. J . PILLING AND P. C . SPENCER 247where y is Euler's constant and wg = AL2exp(-2R/L)/D. A and I, are parametersin the exchange interaction which is presumed to mediate the quenching processand lo(wo) and Ko(wo) are, respectively, modified &st and second type Bessel functionsof zero order and argument wo. Fig. 2 shows the effect of such an increase onthe rate constant of a reaction involving molecules which obey the Stokes-Einsteinrelation. Typical exchange parameters have been assumed ( A = lo8 s-l, L =0.14 nm). The increase in k is quite small and it cannot account fully for the observed- 4 8 12 16 2 0 24 28 32103 K/(T- 158)FIG.2.-VTF plot of log(k/T) against (T-To)-l, with To = 158 K, for phenanthrene+stilbene.-, least squares fit ; - - - , Stokes-Einstein plot with ky = 4kT/y ; . . ., Stokes-Einstein plot witha viscosity dependent encounter distance (see text).deviation from Stokes-Einstein behaviour. It does however modify the B-value forthe phenanthrene+stilbene system. The other two systems were studied over amore restricted temperature range and the modification to B is smaller (table 1).Alwattat et al. recently reviewed the viscosity dependence of diffusion-controlledreactions of electronically excited aromatic molecues. They demonstrated an almostuniversal deviation from Stokes-Einstein behaviour (in the same sense as that shownin fig.2) although the range of visocisities studied was in no case as wide as thatreported here. They used the microfriction theory of Gierer and Wirtz 21 to rational-ise the empirical relation :krl - = a + b t f . T (4248 VISCOSITY DEPENDENCE OF TRIPLET ENERGY TRANSFERFig. 3 shows a plot of log (ky/T) against log ti. At low viscosities, log (kq/T) showsno tendency to reach a constant value, as required by eqn (4). The value of asuggested by the theory of Gierer and Wirtz may be calculated from the relative radiiof the solute and solvent. Taking these as the van der Waals radii (0.34nm forphenanthrene and stilbene and 0.32 nm for MPD), a = 66 J mol-1 K-l, cf. the Stokes-Einstein value of 33 J mol-1 K-l.The experimental value of ky/T falls below boththese values ; as was suggested above, this probably arises from the onset of transfer-control. For y > I kg m-l s-l? ky/T $ a and a may be omitted from eqn (4).Fig. 3 shows that log(ky/T) is not linear in log q in this region, i.e., x depends on '7.'r /+I I I I I I I 1- 4 -2 0 2 4 6 8 10log(v/kg rn-l s-l)FIG. 3.-Plot of log(k7lT) against log 9 for phenanthrene+ stilbene.The origin of the temperature dependence of viscosity and of viscosity relatedparameters is complex and Goldstein 22 has urged caution in oversimplifyingtheoretical descriptions. The theory of Gierer and Wirtz represents such an over-simplification and its main attraction lies in the possibility it presents of ;I semi-empirical classification of diffusive behaviour.The present work, which hasinvestigated the temperature dependence of the rate constant of a well characterizeddiffusion-controlled reaction over nine decades, suggests that the empirical para-meters contained in eqn (3) are viscosity dependent and, therefore, of limited value.A. H. Alwattar, M. D. Lumb and J. B. Birks, Organic Molecular Photopliysics, ed. J. B. Birks(Wiley, London, 1973), vol. I, p. 403.J. B. Birks, Photophysics of Aromatic Molecicles (Wiley, London 1970).N. J. Turro, N. E. Schore, H.-C. Steinmetzer and A. Yekta, J. Amer. Chem. Soc., 1974, 96,1938.M. J. Pilling and S. A. Rice, J.C.S. Famday II, 1976, 72, 792.E. J. Marshall, N. A. Philipson and M. J. Pilling, J.C.S. Faraday 11, 1976, 72, 830.W. C. Ferneiius and J. E. Blanck, Inorganic Syntheses, 1957,5, 130.P. A. Egelstaff, An Introduction to the Liquid State (Academic Press, London 1967).G. Adam and J. H. Gibbs, J. Chern. Phys., 1965, 43, 139.lo C. A. Angel1 and K. J. Rao, J. Chent. Phyx., 1972, 57, 470.l 1 M. S. Henry, unpublished results.l2 A. H. White and S. 0. Morgan, Physica, 1932, 2, 313.' J. B. Birks, Organic MoZeculnr Photaphysics, ed. J . B. Birks (Wiley, London 1973), vol. I, p. 1F. s. DAINTON, M. s. HENRY, M. J . PILLING AND P. c . SPENCER 249l 3 G. E. McDuffie and T. A. Litovitz, J. Chem. Phys., 1962, 37, 1699.l4 C, Haranadh, Trans. Furuhzy Soc., 1963, 59,2728.l6 F. S. Dainton, B. J. Holt, N. A. Philipson and M. J. Pillling, J.C.S. Faraday I, 1976, 72,257.17J. T. Edward, J. Chem. Educ., 1970,47,261. '* C. J. Skipp and H. J. V. Tyrrell, J.C.S. Furaday I, 1975, 71, 1744.l9 W. R. Ware and T. L. Nemzek, J. Chem. Phys., 1975,62,477.'O B. Nickel and U. Nickel, Ber. Bunsenges Phys. Chem., 1972, 76, 584.22 M. Goldstein, Faraduy Symp. Chem. Soc., 1972, 6,44.M. Goldstein, J. Chem. Phys., 1969, 51, 3728.A. Gierer and K. Wirtz, Z. Naturforsch., 1953,8a, 532.(PAPER 6/965
ISSN:0300-9599
DOI:10.1039/F19777300243
出版商:RSC
年代:1977
数据来源: RSC
|
30. |
Reaction of Cob(I)alamin with nitrous oxide and Cob(III)alamin |
|
Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases,
Volume 73,
Issue 1,
1977,
Page 250-255
Robert Blackburn,
Preview
|
PDF (477KB)
|
|
摘要:
Reaction of Cob(1)alamin with Nitrous Oxide and Cob(1II)alaminofinisBY ROBERT BLACKBURN AND MAUNG KYAWDepartment of Chemistry and Applied Chemistry,University of Salford, Salford M5 4WTANDA. JOHN SWALLOW*Paterson Laboratories, Christie Hospital and Holt Radium Institute,Manchester M20 9BXReceived 7th July, 1976Cob(1)alamin (vitamin Blzs) has been generated by pulse radiolysis of NzO-saturated solutions' cob(n)alamin (vitamin Blzr) containing sodium formate. It reacts with NzO with rate constantsthe range 200-1200 dm3 mol-I s-l, depending on pH and buffer concentration. The final productcob(u)alamin, formed in 100 % yield. The initial product is thought to be cob(m)alamin(vitamin BlZa or Blzh) but this does not build up to significant concentrations. Cob(m)alamin(vitamin Blza or Blzh) reacts with cob(1)alamin with a rate constant of 3.2 x lo7 dm3 mol-' s-lindependent of pH in the range 5.8-11.0. Cyanocobalamin (vitamin BIZ) does not react withcob(1)alamin.The lower oxidation states of certain transition metal complexes are known toreact with nitrous oxide.The reaction of the cobalt@ compound cob(1)alamin(vitamin B1& is of special interest for biological systems since N,O is availablefrom enzymatic sources. Moreover, the inhibition of B ,,-catalysed enzymaticreactions by N20 has been used as a test for the involvement of the Co' state, as forinstance in the case of ethanolamine ammonia-lya~e.~The reaction of Col with N,O obeys the stoichiometry :l2 Co1+N20 + 2 Co"+N,.This rules out the possibility that the initial step is the formation of Co" and a freehydroxyl radical, since hydroxyl radicals are unspecific in their attack on corrinoidsystems and could not lead to quantitative oxidation or reduction even if they were toattack the corrin~id.~ On the other hand, the mechanism could involve an initial two-electron transfer mechanism,l which could be represented :2H+CO' + N20 + CO"' + H,O + NzCO"'+ CO' + 2 CO".followed by :We have now used pulse radiolysis to study these reactions over time scales of0-2 s.Cob(1)alamin has been produced in the presence of N,O by irradiation ofsolutions containing cob(I1)alamin (vitamin B,,,) and sodium formate. In thesesolutions the initially produced e&, OH and H react according to :6H20e,;+N,O -+ OH+OH-+N2, k = 8 .7 ~ lo9 dm3 mol-1 s-Ik = 3.5 x lo9 dm3 mol-1 s-lk = 2.3 x lo8 dm3 mol-l s-l.OH + HCOO- 3 H20 + CO,,H + HCOO- + H, +COY,25R. BLACKBURN, M. KYAW AND A. J . SWALLOW 25 1CO; is known to react with cob(1r)alamin to produce cob(1)alamin in 100 % yield :'The cobf1)alarnin is then available for reaction with N20. In other experiments COzhas been used to convert e i into COT :6COT can reduce Co(II1)alamin when in the form of vitamin B12a,5 as well as cob@)-alamin :COT + CoI1 + CO, + Co', k = 8.2 x lo8 dm3 mol-l s-l.e,;+CO, + COT, k = 7.7 x lo9 dm3 mo1-l s-l.CO; +Corl' -+ CO, +Co",Propan-2-01 has been used to convert OH radicalsCO"OH + (CH3),CHOH + H 2 0 + (CH3),C0H,Hydrogen atoms also give reducing radicals from propan-2-01.k = 1.5 x lo9 dm3 mol-1 s-I.into radicals which can reducek = 2.2 x lo9 dm3 mol-1 s-Ik = 8.2 x lo8 dm3 mol-l s-l.(CH3),C0H+Co" + (CH,),CO+Co'+H+,EXPERIMENTALCob(n)alamin was prepared as previously des~ribed,~ all solutions being handled in anitrogen-purged glove box to prevent atmospheric oxidation. Traces of cob(m)alamin inthe solutions were normally removed by pre-irradiation. Other compounds and gaseswere of the highest available purity, propan-2-01 from B.D.H. being of special purity withregard to aldehydes (< 50 ppm combined aldehyde and ketone). pH-values in the range3.5-5.1 were adjusted using perchloric acid and monosodium phosphate, in the range 5.5-8.2using mono- and di-sodium phosphate and at 9.2, 9.9 and 11.0 using sodium tetraboratewithout and with sodium hydroxide.Pulse radiolysis experiments were carried out on the Paterson Laboratories electronaccelerator in the usual way.'^ Spectroscopic changes were observed in cells of opticalpath length 2.5cm using a continuous xenon lamp.Temperatures were in the region18-22°C. A Bausch and Lomb single monochromator was used for most experiments, but aJarrell-Ash monochromator was used for investigations at the isosbestic points for cob(1)-alamin/cob(n) alamin. Dosimetry was carried out using oxygen-saturated aqueous thio-cyanateYg the yield of species absorbing at 500nm being taken to be 2.9. Cobalt-60y-irradiations were performed in the dose rate range 3-54x lov2 W kg-l as measured byFricke dosimetry.RESULTSREACTION OF COB(I)ALAMIN WITH N20mol dm-3 cob(I1)alamin and0.1 mol dm-3 sodium formate, at pH 3.5-8.5, were given single pulses of 5.8-7.8 J kg-l.Assuming the radiation produces COT with G = 6.5,5 the maximum concentration ofcob(1)alamin formed would be 4.5 x mol dm-3.At every pH, the opticaltransmission at 385 nm [a wavelength of maximum absorption for cob(r)alamin] wasfound to decrease as a result of the pulse, and then to increase to the original value ina first-order manner, i.e., the optical density first increased then decreased. Simul-taneously the optical density at 313 nm [a wavelength of maximum absorption forcob(~~)alamin] decreased as a result of the pulse and then increased to the originalvalue in a first order manner.Drawings of typical traces (for pH 8.5) are shown infig. 1. Cob(1)alamin is known to be relatively stable in neutral or slightly alkalinesolution in the absence of N,O.1° Cob(1I)alamin solutions were pulsed at pH 3.6and 3.7 in the absence of N,O, but with 1 moldm-3 sodium formate instead ofNitrous oxide saturated solutions containing 2 252 REACTIONS OF COB(1)ALAMIN0.1 ml dm-3 so as to ensure the conversion of hydrogen atoms produced by reactionof e& with H+ into COT. The absorption at 385 nm produced by the pulse wasabout the same as in the presence of NzO, but the subsequent rate of change inabsorbance was only about 15 % of the rate in the presence of NzO. Thesetime/units of 0.2 sFIG. 1.-Typical oscilloscope traces obtained at 385 nm (lower) and 313 nm (upper). Solutions weresaturated with NzO and contained 2 x rnol dm-3 cob(i1)alamin and 0.1 rnol dm-3 sodiumformate.They were unbuffered. Dose was 5.84 J kg-’.3 4 5 6 7 8PHFIG. 2.-Rate constant for reaction of cob(r)alamin with NzO to form cob(x~)alamin ; 0, calculatedfrom changes at 385 nm ; x , calculated from changes at 313 nm. Solutions were saturated with NzOand initially contained 2 x mol dm-j cob(zI)alamin and 0.1 mol dm-3 sodium formate. Thebuffer concentration was 0.1 rnol dm-3.observations are consistent with cob(1)alamin reacting with N20 with the ultimateformation of cob(I1)alamin and with the yield of the reaction being 100 %. Furtherindications that the yield is 100 % were that oscilloscope traces obtained from solutionswhich had been subjected to as many as six pulses were the same as those from freshsolutions given a single pulse, and that y-irradiation to doses as high as 266 J kg-R.BLACKBURN, M. KYAW AND A. J. SWALLOW 253of nitrous oxide saturated solutions containing about 6 x mol dm-3 cob(I1)-alamin and 0.1 mol dm-3 sodium formate at pH 6-9 produced no change in absorptionspectrum.In none of the experiments were changes seen at any wavelength to suggest theappearance of detectable amounts of cob(1n)alamin. For instance, pulses deliveredto solutions of pH 5.8 produced negligible change in optical density at the cob(I)alamin/cob(1i)alamin isosbestic points which are close to 310, 340, 415 and 530 nin. Thissuggests that the rate-determining step in the overall reaction is the reaction o f Co'with N20, to form CoIII, which then reacts rapidly with residual Co' forming theCo".Further evidence that the reaction with N20 is rate-determining was obtainedby adjusting the N20 concentration by dilution to 7.14 x mol dm-3 and 12.5 xmol dm-3, the N20 concentration in the original solution being taken to be2.5 x mol dm-3 (pH 5.1). The first order rate constants for the changes at 385and 313 nm were found to be directly proportional to the N20 concentration.Second order rate constants for the reaction of cob(1)alamin with N20 to form(ultimately) cob(I1)alamin were calculated from traces like those in fig. 1 and areshown as a function of pH for solutions of buffer concentration 0.1 rnol dm-3 infig.2. At pH 6.1 and 8.0 the rate constants for solutions of buffer concentrationmol dm-3 were 2.3 x lo2 and 1.6 x lo2 dm3 mol-1 s-l respectively.REACTION OF COB(I)ALAMIN WITH COB(III)ALAMINCarbon-dioxide saturated solutions containing 6 x 1 0-5 mol dm-3 cob(II)alamin,6 x mol dm-3 cob(II1)alamin (vitamin B12a), 0.1 mol dm-3 propan-2-01 and0.1 m ~ l dm-3 phosphate buffer (pH 5.8,7.6) or 0.01 mol dm-3 borate buffer (pH 9.9,11.0) were given single pulses of between 1.6 and 1.9 J kg-l. In these experimentsmost of the COT and organic radicals would reduce cob(II)alamin,ll so producingapproximately mol dm-3 cob(1)alamin. First order changes in optical densitywere seen at 390 and 460nm, corresponding to reaction of this cob(1)alamin withexcess cob(II1)alamin.Rate constants for this reaction were calculated from theresults at 390 nm to be 2.4, 3.3,2.5 and 4.1 x lo7 dm3 mol-I s-l at pH 5.8,7.6,9.9 and11.0 respectively and froin the results at 460 nm to be 3.0,3.7, 2.5 and 4.3 x lo7 dm3mol-1 s-l at the same pH values. Taking into account the experimental errors, therate constant for the reaction between cob(1)alamin and cob(rr1)alamin (vitamin B1 2a)is taken to be 3.2 x lo7 dm3 mol-l s-' independent of pH in the range studied.The experiments were repeated using cyanocobalamin (vitamin BI2) as the formof cob(1II)alamin. No reaction was seen after the initial formation of cob(r)alamin,showing that cob(1)alamin does not react with cob(II1)alamin when this is in the formof vitamin B12.DISCUSSIONIn neutral aqueous solution cob(1)alamin exists mainly in the unprotonated form,with pK, - 1.0, producing a form of but in acid solution protonation takescob(1)alamin which may be represented :I3Co'H+ + Co"H $ CoII'H-.It seems possible that the protonated form of cob(1)alamin may react with N20 byhydride ion transfer, as suggested for some enzyme reactions :I4Co"'H-+N,O --+ Co"'+N,+OH---this reaction being followed by fast disproportionation between Co"' and CO'in its unprotonated or protonated form, to yield Cox' as the observed reaction product254 REACTIONS OF COB(I)ALAMINIf this reaction has a rate constant greater than about lo5 dm3 mol-1 s-l it couldaccount for the increase in observed rate constant as the pH decreases below pH - 5(fig.2). It may be noted that the pH dependence shown in fig. 2 is similar to that forthe stability in the absence of N20 of certain species considered to be C0V5 At verylow pH it would be expected that Col would react so rapidly with N20 that nonewould remain to disproportionate with Co"'. The reaction product would thenbecome Co"' instead of Co". There was no evidence for this in the present experi-ments, but in the reaction of other Co' complexes with N20, where the rate constantis about lo7 dm3 mol-1 s-l, there were indications of the formation of ColI1 ratherthan CO".~' However, no hydroxyl radical scavenger was present in those experi-m e n t ~ , ~ ' so interpretation is complicated by the possible presence of productsresulting from hydroxyl attack on the corrin ring5The inflexion at pH 7 in fig.2 could be due to complexing of CO' with the acid formof the phosphate buffer, on the acid side of pH 7, followed by a reaction of the hydrideion transfer type. Buffers have previously been found to accelerate the spontaneousoxidation of cob(1)alamin in aqueous so1ution,16 and the reason could be similar.The much greater decrease in rate constant for reaction with N 2 0 when the bufferconcentration was decreased to mol dm-3 at pH 6.1 than at pH 8.0 is consistentwith this view.On the alkaline side of pH 7, the rate of the reaction of cob(1)alamin with N20may be governed by the rate of entry of N,O into the coordination sphere of the cobalt,which may be represented :There is evidence of other complexes of N20, e.g., [(NH3)5R~N20]2+ in aqueoussolution l7 and in the solid state.18 An analogous complex is believed to be anintermediate in the complex oxidation of cob(I1)alamin to cob(m)alamin : l9It may also be noted that the rates observed in the present reaction with N20 aretypical of those for the substitution of various ligands into aquocobalamin.20If reaction of Co' with N20 in mildly alkaline solution is initiated by the formationof an inner-sphere complex, the subsequent formation of ColI1 could be representedby a scheme such as :Co'+N20 + Co'N20 + CO~~N,O-.cor1+02 Co"'0;.0 N2 OH1 ON2 1 i"/ Bzrn Bzmc$ cH20a cy + N,+ OH-Bzm J OH-IBzm-where the coordination with N20 is via the oxygen atom, as in the similar rutheniumc~mplex,~' the NO bond distance being enlarged.Other formulations however arenot excluded. Formation of cob(Ir1)alamin would be followed by rapid reaction witR . BLACKBURN, M. KYAW A N D A . J . SWALLOW 255residual cob(1)alamin giving cob(I1)alamin as sole product. The pK, for aquo-cobalamin (BIza) and hydroxocobalamin (BIZJ lies between 6.9 and 7.8.21-25 Sinceour experiments show that the rate of reaction of cob(rI1)alamin with cob(1)alamin isindependent of pH between 5.8 and 11.0 it seems that aquocobalamin and hydroxo-cobalamin react with cob(1)alamin at equal rates. Cyanocobalamin does not howeverreact with cob(r)alaniin, and this is in agreement with the available, rather uncertain,redox data.I2This work was in part supported by grants from the Cancer Research CampaignOne of us (M.K.) is indebted to the British and the Medical Research Council.Council for financial support.R. G. S. Banks, R. J. Henderson and J. M. Pratt, J. Chem. Soc. A, 1968, 2886.T. H. Finlay, J. Valinsky, A. S. Mildvan and R. H. Abeles, J. Biol. Chem., 1973, 248, 1285.B. A. Fry, The Nitrogen Metabolism of Micro-organisms (Methuen, London, 1955).G. N. Schrauzer and E. A. Stadlbauer, Bioinorg. Chem., 1974, 4,185.R. Blackburn, A. Y. Erkol, G. 0. Phillips and A. J. Swallow, J.C.S. Faraday I, 1974,70,1693.A. J. Swallow, Radiation Chemistry (Longman, London, 1973), pp. 149-153.J. P. Keene, J. Sci. Instr., 1964, 41, 493.J. P. Keene in Pulse Radiolysis, ed. M. Ebert, J. P. Keene, A.J. Swallow and J. H. Baxendale(Academic Press, London, 1965), p. 1.G. E. Adams, J. W. Boag, J. Currant and B. D. Michael in Pulse Radiolysis, ed. M. Ebert,J. P. Keene, A. J. Swallow and J. H. Baxendale (Academic Press, London 1965), p. 117.lo S. L. Tackett, J. W. Collat and J. C. Abbott, Biochem., 1963, 2,919.l 1 R. Blackburn, M. Kyaw, G. 0. Phillips and A. J. Swallow, J.C.S. Faraday I, 1975, 71, 2277.l2 J. M. Pratt, Inorganic Chemistry of Vitamin BIZ (Academic Press, London, 1972), p. 109.l3 D. Lexa and J. M. Savdant, J.C.S. Chem. Comm., 1975, 872.l4 H. R. Mahler and E. H. Cordes, BiologicaZ Chemistry (Harper and Row, New York, 19661,l 5 A. M. Tait, M. Z . Hoffman and E. Hayon, J. Amer. Chem. Soc., 1976,98,86.l6 P. K. Das, H. A. 0. Hill, J. M. Pratt and R. J. P. Williams, J. Chem. Soc., A 1968, 1261.l7 J. N. Armor and H. Taube, J. Amer. Chem. Soc., 1969, 91,6874.l S A. A. Diamantis and G . J. Sparrow, Chem. Comm., 1970, 819.l 9 J. H. Bayston, N. Kelso King, F. D. Looney and M. E. Winfield, J. Amer. Chem. Soc., 1969,2o W. C. Randall and R. A. Alberty, Biochem., 1967, 6, 1520.21 R. P. Buhe, E. G. Newstead and N. R. Trener, Science, 1951, 113, 62.z2 E. L. Smith, K. H. Fantes, S. Ball, J. G. Waller, W. B. Emery, W. K. Anslow and A. D.2 3 G. I. H. Hanania and D. H. Irvine, J. Chem. Sue., 1964, 5694.24 G. I. H. Hanania and D. H. Irvine, Proc. 8th Int. Conf. on Coordination Chem., Vienna 1964,2 5 G. C. Hayward, H. A. V. Hill, J. M. Pratt, N. J. Vanston and R. J. P. Williams, J. Chem. SOC.,p. 358.91, 2775.Walker, Biochem. J., 1952, 52, 389.p. 418.1965, 6485.(PAPER 6/1325
ISSN:0300-9599
DOI:10.1039/F19777300250
出版商:RSC
年代:1977
数据来源: RSC
|
|