摘要:

Raman Spectra of Thallium(x) Nitrate Solutionsin Liquid AmmoniaBY DEREK J. GARDXNER"Newcastle upon Tyne NEl 8STALI H. HAJI AND BRIAN P. STRAUGHANDepartment of Inorganic Chemistry, University of Newcastle upon Tyae,Newcastle upon Tyne NE1 7RUDepartment of Chemistry, Newcastle Upon Tyne Polytechnic,ANDReceived 17th April, 1975Raman spectra are reported for T1NO3 solutions in liquid ammonia, ranging in composition fromT1N0,/3NH3 to T1N03/26.7NH3 mole ratios. All the spectra were obtained from solutions atambient temperatures. Intensity changes in the N-H stretching region are correlated with theeffects of the electrolyte on the solvent structure. An intense broad band at -1110 cm-' has beenassigned to the symmetric deformation mode of NH3 engaged in strong a-interactions with T1+ ions.The DSh selection rules for the unperturbed NO, anion break down in concentrated solutions ofTINOJ in NH3 and spectral evidence is presented to support a new n-interaction model for TI+-NO; ion-pairs.Raman studies on aqueous solutions of electrolytes have provided much informa-tion on the nature of these solutions.1 Recently the technique has been applied toliquid ammonia solutions of electrolytes.2-5 The results have been interpretedgenerally by analogy with the well-documented aqueous systems. Salts of singlycharged cations and singly charged polarisable anions are usually readily soluble inliquid ammonia and hence the use of nitrates, thiocyanates and iodides in this typeof study. Raman studies on molten thallium(1) nitrate 6 * ' have revealed anomalouslystrong ion-pair interactions. Thallium(1) nitrate is only sparingly soluble in wateryet very concentrated solutions can be prepared in liquid ammonia.This differencein solubility further highlights the anomalous behaviour of thallium(r) ions and thepurpose of this work is to determine the nature of the interactions present in thallium(I)nitrate solutions in liquid ammonia.EXPERIMENTALRaman spectra were recorded on a Cary 81 laser modified Raman spectrometer. The530.8 nm line from a Coherent Radiation model 52 krypton ion laser was used for excitation.The polarisation data were measured on a Spex Ramalog (4) instrument. The spectra wererecorded at ambient temperatures and they were calibrated against carbon tetrachloride andindene.B.D.H. thallium nitrate and K. and K. Laboratories Inc. thallium thiocyanate wereused without further purification. Each sample was oven dried and checked for dryness byinfrared spectroscopy. Air Products Ltd. 99.98 % anhydrous ammonia was used withoutfurther purification.Samples were prepared in -2 mm diameter Pyrex capilliary tubes having a flat end toreceive the laser beam axially. The scattering geometry was 180". All samples were handledon a vacuum line or in a nitrogen filled glove bag to preserve their anhydrous condition.994 RAMAN SPECTRA OF T1NO3RESULTSRaman spectra were obtained from thallium(1) nitrate solutions in liquid ammoniaranging in composition from 1/3 to 1/26.7 TlN03/NH3 mole ratios.A TlSCN/6.5 NH3 solution of thallous thiocyanate was studied also. The spectra may bediscussed most easily in terms of bands arising from NH3 and NO: ion vibrations.Any additional features may result from interaction processes in the solutions.NH3 VIBRATIONSThe N-H stretching region, 3500-3100 cm-l, consists principally of three bandsat - 3400, -3300 and -3200 cm-l. It is apparent from fig. 1 that the band at - 3200 cm-1 becomes markedly less intense whilst the - 3400 cm-i band increasedvery slightly in intensity as the TlN03 concentration increases. There are no markedfrequency changes in this region as a function of concentration. Table 1 includes thepositions of these bands and fig. 1 shows the N-€3 stretchmg region at the twoextremes of the concentration range examined.TABLE L-RAMAN BAND POSITIONS (F Cm-') AND ASSIGNMENTS FOR THE 03SERVED BANDS OFTlNOj IN LIQUD NH3 SOLUTIONS FOR A RANGE OF CONCENTRATIONS AT AMBIENT TEMPERATURENO: NH3mole ratio v 1 v2 v3 v4 2v2 V l v2 v3 v 4 + 2v41:3.1 1043 828 1374 709 1659 3297 1096 3387 1635 32261:3.9 1043 830 1370 710 1658 3299 1096 3387 1637 32221:4.5 1045 830 1363 710 1660 3302 1096 3389 1630 32251:4.7 1044 829 1368 709 1659 3300 1098 3388 1636 32251: 5.0 1044 830 1368 710 1659 3302 1100 3388 1633 32211:5.2 1044 830 1367 710 1658 3300 1098 3388 1638 32231: 5.4 1044 830 1367 710 1660 3300 1097 3388 1637 32241: 5.9 1045 830 1367 710 1660 3300 1101 3388 1638 32221: 8.4 1045 830 1364 710 1660 3301 1101 3388 1636 32211 : 11.1 1045 832 1370 710 1665 3304 1100 3388 1636 32221 : 17.7 1045 - 1360 710 1660 3305 1097 3388 1638 32201 :26.7 1045 - 1360 710 1660 3303 1098 3388 1630 3220NH3 3306 1046 3389 1640 3221* Approximate values.Fig.2 shows the lower frequency region for the T1NO3/3NH3 mole ratio solutionand a broad and intense feature at - 1100 cm-' is observed. The intensity of thisband is concentration dependent as the graph of fig. 3 shows. Here the - 1100 cm-lband intensity, normalised by ratio with the intensity of a known NO; vibration at-710 cm-l, is plotted against the solution mole ratio. This feature appears also inthe Raman spectrum of the TlSCN solution which we studied. One further bandassociated with the NH3 vibrations appears as a weak feature at - 1630 cm-l.Table 1 includes the observed frequencies of all of these bands.No; VIBRATIONSBands arising from the NO, ion vibrations appear at -710, 830, 1045, 1360 and1660 cm-l.The latter band appears as a sharp but weak feature in all of the spectraand shows little variation in frequency. The band at 1045 cm-l is very intense andsharp, its frequency does not vary but its bandwidth increases with concentrationfrom - 5 to -7 cm-' across the range of solutions studied. The -710 cm-l banD . J . GARDINER, A . H . HAJI AND B . P. STRAUGHAN 95/, I i I t J3400 3300 3200i/crn-'FIG. I.-The N-H stretching region for TIN03 in liquid ammonia at two different concentrations.;/crn-'FIG. 2.-Raman spectrum (200-1500 cm-l) of a T1NO3/3NH3 mole ratio solution.1 3 5 7 9 I I 13 I5 17 19 21 23mole ratio NH3/TINO3FIG.3.-Plot of the band intensity ratio v2(NH3)/v4(NOJ) against the solution mole ratio96 RAMAN SPECTRA OF TlNO3behaves similarly in that it shows very little change in frequency and remains sym-metrical in shape in all of the solutions. In the 1360 em-' region a band is presentwhich is clearly asymmetric in shape (see fig. 2) and at least two components areanticipated. The frequency separation of the components appears to be concentra-tion dependent. Complete resolution of the components was not possible due totheir inherent band-widths. We have measured instead the total half band widthof this feature as a function of concentration and expect this to reflect any variation0.23-0.210.1 9 -n 0.17- 0" zP u 0.15---- Na 0 .1 3 -0.11.0.09--1 1I 3 5 7 9 Itmole ratio NH3 /TIN03FIG. 4.-Plot of the band intensity ratio uz(NO~)/u4(NO~) against the solution mole ratio.1 860 840 820 800v/cm-lFIG. 5.--Rarnan spectra showing the polarisation of the vZ(NO;) mode for a TIN03/3NH3 moleratio solution.in frequency separation of the components. The results of these measurements showthat the half band width increases by 35 cm-l over the total concentration rangeexamined.In the more concentrated solutions a band was observed at 830 cm-l (see fig. 2).The graph of fig. 4 shows that this band increases in intensity as the concentrationincreases. Furthermore, polarisation studies showed this band to be polarised, andan example of this behaviour is given in fig.5 for the TlN03/3NH3 solutionD. J . GARDINER, A . €3. HAJI AND B. P. STRAUGHAN 97The only other notable feature in these spectra was a very broad band in the300-500 cm-1 region. This was observed in all the spectra and changed little in shapeor position with concentration.DISCUSSIONThe unperturbed NH3 molecule has CJV symmetry and hence gives rise to a vibra-tional representation : rvib = 2A1 +2E with all modes infrared and Raman active.A band appearing at 3200 cm-l has been recognised as 2v4 in Fermi resonance withvl.* Birchall and Drummond have studied spectra of NH3 and ND3 in the con-densed phase at various temperatures and suggest that the assignments of 2v4 and v1should be reversed. A similar conclusion is drawn by Schwartz and Wang lo aftersome temperature variation studies.However, Gardiner, Wester and Grossmanhave reported results supporting the original assignment. These workers showedthat in dilute solutions of ammonia in carbon tetrachloride and acetonitrile, whereNH3. . .NH3 interactions are reduced to a minimum, the intensity of the 3300 cm-lband is several times that of the 3200 cm-l band and hence assigned the former to vl.Curve resolution analysis of this complex N--I3 stretching region reveals the presenceof a broad underlying band at -3260 cm-l which Gardiner et al. assign to N-Hstretching of associated ammonia molecules. The presence of this band has beencorroborated by other investigators who assign it similarly.12 In the light of thisevidence we prefer to adopt the original assignment for o m spectra (see table 1).The lowering in intensity of 2Vq(NH3) as the electrolyte concentration increaseshas been observed previously.2 We are confident that this behaviour arises fromdisruption of the H-bonded structure of liquid ammonia and the creation of newinteractions with the electrolyte ions.This change in structum would affect theintensity of the broad band underlying the N-H stretching region and with anaccompanying reduction of v1 /2v4 resonance interaction, the observed drop in 2v4intensity can be accommodated. The slight increase in v3(NH3) intensity with con-centration indicates further restriction of NH, rotational motion.The most notable feature in the Raman spectra of NH3 in these solutions is thepresence of an intense and broad band at - 1100 cm-l.We have assigned this bandto v2(NH,) which in pure liquid ammonia appears at 1046 cm-l. Unfortunately astrong band due to nitrate ion masks the 1046 cm-1 position in the TlN03 solutionsand therefore to be sure that we were not observing an additional band, the Ramanspectrum of a TlSCN/6.5 NH3 mole solution ratio was studied. In this latter solutionthere was no evidence for a band at 1046 cm-l, but the intense band at - 1100 cm-Iwas present. In LiN03/NH3 solutions, a band at - 11 10 cm-1 has been reportedthough it was extremely weak. It seems likely that this intense band is due primarilyto NH3 interactions with Tlf ions. Furthermore, the interaction must be strongerthan, or completely different in nature to, the Li+-NH3 interaction.The Tlf ion isknown to engage in strong a-bonding and this fact has been used to account foranomalous intensities in the Raman spectra of TlN03 melts.6* This o-bondingcharacteristic together with the greater polarisability of the Tlf ion compared withthe Lit- ion will contribute to the intensity of the 1100 cm-l band. Furthermore, thelarge increase in frequency of v,(NH3) from 1046 cm-l in liquid ammonia to - 1100cm-l in the electrolyte solutions is expected from comparisons with transition-metalammine spectra.These arguments and observations lead us to assign the band at - 1100 cm-l tothe symmetric deformation mode of ammonia molecules (v,) engaged in stronga-interactions, with Tlf ions, via the nitrogen lone pair.The band due to the antisymmetric deformation mode of NH3 at - 1630 cm-I isI98 RAMAN SPECTRA OF TIN03little affected by the presence of Tlf, NO; or SCN- ions.This is in accord withearlier observations.2The free NO? ion has D3,, symmetry and gives rise to a vibrational representationr v i b = Ai(R)+A;(i.r.) +2B'(i.r. and R) where infrared and Raman activities areshown in parentheses.The antisymmetric stretching mode of NO; appears at - 1360 cm-I. In thedilute solutions there is some slight splitting of this mode evidenced by its asymmetricshape. The splitting probably increases at higher concentrations of TlN03, but thetwo components are never clearly resolvable except in the most concentrated solution.If the increase in total half band width of this feature is a measure of the extent ofthe splitting then an increase of about 35 cm-l is indicated over the concentrationrange.The splitting of this feature over a comparable concentration range inLiN03/NH3 solutions has been reported as 56 cm-l. These observations indicatea small Tl+-NOi interaction and little perturbation of the NO; ion by NH3 solution.T h i s latter conclusion is in agreement with earlier work.14Splitting of the v4(E') deformation mode in aqueous nitrate solutions has beentaken as an indication of ion pair formati~n.'~ Removal of the degeneracy fromthis mode requires ion-pair interactions to occur through 0 atoms of nitrate producinga C,, or C, species. In all of the solutions studied here, there was no evidence ofsuch a splitting.Following this interpretation we conclude either that there are noion-pairs formed or that a species of higher symmetry is produced.Evidence for the latter explanation is convincingly provided by the appearance inthe more concentrated solutions of the NO; out-of-plane deformation mode at - 830 cm-l. Furthermore, this mode, which is Raman inactive for D3h NO, belongsto a totally symmetric character species as shown by its polarisation behaviour. Amodel to explain these observations would be a nitrate ion belonging to point groupc30r v i b = 2A1(i.r. and R)+2B(i.r. and R).However, considering the possible types of interaction that may be present in thesesolutions, we feel it unlikely that sufficient energy could be provided to distort thenitrate ion from its planar nn-bonded configuration. It seems more likely that thenitrate ion in our more concentrated solutions interacts with TI+ to produce a C,,ion-pair.This can be achieved by allowing Tlf to interact with the n-cloud of NO;along the principal axis of the anion. The following schematic depicts such aninteraction and shows how solvating NH3 molecules may be accommodated in themodel to produce a C3, species. Alternatively, the solvating NH3 molecules maylower the symmetry to C 3 ; this point group still allows for the observed spectra.NH3 NH3 NH3\ I /T1+We are left, however, with the slight inconsistency that the band at 136Ocm-1retains some doublet characteristics while the degenerate band at -710 cm remainsunsplit.This observation may be accounted for if an asymmetric solvent cage orion-pair environment is present. This would cause only minor perturbations anD. J . GARDINER, A . H. HAJI AND B . P . STRAUGHAN 99would affect only the most symmetry sensitive vibrations. Force field perturbationcalculations l6 indicate that the antisymmetric stretching mode of NO; is by far themost susceptible to such situations.CONCLUSIONSOur results lead to a model for the structure of T1NO3/NH3 solutions. In thedilute systems each of the ions appears to be fully solvent separated. As the concen-tration of TlN03 increases the H-bonded structure of liquid ammonia breaks up andion-pair interactions take over.However, with Tlf ions the interaction proceeds withthe NO; ion via its z-cloud. The presence of this type of interaction though theoreti-cally quite feasible, has not been demonstrated previously for liquid NH3 or aqueoussysterns.l8 Raman spectra from molten TlN03 reveal a band at 813 cm-’ but nopolarisation data are available.” The questions must arise as to whether thissituation obtains in concentrated aqueous solutions and whether other heavy metalions interact similarly. Experiments directed towards these problems are beingcarried out.We thank Dr. H. Hallam of Swansea University for use of the Spex Ramalog (4)instrument.D. E. Irish in Zonic Interactions Vol. ZI, ed. C. Petrucci (Academic Press, London, 1971), chap. 9.D. J. Gardiner, R. E. Hester and W. E. L. Grossman, J. Chem. Phys., 1973, 59, 175.P. Gans and J. B. Gill, Chem. Coim., 1973,23,914.K. R. Plowman and J. J. Lagowski, J. Phys. Chem., 1974,78,143.A. T. Lemley and J. J. Lagowski, J. Phys. Chem., 1974,78,708.D. W. James and W. H. Leong, Trans. Faraday Soc., 1970, 66, 1948.C. A. Plint, R. M. B. Small and H. L. Welsh, C a d . J. Phys., 1954, 32, 653.T. Birchall and I. Drummond, J. Chem. SOC., 1970,1859.D. J. Gardiner, R. E. Hester and W. E. L. Grossman, J. Raman Spectr., 1973, 1, 87.’ D. W. James, R. D. Carlisle and W. H. Leong, Austral. J. Chem., 1970,23, 1779.lo M. Schwartz and C. H. Wang, J. Chem. Phys., 1973, 59, 5258.l2 A. T. Lemley, J. H. Roberts, K. R. Plowman and J. J. Lagowski, J. Phys. Chem., 1973,77,2185.l3 D. M. Adams, Metal-Ligand and ReZated Vibratiom (Arnold, London, 1967).l4 N. Smyrl and J. P. Devlin, J. Phys. Chem., 1973, 77, 3067.l5 D. E. Irish, A. R. Davis and R. A. Plane, J. Chem. Phys., 1969,50,2262.H. Britzinger and R. E. Hester, Znorg. Chem., 1966, 5,980.G. J. Janz, T. R. Kozlowski and S. C. Wait Jr., J. Chem. Phys., 1963,39, 1809.T. C. G. Chang and D. E. Irish, J. Solution Chern., 1974, 3, 161, have presented evidence for a“ roll-on ”, off axis Ag+--No, ion-pair configuration in acetonitrile

ISSN:0300-9599    

DOI:10.1039/F19767200093

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Kinetics of the Gas-phase Reaction between Iodine andTrimethylsilane and the Bond Dissociation EnergyD( Me, Si-H)?BY ROBIN WALSH" AND JEAN M. WELLSDepartment of Chemistry, The University of Reading,Whiteknights, Reading RG6 2ADReceived 18th April, 1975The title reaction has been investigated in the temperature range 519-618 K. The only products,formed in equaI quantities, were trimethylsilyl iodide and hydrogen iodide, Rates were found to besurface sensitive below about 560 K, but not so in the range 567-618 K where the rate lawd[L1 - k[I&Me,SiH] ---dt 1 +k'[HI]/[I,]was obeyed over a wide range of iodine and Me3SiH pressures. This expression is consistent withan iodine atom abstraction mechanism and for the step.I' + Me,SiH -+ Messis +HIlog (kl/dm3 mol-1 s-l) = 10.9- 82.3 kJ mol-'/RTIn 1 0has been deduced.From this the bond dissociation energy D(Me,Si-H) = (376k 11) kJ mol-I(90 kcal mol-l) is obtained. The implications of this value for the pyrolyses of organosilanes arediscussed.Reliable free radical thermochemistry is crucial to an understanding of gas phasekinetics and mechanisms. Whereas for many organic species good thermochemistryis available and kinetics and mechanisms are reasonably well understood,l* in thearea of organosilicon chemistry this is not so. Organosilane pyrolyses appear tooccur via free radical pathways in many cases and a complete understanding oftheir mechanisms depends heavily on a knowledge of Si-H and Si-Si bond dissocia-tion energies. We decided therefore to attempt a measurement of D(Me,Si-H).The most recent value,4 prior to our work, for D(Me,Si-H) was 340 kJ mol-'.That .this figure might be somewhat low was suggested by Whittles in the light of therelative rates of methyl radical abstraction from SiH4 6* and Me,SiH.6* RelativeHT yields from hot tritium atom abstractions also indicate a higher figure.A recentredetermination by Davidson and Howard lo of D(Me,Si-SiMe,) used in com-bination with the appropriate thermochemistry leads to D(Me3Si-H) = 364 kJ mol-l.Detailed kinetic studies of the gas phase reactions of iodine with many organicmolecules have been exploited by Benson and co-workers to provide reliable bonddissociation energies. We extend this technique into the organosilicon field here bystudy of the reaction of iodine with trimethylsilane.EXPERIMENTALAPPARATUSThe apparatus consisted of a static vacuum system, and quartz reaction vessel located inan electrically heated metal block furnace, designed to allow passage of a light beam throught A preliminary report of this work has appeared R.Walsh and J. M. Wells, Chem. Comm., 1973,513.10R. WALSH AND J . M. WELLS 101the vessel. The vessel contents could be continuously monitored by a single beam spectro-photometer (Hilger-Watts) the optics of which were modified for the purpose. The vessel,of path length N 17 cm, was connected to the vacuum system via a three-way tap the thirdarm of which was attached to a pressure transducer (Bell and Howell type 4-3274003) whichwas heated to -lW"C, along with that section of the line used for handling I2 vapour.Conventional silicone greased taps had to be used in this part of the line but otherwisegreaseless valves (Springham) were employed.Temperatures were measured with one of several chrome1 alumel thermocouples locatednext to the vessel.The calibration of the thermocouples was checked against an N.P.L.standard platinum resistance thermometer. Reaction vessel temperatures were controlledby an AEIRT3R proportional controller and were uniform in both space and time towithin &I*C.Both the unpacked quartz vessel of S/V = 0.91 cm-l, and an alternative one packed withPyrex tubes (which still allowed passage of a narrow light beam) of S/Y= 3.55 ern-', werewashed with a dilute solution of silicone oil DC703 in CCI4 prior to use, to attempt torender their surfaces inactive.12 Dead spaces were -0.6 % and were neglected.MATERIALSTrimethylsilane was prepared by the LiAlH4 reduction of trimethylsilyl chloride in di-n-butyl ether under nitrogen.Slight deterioration of the purified gas occurred over longperiods leading to formation of small quantities of higher molecular weight materials (notidentified). Because of this it was always redistilled through a -78°C trap prior to a day'sexperiments. It was identified by its i.r. spectrum l3 and contained no gas chromato-graphically detectable impurities (after redistillation).Trimethylsilyl chloride was a gift from Midland Silicones.Trimethylsilyl iodide was prepared l4 by the reaction of aluminium iodide with hexa-methyl disiloxane (Koch Light).After distillation the fraction boiling at 106-107°C wascollected. The product proved still to have substantial contamination with hexamethyl-disiloxane (which could either be unreacted starting material or a hydrolysis product). Theiodide was identified by its n.m.r. spectrum l5 (single line absorption at 9.28~). Furtherpurification was not attempted because of the ease of hydrolysis of the iodide. The impureiodide was adequate for identification purposes since the contaminant does not absorb inthe U.V. (above 210 nm).Iodine (Koch Light 99.998 % pure grade) was degassed before each experiment.Hydrogen iodide gas was prepared by adding an HI solution (Fisons AnalaR) dropwiseon to Pz05.It was dried by passage through a -78°C trap and collected at - 196°C.It was stored at room temperature in a blackened bulb and degassed before use.PROCEDUREPrior to a kinetic run, iodine (at - 20°C) and trimethylsilane (at - 196°C) were thoroughlydegassed. It was evaporated into the reaction vessel at a known pressure and, in mostexperiments its absorbance at a wavelength of 484 nm was recorded. A run was initiatedby the sharing of a known pressure of trimethylsilane into the vessel. A continuous traceof the absorbance variation with time was recorded during runs which varied (according toconditions) from minutes to several hours. The pressure transducer was used to recordpressures at intervals during a run, but was not left continuously in contact with reactingmixtures because of small losses due to slow absorption of I2 into the transducer. In someruns absorbance traces were obtained at other wavelengths in both the visible and U.V.regions.The 484 nm absorbance traces represent a direct record of iodine consumption as a functionof time since in the reaction mixture only iodine absorbs at this wavelength. Calibrationsof the spectrophotometer showed that Beer's law was obeyed by iodine at 484nm up to2.0 absorbance units.After a run the products were usually frozen out of the vessel and removed; they wereonly tetained in preliminary identification experiments. Some runs were fairly short and soa test was devised to ensure that mixing times (for I2 with Me,SiH) were still negligible102 D(Me,Si-H) BOND DISSOCIATION ENERGYBecause Iz vapour is visible this was easily done by eye in an identical vessel (at about 50°C)outside the oven.Even at 1 atm of added air mixing times were less than 3 s. Lowerpressures and higher temperatures reduce this time.In runs designed to study the inhibition by HI, the procedure was identical apart frominitial addition of HI to the reaction vessel, after the I3 but before Me3SiH.PRODUCT IDENTIFICATIONQualitative identification of the products as HI and Me3SiI was achieved simply byevaporating them sequentially into the vessel after a run, and recording their U.V. spectra.The U.V. spectra were in accord with those of authentic samples (see fig. 1). A more quanti-tative identification was made by attempting to monitor the absorbance at several U.V.wavelengths both during and after a run at 546 K.The absorbances at these wavelengthswere calibrated against those of HI and Me3SiI. The former was a pure sample. The latter,because of contamination of the prepared sample, was the triply distilled recovered samplefrom several actual runs. Fig. 1 shows a comparison between the runs of the absorbanceslpressure of HI and Me3SiI (in Torr, 1 Torr = 133.2 N m-2) and the final absorbance (basedon pressure of I2 consumed) of a particular run. The matching is reasonable although notperfect. The time evolution of U.V. absorbance at 220,230,250,270 and 490 nm shows that[HI] and [Me3SiIl must be within 10 % both of each other and of the decrease in [I2].I220 24 0 260 280 30(A/nmFIG.1.-Comparison of the U.V. spectrum of reaction products at 546 K with the sum of U.V. spectraof HI and Me3SiI. 0, reaction products ; - - - , HI spectrum ; - - - , Me3SiI spectrum ; -,sum of spectra.Several species were specifically ruled out from consideration. Since the reaction productwas completely condensible (G0.5 % of total products remain in the vessel), hydrogen andmethane were absent at this level. Methyl iodide could not have exceeded a few percentof the products based on chromatographic analysis (4 m length 15 % ppG column at 6O"C),and substantial amounts of other iodides (such as Me2SiHCH21) could not have been presentor a U.V. absorption peak at -2260 nm would have been observed.Gas chromatographicanalysis of the -778°C condensible reaction product on the column showed the only sub-stantial peaks to be due to Me3SiH (unreacted) and Me3SiI. Some minor peaks totallinga few percent in area were present. However, as the iodide is partially hydrolysed in thesampling system and co-elutes with (CH&3iOSi(CH3)3 the chromatographic andysis is notparticularly reliableR. WALSH AND J . M. WELLS 103RESULTSPRELIMINARY EXPERIMENTSThe quantitative identification of the products points to the chemical processThis is supported by pressure measurements which indicate no pressure change (towithin k0.3 Torr or better) during reaction. No secondary reaction between tri-methylsilane and HI occurs as evidenced by a complete lack of any absorbancechanges (in both visible and u.v.) when they were mixed at typical pressures at 547 K.Trimethylsilyl iodide was thermally stable under the conditions of our experiments.Table 1 shows an example of some iodine decay data in an early run at 546 K,using 7.02 Torr I2 and 21.3 Torr Me,SiH.(CH,),SiH + I2 (CH3),SiI + HI.TABLE 1 .-IODINE DECAY w r r ~ TIMEtimelmin 0 4 8 12 16 20 24 2812/Torr 7.02 6.26 5.73 5.24 4.74 4.37 4.01 3.71timelmin 32 36 40 44 48 52 56 60I~/ToIT 3.37 3.06 2.75 2.46 2.14 1.87 1.65 1.42These data were tested to find the order with respect to iodine, by the van’t Hoffmethod. It was assumed that the reaction was first order in Me,SiH, although sinceMe,SiH was in excess this assumption was not very critical.A plot was made oflog((AI,)/[Me,SiH],} against log[I,], where A12 is the loss in I2 over an 8 min interval(i.e., is an approximation to the rate) and [Me,SiH], and [IJr were the instantaneousconcentrations at the middle of the interval, [Me,SiH], being determined from stoichio-metry and the I2 decrease.This plot is shown in fig. 2. Such plots employingdifferences are always scattered but a forced straight line fit gives a slope (i.e., anorder) of 0.47k0.16. A further test of this data was made by attempting to fit itto the integrated form of the rate equation on the assumption of half and first orderdependence of the rate on I2 and Me3SiH respectively. The appropriate equation(see appendix) predicts a linear fit of tan-l(fj) with time wheref = [12],/([Me3Si€€Jo -[I& Fig.3 shows the plot for the data of table 1. The fit is reasonable althoughslight departures from linearity at long times were observed in many runs.DETAILED KINETIC MEASUREMENTSFurther tests to establish the rate law consisted in measuring the three halvesorder rate constant, k3, defined byby means of integrated plots and then examining the dependence of k* on startingconditions (particularly the ratio [Me,SiHJ/[I,]). Most of the integrated plots showedslight curvature, indicating inhibition of the reaction at high conversions. This wassubsequently allowed for explicitly but to avoid complications in these tests the decayplots were not examined beyond 50 % conversion of 12. Good linear fits to thetan-l(f*) against time plots were usually obtained up to this point.Table 2 showssome rate constants obtained at about 595 K.Apart from one run the k* values scatter to a maximum of & 15 % around the mean,and cover a range of values of the ratio [Me,SiHI/[I,] from 5-50. At all temperaturesruns with this ratio less than two tended to give high values for k+. Between values- d[I,]/dt = k3[12]+[Me3SiH] (104 D(Me3Si-H) BOND DISSOCiATION ENERGY0 0.2 0.4 0.6 0.8loslo{[Izlr/(Torr)~FIG. 2.-Log-log plot of rate against concentration to determine reaction order with respect to iodineat 546 K. Error bars represent uncertainty limits from spectrophotometric trace.3timelmindefinition off.FIG. 3.-Plot of the rate data at 546K according to the integrated rate equation.See text forTABLE 2.-A SELECTION OF RATE DATA AT 595 K[Id/ Tom Me3SiH/Torr 104k+/T0d S-12.03 64.6 3.594.95 45.3 4.1 11.30 22.3 3.491.55 7.7 3.214.66 22.5 4.198.28 10.5 5.732.29 117.1 3.5R. WALSH AND J . M. WELLS 105of 3 and 200, k, was constant within the scatter. This moderately high scatter couldbe attributed to instability and noise in the single beam spectrophotometric system.At the lower two temperatures the scatter became greater but this was shown to arisefrom surface effects. Within these limitations the rate data supported the threehalves order expression (A) reasonably well.Several runs were performed in the packed vessel to test for heterogeneity. Table 3shows a selection of the results. It is clear from these data that substantial surfacecatalysis is occurring at 522 and 533 K.To try to reduce this effect both vessels wererewashed with silicone oil solution and the unpacked vessel with hexamethyldisilazanealso. These treatments had little or no effect. At 581 K and higher temperaturesthe surface effect seemed to have disappeared and it probably was not serious in theunpacked vessel above 500 K.TABLE 3.-cOMPARISON OF RATE CONSTANTS ‘ OBTAINED IN DIFFERENT VESSELS, WITH VARIOUSSURFACE TREATMENTStemperature/Kvessel, treatment 522 533 581unpacked, untreated 0.65 1.03 16.09unpacked, silicone oil washed 0.66 1.14 15.89unpacked, HMDS washed 0.78 1.24 14.30packed, untreated 4.86 5.32 22.51packed, silicone oil washed 3.52 4.24 16.94a 105kq/Torr-* s-l, b hexamethyldisilazane.All rate constants were corrected for inhibition which raised their values by about10 %. They were then put into an Arrhenius plot which is Shawn in fig.4. Theline shown is a least squares fit to all the data (over 100 runs in the temperature range519-618 K) apart from runs in the packed vessel and runs with low values for the ratio[Me3SiH]/[12]. This yieldsSince at lower temperatures surface contributions in the unpacked vessel occurred,an alternative calculation was done for the limited temperature range 567-61 8 Kyieldinglog(k+/Torr-* s-l) = (12.50 & 0.30) - (1 8 1.6 & 3.4 kJ mol-l)/RT In 10.The error limits quoted are one standard deviation (68 % confidence level).log(k+/Torr-* s-l) = (9.07 0.20) - (142.8 2.2 kJ mol-l)/RT In 10.To examine the inhibition phenomenon in more detail a series of runs was per-formed in which HI was added initially to the reaction mixtures.the effect and to try to minimise the errors associated with theiodine decay with time were obtained as before but these weresistency with the expressionThis was to enhanceanalysis. Traces ofnow tested for con-shown later (see mechanism section) to fit the proposed mechanism.Rearrangementof eqn (B) gives[12]3[Me,SiH] 1 k’ [HI] = -+--4CIzllclt k, k, CblFig. 5 shows a test of this equation in which - d[I,]/dt is approximated by the change106 D(Me,Si-H) BOND DISSOCIATION ENERGY1.60 1.70 1.80 1.90103 KITFIG. 4.-Arrhenius plot for k+sol I I I I I0 2 4 6 8 10[ K I I I E I Z IFro, 5 .P l o t showing a test of the inhibition eff' of HI on the reaction at 568 RR . WALSH AND J. M. WELLS 107-A12, over a h e d but small time interval (4 min). To use such plots to obtain thevalues of k' (= slope/intercept), the intercepts were fixed by taking an average valueof k3 appropriate to the temperature in question. The data fit linear plots within thescatter where the error bars on each point represent maximum errors arising fromuncertainties in the experimental traces. Fig. 6 shows an Arrhenius plot of the derivedvalues of k' from such experiments. The temperature range here was limited to-1.6 5 1.70 I .75103 KITFIG. 6.-Arrhenius plot for k'. See text for explanation of the two lines.567-618 K to avoid any risks of surface contributions.The scatter is again largeand the highest temperature point appears somewhat out of line with the others.Since these values depend somewhat on k3 used in deriving them it is interesting tonote that a lower k3 value (which would bring it closer to its Arrhenius line) wouldreduce k' at this temperature and bring it closer to the line suggested by the otherpoints. The two lines drawn correspond to inclusion and exclusion of this point.The corresponding Arrhenius lines are, respectively,andlog k' = (1.28kO.56)-(22.3k6.3 kJ mol-')/RTln 10log k' = (- 0.19 & 0.59) - (6.1 k 6.5 kJ mol-l)/RT In 10.REACTION MECHANISM AND THE BOND DISSOCIATION ENERGYTheand so,likely ;observed kinetics parallel very closely those between I2 and hydrocarbons loby analogy, the iodine atom abstraction chain mechanism below seems mostI2( + M) * 214 + M)123I-+Me,SiH * Me,Si-+HIMe&+ I2 + Me,SiI + I-.A stationary state treatment of this mechanism leads t108 D(Me,Si-H) BOND DISSOCIATION ENERGYThus by comparison with the experimental results both the order dependence andinhibition effect are accounted for.In addition k+ can be identified with k1@ andk' with k2/k3. From the known values l6 of Kr, the following expressions for kl areobtained from either all the data, or only the higher temperature selection, respectively,log(kl/dm3 mol-l s-l) = 9.90-71.2 kJ mol-l/RTln 10orThese figures are unfortunately rather far apart and furthermore the more limited(second) set of parameters, apparently free from surface effects, gives an A factorrather higher than the collision number ( N 1 0 1 1 e 3 dm3 mol-l s-l).The total set ofdata probably gives an A factor which is too low because of the inclusion of surfaceeffects. In the light of this, probably the best procedure is to select A = 1010*9 dm3mol-1 s-I by analogy with the hydrocarbon case,l1. viz., I-+i-C4Hlo --+ But-+ HI.By compensation of parameters this leads tolog(kl/dm3 mol-1 s-l) = 10.90-82.3 kJ mol-l/RTln 10.The Arrhenius line corresponding to these parameters is very similar to one obtainedby omitting the rate constants at the single highest and the two lowest temperatures.However, there is no compelling experimental reason for making such a selection.suggests that of the two possibili-ties for the inhibition constant the one where the highest temperature data are omittedis more likely.Hencelog(kz/k3) = -0.19-6.1 kJ mol-l/RTln 10.It seems reasonable to suppose that E3 = 0 since alkyl radicals react with Iz withno activation energy and Si-I bonds are probably stronger (see discussion) thanC-I bonds., Hence Ez = 6.1 kJ mol-l. Thus, AH1,,(593 K) = El--& = 76.2kJ mol-l. Correction of this enthalpy change to room temperature using he; =- 8.8 J mol-I K-l, estimated by thermochemical methods, yields AH;,* (298 K) =78.8 kJ mol-I. Since AHf.2 = D(Me,Si-H) - D(H-I), then from the knownvalue'lof D(H-I) = 298 kJ mol-1D(Me,Si-H) = 376 kJ mol-l.Error limits have not been included in the above figures because of the assumptionabout the A factor, which lay well outside the apparent precision of the data.Itwould seem most reasonable to assume that the A factor is within a factor of 10 ofthe true value in which case the uncertainty in D(Me,Si-H) is f 11 kJ mol-I.13.35- 110.2 kJ mol-l/RTln 10.Again comparison with hydrocarbon systemsDISCUSSIONDespite the uncertainties in the data, the kinetics found for this reaction offerstrong support for the atomic chain mechanism proposed. The only differencebetween this reaction and its hydrocarbon counterpart is the lack of reversibility ofthe overall process undoubtedly due to the more favourably negative enthalpy changewhich in turn arises from the increased strength of the Si-I bond relative to C-I.A possible alternative atomic mechanism is worth considering ;Iz( + M) * 21*( + M)I. /I-+ Me,SiH -+ (Me3% ) + Me,SiI +HR.WALSH AND f . M, WELLS 109The key step in this sequence is the displacement step (via a penta co-ordinated inter-mediate or transition state). However, it seems much more likely that if a displace-ment were to occur, methyl radicals rather than hydrogen atoms would be theproduct. No methyl iodide was detected amongst the products and so this mechanismappears unlikely.Bond dissociation energies determined by this method l1 are usually amongst themost reliable and so it is unfortunate that the error associated with the present deter-mination is so large. Nevertheless our value of 376 kJ mol-1 is in reasonable agree-ment with a figure of 364 kJ mol-l which may be derived from the latest determinationof D(Me,Si-SiMe,)l O and the appropriate heats of formation.20 Davidson andHoward O also quote separate unpublished electron impact measurements leading toa value of 368 kJ mol-l.Clearly the earlier figure of 340 kJ mol-l, obtained fromdata on the pyrolysis of Me,SiH is too low. Our figure substantially reduces thediscrepancy in relative methyl radical abstraction rates from SiH4 6 p ' and Me3SiH.6*On a per Si-H bond basis the rates are closely comparable, and the activationenergies are extremely close, being 29 and 33 kJ mol-1 respectively. The bondstrengths for the two Si-H bonds, with D(SiH3-H) = 397+5 kJ mol-l,18 differby 57 kJ mol-l on the basis of the old value for D(Me,Si-H) but by only 21 kJmol-1 from the present figure.Unfortunately it is not possible to predict methylradical abstraction rates from bond dissociation energies or the converse, but theyare usually c~rrelated.~~ l9 It is still slightly surprising that, on a per bond basis,methyl radicals react faster with SiH4 with its stronger Si-H bond than with Me,SiH.Perhaps D(SiH3-H), determined by electron impact studies,l is slightly too high.The correlation of HT yields from recoil tritium abstraction reactions has ledHosaka and Rowland to suggest that D(Me3Si-H) is around 356 kJ mol-l. How-ever, the correlation line is curved and the present value is not incompatible with thecurve. Yields of HT from Me,SiH, and MeSiH, imply that D(Me,SiH-H) andD(MeSiH,-H) are about 2-3 kJ mol-1 greater than D(Me,Si-H).We intend toinvestigate this point later.Heat of formation data for silicon containing compounds have been somewhatunreliable in the past. A recent compilation 2o gives AHf"(Me,SiH) = - 156 kJmol-lwhich, taken in conjunction with our value for D(Me,Si-H), leads to AHf"(Me,Si*) =+2 kJ mol-l. This figure is only on the borderline of compatibility with the valueof - 11 kJ mol-1 derived by Davidson in the Me,SiSiMe, pyrolysis work usingAHf"(Me,SiSiMe,) = - 359 kJ mo1-1.20 This suggests a possible remaining incon-sistency between AHf"(Me,SiH) and AHf"(Me,SiSiMe,). Used in conjunction withother molecular heats of formation20 our value for AHf"(Me,Si) leads to higherfigures for several bond dissociation energies.Noteworthy amongst these is thevalue of 380 kJ mol-l for D(Me,Si-Me). Although there is undoubtedly still someerror associated with it, this value is quite a bit higher than that for the analogouscarbon-carbon bonds. Apart from being surprising in itself, t h s fact, if true forother Si-C bonds, means that the initiation step in free radical organosilane pyrolyseswill be considerably slower than was tho~ght.~. This in turn implies that chainsequences of the Rice-Herzfeld type, previously ruled are likely to occur. Forexample, in Me,SiH pyrolysisCH; + Me,SiH + CH4 + =CH,SiHMe,-CH,SiHMe, + CH,=SiHMe + CH;is a probable cycle. Because some products are common to both propagation andtermination, the determination of chain lengths in these pyrolyses is not straightforward.The participation of divalent silicon intermediates lo* 2 2 is a possible added complica110 D(Me,Si-H) BOND DISSOCIATION ENERGYtion and it is clear that these pyrolyses will need careful reinvestigation if theirmechanisms are to be understood.Another important thermochemical quantity, the n-energy in silico-olefins, wasevaluated 23 from earlier kinetic studies with organosilanes, as lying between 119 and158 kJ moP.In the light of the new Si-C bond dissociation energies this requiresupward revision. A value close to 200 ( & 20) kJ mol-' may be estimated using similarthermochemical argurnent~.~~ This new value means that these n-bonds have about80 % of the strength of the n-bonds in olefins, considerably more than had been24 *APPENDIXIntegration of the approximate rate equation (A) givestan-l(f+)- tar1(& = *(b- a ) + ~ (C)wheref= [12]/(b-a), a = [I2lO and b = [Me3SiHIo.The integrated form of eqn (B) containing both unknown constants k , and k', gave veryimprecise values for them when fitted to the data.The method, described in the resultssection, of enhancing inhibition by HI addition to reaction mixtures gives inherently moreprecise values for k'. The implication of inhibition by HI is that the plots of eqn (C) shouldbe slightly curved. However, although slight curvature is observed, allowance for curvatureis best made by correcting the k, from the linear fit by an aoerage inhibition term (- 5-10 %).The data are insufficiently precise to warrant any more sophisticated treatment.The authors acknowledge the provision of both an equipment grant and a main-We thank Mrs.Diane King for help tenance grant (to J. M. W.) from the S.R.C.with some of the preparative work.S. W. Benson, Thermochemical Kinetics (Wiley, New York, 1968).I. M. T. Davidson, Quart. Reu., 1971, 25, 111.I. M. T. Davidson and C. A. Lambert, J. Chem. SOC. A, 1971,882 ; Chem. Comm., 1969,1276.E. Whittle, Chemical Kinetics, ed. J. C. Polanyi (M.T.P. International Review of Science,Physical Chemistry, Series 1, 1972), vol. 9, p. 75.E. R. Morris and J. C. J. Thynne, J. Phys. Chem., 1969, 73, 3294.552.A. Hosaka and F. S. Rowland, J. Phys. Chem., 1973,77,705.69.D. M. Golden and S .W. Benson, Chem. Rev., 1969,69,125.' H. M. Frey and R. Walsh, Chem. Rev., 1969, 69,103.' 0. P. Strausz, E. Jakubowski, M. S. Sandhu and H. E. Gunning, J. Chem. Phys., 1969, 51,* J. A. Kerr, D. H. Slater and J. C. Young, J. Chem. SOC. A, 1967, 134.lo I. M. T. Davidson and A. B. Howard, Chem. Comm., 1973, 323 ; J.C.S. Faruduy I, 1975, 71,l2 K. W. Egger and S. W. Benson, J. Amer. Chem. Suc., 1965,87,3313.l3 D. F. Ball, P. L. Goggin, D. C. McKean and L. A. Woodward, Spectrochim. Acta, 1960, 16,l4 M. G. Voronkov and Yu. I. Khudobin, Izvest. Akud. Nuuk S.S.S.R. Ser. khim., 1956, 713.l5 E. A. V. Ebsworth and S. G. Frankiss, Trans. Furudzy Soc., 1967, 63,1574.l6 J.A.N.A.F. Thermochemical Tables, ed. D. R. Stull and H. Prophet, NSRDS-NBS 37 (Nat.Bur. Stand., Washington, 2nd edn., 1971).H. Teranishi and S. W. Benson, J. Anzer. Chem. Suc., 1963,87,2887 ; see also J. H. Knox andR. G. Musgrave, Trans. Furaduy SOC., 1967, 63,2201.W. C. Steele, L. D. Nicholls and F. G. Stone, J. Amer. Chem. Suc., 1962, 84, 4441 ; see alsoF. E. Saalfeld and H. J. Svec, J. Phys. Chem., 1966,70, 1753.P. Gray, A. A. Herod and A. Jones, Chem. Rw., 1971,71,247.1972).1358.2o J. B. Pedley and B. S. Iseard, CATCH Tables for Silicon Compuunds (University of Sussex,* Some of these points are also contained in a recent review 2 5 which was not available to us whenthis paper was submittedR. WALSH AND J. M WELLS 11121 R. P. Clifford, B. G. Gowenlock, C. A. F. Johnson and J. Stevenson, J. Organometallic Chem.,22 M . A. Ring, M. J. Puentes and H. E. O’Neal, J. Amer. Chem. SOC., 1970,92,4845.23 R. Walsh, J. Organometallic Chem., 1972, 38, 245.24 R. A. Jackson, Essays on Free Radical Chemistry (Chem. SOC. Spec. Publ. No. 24, London,25 I. M. T. Davidson, Reaction Kinetics (Chem. SOC. Spec. Periodical Rep., London, 1975), vol. 1,1972, 34, 53.1970), p. 295.p. 212.(PAPER 5/738

ISSN:0300-9599    

DOI:10.1039/F19767200100

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Kinetic Isotope Effects and Tunnelling Corrections in theProton-transfer Reactions between 4-Nitrophenylnitromethaneand some Tertiary Amines in Aprotic SolventsBY EDWARD F. CALDIN” AND SALVADOR MATEO~University Chemical Laboratory, Canterbury, Kent CT2 7NHReceived 24th April, 1975Kinetic isotope effects for the proton-transfer reaction between 4-nitrophenylnitromethane andthe tertiary amine bases tri-n-butylamine, triethylamine and quinuclidine, in toluene, have beendetermined at temperatures between - 15°C and + 35°C by the stopped-flow technique. The rateratioskH/kDforthese basesat 25”Carerespectively 14+ 1,11.050.7and 15.6k0.6. Theratiosofthepre-exponential factors AD/& are respectively 1, 4 and 9. These values are interpreted in terms oftunnelling. Bell’s equations for an unsymmetrical parabolic barrier give almost the same barrierwidth for all three bases.Earlier results for such reactions in acetonitrile are interpreted in termsof coupling of the motions of solvent molecules with that of the proton.In previous communications we reported an investigation of the kinetic isotopeeffect and its temperature-variation in the proton-transfer (1)kikbNO2=C6H4.CH2NO2 + B $ NO2-C6H,=CHN0;. .HB+(ion pair) (1)between 4-nitrophenylnitromethane (4-NPNM) and the strong base tetramethyl-guanidine [TMG ; HN=C(NMe,),] in a series of aprotic solvents. The isotopic rateratios kH/kD and the ratios of the pre-exponential factors AD/AH in the less polarsolvents were large, and several of the Arrhenius plots were non-linear.The resultswere interpreted in terms of tunnelling. By fitting them to Bell’s equations for aparabolic barrier, assuming that only the motion of the proton or deuteron isrelevant so that the effective mass is mH = 1 or mD = 2 in atomic mass units, the bestvalues for the barrier dimensions were found. In five solvents, all with dielectricconstants below 6, the barrier width was nearly constant (78-82pm). In the morepolar solvents, with dielectric constants from 7 to 37, the calculated barrier widthswere markedly larger (90-96 pm) ; but on repeating the calculations allowing theeffective masses to vary, the best values of the barrier width were considerably smaller(78-80 pm) and agreed with those for the less polar solvents, while the apparent massesmH rose to 1.17-1.27 atomic mass units.These results were interpreted by a model in which, in the less polar solvents, thechange in solvation due to the charge-separation on forming the transition state iseffected by electronic polarisation without appreciable movement of solvent molecules,while in the more polar solvents the motion of the proton is coupled to motions ofsolvent molecules, so that the effective mass is raised.The reaction (1) is exceptionally suitable for a systematic study of tunnelling effects ;it is kinetically a simple reaction, which can be accurately followed since the productt permanent address : Universidad de Carabobo, Facultad de Ingenieria, Dep.de Quimica,Barbula, Edo. Carabobo, Venezuela.11E.F. CALDIN AND S . MATE0 113is intensely coloured, and the rates are conveniently in the range for stopped-flowtechniques.We now report on the corresponding reactions of 4-NPNM with three tertiaryamines, tri-n-butylamine, triethylamine and quinuclidine, in toluene. These form aseries in which the steric crowding of the lone pair on the nitrogen atom progressivelydecreases. The reactions of tri-n-butylamine and triethylamine in acetonitrile hadalready been ~tudied.~EXPERIMENTALMATERIALSThe preparation and deuteration of 4-NPNM, and the purification of toluene, aceto-nitrile, triethylamine and tri-n-butylamine were as before.2* Quinuclidine was obtainedfrom the hydrochloride by treatment with NaOH (aq.) and extraction with ether; theextract was dried over CaS04, the ether was distilled off, and the quinuclidine purified bysublimation in a vacuum at room temperature (m.p.153°C ; lit. 153°C) and used immediately.APPARATUS AND PROCEDURESThe stopped-flow apparatus and the experimental procedures 2* have been described.Concentrations were corrected for change of density with temperature. The kinetic experi-ments were performed under pseudo-first-order conditions with the base in large excess, atleast 25-fold and usually lo2 to lo3-fold. The first-order rate constant k(obs), determinedfrom the oscillogram by means of an exponential generat~r,~ is related to the rate constantsof eqn (1) by the equation :where b is the concentration of base after mixing of the reactant solutions.Plots of k(obs)against 6 were linear ; the number of points used was 4-6 (usually 5). Values of kf and kbwere derived by fitting the best straight line; this was done by a computer using a weightedleast-squares program. Arrhenius plots were treated in the same way. Uncertainties arestandard deviations.k(obs) = kfb+ kb (2)RESULTSABSORPTION SPECTRA AND EQUILIBRIUMThe product of the reactions, as with TMG, is a yellow solution whose absorptionspectrum shows a single broad peak with Amax = 420 nm. The absorbance measure-ments agree with the equation applicable to an ion-pair; if a = initial concentra-tion of 4-NPNM, b = concentration of base (in large excess), A = absorbance,E = molar extinction coefficient of the ion-pair, and K is the equilibrium constantfor eqn (l), the equation for the absorbance is :Good linear plots of a/A against llb, corresponding to eqn (3), were obtained.In tabb 1 are shown the results obtained from the spectrophotometric measure-ments for K.The values are in good agreement with those derived from the kineticresults in table 3 (K = kf/kb).Standard enthalpy and entropy changes may be calculated for the proton-transferreactions of triethylamine and tri-n-butylamine, though the errors are large ; the valuesare, respectively, - AH"/kcal mol-1 = 12 f 2 , 5.0 f 1.3 and - AS"/cal K-l mol-1 =37+8, 16+8. These may be compared with the more precise values derived fromthe kinetic results (table 3), which in the same order are 10.2+0.5, 7.8+0.2; 32+ 1,26+ I ; the agreement is adequate.alA = (1 /Kbe) + (1 I&). (3114 TUNNELLING I N PROTON-TRANSFER REACTIONSTABLE 1 .--EQUILIBRIUM CONSTANTS FROM SPECTROPHOTOMETRIC MEASUREMENTS FOR THEREACTION OF 4-NPNM WITH BASES IN TOLUENEbase temp./"C 104 ufmol dm-3 102 b/mol dm-3 A KH/dm3 mol-1(a) proton- transfer react iontriethylamine 15.0 0.7-2.9 0.7-3.1 0.05-0.84 6.3k0.620.0 1.1-3.6 1.2-3.8 0.10-0.97 5.7+ 0.625.0 1.1-3.6 1.2-3.8 0.08-0.74 4.2+ 0.330.0 1.5-4.5 1.6-4.8 0.09-0.86 2.6k0.1tri-n-butylamine 15.0 2.7-6.5 2.3-5.5 0.17-0.89 3.5+ 1.520.0 2.7-8.2 2.3-6.9 0.12-1.02 1.7k0.225.0 2.7-8.2 2.3-6.9 0.10-0.78 1.5+ 0.430.0 2.7-10.2 2.3-8.6 0.08-0.88 1.3k0.4quinuclidine 25.0 0.376 0.2-1.8 0.05-0.26 45+ 3(b) deuteron-transfer reactiontriethylanline 25.0 1.5-4.8 1.2-3.8 0.10-0.94 4.1k0.7tri-n-butylamine 25.0 3.5-8.5 3.2-7.7 0.14-0.79 2.3+ 1.0A = absorbance ; a = initial concentration of 4-NPNM ; b = initial concentration of base.TABLE 2.-KINETICS OF THE REACTION BETWEEN 4-NPNM AND BASES IN TOLUENEbasetrieth ylaminetri-n- butylaminequinuclidinebasetrieth y faminetri-n-butylaminequinuclidinetemp. /"C- 14.0 - 10.0- 5.00.05.010.0- 15.0 - 5.00.05 .o10.0- 15.0- 10.0- 5.00.05.010.0tcmp.l°C10.020.025.030.035.040.010.015.020.025.030.00.010.020.025.030.0(a) proton-transfer reaction104 a/ 103 b/ k(obs)/mol dm-3 mol dm-3 S-11.42 4.0-29.0 1.46-2.561.42 4.0-21.0 2.1-291.41 4.0-29.0 3.1-4.52.25 4.0-37.0 5.3-7.42.24 5.0-20.0 8.0-9.22.23 4.0-36.0 12.5-15.32.212.192.82.150.3840.3820.3800.3820.3800.3790.43-2.73.0-55.03.0-54.03.0-48.03.0-53.01 .O-6.91 .o-5.91 .O-6.81 .O-5.81.0-5.81.0-6.73.0-49.01.1-1.52.9-3.64.6-5.47.4-8.411.6-12.82.2-5.63.1-6.64.9-9.97.9-12.611.2-17.018.2-26.6(b) deuteron-transfer reactionlo4 a/ 102 blmol dm-3 mol dm-36.19 1 S-4.66.16 1 S-4.66.12 1 s-4.56.06 1.5-3.86.02 1.5-4.52.23 1.4-5.46.26 1.6-4.72.21 1.4-7.12.20 2.7-6.72.19 2.1-6.72.1 8 2.1-6.90.454 0.10-0.390.375 0.10-0.380.367 0.06-0.380.365 0.06-0.380.363 0.06-0.3844.3fl.047.9 f 3.458.4f0.265.5f 1.578.6f1.187.5f0.59.0f0.314.410.217.4 *0.221 Jf0.226.412.558517721 f 8855f4985A231201 f141438rt191.29f0.021.89fO.052.83 fO.015.05 *0.037.63 fO.0 112.14 fO.0 11.03 fO.O12.85fO.014.50&0.017.35f0.011 1.47 f0.071.65f0.022.40f0.024.05&0.016.96f0.0710.08f0.0416.86f0.07k(obs)/s-1 kyldm3 mol-1 s-10.11-0.28 5.7510.260.17-0.48 10.04fO.090.19-0.55 12.0 f0.50.20-0.60 129f0.30.24-0.61 15.5f0.60.34-1.07 23.2&0.40.03-0.09 1.48f0.220.04-0.15 1.99 f0.100.07-0.16 2.37f0.100.07-0.20 3.08 10.140.09-0.26 3.85 f 0.130.060.19 43.210.70.09-0.30 7 1.2f 1.00.09-0.46 11 5.9 f 1.20.09-0.56 145.2A3.60.11-0.67 184.6f3.3a = initial concentrations of 4-NPNM; b = initial concentrations of base; k(obs) = pseudofirst-order rate constantE.P. CALDIN AND S. MATE0 115KINETIC RESULTSThe rate constants found for the proton- and deuteron-transfer reactions aresummarised in table 2.To save space, only the ranges of values of k(obs) and of theconcentrations are recorded, with the derived values of kf and kb and their standarddeviations ; complete data for triethylamine and tri-n-butylamine are in the originalthesis.6 The proton-transfer reaction is too fast to be measured with the stopped-flow apparatus at temperatures above 1 O"C, while the deuteron-transfer reaction istoo slow below 10°C. The oscillograph traces were always exponential, and thepseudo-first-order conditions were checked in many cases by varying the initialconcentration of 4-NPNM ; no discrepancies were found.TABLE 3.-ISOTOPE EFFECTS, EQUILIBRIUM PARAMETERS, AND ACTIVATION PARAMETERS FOR THEREACTIONS OF 4-NITROPHENYL"IROhfE"HANE WITH BASES IN TOLUENEkF(25"C)/dm3 mol-' s-IAHzD/kcal mol-l1og,,~F/c1m3 mol-1 s-1-AS:~/C~I K-, mo1-1kr / kF(25 "C)M:D-AHffH/kcal mol-Ilog, OAFIAYA 3 A Fkf(25"C)/s-'AHzH/kcal mol-1logA:/s-'-AS: /CEJ K-1 mol-1K"(25OC)/dm3 mol-'*-AHoH/kcal mol-'*-AS"H/cal K-I mol-l*pK in water (25°C)tetramethyl-guanidinet2290f 93.62 + 0.056.45+ 0.0431.0k0.251 k2.57.9+ 0.27.95+ 0.1524.1 0.745224.3 f 0.31.SO+ 0.1 931f1412.7f0.514.2+ 1.211.9+ 0.96.0f 3.8179+ 610.6+ 1.725+413.6quinuclidine2200& 154.54f0.097.1 1 st: 0.0728.0f0.3145_+47.33 f 0.077.97k0.0624.0_+ 0.215.6f0.62.79+ 0.160.86f0.127.2+ 254f 212.7 f 0.511.5+0.4S.l+l.541+28.2+ 0.620& 210.95triethylamine132&23.51 f0.075.11 f 0.0637.2+ 0.212.0+ 0.55.7k0.55.7k0.434f 2ll.Ok0.72.2f0.60.6+ 0.454+233+ 11 3.7 & 0.412.1k0.35.2f 1.24.0k0.110.2+ 0.532+ 110.87tri-n-buty lamine43.0& 0.45.61k0.146.21 kO.1132.1k0.53.08 & 0.147.3 0.46.25f0.332+ 1.314+ 11.7+ 0.50.04f 0.401.139.8 f 0.513.38 & 0.061 1.84+ 0.056.4f 0.2l.l+O.l7.8k0.226f 110.60* Values from kinetic measurements ; t values from ref.(2) ; 1 cal = 4.18 J.In table 3 are summarised the isotope effects and activation parameters, and thethermodynamic parameters derived from the kinetic results, for the reactions of4-NPNM with quinuclidine, triethylamine and tri-n-butylamine, in toluene ; thevalues for TMG are added for comparison.The values for the backward deuteron-transfer reactions are not accurate enoughto justify their use in deriving activation parameters or isotope effects116 TUNNELLING I N PROTON-TRANSFER REACTIONSDISCUSSIONEQUILIBRIUMThe reactions of 4-NPNM with quinuclidine, triethylamine and tri-n-butylamineare in general similar to those with TMG, where the product is shown by its p.m.r.spectrum (in acetonitrile) to be NO2-C6H4.CHN0;.The visible absorption spectrawith the three tertiary amines are similar to those with TMG. The equilibriumconstant determinations show that the product is an ion-pair. There is thus noreason to doubt that the reactions are represented by eqn (1).The equilibrium constants in toluene (tables 1 and 3) are in the same order asthe pKvalues in water, but cover a smaller range.The values of AHo and AS" varysomewhat irregularly, but in all cases AHo is negative and is partly compensated bya considerable negative value of AS".KINETICSGENERALThe kinetics of the reactions with the three amine bases resemble generally thoseof the reaction with TMG and follow the simple scheme given in eqn (1). There areconsiderable negative entropies of activation, for deuteron-transfer as well as proton-transfer, indicating an increase of solvation in the transition state; the values aremarkedly more negative for triethylamine and tri-n-butylamine than for quinuclidine(and TMG). There is no regular relation between AH: and AS:.The Brranstedplots of logkH and logkD against logKH for the three amine bases are fairly goodstraight lines, with slopes of about unity (the point for TMG lies below the line ineach case).KINETIC ISOTOPE EFFECTS FOR REACTIONS I N TOLUENEThe isotopic rate ratios for the forward reactions in toluene (table 3) are all largerthan can be explained on the semi-classical model by loss of the zero-point energyof the stretching C--PI vibration in the transition state ("7) ; so are the values ofThe values of AF/Ap vary from about 1 for tri-n-butylamine to 7 2for quinuclidine; the latter value is well outside the semi-classical maximum, andsuggests that there is an important tunnelling correction.-CALCULATIONS ON TUNNELLING FOR REACTIONS I N TOLUENE(a) We have applied Bell's equation 7-9 for tunnelling through an unsymmetricalparabolic barrier, as in our previous paper,2 taking the first four terms of the equation,and using a computer program.It was assumed first that the effective masses mHand mD are 1 and 2 a.m.u. respectively, and that the barrier width 26 is the same forprotons and deuterons. The limitation was also imposed that the barrier heights forH and D do not differ by more than the physically-reasonable maximum of 1.2 kcalmol-'. The barrier dimensions that gave the best agreement with the isotopic rateratios over the whole temperature range were then found. The results are given intable 4, with those for TMG for comparison. The values shown reproduce theexperimental isotopic rate ratios within +1 % for triethylamine and kl2 % fortri-n-butylamine over the whole temperature range.It is seen (i) that the calculatedbarrier width varies little for the three amine bases (0.96-0.97A), and is larger thanfor TMG (0.79A); (ii) that the calculated barrier height E decreases in the orderquinuclidine > tri-n-butylamine > triethylamine ; correspondingly the curvature vand the tunnelling correction Q at 25°C decrease in the same order: and (iii) thaE. F. CALDIN AND S. MATE0 117the values of (ED--EH) are all, within experimental error, equal to 1.1 kcal mol-I, thevalue expected for complete loss of the stretching C--H vibration in the transitionstate.(b) Calculations were performed secondly (as in our previous work), on theassumption that the effective masses may vary, the only constraints being that(rnb-rnfI) = 1 a.m.u., and that (ED-EH) does not exceed 1.2 kcal mol-l.For thereaction with TMG in toluene, the resulting optimum value for the barrier width (2b')was unchanged from its former value, and the effective mass was still 1 a.m.u. Forthe reactions of the three amine bases, the optimum values of 2b' were close to thatfor TMG, while those of the effective mass rnfr were appreciably larger than 1 a.m.u.,and fairly constant, that for quinuclidine being slightly larger than for triethylamine(table 4). The values reproduce the experimental isotopic rate ratios somewhat betterthan those found by calculation (a) ; but this fact is not necessarily significant, becausethe barrier shape is not realistic.TABLE 4.-TU"ELLING CORRECTIONS AND PARABOLIC-BARRIER PARAMETERS FOR THE REAC-TIONS OF ~NITROPHENYLNITROMETHANE WITH VARIOUS BASES IN TOLUENEtetramethyl-&dine* quinuclidine triethylamine tri-n-butylamineEH/kcal mot1Plkcal mol-l2b/AuH(25"C)uD(25OC)Q"(25"C)Q"(25"C)vH/cm--lvD1crn-lED- FI/kcal mol-'E 2 I PE3ED2bf /A8.60f 0.059.75f 0.050.788 f 0.0026.865.0628f24.3k0.1142010471.152 0.10.490.870.7887.00+0.058.20+0.050.96oIf: 0.0025.033.764.082 0.071.97k 0.0310407661.20+ 0.10.730.970.7905.301 0.056.35* 0.050.9721 0.0024.773.553.2850.051.821 0.029907391.05 0.10.770.990.7886.65 k 0.057.801 0.050.970+ 0.0024.873.643.62f 0.051.88*0.0210107551.15f 0.10.931.010.778m;I/a.m.u.1.00r40.01 1.32k0.01 1.30+0.01 1.3510.01* from ref. (2). All symbols as in ref (2). 1 cal = 4.18 J. Ea = Arrhenius activation energy.BARRIER DIMENSIONS FOR THE REACTIONS I N TOLUENEThe barrier width is the distance that the proton moves in the change from&Ha - .NH< to %C- .HNH<. It must, therefore, depend on the length ofthe +N-H bond and the CH. * .N distance at which proton-transfer occurs, whichwill depend on the electron-distribution around the nitrogen atom in the base mole-cule. We can therefore understand the results of calculation (a) above, in which thecalculated barrier width 2b is about the same for the three amine bases, where theorbitals of the N atom are sp3 hybridised, but smaller for TMG, where the hybridisa-tion is sp2 and the more compact orbital on the N atom will allow closer approachof the acid molecule before reaction.The apparent barrier width 2b calculated as above is sensitive to any motions ofheavy nuclei that are coupled to the transfer of the proton.Solvent motions mightbe so coupled. With TMG as base, however, the calculated banier width in tolueneand other solvents of low polarity has been found to be nearly constant,2 and it appears118 TUNNELLING IN PROTON-TRANSFER REACTIONSthat effects of coupling are not appreciable in these solvents (though they are probablyresponsible for the variations of barrier width found in more polar solvents such asacetonitrile; see also below).There seems no reason to expect coupling of solventmotions for the reactions of the other bases in toluene.Changes of configuration in the reactant molecules (from tetrahedral towardsplanar, for example) accompanying proton-transfer might also be coupled with themotion of the proton; this would again lead to an increase in 2b. If we were toadopt this interpretation, the results of calculation (b) would imply (i) that the barrier-width 26' is nearly the same for all four bases; (ii) that the effects of the coupledconfigurational changes, which are measured by the values of the effective mass mfr,are considerably greater for quinuclidine (mfI = 1.32 a.m.u.) than for TMG (forwhich they are negligible ; mfI = 1.00 a.m.u.) ; and (iii) that these effects are aboutequal for quinuclidine, triethylamine and tri-n-butylamine (mfI = 1.32, 1.30 and 1.35a.m.u.respectively). Of these conclusions, (ii) and (iii) seem unlikely, since protona-tion of TMG leads to a change of hybridisation which will affect bond lengths andangles, and protonation of NEt, or NBu, will likewise lead to a change in bondangles,'O while for quinuclidine these configurational changes are restricted by thering system. We therefore regard as more likely the interpretation based on calcula-tion (a), i.e., that the barrier width for TMG is smaller by about 0.2A than for theamine bases, and that the effects of any configurational changes in the amine moleculesare relatively small. The larger isotope effect with TMG would then be due to thenarrower barrier ; the difference between quinuclidine and the other amine bases wouldbe attributed to the greater height (and hence curvature) of the barrier for quinuclidine.Experimental data on a wider variety of bases are required before the role of coniigura-tional change can be assessed.The barrier height might be affected by the strength of the base, by solvation ofthe base, or by steric factors.Bulky substituents in the reactant molecules l1-I4could affect it, either because of increased repulsion forces or because crowding ofthe reaction site leads to exclusion of solvent molecules and so prevents lowering ofthe energy barrier by solvent reorganisation." Either effect would explain why thebarrier is higher for NBu3 than for NEt3 ; the former explanation is the more likelyof the two, since the entropies of activation for deuteron-transfer are nearly the samefor the two bases, suggesting that the solvation changes are not greatly different.(The entropies of activation for proton-transfer give less direct evidence on solvationchanges than those for deuteron-transfer, since they are complicated by differingtunnelling corrections).With quinuclidine, the barrier is even higher, although stericinterference near the reaction site must be less ; some other factor must be responsible.This factor could be a steric effect of a different kind, on the change of solvation onforming the transition state. The entropy of activation for quinuclidine is markedlyless than for NEt3 and NBu3, and suggests a smaller increase in solvation.Modelsshow that when the base is quinuclidine its cage structure makes it impossible forsolvent molecules to approach the nitrogen atom (where the incipient positive chargewill be concentrated) in the reaction complex +C- - -He - .N on the side remote fromthe proton, whereas with ethylamine there is much less hindrance from the shortcarbon chains. These interpretations are necessarily somewhat speculative ; moreexperimental evidence is required.+CALCULATIONS ON TUNNELLING FOR REACTIONS IN ACETONITRILEThe kinetic isotope effects for the reactions of triethylamine and tri-n-butylaminein acetonitrile have already been deterrnh~ed.~ They are much smaller than in toluenE .F. CALDIN AND S . MATE0 119(kH/kD = 2 to 3), but again show values of A"/AH that are greater than unity andthus suggest tunnelling. We have, therefore, applied Bell's equations to the results,in the way described above, and found the best values of the barrier height and thewidth 2b, assuming in the first place that mH = 1 and mD = 2 atomic units. (Sinceno values were available for AH", it was assumed to be zero.) The results are givenin table 5, with those for the reaction with TMG in acetonitrile for comparison.They are, of course, less reliable than those for the reaction in toluene, where theeffects are larger and the value of AH" is known.The values found for the barrier width 2b are much larger in acetonitrile than intoluene (as was found for the reaction with TMG) and they are markedly differentfor NEt, and NBu,.It is hard to see how a change of solvent could bring aboutsuch large and different changes in the width of the barrier. We have thereforesupposed that the tunnelling correction is affected by coupling of the motions ofsolvent molecules with that of the proton, which will increase the effective mass, andhave repeated the calculations, as before [calculation (b) above), taking the barrierheights already calculated as correct, but allowing the effective masses mfi and m6 tovary, with the constraints that (mb-mfI) = 1 and that (ED-EH) #- 1.2 kcal mol-'.The best values of the effective masses and of the barrier width 2b were then found.These values reproduce the experimental isotopic rate ratio within +2 % over thewhole temperature range.The results of these calculations (table 5) show that there are two distinct sets ofvalues of the barrier width and effective mass that are compatible with the experi-mental results.(i) In one set, the values of the barrier width (2b') for NEt, and NBu, in aceto-nitrile are nearly equal to each other and to the values of 2b in toluene.This resultis in accordance with the view that the barrier width is nearly the same for both basesin both solvents, while the effective mass of the particle transferred (mfi) in acetonitrileis greater than in toluene for all bases, and somewhat greater for NBu, than for NEt,.The values of mfI (1.27-1.52 a.m.u.) are all appreciably greater than unity, and maybe attributed to coupling of molecular motions with that of the proton during itstransfer.(ii) In the other set, the values of the barrier width (2b" in table 5) in acetonitrilefor NEt3 and NBu, are again nearly equal, but they are appreciably smaller thanin (i), and nearly equal to the value for TMG and to the value of 2b' in toluene(table 4).The fit is slightly better than for (i), but as we have noted this is notnecessarily significant. These values would imply that the barrier width is nearlyconstant throughout, independent of base as well as of solvent. We have, however,seen reason to doubt such a conclusion (see above). We therefore incline to the viewthat the more likely state of affairs is (i), i.e. the true barrier width for each base isthe same in acetonitrile as in toluene, the smaller isotope effects in acetonitrile beingdue to coupling of the motions of solvent molecules with that of the proton.Whichever values we adopt, the values of the effective mass (mfi or mff, in table 5)suggest that in acetonitrile solution the coupling of molecular motions with that ofthe proton is greater than in toluene, and varies with the base, being greatest withtri-n-butylamine and least with TMG.If cofigurational effects are small, as theresults in toluene suggested, the variations in the effective mass will reflect differencesin the reorganisation of the solvent molecules in the field of the polar transition state.Beyond this point, interpretations can only be somewhat speculative.Less solventreorganisation for TMG might be due to less solvation of the transition state; thecharge density at the nitrogen centre will be less than for the tertiary amines becausethe positive charge is delocalised. The larger effective masses wit2.i NBus compare120 TUNNELLING I N PROTON-TRANSFER REACTIONSwith NEt, are puzzling ; the steric effect of the alkyl chains would be expected to giverise to less solvent reorganisation for NBu, than for NEt3, contrary to observation,since the longer n-butyl chains will tend to exclude solvent molecules from the reactionsite more effectively than the short chains of NEt,.TABLE 5.-KINEnC ISOTOPE EFFECTS, TUNNELLING CORRECTIONS AND PARABOLIC-BARRIERPARAMETERS FOR THE REACTIONS OF 4-NITROPHENYLNITROMETHANE WITH VARIOUS BASES INkYlkF(25"C)(AH? D- AH," H)/kcaI mo1-IAD/AH-AHo/kcal mol-1QY25"C)QD(25"C)vH/cfnlvD/cm-lEH/kcaI mol-'EDJkcal moV12b/AlogloAD/AHE,Hf/EH& / E D(ED- ER)/kcal mol-'v'H/cm--lv'D/cm-rn;I/atomic units2b'/A&/atomic units2b'/AACETONITRILEte tramethylguanidine$ triethylaminet1 l.8+ 0.3 3.12 + 0.241.46k0.2 1 .00+ 0.2O.O+ 0.08 0.25+ 0.071 .Of 0.2 1.8k0.34.3 *O2.56f 0.1 1 -67 rfr 0.021.64f 0.1 1.30rt0.02890 685675 5015.85 5 0.05 7.705 0.057.00-t 0.05 8.20+ 0.050.960+ 0.002 1.232+ 0.0020.85 0.950.92 1 .o1.15 0.5956 739786 5801.27+ 0.01 1.39& 0.010.794 0.9701.27k0.01 1.825 0.010.794+ 0.002 0.770+0.002tri-n-butylaminet2.17+ 0.070.80f0.20.23 & 0.061.7f 0.3*O1.42& 0.021.20 rfr 0.025804226.85$.0.057.25+ 0.051.380+ 0.0020.961 .o0.46575251.52*0.010.9922.05f_ 0.010.788 rfr 0.002* assumed ; t experimental data from ref.(3) ; $ from ref. (2). 1 cal EZ 4.18 J. Ea = Arrheniusactivation energy.CONCLUSIONThis study is complementary to previous work on the reaction of 4NPNM withtetramethylguanidine. With tertiary aliphatic amines in toluene, the isotopic effectsare large (though smaller than with TMG), and the tunnelling corrections are con-siderable ; their variation from base to base appears to depend on solvation differencesas well as steric effects, but there is no clear evidence for effects due to configurationaldifferences. The effects of changing the solvent to acetonitrile are similar to thosefound with TMG but smaller.We acknowledge helpful discussions with Professors R. P. Bell and J. R. Keeffe,and with our colleagues Drs. N. J. Bridge, B. H. Robinson and C. J. Wilson. Oneof us (S. M.) acknowledges support from the University of Carabobo, VenezuelaE . F . CALDIN AND S . M A T E 0 121E. F. Caldin and S . Mateo, Chem. Comm., 1973, 855.E. F. Caldin and S. Mateo, J.C.S. Furuduy I, 1975, 71, 1876.E. F. Caldin, A. Jarczewski and K. T. Leffek, Trans. Furachy SOC., 1971,67, 110.G. Davies, Inorg. Chem., 1971, 10,1155.J. E. Crooks, P. A, Tregloan and M. Zetter, J. Phys. E, 1970, 3,73.S. Mateo, Ph.D. Thesis (Kent, 1974). ' R. P. Bell, Trans. Furuduy SOC., 1959, 55, 1.13 R. P. Bell, W. H. Sachs and R. L. Tranter, Trans. Faruhy SOC., 1971, 67, 1995.R. P. Bell, The Proton in Chemistry (Chapman and Hall, London, 2nd edn., 1973), chap. 12,p. 275.London, 1975), chap. 10, p. 333.lo H. C. Brown, H. Bartholomay and M. D. Taylor, J. Amer. Chem. SOC., 1944, 66,435.l1 E. S. Lewis, in Proton-transfer Reactions, ed. E. F. Caldin and V. Gold (Chapman and Hall,l2 E. S. Lewis and L. H. Funderburk, J. Amer. Chem. Suc., 1967,89,2322.l 3 H. Wilson, J. D. Caldwell and E. S. Lewis, J. Org. Chem., 1973, 38,564.l4 E. Grovenstein and F. C. Schmalstieg, J. Amer. Chem. Soc., 1967, 89, 5084

ISSN:0300-9599    

DOI:10.1039/F19767200112

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Critical Properties of Binary MixturesBY COLIN P. HICKS*?Chemistry Department, University of the West Indies, Kingston 7, JamaicaCOLIN L. YOUNG$Department of Physical and Inorganic Chemistry, University of New England,Armidale, N.S.W. 2351, AustraliaANDReceived 25th March, 1974Gas-liquid critical pressures of mixtures of octamethylcyclotetrailoxane with neopentane, 2,3-dimethylbutane, cyclopentane and tetramethylsilane, critical temperatures of octamethylcyclotetra-siloxane + tetramethylsilane mixtures and the vapour pressure of tetramethylsilane near the criticalregion are reported. The critical temperature and pressure data for these and other mixtures arecompared with values calculated from the van der Waals one fluid theory assuming different combiningrules and equations of state.The limitations of the criticality condition solutions are discussedbriefly.We have previously lm3 reported measurements of the mixture critical temperaturesfor a wide range of quasi-spherical molecules of different sizes. In this paper wereport critical pressure measurements on mixtures of the large quasi-spherical moleculeoct amethylcy clo te trasiloxane (0 MCTS) with three quasi-spherical alkanes and tetra-methylsilane. A comparison of experimental critical temperatures and pressures ismade with values calculated from the van der Waals one fluid theory for these systems,and for values taken from the literature for hydrocarbon + n-alkane systems.The prediction of gas-liquid critical properties of hydrocarbon + n-alkane mixturehas been investigated by several Often a random mixing lo or relatedmodel has been used together with a modification of the approximate solution of thecriticality conditions given by Rowlinson.OSpear, Robinson and Chao used a more general solution to the criticality condi-tions based on earlier work by Kuenen,l together with the Redlich-Kwong equationof state.12 Kay and co-workers 5 * used an iterative solution to the criticalityconditions similar in concept to that used here, but with important differences. Theycompared the predictions of the Redlich-Kwong, Dieterici and Redlich-Ngo l4equations of state. Teja and Rowlinson studied the critical properties predicted whenthe Vennix and Kobayashi 37 equation of state for methane is used in an iterativesolution to the criticality conditions.In this work we have used a general solutionto the criticality conditions with two different equations of state.The theoretical interpretation of the critical properties of non-polar mixtures as afunction of composition is usually achieved by the assumption that there exists ahypothetical equivalent (pure) substance, which has the same configurational freeenergy as the mixture for specified conditions.present address : Division of Chemical Standards, National Physical Laboratory, Teddington,Middlesex TWl1 OLW$ present address : Department of Physical Chemistry, University of Melbourne, Parkville,Victoria 3052, Australia.12C. P . HICKS AND C. L. YOUNG 123The combinatorial contribution to the free energy is assumed to be separable andindependent of volume or pressure. A number of further assumptions are then madeconcerning the equivalent substance in order that its configurational free energy maybe evaluated for any specified conditions : (1) that the equivalent substance obeys aparticular reduced equation of state ; (2) that the reducing parameters for the equationof state may be obtained from a particular " recipe " or " prescription " which dependsonly upon composition and the energy and volume parameters characterising inter-actions between like and unlike molecules ; (3) that the parameters characterising theinteraction between unlike molecules may be obtained from combining rules appliedto the parameters for the interactions of like molecules.The equivalent substance, and associated assumptions, enable the criticalityconditions lo(a2G/i3x2),,, = 0 (1)(i33G/i3~3)p,, = 0 (2)to be solved for the critical temperature, volume and pressure.There is good evidence that the van der Waals one fluid prescriptionTisV& = C xixjTf,V,cii j( 3 4is a reasonable assumption for molecules of similar sizes lS* l6 (i.e., size ratios ofV2/V1 < 2).In this pair of equations Tts and VEs are the critical temperature andvolume of the equivalent substance, while if i = j the critical properties are those ofthe pure components, and if i # j the hypothetical critical properties are calculatedby means of the combining rules.The most used combining rules for the critical temperature and volume are theLorentz-Berthelot combining rules5 = 1The parameters 5 and q, or their equivalents, have been introduced by various workersto increase the usefulness of the combining rules.lo* 15-23 t is usually less than unity even for simple systems.15* 2o For more complicatedsystems is still usually less than unity 1-3* 6* ' but there are several cases where itis unity or slightly greater.2* 23The solution of the criticality conditions is not straightforward, but for not toodissimilar species a first-order approximation O gives a solution for gas-liquid criticalpoints which is satisfactory for most purposes.24 Previously we have used the van derWaals equation of state in this first-order approximation to calculate critical tempera-tures.l* 2* 23 Several other equations of state give broadly similar values.Howeverthis is not the case for critical pressures and volumes, which are sensitive to the reducedequation of state assumed for the equivalentIn this work we consider three different solutions of the criticality conditions. Oneof these solutions uses the approximate solution with van der Waals' equation of statefor the equivalent substanCe,'O and two use a general (rigorous) solution. One of thegeneral solutions uses a double Taylor expansion 25 about the critical point of theequivalent substance, and the other uses the equation of state proposed by Bjerreand B a l ~ . ~ 124 CRITICAL PROPERTIES OF BINARY MIXTURESEXPERIMENTALC H E MI C A LSOMCTS was prepared by the hydrolysis of dichlorodimethylsilane '' and purified on a1 m spinning band distillation column and then by fractional crystallisation.Gas chromato-graphy revealed only about 0.02 mol % impurity (probably decamethylcyclopentailoxane).Neopentane, 2,2-dimethylpropane, was puriss grade obtained from Fluka. The statedpurity was 99.92 mole %. 2,3-dimethylbutane was a National Physical Laboratory sampleof purity 99.89 mole %.Cyclopentane was a Phillips sample with a minimum purity 99.96 mole %.Tetramethylsilane was n.m.r. grade obtained from PCR Inc., Gainsville, Florida.APPARATUSThe sealed tube method was used to measure the critical temperatures of the OMCTS+tetramethylsilane mixtures.The apparatus for measuring critical pressures was a modification of that described else-where.28 In that publication, its operation for measuring the vapour pressure of pure sub-stances as a function of temperature was described.In the present work the sample wasconfined over mercury in a J-tube. The J-tube used in the present work had a demountableglass-to-metal-to-glass high pressure seal between the portion (A) which was heated in thefurnace and the portion kept at room temperature (B) (see fig. 1).28 The two componentswere injected into the part (A) using syringes with long Teflon needles. The total samplelength was about 15 mm. The seal was assembled and the J-tube attached to the vacuumline. The sample was frozen in liquid air, the J-tube evacuated, the sample melted and re-frozen and the J-tube evacuated again.Mercury was then poured on the sample as in thecase of determining the vapour pressure of pure substances. The J-tube was topped up withhydraulic fluid coupled to the dead weight tester and aligned in the furnace (while cold).The furnace was then switched on and the temperature rapidly raised to about 20 K belowthat of the lowest temperature studied. The stirrer was started. It consisted of a small ballbearing with about 0.1 mm clearance between the capillary tube walls. The ball was movedup and down through the sample by a pair of magnets at the rate of 30 strokes per minute.The temperature was raised slowly at the rate of 0.1 K min-I and the pressure adjusted sothat there was a minute vapour phase. The pressure was then increased slightly by addingsmall weights to the dead weight tester until the vapour phase disappeared.The temperaturewas measured when the vapour phase reappeared with the dead weight tester weights balancedby the vapour pressure of sample and hydrostatic heads of mercury and hydraulic fluid. Thisprocedure was repeated about 10 times. Dew points were observed by measuring the tem-perature at which liquid phase first appeared on lowerihg the temperature at a similar rateto that used on the heating cycle, the pressure being constant. At temperatures in excessof the critical temperature two dew points were observed at the same temperature, i.e., retro-grade condensation occurred [see ref. (lo), (29) and (30) for a discussion of this phenomenon].RESULTSThe critical properties of the pure components are given in table 1, and the criticaltemperatures of OMCTS + tetramethylsilane mixtures in table 2.The composition of the mixture was not accurately known in the work on criticalpressures because it was difficult to degas the sample effectively and to be sure thatnegligible sample was lost, the total sample size being of the order of mol.Thecomposition could, however, be estimated satisfactorily from the known compositiondependence of Ti.2* (We assume that Tk is that of a binary system and not affectedby having mercury in the heated zone.)The measured pressures were corrected for the difference between the earth'sgravitational field at Armidale (9.791 11 m s - ~ ) and that for which the dead weighC.P. HICKS AND C . L. YOUNG 125TABLE l.-cRITICAL PROPERTIES OF THE PURE COMPONENTSsubstana Tc/K 10-5 m-2 Vc/cmJ mol- 1OMCTS 586.5 (28) 13.40 (28) 984 (28)c yclopent ane 511.6 (38) 45.08 (38) 260 (38)neopen t ane 433.75 (38) 31.99 (38) 303 (38)2,3-dimet hyl but ane 499.93 (38) 31.27 (38) 358 (38)tetramet hylsilane 450.4" 28.14* 357 -fReference numbers are given in brackets.* this work ; estimated assuming ZC = 0.269.TABLE 2.-cRITICAL TEMPERATURES OF OMCTS( 1) + TETRAMETHYLSILANE(2) MIXTURESx1*0.092 90.198 20.233 50.402 9TcIK t475.8498.0506.9535.1X10.445 00.543 00.693 40.829 9T a l e537.8548.6568.0575.9* estimated error k 0.001 ; t estimated error t- 0.3 K.TABLE 3 .-CRITICAL TEMPERATURES AND PRESSURES OF OMCTS MIXTURES0.8650.6450.4900.2550.1 150.9550.6750.3200.2450.2100.1980.1 700.9250.41 50.3250.21 50.1600.9500.5550.3950.3420.2820.145Tc/KOMCTS(l)+ cyclopentane(2)583.7 19.09573.8 25.90565.8 30.16547.4 38.77OMCTS(l)+ neopentane(2)585.3 15.03565.6 23.40526.5 33.29509.7 35.93501.8 37.01497.7 37.21490.5 37.35529.8 43.8618.7325.6029.9138.6243.7614.6623.1633.1935.8736.9637.1637.310 MCTS( 1) + 2,3-dimet hyl but ane(2)583.6 16.45 16.10555.0 26.73 26.54545.9 28.69 28.53532.2 31.14 3 1.02524.6 32.51 32.42OMCTS( 1) + tetramethylsilane(2)585.0 15.13551.2 25.16533.7 28.69527.5 29.76518.1 31.18488.1 33.4914.7624.9928.5729.6731.1133.46(i) without mercury vapour correction ; (ii) with mercury vapour correction given by eqn ( 5 )aestimated from independent x, Tc measurements ; b estimated error k0.5 K ; =estimated errorIf: 7 kN m-2126 CRITICAL PROPERTIES OF BINARY MIXTUREStester was calibrated (9.806 65 m s-~).It was also necessary to make some estimateof the effect of mercury on the measured pressures. This effect has been investigatedby Pak and Kay,6 who used gallium in place of mercury as the pressure transferringmedium in the heated zone. The vapour pressure of gallium is negligible at thecritical temperatures of organic compounds. They found that the partial pressureexerted by the mercury in the hydrocarbon was about 10-14 % less than the vapourpressure of pure mercury at the same temperature, and proposed that the partialpressure of mercury (PHg) be calculated from the equation(5) log,,(P,,/N m-2) = 9.765 72 - 3037.6 (K/T).TABLE 4.-vAPOUR PRESSURE, p , OF TETRAMETHYLSILANET/K lO-sp/N m-2 TIK 10-5plN m-2450.4448.1444.3442.0439.2437.728.14 434.5 22.2527.16 43 1.7 21.2726.18 429.0 20.2925.20 425.3 19.3124.20 422.7 18.3323.22 419.1 17.35We have used this equation, although it is doubtful whether it gives PHg to withinbetter than about +25 % over the temperature and pressure range studied here.The critical temperatures and pressures of the four mixtures studied are given intable 3, together with the compositions estimated from the known compositiondependence of T'L.The vapour pressure of tetrainethylsilane was also measured in this work, and isgiven in table 4.THEORYTo solve the criticality conditions we assume that the free energy of the mixturecan be divided into a combinatorial part given by the Flory theory,31 Acb, and aninteractional part, AT,, which is the configurational free energy of an assumed equi-valent substance.A = (6)(7) = A:,+RT(x, In 4' +x2 In &)where#, = x,/(x, +rx,) = 1-42.If we make the corresponding states assumption thatA2 = A:S(T/fes VIhes) (9)where fes and he, are temperature and volume reducing parameters respectively, thenwe can proceed to solve the criticality conditions in terms of the properties of theequivalent substance.APPROXIMATE SOLUTIONRowlinson lo has presented an approximate solution to the criticality conditionswhich we have discussed and used previously.'.2 * 2 3 For the mixture critical tem-perature, Tk, and critical pressure, p;, the solutions arC. P. HICKS AND C. L. YOUNG 127where the equation of state properties on the right hand side of these equations arefor the equivalent substance at its critical point; TZs, & and VSs are the criticaltemperature, pressure and volume respectively of the equivalent substance, andf:s = (afeslax2) (12)h6s = (aheslax2). (13)Although eqn (10) and (1 1) are only approximate, they have been derived withoutreference to any particular equation of state. To make use of them it is necessary toassume an equation of state for the equivalent substance.GENERAL SOLUTIONOnly a brief outline of the major points of the general solution is given here.Eqn (1) and (2) can be rewritten asDetails of the treatment will be given in a separate p~blication.~~(a2&/aX2)v,T (a2A:s/ax2)v,T ( d P / a X ) Y , d 2 = 0 (14)(a3&/aX3)v,27 4- (a3A,*,/ax3)v,7- 3(a2P/ax2)v,~Q -3(a2p/ax av),~2 + ( a 2 p / a v 2 ) , , , ~ 3 = o (15)Q = (ap/ax)V,T/(ap/av)T,x (16)whereand where Acb and AZs are the separate contributions to the Helmholtz free energy ofthe mixture which were defined earlier.Rewriting eqn (14) and (15) in reduced formenables us to defineso that the critical points of the mixture are now the solutions of the simultaneousequationsa(Gi,T) = 0 (19)/?(RT) = 0. (20)In eqn (17)-(20) = V/V& = T/Zs, P" = P/&128 CRITICAL PROPERTIES OF BINARY MIXTURESo! and /? may be calculated from eqn (17) and (18) if' we know the composition,the mixture prescription, the size parameter fr in eqn (8)j and the reduced equation ofstate for the equivalent substance.and by using a Newton-Raphson iterative procedure,33 starting from the approximate solution given by eqn(10) and (1 1).The iteration failed to converge to a solution in a few cases when theapproximate solugon was inadequate. In these cases we subtracted 0.05 from theinitial value for V . If the iteration still broke down a further 0.05 was subtractedfrom runti1 a solution was found. This technique was based upon the observationthat the initial value for Twas always very near the final solution, but the initial valuefor was_always greater than the final solution if the two differed significantly.Once V and Thad been obtained, the corresponding value of 3 was derived fromthe reduced equation of state.The actual mixture critical properties were thenobtained from the reduced critical properties by use of T:,, VEs and &. pzs is notgiven directly by the prescription for the equivalent substance, and a value for 2" mustbe adopted before it may be calculated from the prescription. During the solutionof the criticality conditions for r a n d T, the value of Z,", assumed was that for therelevant equation of state. However to generate the actual mixture critical propertiesfrom the reduced values, a value of 2" was chosen which would make the calculatedcritical pressures run smoothly from one pure substance critical pressure to the other.This will not happen if the equation of state value is used, and so the inole fractioncombination of the pure substance values for Z",proposed by Pitzer and Hultgren 34 was employed.This practice in which the equation of state is used to generate reduced values ofthe critical properties, after which an experimentally realistic value of 2" is used inthe calculation of the actual critical properties from the reduced critical properties,differs from the practice of previous worker^.^'^In this work we have solved these equations forzc = x,z;1+x22;2 (21)EQUATIONS OF STATE(1) BJERRE A N D BAK EQUATION OF STATEWe have used the two parameter equation of state proposed by Bjerre and Bak.26They tested a number of two parameter equations of state and found that the mostsatisfactory equation for 0.56 < T/T" < 0.95 was10 4; 75 1 p- --3 (P-1/6) 16 T*(v+1/4)22" = 3/10 (23)where, as before, properties with a tilde are reduced with respect to the equivalentsubstance critical properties.(2) DAVIS A N D RICE TAYLOR SERIESWe have also used the double Taylor series expansion about the critical point ofa pure substance proposed by Davis and Rice.25 These workers wrote the reducedpressure of a one component system in the vicinity of its critical point as a Taylorseries expansion in reduced temperature and density, and evaluated the coefficientsof the leading terms in the expansion for argon.The work of Davis and Rice does not provide a theory of the critical point, anddoes not reproduce the known singularities which appear as the critical point iC.P . HICKS AND C. L. YOUNG 129approached. It does, however, give us an accurate reduced equation of state for thenear critical region for the equivalent substance, and is valid as long as there is a pointwithin 6 on each side of the critical point about which such analytic expansions canbe made.The values of the coefficients of the Taylor series expansion obtained by Davisand Rice were used where available, and higher order coefficients were taken fromwork by Hicks 24 where necessary. The Taylor series was taken to the sixth orderof the derivatives of pressure with respect to density, and the temperature derivativeof pressure was expanded to the fourth order with respect to density.The restrictedrange of validity of the Taylor series expansion meant that the configurational energy,which requires an integration from V = GO down to Y = V, could not be calculated.Instead it was calculated using the Beattie-Bridgman equation with the generalcoefficients determined by Su and Chang.36 Calculations showed that small changesin the configurational energy used did not affect the predicted critical propertiessignificantly. The restricted range of validity also meant that solutions to thecriticality conditions could not be found for all mixtures studied here.COMPARISON WITH EXPERIMENTOMCTS MIXTURESThe three methods already outlined have been used to predict the critical propertiesof the OMCTS mixtures which were studied experimentally.In each case the van derWaals one fluid prescription, eqn (3), was used, and Y for eqn (8) was taken to be theratio of liquid molar volumes at 293.15 K.Van der Waals' equation was used to give the equation of state properties requiredin the approximate solution to the criticality conditions.Different pairs of combining rules were used for the critical temperature, T;2, andcritical volume, Y;2, and those used are summarised in table 5.TABLE 5 .-COMBINING RULESno. name(1) Berthelotf Lorentz< and 7 for use in eqn (4)5 = 1 p I = 12(1112)+ 26 ViI v,;(11 +Id w:p+ VC,j)"(2) Hudson-McCoubrey 17+ Lorentz =- = 1(3) Berthelot+ Good-Hope * = I2(Z112)* 26 v;l vg2(11 + 1 2 ) ( q+ Vc,tY(4) Hudson-McCoubrey + Good-Hope = -I I and 1, are the first ionisation potentials of each species.The standard deviations between predicted and experimental critical properties aregiven in table 6.For these mixtures all the methods predict critical pressures whichare below those observed ; a value of < greater than unity would be needed to improveagreement, whereas a value of < less than unity is needed to fit critical temperatures atequimole fraction (assuming q = 1). In spite of this the Hudson and McCoubreycombining rrrle for T;.,, which gives values smaller than the Berthelot or geometricinem rule, gives larger standard deviations in every case.I-130 CRITICAL PROPERTIES OF BINARY MIXTURESN-ALKANE MIXTURESWe now consider the application of the general solution together with the Bjerreand Bak equation of state to n-alkane + n-alkane, n-alkane + cycloalkane and n-alkane + benzene mixtures.The experimental data are taken from a recent compilation.The Davis and Rice Taylor series was not used in this comparison because the criticalpoints of many of these mixtures lie outside its range of validity.TABLE 6.-STANDARD DEVIATIONS OF CALCULATED AND OBSERVED CRITICAL PROPERTIES FOKTHE O M n S MIXTUREScritical temperature TC/KmixturesapproximatesolutionVan der Waalsequation(1) (2)OMCTS-5 cyclopentane 7.7 24.9OMCTS+ iieopentane 3.2 27.8OMCTS+ 2,3-dimethylbutane 6.1 16.3OMCTS+ tetramethylsilane 5.0 16.0general solutionDavis and Rice Taylor series Bjerre and Bak equation(1) (2) (3) (4) (1) (2) (3) (4)t3.8 23.5 4.7 20.4 4.8 26.3 5.9 23.84.7 * * * 2.6 29.0 1.9 26.43.1 15.4 3.4 14.0 4.3 17.3 4.8 16.9 * * * * 3.6 16.2 4.6 14.4critical pressure pC/N m-2OMCTS+ cyclopentane 5.6 7.5 5.9 6.5 4.8 5.0 5.6 7.1 4.5 5.9OMCTS+ neopentane 6.4 8.8 6.0 * * * 5.6 7.7 4.1 6.2OMCTS+2,3-dimethylbutane 4.1 5.4 4.0 4.5 3.4 6.3 3.7 5.0 3.1 4.2OMCTS+tetramethylsilane 5.7 6.8 * * * * 4.9 5.7 4.1 4.9* Solution outside the range of the Taylor series.7 The pairs of combining rules used for eachcolumn are shown by the numbers in parenthesis. See table 5 for details of combining rules.The calculations were made using, as before, the van der Waals one fluid prescrip-tion, and Y as the ratio of molar volumes (extrapolated values were used if necessary)at 293.15 K.(1) n- A LK ANE + n- AL K ANE MIXTURESThe standard deviations between predicted and experimental critical propertiesusing each pair of combining rules are given in table 7.When the molecules are ofsimilar size critical temperatures and pressures are predicted reasonably well, but asthe chain length difference increases the predictions worsen.Overall pair 3 is best and pair 4 is worse for these mixtures.(2) n-ALKANE-kCYCLOALKANE MIXTURESThese mixtures do not form a good test of the combining rules. The standarddeviations between predicted and experimental critical properties using each pair ofcombining rules are given in table 8. Both pairs 1 and 3 predict the critical propertieswith fair accuracy, pair 3 being marginally superior, as in the case of the n-alkane+n-a1 kane mixtures.(3) n-ALKANE+BENZENE MIXTURESStandard deviations between predicted and experimental values are given in table 9.Again the pairs 1 and 3 predict the observed critical properties with the best overallaccuracyC .P. HICKS AND C . L. YOUNG 131CONCLUSIONSOf the pairs of combining rules considered the combination of the Good and Hoperule for VC,, and the Berthelot rule for Ti2 is the best for these mixtures. However,when the size difference of the two molecules is large none of the combining rulespredict the observed critical properties well. This is probably due as much to theinadequacy of the van der Waals one fluid prescription as to the inadequacy of theequation of state used.TABLE 7.-sTANDARD DEVIATION BETWEEN EXPERIMENTAL CRITICAL PROPERTIES AND THOSEPREDICTED BY THE VAN DER WAALS ONE FLUID MODEL, THE BJERRE AND BAK EQUATION OFSTATE AND VARIOUS COMBINING RULES FOR THE n-ALKANE+ n-ALKANE MIXTURES(1)0.83.67.813.611.326.70.73.22.95.75.011.50.20.92.63.07.30.51.42.92.10.92.72.64.87.90.30.32.11.61 .o0.21 .o0.4(2)1.70.83.23.35.618.30.70.83.44.18 .O10.00.51.31.73.58.60.71 .o1.36.26.19.59.811.111.90.31.42.49.011.80.53.20.5TC/K(3)0.83.67.613.215.420.10.70.83.44.18.010.00.21 .o2.73.26.50.51.42.92.10.92.72.54.78.40.30.32.22.10.50.21 .o0.5(4)1.60.73.13 .O2.719.30.70.93.04.48.48.60.51.32.03.49.30.71 .o1.16.26.29.29.311.012.20.31.42.48.811.40.53.20.5(1)2.01.64.77.96.012.40.81.21.83.24.26.20.30.71.52.13.90.8---1.5----0.30.61.73 .O4.10.30.90.7(2)2.12.04.56.79.827.40.81.52.43.94.89.80.30.81.72.35.90.8---1.7----0.30.71.83.34.40.31 .o0.7(3)1.91.33.96.24.98.60.81.11.52.53.45.10.20.61.41.83.5----1.3----0.30.51.52.73.80.30.80.74)2.01.73.65.37.027.40.81.32.13.44.110.00.30.71.72.05.5----1.6----0.30.61.73 .O4.00.30.90.7The two different entries in this table for Cs+ C9 correspond to measurements by differentworkersI32 CRITICAL PROPERTIES OF BINARY MIXTURESThese results lend support to the hypothesis of Good and Hope l 8 that " ageometric mean distance rule " is superior to the arithmetic mean rule.However,in view of the uncertainty in applying any equation of state and mixture prescriptionto the criticality conditions, too much emphasis must not be placed on this observation.TABLE 8.-STANDARD DEVIATION BETWEEN EXPERIMENTAL CRITICAL PROPERTIES AND THOSEPREDICTED BY THE VAN DER WAALS ONE FLUID MODEL, THE BJERRE AND BAK EQUATION OFSTATE AND VARIOUS COMBINING RULES FOR THE n-ALKANE+ CYCLOALKANE MIXTURESmixture TC/K l O - 5 p C I N m-2(1) (2) (3) (4) (1) (2) (3) (4)c-C5 + n-CS 0.9 1.4 0.9 1.4 0.3 0.3 0.3 0.3c-C5 + n-C6 0.4 1.9 0.4 1.9 0.2 0.3 0.2 0.3c-C5 + n-C7 0.8 3.8 0.8 3.8 0.2 0.5 0.1 0.3c-C5+n-Cs 1.4 5.9 1.4 5.7 0.6 1.0 0.3 0.8c-C5+n-Cg 2.2 7.7 2.3 7.5 0.9 1.5 0.5 1.211.7 6.6 12.0 7.1 6.3 6.6 5.1 5.63.5 3.5 3.5 3.50.4 0.9 0.4 0.9 0.3 0.3 0.2 0.30.3 2.3 0.3 2.3 0.1 0.2 0.1 0.20.3 3.9 0.4 3.9 0.3 0.5 0.2 0.41.1 5.2 1.1 5.1 0.3 0.7 0.1 0.51.0 7.6 1.0 7.4 0.7 1.2 0.5 1 .o1.4 14.3 1.6 13.9 1.7 2.5 1.3 2.1- - - -The two different entries in this table for C'C6+n-C8 correspond to measurements bydifferent workers.TABLE ST STANDARD DEVIATION BETWEEN EXPERIMENTAL CRITICAL PROPERTIES AND THOSEPREDICTED BY THE VAN DER WAALS ONE FLUID MODEL, THE BJERRE AND BAK EQUATION OFSTATE AND VARIOUS COMBINING RULES FOK BENZENE(BZ) + n-ALKANESmixture TC/K(1) (2) (3) (4)59.0 56.5 60.8 56.916.4 15.2 16.5 15.35.2 4.6 5.2 4.65.4 2.6 5.4 2.63.6 0.7 3.2 0.74.0 1.0 4.0 1 .o4.4 3.4 4.5 3.26.4 3.4 6.4 3.24.8 5.8 5.0 5.58.6 6.7 8.6 6.55.9 7.4 6.0 7.16.4 15.5 6.3 14.87.5 26.6 8.5 26.3lO-spc/N m-2(1) ( 2 ) (3) (4)18.3 15.7 17.4 15.51.2 1.5 1.2 1.5- - - -0.3 0.2 0.4 0.30.5 0.2 0.7 0.30.5 0.2 0.7 0.20.4 0.5 0.6 0.20.7 1 .o 0.6 0.61.4 2.5 0.8 2.13.1 3.9 2.2 3.3- - - -- - - -The two different entries in this table for some mixtures correspond to measurements bydifferent workers.Financial support from the Queen Elizabeth I1 Fellowship Committee to C.L. Y.is gratefully acknowledged133 C . P. HICKS AND C . L. YOUNG’ C. P. Hicks and C. L. Young, Truns. Furaduy SQC., 1971,67,1598.C. P. Hicks and C. L. Young, Trans. Favaday SOC., 1971, 67, 1605.C. L. Young, J.C.S. Faraday 11, 1972, 68,452. ’ R. R. Spear, R. L. Robinson and K. C. Chao, Znd. Eng. Chem. Fund., 1969,8, 2.W. B. Kay and D. Hissong, Proc. Amer. Petr. Znst. ReJinirtg Div., 1969, 49, 13.S. C. Pak and W. B. Kay, Ind. Eng. Chem. Fund., 1972, 11, 255 and correction, 1974, 13,298.P. L. Chueh and J. M. Prausnitz, Amer. Inst. Chem. Eng. J., 1967, 13, 1107.A. Kreglewski and W. B. Kay, J. Phys. Chem., 1969, 73, 3359.A. S. Teja and J. S. Rowlinson, Chem. Eng. Sci., 1973, 28, 529.l o J.S. Rowlinson, Liquids and Liquid Mixtures (Butterworth, London, 2nd edn., 1969).* J. P. Kuenen, Theorie der Verdumpfug und Verfliiss-igung von Gemischen und der FructioniertenDestillation (Barth, Leipzig, 1906).l 2 0. Redlich and J. N. S. Kwong, Chem. Rev., 1949, 44,233.l 3 J. 0. Hirschfelder, C. F. Curtiss and R. B. Bird, Molecular Theory of Gases and Liquids (Wiley,l4 0. Redlich and V. B. T. Ngo, Znd. Eng. Chem. Fund., 1970,9,287.l6 K. N. Marsh, M. L. McGlashan and C. Warr, Trans. Fur&y Soc., 1970,66,2453.l7 G. M. Hudson and J. C. McCoubrey, Trum. Furuhy SOC., 1960,56,761.l9 C. P. Hicks and C. L. Young, Chem. Rev., 1975,75,119.2o E. A. Guggenheim and M. L. McGlashan, Proc. Roy. SOC. A, 1951, 206,448.21 A. J. B. Cruickshank, M. L. Windsor and C. L. Young, Trans. Furuduy SOC., 1966, 62,2341.22 E. M. Dantzler, C. M. Knobler and M. L. Windsor, J. Phys. Chem., 1968, 72,676.2 3 C. P. Hicks and C. L. Young, J. Chem. Thermodynamics, 1971, 3, 899.24 C. P. Hicks, Pl2.D. Tlzesis (University of Bristol, 1970).25 B. W. Davis and 0. K. Rice, J. Chem. Phys., 1967, 47,5043.26 A. Bjerre and T. A. Bak, Actu Chem. Scund., 1969, 23, 1733.” W. Patnode and D. F. Wilcock, J. Amer. Chem. SOC., 1946, 68,358.28 C. L. Young, J. Chem. Thermodynamics, 1972,4,65.29 W. B. Kay, Acc. Chern. Res., 1968, 1, 344.30 R. H. Olds, in Physicul Chemistry of Hy&ocurbons, ed. A. Farkas (Academic Press, New York,31 P. J. Flory, J. Chern. Phys., 1941, 9, 660; 1942, 10, 51.32 A. J. B. Cruickshank and C. P. Hicks, unpublished work.33 J. F. Traub, Iterative Methods for the Solution of Equations (Prentice Hall, Englewood Cliffs,34 K. S. Pitzer and G. 0. Hultgren, J. Amer. Chem. SOC., 1958, 80, 4794.35 J. A. Beattie and 0. C. Bridgman, J. Amer. Chem. Soc., 1927, 49, 1665.36 G. J. Su and C. H. Chang, J. Amer. Chem. Suc., 1946,68,1080.37 A. J. Vennix and R. Kobayashi, Amer. Inst. Chem. Eng. J., 1969,15,926.jS A. P. Kudchaker, G. H. Alani and B. J. Zwolinski, Chern. Reu., 1968,68,659.New York, 1954), pp. 251-253.T. W. Leland, J. S. Rowlinson and G. A. Sather, Trans. Furaduy SOC., 1968,64,1447.R. J. Good and C. J. Hope, J. Chem. Phys., 1970, 53,540.1953), vol XI, p. 131 et seq.New Jersey, 1964).(PAPER 4/591

ISSN:0300-9599    

DOI:10.1039/F19767200122

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Terminal Solid Solubility of Hydrogen in theZirconium-2.5 weight % Niobium AlloyBY V. K. Sinha-j-Indian Institute of Technology, Kanpur, IndiaReceived 14th June, 1974The terminal solid solubility (TSS) of hydrogen in the Zr-2.5 wt. % Nb alloy has been determinedby measuring the decomposition pressures of several Zr-2.5 wt. % Nballoys containing hydrogen in thetemperature range 425 to 900°C. The TSS boundaries a/(.+ 6) and a/(. + p ) were determinedfrom the points of inflection of the van’t Hoff decomposition pressure against temperature plots.The TSS of hydrogen in Zr-2.5 wt. % Nb alloy is about 1 to 2.5 atom % more than that in unalloyedzirconium (the maximum difference being at the eutectoid temperature of the Zr/H2 system). Thedecomposition pressure against temperature relation in the a and (a+6) phase regions and thetemperature against concentration relation along the TSS boundary have also been derived. Themean relative partial molar enthalpy of hydrogen in the a-phase and the mean enthalpy of formationof 6-hydride from the saturated a-phase in the alloy are -24.6 kcal (mol Hz)-l and -40.9 kcal(mol H2)-’ respectively.The excellent physico-chemical and technological properties 1-9 of the Zr-2.5 wt.%Nb alloy make it indispensable for use as pressure tubes, calendria tubes, fuel cansand other structures in pressurised and boiling water nuclear power reactors. Thealloy, however, readily absorbs hydrogen in water or steam environments and hydro-gen uptake, when it exceeds the terminal solid solubility limit (TSS), markedly reducesthe alloy’s strength and life due to precipitation of a brittle hydride phase.l0’l3Therefore a precise determination of the TSS of hydrogen in this alloy is of utmostimportance.Several investigators 14-20 have attempted to establish the TSS of hydrogen in theZr-2.5 wt. % Nb alloy at temperatures below 550°C using dilatometric cooling curves,hydride layer, decomposition pressure and temperature gradient techniques.How-ever, there is considerable disagreement among the various results and the TSS ofhydrogen in this alloy is still a matter of debate. In addition, there is no publishedinformation on the TSS of hydrogen in the alloy above 550°C. Therefore, the effectof niobium on the TSS a/(a +/3) boundary is also unknown.In the present investigation decomposition pressure studies were made to determinewith greater accuracy the a/(a + 6) TSS boundary and to establish the high temperaturea/(a+P) TSS boundary in the (Zr-2.5 wt.% Nb)/H2 system.EXPERIMENTALThe gas handling vacuum apparatus for determining the equilibrium decompositionpressures as a function of temperature and composition was similar to one describedearlier 2 1 g 22 with necessary modification^.^ A two-stage Mcleod gauge 9* 23 was used tomeasure the pressure of the system in the range of about &Hg to 3 mmHg; thisrange was adequate for the present investigation. The pressure readings of the McLeodt present address : Department of Materials Science, National Institute of Foundry and ForgeTechnology, Ranchi-834003, Bihar, India13V.K. SINHA 135gauge were accurate and reproducible to k0.5, kO.6, f 2 and +6 % at pressures ofmmHg respectively. A Veeco ionization gauge, type RG-2lx, was usedin the pressure range n~rnHg-lO-~ mmHg. The use of either the Buna 0 ring joint 24or a ground glass joint to connect the quartz specimen tube to the vacuum apparatus leadto leakage and a high static vacuum could not be maintained for long periods. Consequently,the specimen tube for each run was fused to the Pyrex apparatus via a graded seal. Thetemperature of the specimen zone was controlled to about +l"C and measured using apotentiometer with a sensitivity of +O.OOl mV. The volume of the entire apparatus wasdetermined by a gas expansion method in terms of the known volume of the bulb using themercury manometer with appropriate corrections. The apparatus could reach a pressureof the order of lo-' mmHg within 3 to 1 h, and when isolated from pumping accessories, itcould maintain a pressure ofandmmHg-10-5 mmHg for several days.MATERIALSReactor grade alloy was used (Bhabha Atomic Research Centre, Bombay, India).Thinstrips were cut from the alloy buttons and these were polished on Emery paper to give brightsurfaces. Samples were about 1 cm long and 0.4 cm wide and weighed approximately 300-400mg. Pure hydrogen was obtained by dissociation of titanium hydride at pressureslower than mmHg.TABLE 1 .-SUMMARY OF TERMINAL SOLID SOLUBILITY (TSS) AND THERMODYNAMIC DATA INTHE (Zr-2.5 wt.% Nb)/H2 SYSTEMTSS dataT/"C €or theequilibrium TSScomposition boundary(H atomicratio) (a+s)/a a/(a+S) phase0.0198 417 802 (a+@0.0284 450 7750.037 5 478 750 a/(a+6)0.046 1 492 7270.051 6 496 - al(a4-p)equationThermodynamic dataliteratureenthalpyl enthalpy/kcalkcal (mol (mol H2)-1 forHz1-l Zr/Hzln(p/N m-2) = -40.92 1 -45.8 2922.42- 40 876cal mol-' /RTln(CT/p.p.m. H2 by wt.) -13.1+0.8 - 8.95 32cal mol-l/RT= 14.94-13 120ln(CT/p.p.m. H2 by wt.) + 20.9Ifi 1 -cal mol-'/RT= - 4.35 + 20 8870.0620 512 7000.0709 523 6790.0906 530 650PREPARATION OF THE HYDRIDE SAMPLEThe weighed alloy specimen was first annealed and outgassed at 900°C in a pressure ofless than - mmHg for 3 to 1 h, and then a known amount of hydrogen was introducedinto the system. The specimen was kept at 400°C for about 24 h to allow a homogeneouszirconium-niobium-hydrogen alloy to form and finally cooled to room temperature. Thecomposition of the hydride specimen was computed from the observed reduction of hydrogenpressure in the apparatus with appropriate corrections ; the maximum uncertainty was 0.1 5to 0.2 atom % hydrogen.Eight alloys of different hydrogen concentration, listed in the firstcolumn of table 1, were prepared and studied to determine their decomposition pressuresas a function of temperature. After each decomposition pressure run, the specimen wasweighed. In each case, the hydrogen content of the specimen tallied within about 1 % tothat computed from the pressure drop136 (Zr-2.5 wt.% Nb)/HZ SYSTEMDECOMPOSITION PRESSURE STUDIESDecomposition pressure measurements were made directly after preparation of eachhydride sample to avoid contamination. For experimental simplicity, the changes of pres-sure occurring with temperature when the zirconium-2.5 wt. %niobium-hydrogen alloywas heated in a closed system, were measured. The temperature interval between measure-ments was 25"C, but near the suspected transition points it was 10-15°C. The decompositionreaction above 500°C was extremely rapid and equilibrium was reached almost instantane-ously. At lower temperatures, however, the reaction was slow and the time for attainmentof equilibrium was 3 to 4 h. The decomposition pressures above 500°C were measuredwhen equilibrium had been maintained for at least 8 h. At lower temperatures, measure-ments were made when equilibrium had been maintained for 15-20 h.Since the volume of the system was finite, the composition of the hydride specimenchanged at the equilibrium decomposition pressure at any temperature.However, thevolume of the system was calibrated and at any instant during the course of an experiment,pressure, composition and temperature were known and hence the relationship betweenpressure and temperature at constant composition and between pressure and compositionat constant temperature could be derived. Experimental runs, for which the decompositionreaction was not fast or if slight leakage in the system was suspected, were discontinuedand a new sample introduced.Decomposition pressure measurements for each of thehydride specimens, when heated in the range 425-9OO0C, took 10 to 15 days. Therefore,cooling runs were also made to check the reliability of the data and the decomposition pres-sures were recorded particularly at lower temperatures. There was agreement between thedata obtained from heating and cooling runs.RESULTS AND DISCUSSIONFor a binary system, it follows from the phase rule that in a single phase regionthe decomposition pressure depends both on temperature and hydrogen concentrationin the specimen. In a two-phase region, however, the total composition can varywithout affecting the decomposition pressure against temperature line. In the van?Hoff plot of lnp against l/T at constant hydrogen concentration, the one-phase andtwo-phase regions would, therefore, be represented by areas and lines respectively.The experimentally determined equilibrium decomposition pressure, composition,temperature data for the eight zirconium-2.5 wt.% niobium-hydrogen alloys in thetemperature range 425-900°C are shown in isothermal plots, fig. 1 and 2. In eachcase the composition of the specimen changed with increasing temperature andpressure. From fig. 1 and 2, the equilibrium decomposition pressure, p , for constantcompositions of the first seven zirconium-2.5 wt. % niobium-hydrogen alloys (table 1)were interpolated and for the eighth alloy extrapolated at several temperatures. Theplots of In p against 1 /T for the eight alloys are summarised in fig.3 where the line ABis based on average data for all alloys. The intersection of the decomposition pressureline for the g-phase with the line AB gave the temperature at which the given composi-tion corresponded to the TSS boundary a/(a+6). Similarly the dotted curve CC'joining the ends of the cc-phase region gave the temperature at which the given com-position corresponded to the a/(a + /?) TSS boundary. In a true binary system, a plotof the In p against 1 /T for any composition beyond the a-phase should be a singleline (like the one shown as CC' in fig. 3) in the temperature range up to which thetwo-phase region (a + 8) exists and then rise steeply in the /?-phase at higher tempera-tures. In the present investigation, the plots for the two-phase region (a +b), however,do not coincide with the line CC' and there is a considerable spread after the end ofthe a-phase region.This suggests that the zirconium-2.5 wt. % niobium-hydrogensystem may not be treated as a true binary system. Similar spread in the lnp againsV . K. SINHA 1371 /T data for zirconium-hydrogen and titanium-hydrogen has been observed by Ellsand McQuillan 2 5 and by McQuillan.26* 27 In both cases the conditions for a binarysystem were not satisfied due to the presence of third elements like oxygen, magnesiumetc. In the present investigation due care was taken to avoid contamination of thespecimens by gases such as oxygen and nitrogen, hence the spread in the plot inthe (ct+P) two-phase region may be attributed to the presence of niobium.Thenon-linearity of the (a + p) two-phase line CC' in fig. 3 is expected because the composi-tion and relative amounts of the a and /? phases in the zirconium-hydrogen systemchange simcantly with temperature 2 5 * 28 , the presence of the small amount ofniobium should not alter the nature of the two-phase region significantly.atomic ratio (H/M)0.0204 0 0 3 0 8 OD417 0.G956 CLG638 0.0'7 31 0.0&70I I i 1 I I I 4I I I I I2.0 3.0 4.0 5.0 6.0 7.0 8.0atom % hydrogenFIG. 1.-Isothermal pressure plotted against atom % hydrogen for 25-2.5 wt. % Nb alloy in thetemperature range 425 to 575°C : A, 425 ; B, 450 ; C, 475 ; D, 500 ; E, 525 ; F, 550 ; G, 575 "C.0, recorded during cooling.The experimentally determined TSS data for the a/(a + 6) and a/(a +B) boundariesof the eight zirconium-2.5 wt. % niobium-hydrogen alloys within an estimated un-certainty of about 2 % are listed in table 1.The TSS a/(a+P) boundary for thehydride alloy with atomic ratio hydrogen/metalO.O51 6 could not be determinedbecause an insufficient number of high temperature data points were recorded. Fromthe TSS data of table I a partial temperature against composition diagram of the(zirconium-2.5 wt. % niobium)/hydrogen system can be drawn (fig. 4) and may becompared with the TSS data of other investigators 2 5 * 30 for the unalloyedzirconium/hydrogen system. The least squares plots of concentration for the (a + 6)laand the a/(a+P) TSS boundaries against temperature are shown in fig.5 and 6.The published results of other investigators for the TSS of hydrogen in 25-2.5 wt. %Nb alloy 15* ' ' 9 31 and unalloyed zirconium 29s 32 are shown for comparison in fig. 5,It is observed (fig. 4) that the TSS of hydrogen is increased by the presence of niobium138 (Zr-2.5 wt. % Nb)/H, SYSTEMlo4 I I I I I I Iespecially in the temperature range 425-700°C. Extrapolation of the TSS boundaryindicates that the TSS of hydrogen in the Zr-2.5 wt. % Nb alloy at the eutectoidtemperature of the Zr/H2 system is 8.1 atom % hydrogen as compared to the estab-lished value 28 of about 6.1 atom % hydrogen in the latter. The results for thea/(a+6) TSS boundary (fig. 4 and 5) are in qualitative agreement with those ofBrown l4 and Erickson l5 both of whom reported an increased TSS of hydrogen inthe Zr-2.5 wt.% Nb alloy. The increased solubility is, however, not as large asreported by them.14* l5There are no published results on the zirconium-2.5 wt. % niobium-hydrogensystem at higher temperatures with which the present TSS data for the a/(a+P)atomic ratio (H/M)I I 0.0630 9.0'54 0.0070 ';.OiC9 0.04 !;f 5.0526 5.CI'Oh10 I I 1 Iatom % hydrogenFIG. 2.-Isothermal pressure against atom % hydrogen for Zr-2.5 wt. % Nb alloy in the temperaturerange 600 to 900°C : H, 600 ; I, 625 ; J, 650 ; K, 675 ; L, 700 ; M, 725 ; N, 750 ; P, 775 ; R, 800 ;S, 825 ; T, 850 ; U, 900 "C. 0, recorded during cooling.boundary may be compared. Comparison with the data for the unalloyed Zr/H,system 1 6 9 2 5 9 30 (fig.4), however, shows anincreasedTSS of hydrogenin the presence ofniobium particularly below 700°C. In an isothermal study, Erickson and Hardie l6reported that unalloyed Zr and the Zr-1.5 wt. % Nb alloy had the same TSS for hydrogenat 750°C. This result is qualitatively supported by the present investigation wherethe effect of 2.5 wt. % Nb in increasing the TSS is insignificant above 700°C (fig. 4).The relative partial molar entropy steadily decreases (becomes more negative)with increasing hydrogen content as is obvious from the intercepts of the isochoresin fig. 3. Such a behaviour of the relative partial molar entropy for hydrogen in thea-phase is due to the decreasing availability of interstitial sites as more hydrogen isintroduced in to the lattice.The mean relative partial molar enthalpy of hydrogenin the a-phase was computed to be -24.6) 1 kcal (mole H2)-l as compared to - 15.7to -28.9 33 kcal (mole HJ-' for the unalloyed Zr/H, system. The variation in thV . K . SINHA 139mean relative partial molar enthalpy value with concentration of hydrogen was withinthe estimated error 1 kcal (mole H2)-l. Therefore it appears that the enthalpy isalmost constant and does not depend on the concentration of hydrogen in the a-phase.The decomposition pressure against temperature relation in the (a + 6) two phaseregion along with the temperature against concentration relation for the (a + @/a anda/(a + /?) TSS boundaries with various enthalpy values were also determined and aresummarised in table 1.The enthalpy values of other investigators 29* 3 2 * 3 3 for theunaIloyed Zr/H, system are also mentioned therein for comparison.temperature/"C104 KITFIG. 3.--Surnma1-y plot of decomposition pressure against reciprocal of tempsraturz for differentzirconium-2.5 wt. % niobium-hydrogen alloys. The number on a plot represents tho, H/(Zr-2.5wt. % Nb) atomic ratio.Supersaturation 1 5 9 18* 34-36 during cooling has been suggested as the possiblemechanism for the increased TSS of hydrogen in zirconium and zirconium alloysIt has been argued,15 therefore, that the equilibrium TSS values of hydrogen inzirconium and zirconium alloys are essentially the same. In the present investigationthe data for heating and cooling runs were consistent, within experimental error, andno hysteresis was observed (fig.1 and 2). Consequently, an increased TSS of hydro-gen in the Zr-2.5 wt. % Nb alloy for both a/(a+6) and a/(a+P) boundaries in thetemperature range 425-700°C may be attributed to the presence of niobium.The zirconium-niobium phase diagram 37-39 shows that in the temperature rangeof the present investigation some niobium will eventually precipitate as a niobium-rich phase. It has been suggested l9 that this niobium dissolves a large fraction ofthe hydrogen present; however, the higher equilibrium pressures of hydrogen inniobium 40 suggest that hydrogen would probably have a lower solubility in this phas140 (Zr-2.5 wt. % Nb)/H, SYSTEMthan in the zirconium matrix. The niobium-rich phase precipitating from the Zr-2.5 wt.% Nb alloy, therefore, seems either to have slightly different characteristicsfrom pure niobium or else the interactions lead to the formation of a zirconium-niobium-hydrogen complex having an increased affinity for hydrogen, like the oneobserved for zirconium-nickel-hydrogen.4 * 4290C80C7 OC60C PE Bs?-4-w 5E 5 0 C400300200atomic ratio (H/M)0-111 9.250 0.428 3.666 1.00 1.50 2.33I 1 I 1 1 Iatom % hydrogenFIG. 4.-TSS boundaries a/(cc+ 8 ) and a/(.+ /3) for the (Zr-2.5 wt. % Nb)/Hz system and theZr/Hz system. 0, present work on (Zr-2.5 wt. % Nb)/Hz system ; x , Ells et al. ; ” 0 , Lagrangeet al. ;30 a, Erickson et al. ;I6 A, Gulbransen et al. Sieverts plot ;29 0, Gulbransen et al. van’tHoff plot.29 -a, a/(.+ p), (Zr-2.5 wt.% Nb)/Hz ; - - - - b, 4(.+P), Zr/Hz; L-C, a/(a+ a), (Zr-2.5 wt. % Nb)/Hz ; - - - - 4 al(a+&), ZrlH2.CONCLUSIONS(1) The terminal solid solubility (TSS) of hydrogen in the Zr-2.5 wt. % Nb alloyis greater than that in unalloyed zirconium over the temperature range 425-700°C.(2) The TSS of hydrogen in the Zr-2.5 wt. % Nb alloy for the u/(a + 6) boundaryover the temperature range 425-550°C is given by,In C, (p.p.m. H2 by wt.) = 14.94- 13 120 cal mol-l/RT.(3) The TSS of hydrogen in the Zr-2.5 we. % Nb alloy for the a/(cc+p) boundaryover the temperature range 550-850°C is given by,In CT @.p.m. H2 by wt.) = -4.35+20 887 cal mol-l/RT.(4) In the Zr-2.5 wt. % Nb alloy the mean relative partial molar enthalpy ofhydrogen in the a-phase is - 24.6 kcal (mol H2)-l and the mean enthalpy of formationof 6-hydride from saturated a-zirconium is -40.9 kcal (mol HZ)-lV. K .SINHA6141I6-temperature/"C525 500 L75 L 50 425I I I I I5.050!?SO 12-90 13.30 13.70 14.10 14501 I I I Ilo4 KITFIG. 5.-F'lot of lnCT against 1/T for the a/(a+ 8) TSS boundary in the (Zr-2.5 wt. % Nb)/H2system and the Zr/H2 system. (Zr-2.5 wt. % Nb)/H2 system : (a), 0, present work; (b), Erick-son ;15 (c), Sawatzky etaL ;17 (d), S l a t t e r ~ . ~ ~ Zr-Hz system : (e), Gulbransen etal. ;2g (f), Kearn~.~'I I I I I I Ii'Y 5 . 2I thank Prof. K. P. Singh, Department of Metallurgy, Indian Institute of Techno-logy, Kanpur for laboratory facilities and for showing an interest in the progress ofthis work, Thanks are also due to the Department of Atomic Energy, Govt.of Indiafor financial assistance1 42 (Zr-2.5 wt. % Nb)/H2 SYSTEMR. S. Ambartsumyan, A. A. Kiselev, R. V. Grebennikov, V. A. Myshkin, I,. J. Tsuprun andA. F. Nikulina, Proc. 2nd Int. Con$ Peaceful Uses of Atomic Energy, Geneva, 1958, 5, 12.0. S. Ivanov and V. K. Grigorovich, ref. (l), p. 34.C. R. Cupp, J. Nuclear Materials, 1962, 6, 241.W. Evans and L. G. Bell, Report No. AECL 1395, 1961.J. H. Foley, Heavy Water Reactor International News Letter, Chalk River Ontario, No. 6,1963.G. F. Slattery, J. Less Common Metals, 1965,8, 195. ' B. G. Parfenov, V. V. Gerasimov and G. I. Venediktova, Corrosion of Zirconium and ZirconiumAlZoys (Israel Programme for Scientific Translations, Jerusalem, 1969), chap.1, p. 4.K. Balaramamurthy, Symp. Non-ferrous Metallurgy, ed. J. E. Manner and P. K. Gupta (N.M.L.Jamshedpur, India, 1968), vol. 111, p. 181.V. K. Sinha, Ph.D. Thesis (Indian Institute of Technology, Kanpur, 1973).lo P. Cotterill, Prog. Materials Sci., 1961, 9, 136.l 1 C. E. Ells, J. Nuclear Materials, 1968, 28, 129.l2 D. G. Westlake, J. Nuclear Materials, 1968, 26, 208.l3 C. Roy and J. G. Jacques, J. Nuclear Materials, 1969,31,233.l4 A. Brown, Ph.D. Thesis (Durham, 1961).l 5 W. H. Erickson, J. Electrockem. Tech., 1966, 4, 205.W. H. Erickson and D. Hardie, J. Nuclear Materials, 1964, 13, 254.l7 A. Sawatzky and B. J. S. Wilkins, J, Nuclear Materials, 1967, 22, 304.l8 A. Sawatzky, quoted as personal communication in J.Electrochem. Tech., 1966, 4, 205.l9 H. H. Klepfer, J. Nuclear Materials, 1963, 9, 77.2o D. E. Coates and W. H. Erickson, unpublished work.21 V. K. Sinha and K. P. Singh, J. Nuclear Materials, 1970,36,211.22 V. K. Sinha and K. P. Singh, Metal Transactions, 1972,3,1581.23 J. H. Leck, Pressure Measurements in Vacuum Systems (Inst. of Physics, Chapman and Hall,24 W. R. Doty, Reu. Sci. Instr., 1959, 30, 1053.2 5 C. E. Ells and A. D. McQuillan, J. Inst. Metals, 1956, 85, 89.26 A. D. McQuillan, Proc. Roy. SOC. A , 1950,204, 309.27 A. D. McQuillan, J. Inst. Metals, 1951, 79, 371.28 R. L. Beck and W. M. Mueller, Metal Hydrides, ed. W . M. Mueller, J. P. Blackledge and G. G.Libowitz (Academic Press, New York, 1968), chap. 7, p. 241.29 E. A. Culbransen and K. F. Andrew, J. Metals, 1955, 7, 136.30 L. D. La Grange, L. J. Dykstra, J. M. Dixon and U. Merteen, J. Phys. Chem., 1959, 63,31 G. F. Slattery, J. Inst. Metals, 1967, 95, 43.32 J. J. Kearns, J. Nuclear Materials, 1967, 22, 292.33 D. G. Westlake, J. Nuclear Materials, 1962, 7 , 346.34 G. Ostberg, J. Nuclear Materials, 1962, 5, 208.35 J. J. Kearns, Report WAPD-TM-147, 1958.36 R. E. Westernman, HW-81269, 1961.37 M. Hansen, Constitution of Binary Alloys (McGraw-Hill, New York, 1958), p. 1023.38 R. P. Elliott, Constitution of Binary Alloys, First Supplement (McGraw-Hill, New York, 1965),39 I. T. Bethune and C. D. Williams, J. Nuclear Materials, 1969, 29, 129.40 W. M. Albrecht, W. D. Goode and M. W. Mallet, J. Electrochem. Soc., 1959, 106,981.41 G. G. Libowitz, H. F. Hayes and T. R. P. Gibb, Jr., J. Phys. Chem., 1958,62,76.42 R. L. Beck and W. M. Mueller, Metal Hydrides, ed. M. M. Mueller, J. P. Blackledge and G. G.London), chap. 1, p. 16.2035.p. 279.Libowitz (Academic Press, New York, 1968), chap. 7, p. 285

ISSN:0300-9599    

DOI:10.1039/F19767200134

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Electron Spin Resonance Evidence of the Photoionizationand Photoisomerization of Free Radicals from y andUltraviolet Irradiated Pyridi nium CationsBY J. P. QUAEGEBEUR, H. OFENBERG AND A. LABLACHE-~OMBIER*Laboratoire de Chimie Organique Physique, Universi tC des Sciences et Techniquesde Lille, B.P. 36, 59650 Villeneuve d'Ascq, FranceANDJ. C. RONFARD-HARET AND C. CHACHATY"Service de Chimie Physique, C.E.N. de Saclay,B.P. no. 2, 91190 Gif-Sur-Yvette, FranceReceived 4th October, 1974The analysis of the e.s.r. spectra of U.V. illuminated (250 < X < 400 nm) pyridinium cation andof several of its derivatives in an HCl+H20 glassy matrix shows the presence of an allyl radical(IIId).This radical is also formed under U.V. illumination of the azacyclohexadienyl radical (I) producedby y irradiation of glassy solutions of pyridinium cations.Radical (I) is not the direct precursor ofradical (IIId). Evidence is presented to support the sequence :hvradical (I) -+ (I)++e-hve- + pyridinium cation -+ pyridinyl radical -+ azaprisinane radical + allyl radical (1IId).The transient formation of an ammonium like azaprismane radical is suggested by the photo-lysis of some methyl substituted pyridines which indicates an important rearrangement of the carbonbackbone of the pyridinyl radical leading to the radical (IIId).The U.V. photolysis (250 < ;1 < 400nm) of poly-2- and -4-vinylpyridines (P2VPand P4VP) in glassy solutions of HCl 10 mol dm-3 at 77 K gives rise to a radicalwhose e.s.r. spectrum cannot be assigned to any of the radicals resulting from hydrogenabstraction or addition to these polymers (fig.1) which have been considered or identi-fied previously.' To identify this new radical and to understand the mechanism ofits formation, we have performed a complementary study of the photolysis of smallmolecules related to PVP such as pyridine and some alkylpyridines; this has beenreported in a preliminary note.2 Some experiments have been performed in whichazacyclohexadienyl and pyridinyl radicals, likely to be produced in the first step ofthe photolysis, have been selectively formed by y irradiation * and then photolysed.In several cases it appears that the photoinduced radicals, observed by e.s.r. spectro-scopy of a rigid matrix, are not the primary radicals.Secondary processes may occursuch as photoisomerization or photoionization of primary radicals or even theirreaction with a molecular compound.14I44 IRRADIATION OF PYRIDINIUM CATIONSP2VP V----L ,vFIG. 1 .-l3.s.r. spectra of y irradiated and U.V. illuminated polyvinylpyridines at 77 K in HCl + H20matrix after removal of 61;. The hyperfine coupling constants given in table 2 for the radical (IIId)(optimized values) have been used in the simulation of the e.s.r. spectra, neglecting the contributionof the methine proton of the chain (see Reaction Mechanisms section).EXPERIMENTALThe e.s.r. experiments were carried out with a Varian V4502 spectrometer operating at9.2-9.3 GHz with a modulation frequency of 100 kHz.The output from a high pressure500 W mercury arc (Philips SPSOO) was focused on the sample in the cavity of the spectro-meter. The light illtensity received by the sample was 5 x lot7 quanta s-I cnr3, determinedby a ferrioxalate actinometer. The samples, degassed in silica tubes, were either directlyU.V. irradiated at 77 K or firsty irradiated then warmed at 140 K to remove the Cl.; radicalsand then U.V. irradiated at 77 I< in the cavity.[4-2Hl] Pyridine, [2,6-*H2] pyridine and [3,5-'H2]pyridine were synthesized accordingto the method of Bak by reduction with Zn+D2S04 of 4-chloropyridine and 2,6- and3,5-dibromopyridine respectively. The latter was prepared by the method of Englert andM~Elvain.~ The isotopic purity of the deuterated pyridines was checked by high resolutionn .m .r.spectroscopy.RESULTSASSIGNMENT OF E . S . R . SPECTRAThe new radical produced by photolysis of the pyridine-like compounds inHC1+ H20 matrix, which will be denoted (111), the azacyclohexadienyl and pyridinylradicals being designated (I) and (II), is probably formed with a low quantum yieldsince under our U.Y. irradiation conditions its limiting concentration never exceedsca. 10IJ radical ~ r n - ~ . On the other hand when the same solutions are y irradiatedat 77 K with a dose of 1 Mrad and then photolysed at this temperature (250 < 3, <400 nm) the radical (I) initially formed (generally the 4-azacyclohexadienyl radicalexcept with the 4 picolines and P4VP where 2 azacyclohexadienyl is formed) dis-appears after a few minutes giving radical (111) with a concentration of 101'-101 * spinc111-~ (fig.2). The assignment of the spectrum of radical (111) has been performed bycomparing the spectra of the pyridinium ion successively y and U.V. irradiatedwith those of several substituted hornologues including the PVP radical. The differentcoupling constants have been deduced from the total widths and the second momentsof the poorly resolved spectra generally obtained under our experimental condition.The analysis of the e.s.r. spectra has been described previously.'* The validity oQUAEGEBEUR, OFENBERG, LABLACHE-COMBIER, RONFARD- HAHET AND CHACHATY 145the assignments was checked by INDO calculations lo of the coupling constants andspin densities, then by computer simulations of the spectra using the program ofLefebvre and Maruani.u - c/FIG.2.-E.s.r. spectra from pyridinium ion in HCl+ H20 matrix. Azacyclohexadienyl radical aftery irradiation (A), allyl radical after U.V. illumination subsequent to y irradiation (B), simulated spec-trum (C), of the allyl radical (IIId).We may first rule out the assumption of the forniation of radical (111) by ringopening according to reactions such as (1) or (2) :CH~=CH -CH=~H-CH=C* \'HThe e.s.r. spectrum of radical (111) being clearly not consistent with that of pyrryl l 2[reaction (l)] or vinyl l 3 radicals [reaction (2)]. The signal has an average hyperfinesplitting constant of 10-13 G which is characteristic of an allyl radi~a1.l~ We there-(IIIa) (IIIb) (IIIC) (IIIdI46 IRRADIATION OF PYRIDINIUM CATIONSfore consider the allyl radicals (IIIa, b, c, d) as possibly resulting from the photolysisof radical (I).The nitrogen hyperfine coupling constant cannot be deduced directly from thestructure of the spectra and has been estimated from the second moment of radical(111) from pentadeuteropyridine (fig.4) in a DCl+ D20 matrix, neglecting the contri-bution of the deuteron hyperfine coupling (a; = 0.023 ah), and assuming that thecontribution of nitrogen is dominant.As in radicals (IIIa) and (IIIb) the nitrogen is included in the conjugated systemand therefore sp2 hybridized, the isotropic aN and dipolar bN coupling constantsshould be given approximatively by 15*UN = 28pk, b N = 17&.(3)Hence the predominant contribution of nitrogen to the second moment of thespectrum of the fully deuterated radical (111) will be :(4)giving aN = 6.7 G for a second moment of 51 G2 and a:H z - 8 G since a&,aN w- 1.2 for sp2 hybridized nitrogen. A difference of ca. 8 G should therefore beobserved between the total width of the spectra of radical (111) formed in HCl+ H20and DCl+D,O matrices. Since it remains virtually the same in both matrices theformation of radicals (IIIa) or (IIIb) may be excluded.An INDO calculation performed on radical (IIIa) in its minimum energy con-formation gives a value of aEH different from the experimental one (table 2). Notethat in radicals (IIIa) or (IIIb) the coupling of the proton bound to carbon 4 shouldbe much smaller than the actual value of 13 G given by the difference in width betweenthe spectra obtained with pyridine and [4-2H,]pyridine (see table 1).We thus ruleout the assumption of the photoisomerization of the azacyclohexadienyl radical byan electrocyclic reaction such as :MZN = $(A&,,+2AkL) x 1.15aihv - (IIIa) (5) (1)similar to that proposed by Monge and Schott l7 to account for the formation of anallyl radical from the cyclohexadienyl radical formed by y radiolysis of durene. Thesame is true for an electrocyclic reaction followed by a sigmatropic rearrangement l8leading to radical (IIIb) :4 (111 b).We now consider the formation of radicals (IIIc) and (IIId) where nitrogen is oneor two bonds away from the allyl fragment and assume therefore that its hyperfinecoupling is nearly isotropic, with a contribution to the second moment :M,, = +a; (7)which gives aN M 8-9 G for M2 = 51 G2.a sigmatropic rearrangement of radical (IIIa).However, in radical (IIIc) the nitrogen hyperfine coupling is induced mainly by spinpolarization and it may be estimated, using the CO2 -cH-NH,f radical l9 as anexample, that laNl = 3.5 G for p: = I , with principal values of the dipolar tensorsmaller than 1 G.The spin density on the outermost carbon of the allyl fragmentis of the order of 0.5-0.6 so that a maximum hyperfine coupling of 1.5 G would beexpected for nitrogen, much smaller than the actual value of ca. 8 G. This point isThe radical (IIIc) could also result from(IIIc) (8) (IIIa) __QUAEGEBEUR, OFENBERG, LABLACHE-COMBER, RONFARD-HARET AND CHACHATY 147confirmed by the INDO calculated spin densities and coupling constants of radical(IIIc) (table 2).TABLE 1.-OVERALL WIDTH AND SECOND MOMENTS OF THE E.S.R.SPECTRA OF ALLYL RADICALScompound matrix mIG <AH2)/G2*pyridine[4-2Hl]pyridine[2,6-2H2]pyridine[3 ,5-2H2]pyridineperdeuteratedpyridine2-picoline3-picoline4-picolineP2VPP4VPHC1+ H2OHCl+HzOHCl+ H2OHCl+H20DCl+D20HCI+ H2OHCl+ HzOHCl+ H2OHCI+H20DCl+DzOHCl+ H2ODCI+DzO71595946(structurelessspec tr urn)816098696958582131761 67145512491803782172101731 70* measured at 77 K with respect to g = 2.0028.TABLE 2.-INDO CALCULATED SPIN DENSITIES AND COUPLING CONSTANTS OF ALLYL RADICALSDERIVED FROM THE PYRIDINIUM CATIONSradical position? n spin densities hyperfine coupling constants!G(IIIa) 1 - 0.24 CZN = -7.7, = 5.6296 0.55 = -14.03Y5 - 0.03 = 11.54 0.01 afI = -2.7, ah = -3.7(IIIC) 1 - 0.02 a~ = -3.0, aEH = 18.12 - 0.04 a, = 29.54 0.01 a;I = -3.6, ak = -2.93 0.55 = -12.75 - 0.25 = 4.86 0.57 = -13.2(IIId)1INDO optimized*8.8afr = - 1.6air, = - 3.00.01 } u ~ I = - 1.6ah = - 3.12 - 0.25 = 4.2 = 4.53,4 0.55 = -13.5 = -12.5596 - 0.02 CZH = 11.0 aH = 10.8* by computer simulation of the e.s.r.spectrum. 7 For numbering see text.N.B. The interatomic distances and bond angles used in the INDO calculations for radicals(IIIa) and (IIIc) are given in ref. (26) for the cyclopentadienyl and cyclopropyl rings assuming anC(4) angle of 120" and an angle of 45" between the planes of the three and five membered rings.The assumed geometry of radical (IIId) is given in fig.3./H'€148 IRRADIATION OF PYRIDINIUM CATIONSIt appears therefore that radical (IIId) is formed by photolysis of pyridiiiium ions.The approximate values of the different hyperfine coupling constants of radical (IIId)have been deduced from the differences in the overall widths and second moments ofthe e.s.r. spectra given in table 1 for the radical derived from pyridine and its deuteri-ated and methylated homologues. The analysis of the e.s.r. spectra of the lattercumpounds enables the positions of all the carbon atoms of the initial pyridiniumcations (or of the intermediate 4-azacyclohexadienyl radical) in the allyl radical (IIId)to be determined.For a methyl group attached to an allylic carbon, the overall width of the e.s.r.spectrum will be increased byAH = 3aEH3-a! (9)where a! is the coupling constant of a proton directly attached to the same carbonin the non methylated radical.The substitution of a proton by a methyl group doesnot appreciably change the spin density pi on the adjacent carbon so that :(10)taking QF = -23 G and QcH3 = 28 G, la!J is estimated to be 0.4 AH.Pi = arIQ,H = ~ F H ~ I Q E H ~ ;23' \; \FIG. 3.-Geometry assumed for the allyl radical (IIId) with INDO calculated spin densities andcoupling constants in parentheses. The following interatomic distances and angles were used :C(3)-C(2) q.nd C(4)-C(2) = 1.42 A, C(3)-C(6), C(4)-C(5) and C(5)-C(6) = 1.52 A, C ( 6 j Nn nand C(5)-N = 1.49 A, C-H = 1.09 A, N-H = 1.01 A, C(3)C(2)C(4) = 114", C(2)C(3)C(6) =n n nC(2)C(4) C(5) = 106", C(3)C(6)C(5) = C(4)C(5)C(6) = 107".C(WH(6) and C(5 j H ( 5 ) bonds aren nlocated respectively in the bissector planes of C(3)C(6)C(S) and C(4)C(5)C(6).If, on the other hand, the CH3 group is located one or two bonds away from theallyl conjugated system apH3 becomes negligibly small and the total width of thespectrum will be diminished by uF, the coupling constant of a proton bound to acarbon of the three membered ring in the corresponding non methylated radical.Keeping the same numbering of the carbons as in the initial pyridinium cation,two equivalent carbons of this cation (2 and 6, 3 and 5 ) become inequivalent in theradical (IIId) (table 2).The effect of the position of the methyl group on the overallwidth of the spectra of irradiated picolines indicates that carbon 4 and one of carboQUAEGEBEUR, OFENBERG, LABLACHE-COMBIER, RONFARD-HARET AND CHACHATY 1493 or 5 are located at the outermost position of the allyl fragment, carbon 2 (or 6)being in the middle position, whereas carbon 5 (or 3) and carbon 6 (or 2) are includedin the three membered ring.20 gauss-7-Fro. 4.-Experimental and calculated spectra of (IIId) allyl radicals derived from y and U.V. irradiatedpyridinium and deuterated pyridinium cations with the corresponding second moments. Thevertical lines correspond to g = 2.0028.2 - p I c oli ne 3 p i c o l i n e I;-picoIineFIG.5.-Experimental and calculated e.s.r. spectra of allyl radicals fromy and U.V. irradiated picolineswith the corresponding second moments.To assess more accurately the nitrogen and proton coupling constants in theradical (IIId) issued from C5H5NH, we have first performed an INDO calculation,varying the angles 8 and 4 (fig. 3) to find the minimum energy configuration of150 IRRADIATION OF PYRIDINIUM CATIONSthis radical. The calculated coupling constants were then slightly modified untilsatisfactory agreement was obtained between the computer simulated spectrum andthe experimental one (fig. 1,4, 5), taking into account the anisotropy of the hyperfinecoupling of allyl protons.14 The optimized coupling constants are compared intable 2 with those given by INDO calculations.They give a second moment of210 G2 and a total width of 72 G in good agreement with the experimental data(table 1). They have also been used to simulate the e.s.r. spectra of allyl radicalsfrom deuterated pyridines (with a, = 0.154 aH) and picolines which are shown infig. 4 and 5 with their corresponding second moments.REACTION MECHANISMSWhatever the mechanism for the photolysis of the pyridinium ion is, it seems thatthe allyl radical (IIId) cannot result directly from the isomerization of radical (I).After y and U.V. irradiation of the pyridinium cation in a DC1-t D20 matrix, the e.s.r.spectrum of the radical (IIId) is virtually the same as for irradiation in an HCl+H,Omatrix since the contribution of protons or deuterons attached to the nitrogen is small.On the other hand when the [4-2H,]pyridinium cation is irradiated in a H,O+HCImatrix the width of the (IIId) spectrum will be 13 G narrower than the one resultingfrom irradiated C5H5NH or C5H5ND, this difference corresponding to a:.In bothcases however, the same radical (I) is first formed under y irradiation and it is difficultto explain why in the course of the formation of the ally1 radical either a proton ora deuteron is detached from carbon 4 depending upon whether one starts from[4-2H,]pyridinium in HC1+ H,O or from C5H5NH+ in DCl + D,O.Thus the azacyclohexadienyl radical is clearly not the direct precursor of radical(IIId), so that we may eliminate definitely the assumption of an electrocyclic reactionmentioned above [reaction (5)] and also that of another intermediate transient suchas the azoniabenzvalene cation considered by Wilzbach and Kaplan 2o for the photo-hydration of the methylpyridinium cation, which could be formed by the reaction :+ +1followed by a hypothetical rearrangementand 3.reversing the position of carbon atoms 2In the course of the photolysis of radical (I) formed under y irradiation, an increaseof the signal of Cl-, is observed simultaneously with the formation of the allyl radical(IIId).We suggest therefore that the azacyclohexadienyl radical, which absorbs inthe near U.V. and visible regions of the spectrum as the cyclohexadienyl radical,21may be photoionized, giving a diamagnetic compound which has not been character-ized.It is known that the ionization potential of free radicals may be appreciablysmaller in the condensed phase than in the gas phase 4a9 22 and are thus of energy2-5 eV, i.e., in the region of visible and near U.V. energy. The electrons released byphotoionization of radical (I) are partially scavenged by Cl, formed during radiolysisor more likely react with the matrix according to reactions :HCIHCl +e- 4 HCl- + Cl-5 + HQUAEGEBEUR, OFENBERG, LABLACHE-COMBIER, RONFARD-HARET AND CHACHATY 15 1the remainder being scavenged by pyridinium cations with the forination of pyridinylradicals, which also absorb in the near U.V. and visible regions 2 3 and would bephotolysed in turn to give the allyl radicals (IIId).hvH f___, (IIId) (11)This last reaction may account for the similarity of the spectra of the allyl radicalsobtained in HCl+ H20 and DCI + D20 matrices.To verify this assumption we subjected HCI + H20 glasses containing the pyridi-nium ion at concentrations up to 4 mol dm-3 to y irradiation.The spectrum ofradical (I), selectively formed at low solute concentrations is progressively replacedby that of the pyridinyl radical and virtually disappears above [C5H5NH] = 2 moldm-3. The electrons released by y irradiation are scavenged by high concentrationsof pyridinium cations to give pyridinyl radicals. These radicals were subjected tophotolysis at A > 300nm to avoid contributions from the photolysis of pyridiniumcations which absorb at shorter wavelengths. Under our conditions of t1.v.irradia-tion, the pyridinyl radicals disappear in a few minutes giving the same allyl radical(IIId) (fig. 6) as in the photolysis of the azacyclohexadienyl radical, suggesting thatradical (11) is actually an intermediate species between radicals (I) and (111). Thedecrease in the total concentration of radicals when the pyridinyl radicals are phcto-lysed in HCl+H,O matrices may be attributed to an inter-radical electron transferinducing a long range recombination, as suggested iD a previous paper.lThe rearrangement of the carbon backbone evidenced by the analysis of the spectraof allyl radicals derived from picolines may be accounted for by assuming the inter-mediate formation of an azaprismane ammonium radical :tAzaprismane derivatives have been isolated in the photolysis products of severalsubstituted p y r i d i n e ~ .~ ~ ~ The information obtained from the photolysis of the pyridi-nium cation and its methylated derivatives allow the identification of free radicalsproduced under U.V. irradiation of polyvinylpyridines quaternized in an acidic matrix.It is assumed that P2VP and P4VP give rise respectively to the radicals :Although the methine groups of the chain are linked to an allyl carbon, the couplingof the protons of these groups is not observed showing that they are probably locatedin the plane of the conjugated system. A similar observation was made for pyridinylradicals derived from these polymer^.1 52 IRRADIATION OF PYRlDINIUM CATIONS/- I' //2 0 gauss ivi * YFIG.6.-E.s.r. spectrum of the pyridinyl radical (11) trapped in HCl+HzO matrix at 77 K after yirradiation (A). E.s.r. spectra recorded after 6 min (B) and 10 min. (C) U.V. illumination of the pyri-dinyl radical, compared with the spectrum (D) of the (IIId) allyl radical formed by photolysis of the4-azacyclohexadienyl radical (I). The relative amplification of signals is indicated on the right handside of each spectrum.CONCLUSIONThe photolysis of pyridinium and alkylpyridinium cations and of the relatedpolymers in acidic matrices may be interpreted by a four step process : (1) formationof initial radicals which may be photoionized by irradiation at 250-400nm such aspyridyl or azacyclohexadienyl radicals ; (2) photodetachment of an electron fromthese radicals ; (3) scavenging of the photo released electrons by pyridinium cationswith formation of pyridinyl radicals ; (4) photoisomerization of the pyridinyl radicalto the allyl radical (IIId) with the transient formation of an ammonium radicalderived from azaprismane.This work therefore presents new evidence for the importance of the photo-ionization of free radicals in the solid state photolysis of organic molecules.We thank the DCl6gation GCnCrale A la Recherche Scientifique et Technique forits financial support.J. C.Ronfard-Haret, A. Lablache-Combier and C. Chachaty, J. Phys. Chem., 1974, 78, 899.C. Chachaty, J. C. Ronfard-Haret, A.Lablache-Combier, J. P. Quaegebeur and H. Ofenberg,Chem. Comm., 1974, 579.(a) C. Chachaty and A. Forchioni, Compt. rend. C, 1967,264,1421 ; (6) H. Bower, J. A. McRaeand M. C. R. Symons, J. Chem. SOC. A , 1968, 1918.(a) Ye. I. Finkelshtein, Vysokomol. Soed., 1967, A9, 7 ; (b) C. Chachaty, A. Forchioni and J.Desalos, Compt. rend. C, 1970,270,4.49 ; (c) M. Shiotani and C. Chachaty, Bull. Chem. Sac.Japan, 1974, 47,28.C. Chachaty and A. Forchioni, J. Chim. phys., 1969, 66,929.B. Bak, L. Hansen and J. Rastrup-Andersen, J. Chem. Phys., 1954, 22, 2013,G. Vincow and P. Johnson, J. Chem. Phys., 1963, 39, 1143.C. Chachaty, J. Chini. phys., 1967, 64, 608.' S. M. E. Englert and S. M. McElvain, J. Amer. Chem. SOC., 1929, 51, 863QUAEGEBEUR, OFENBERG, LABLACHE-COMBER, RONFARD-HARET AND CHACHATY 1 53l o J.A. Pople, D. L. Beveridge and P. A. Dobosh, J. Anzer. Chem. Soc., 1968, 9% 4201.l i R. Lefebvre and J. Maruani, J. Chem. Phys., 1965, 49, 1480.l 2 S. M. Blinder, M. L. Peller, N. W. Lord, L. C. Aamodt and N. S. Ivanchukov, J. CCienr. Phys.,l 3 P. H. Kasai and D. MzLeod, J. Amer. Cliem. Soc., 1972, 94, 720.l4 (a) C. Heller and T. Cole, J. Chenz. Phys., 1962, 37, 243 ; (6) C. Chachaty and J. Maruani,1962, 36, 540.Canad. J. Chem., 1966,44,2631.J. P. Malrieu and B. Pullman, Theur. Chim. Acta, 1964, 2, 307.1967), p. 21.(a) J. A. Berson and M. R. Wiilcott, J. Anzer. Chem. Soc., 1965, 87, 2751 ; (b) J. A. Berson, P.W. Grubb, R. A. Clark, D. R. Hartte and M. R. Wiilcott, J. Amer. Chem. Sac., 1967, 89,4076;(c) P. Vogel, M. Saunders, N. M. Hasty and J. A. Berson, J. Anier. Chem. Suc., 1971, 93,1551.l 6 F. W. Atkins and M. C. R. Symons, The Structure of 1nnr:pnnic Rociicols (Elsevier, Amsterdam,l 7 J. L. Monge and M. Schott, J. Chim. phys., 1973, 70, 1555.l9 D. K. Ghosh and D. H. Whiffen, Mol. Phys., 1959, 2, 285.2o L. Kaplan, J. W. Pavlik and K. E. Wilzbach, J. Amer. Chem. Soc., 1972,94, 3283.21 T. Shida and I. Hanazaki, Bull. Chem. Sac. Japan, 1970,43,646.2 2 G. Gutmann and L. E. Lyons, Organic Semiconductors (Wiley, New York, 1967).23 M. Itoh and S. Nagakura, J. Amer. Chem. Soc., 1967, 89, 3959.24 (a) K. E. Wilzbach and L. Kaplan, J. Amer. Chem. Soc., 1965,87,4004 ; (b) D. M. Lemal andJ. P. Lokensgard, J. Amer. Chem. Soc., 1966, 88, 5935 ; (c) D. M. Lemal and J. V. Staros, J.Amer. Chem. SOC., 1969, 91, 3373; (d) M. G. Barlow, R. N. Haszeldine and R. Hubbard,Chem. Comm., 1969, 202.25 C. Chachaty, A. Forchioni and J. C. Ronfard-Haret, Mukromol. Chcm., 1973, 173, 213.26 Tables of Interatomic Distances and Configuration in Molecules and Ions (Chem. SOC., London,1958).(PAPER 412058

ISSN:0300-9599    

DOI:10.1039/F19767200143

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Proton Mobility in SolidsPart 5-Further Study of Proton Motion in Decationated Near-faujasite H-sieves byPulse Nuclear Magnetic Resonance.BY M. M. MESTDAGH," W. E. E. STONE AND J. J. FRIPIAT~Laboratoire de Physico-Chimie Minerale, UniversitC Catholique de Louvain,Place Croix du Sud 1, B-1348 Louvain-la-Neuve, BelgiumReceived 25th November, 1974Proton relaxation times have been measured at 30 and 60 MHz at temperatures between 360 and- 180°C for two near-faujasite H zeolites with different iron contents, 700 and 75 p.p.m. Ammoniawas removed from these samples by the " thin bed " procedure. The Fe3+ cations, probably on theI' exchange sites, have an electronic relaxation time Tl(Fe) m T2(Fe) 5 x lo-* s ; T'1(.27Al) ms at 20°C. The temperature range investigated may be divided mto tworegions.From -180" to +180"C, the proton relaxation times TIP and T2,, remain constant.Tlp is determined by a spin energy diffusion process, the paramagnetic impurities being the sinktowards which the proton spin energy moves by a flip-flop mechanism. From 180 to 360"C, the pro-ton longitudinal relaxation time Tlp falls sharply as the temperature is increased. Protons go fromone oxygen atom to another and therefore they move to oxygen atoms which are alternately insidethe hexagonal prism, inside the cubo-octahedron and inside the supercage. This motion modulatesthe proton-paramagnetic interaction which is shown to be the most efficient mechanism for the longi-tudinal relaxation.The activation energy for the proton jumps is 19 kcal mol-'.It is lower for partially dehydroxy-lated " deep bed " samples.s and ?'2(27Al) =The previous paper on proton motion in the near-faujasite HY zeolites (" thinbed " calcined) arrived at the following conclusions.(I) The sharp decrease or increase observed in the longitudinal (Tl) or transverse(T2) relaxation times respectively as the temperature is increased is due to an isotropicdiffusion process of the proton with respect to a continuum of Fe3* paramagneticimpurities.(2) The magnetic interaction p-27Al produces the main contribution to the secondmoment.The main ambiguity in the previous measurements results from the fact that the mini-mum expected in the variation of T1 with temperature for a single relaxation mechan-ism would occur at a temperature exceeding the limit of the thermal stability of thedecationated sieve.In addition no measurement of the relaxation times of the Fe3f spin was available.Moreover it was assumed that all the iron atoms were tetrahedrally coordinatedby the zeolite framework.More work was thus needed to obtain information about the location of the para-magnetic impurities and on their influence on the proton relaxation times.An e.p.r.study of the same samples as used for the n.m.r. work showed thepresence of Fe3+ (a) tetrahedrally coordinated (Fez+), (b) on exchange positions(Fez+), possibly on the SI sites and (c) in an iron oxide-like phase (Fei+) characterizedf present address : Centre de Recherche sur les solides a Organisation Cristalline Iniparfaite, rueQe la Ferollerie, 45045 Orleans Cbdex, France.15M.M. MESTDAGH, W. E . E . STONE AND J . J . FRIPIAT 155by a strong exchange spin-spin interaction. These conclusions are in general agree-ment with recent data published by McNicol and P ~ t t . ~ By combining chemicalextraction experiments and e.p.r. results, the contents Fe,3+, Fe?+ and Fe;+ could beevaluated. Finally, an approximate value for the electronic longitudinal relaxationtime of Fez+ was determined.To reduce the uncertainty in determining the correlation time in the absence ofan observable minimum in T I , relaxation times TIP and T2p were measured at twofrequencies and for two samples with different iron contents. These new data werealso necessary to interpret the longitudinal relaxation mechanism at low temperature.In addition the TI and T2 of 27Al at room temperature became available.This paper is devoted entirely to fully " decationated " Y sieves (" deamination ">90 %).The results obtained for the initial and partially deaminated samples,important from the view point of catalysis, are published el~ewhere.~EXPERIMENTALMATERIALSTwo different Na-Y sieves were used, YI (I for " impure ") commercial sample (LindeY SK 40,3606-282) and Yp (P for '' pure ") made in this laboratory. The starting materialswere respectively 200 Degussa Aerosil and a sodium aluminate solution, obtained by dissolv-ing very pure metallic aluminium in a 16 N NaOH s~lution.~The chemical compositions of these samples, dried at 40°C for 24 h, after being NH,fexchanged by repeated contacts with a 1 N NH4CI solution, were as follows :Yp : (NH4+)56Na18A174Si1180384.220 H2O.YI: (NH4+)38Nal6Al54Sil380384.204 H2ODecationation was theii carried out in a Pyrex tube, 12 mm internal diameter, placed horizon-tally in a furnace, in such a way that the thickness of the powder bed on the wall was about1 mm. The temperature was increased at the rate of 1.8"Cmin-' up to 300 or 360°C.The sample was maintained at the highest temperature for at least 1 h under a residualvacuum of Torr. The tube was then sealed and allowed to cool at room temperature.sample is obtained. Chemical analysis revealedthat at 300"C, YI and Yp had been deaminated by 95 % and 98 % respectively, 36 or 55 OHper unit cell (u.c.) being left in the lattice.The different OH contents are due to the largerinitial NHZ content in Y,, because of the lower Si/Al ratio.After treatment at 360"C, deamination of YI reached 98 % and the number of OH perunit cell was 37.2. Since HY, has a lower thermal stability than HYI, it was not treated attemperatures higher than 300°C. The apparent discrepancy between the number of OHgroups indicated here and those given in table 1 originates from the fact that some protonsare still present as residual H,O and NHt.To obtain some information about the thermal stability of " deep bed " (DB) samples,NH4-Y1 was heated directly to 400°C over about 10 h in a static vacuum, 1 to 2 g of powderbeing loosely packed in a vertical 12 mm (i.d.) tube.Under these conditions, the deamina-tion reached 30 %. The behaviour of this DB (NH4)26,H12-Yi sample is compared to thatof the TB (NH4)22,H16-Y~ sample obtained by the thin bed (TB) procedure, but heated to200°C. Deamination of the latter was 40 %.In no case was the temperature at which the n.m.r. data were obtained higher than thepretreat men t temperature.With such a treatment a " thin bed "DISTRIBUTION OF NUCLEIThe distribution of the iron impurities among the different species has been reportedelsewhere.= Even in YI the Fe,"+ content was negligible. The HY, sample contained nodetectable Fe:+ whereas in HYI, 523 p.p.m. by weight of the sample was assigned to thisspecies, Siace the oxide type impurity probably crystallizes Qut as a separate phase, it haI56 PROTON MOBILITY IN SOLIDSbeen assumed that it does not influence the relaxation mechanism of the zeolitic protons.Therefore only the F e 2 content, 700 p.p.m.and 75 p.p.m. for YI and Yp respectively, needbe considered in this context.Homogeneous distribution of A13+ and Fes+ cations and of protons was considered.The existence of paired protons has been suggested by Freude et aL7* * but when the proton-aluminium magnetic interaction is no longer neglected a p-p distance of the order of thatexpected for a homogeneous distribution is obtained. The average proton-proton inter-nuclear distance is therefore assumed to be the same as the average AI-A1 distance. Theinternuclear distances obtained are shown in table 1.Actually there are still protons com-bined as water or ammonium eveii after outgassing at 300°C but they constitute less than18 % of the total, therefore the observed proton signals are assumed to be entirely due tostructural OH.The contributions of ''0 and 29Si have been ignored as their natural abundances are solow. Theresidual Na+cations are assumed to be on sites I. Their contribution to the protonrelaxation mechanism is negligible.TITId' I I I I JI 2 3 4103 K/TFIG. 1 .-Variation of the longitudinal (upper points) and transverse (lower points) proton relaxationtimes with temperature. Resonance frequency : 60 MHz : 0, H-Yp ; e, H-Yr pretreated at 300°C ;x , H-YI pretreated at 360°C. A, spin-echo Tze obtained for H-YI pretreated at 300°C.Resonancefrequency : 30 MHz : 0, H-YI pretreated at 300°C. The gaussian signals are indicated by Q or Q.The experimental data published previously and obtained without averaging the signals are shown byA : their positions should be compared with those of 0.N.M.R. MEASUREMENTSThe 60 MHz measurements were performed with a modified Briiker BKR 302 spectro-meter fitted with Tektronic 162 and 163 pulse generators. monitored by an electronic timM. M. MESTDAGH, W . E . E . STONE AND 3 . J. FRIPIAT 157counter. The 90" pulse duration was of the order of 2 0 ~ s while the magnetic fieldheterogeneity at the sample was 28+ 13 x G . The 30 MHz data were obtained usinga basically similar instrument fitted with a more powerful high frequency generator producing90" pulses of - 4 x lod6 s.To increase the signal to noise ratio, a Nicolet 1072 time averagerwas used with both instruments.In most cases, the proton longitudinal relaxation time (TIp) was obtained from the foIlow-ing pulse sequence : four 90" pulses, t, one 90" pulse. The decay of the magnetization wasfollowed as a function of t. The integrated amplitude of the last 90" pulse was recorded.The proton transverse relaxation time (Gp) was computed from the decay of the free induc-tion signal, following a 90" pulse. In most cases spin-echo was observed at 60 MHz. Thecorresponding T,, was derived from the exponential decay of this echo.Using a SXP4-100 Briiker spectrometer, the relaxation times of the 27Al nuclei at 21 MHzand of the protons at 30, 60 and 90 MHz were obtained at room temperature,RESULTSThe variations of T I , and T2p with temperature are shown in fig.1. In all cases,the recovery of the longitudinal magnetization can be accounted for by a single ex-ponential function. The situation is more complicated for the transverse component.In the region where TZp was constant, referred to as the '' low temperature " region,the free precession signal decays as an exponential function of the square of the time,corresponding to a gaussian signal. In the region where both TI and T2 are tem-perature-dependent, referred to as the " high temperature " domain, the free preces-sion signal decays as an exponential function of the time, corresponding to a lorentzian-shaped signal.In contrast to TZp, TIP is frequency dependent and is markedlyinfluenced by the iron content.H:/106 GZI oc 200' I00 ' 15030 6350Hk fG+frequency /M HzFIG 2.-Frequency dependence of T , , at 20°C.In fig. 2 the low temperature TIP signal is shown to be either a function of the squareof the static magnetic field Ho for H-Y, or of the square root of H,, for H-Y,.Finally the longitudinal relaxation times TIP of the partially deaminated deep bedsample is compared to that of its thin bed homologue in fig. 3I58 PROTON MOBILITY IN SOLIDSAt 20°C and 21 MHz, Tl(Al) and T,(Al) are about and s respectively.At this temperature, T,(Fe) = T2(Fe) z 5 x lo-* s as derived from the e.p.r. study.2d - 1103 KITFIG. 3.-Comparison between deep bed : DB (NH&,HI2-Y; sample (0) (activation energy, 6 kcalmol-’) .and thin bed deaminated : TB (NH&,HI~-YI sample (0) (activation energy, 19 kcal mol-I)showing the slope of TIP (60 MHz) as a function of temperature.DISCUSSIONIn the low temperature region, the proton correlation time z, is very long withrespect to the relaxation time of the paramagnetic spin, and perhaps of the 27A1nucleus.Therefore the modulation of the dipolar p-S interaction is provoked by therelaxation process of spin S (either electronic or Al). In such a case,9 the longitudinalrelaxation time isfor a random orientation of the internuclear p-S vector of length r. In (l), coo isthe proton resonance frequency. Taking for ys either the gyromagnetic ratio of 27Al,0.697 x lo4 rad G-l s-l (S = 5/2) or of the Fegf electronic spin, 1.76 x lo7 rad G-Is-I ( S = 5/2), gives Ti-pl(~-~’Al)/T<~~ (p-electronic spin) = This ratio isobtained assuming that P is 2.3 A for the p-27Al interaction or that r = 7 A for thep-Fe$* interaction.The latter distance is deduced from the radius of the sphereinside which the proton resonance frequency is shifted well away from single frequency.lThis critical radius is (iip/6Hmax) where 6Hm,, is the peak to peak line width measuredon the first derivative of the n.m.r. absorption signal and where ji, is the paramagneticmoment averaged for the reorientation process experienced by this spin. The relaxa-tion process via the paramagnetic impurities is therefore more efficient than that viathe aluminium nucleus.The relaxation could then occur either by direct contact or by a spin energydiffusion process.To decide between the two, the variation of the low temperatureT I , with frequency is of value (fig. 2).Since the iron content is low relative to the proton content (table l), each protonexperiences a particular field, depending on its position with respect to the Fe;+ ionsM. M. MESTDAGH, W. E . E. STONE AND J . J . FRIPIAT 159The relaxation probability of the longitudinal component of the proton spin at adistance r from the paramagnetic cation (S) may be expressed as C r 6 , whereThis relaxation probability averaged for the protons outside the critical sphere shouldbe equal to Ti: for direct paramagnetic contact.Previously,' in the absence of infor-mation about Tl(Fe), T2(Fe) and about the distribution of the paramagnetic impurities,it was suggested that the low temperature Tlp value was determined by this mechan-ism. TIP was then proportional to NFe and to H;.The proportionality in H i is found for H-Y,, as shown in fig. 2, but if the observedvalues of NFe and of Tls are used to compute the longitudinal relaxation rate assumingdirect contact, the calculated Tc: is two orders of magnitude higher than the experi-mental one. This therefore suggests the possibility of a spin energy diffusion processoperating for Ti:, i.e. two adjacent proton spins must invert simultaneously in oppo-site directions to keep the Zeeman energy of the proton population constant.Thisrelaxation mechanism is very often encountered in solids.'' The probability of aninduced " flip-flop " of spins is not negligible for protons of similar resonance fre-quency and the rate at which this mechanism propagates the spin energy towards theparamagnetic impurity may be defined for a powder by a diffusion coefficienta2 D =30 T,( H-H)where a is the proton-proton distance and [T;'(H-H)] the transverse relaxation rateof the protons dipolar interaction.Suppose that a proton at a distance r from a paramagnetic centre is relaxing witha correlation time z1 = r6/C. Since it takes a time r2 = r 2 / D for the " flow " ofspin energy to reach this proton, z1 = z2 for r = b = C1'4D-114 . For Y < b,relaxation occurs by direct paramagnetic contact while for r > b, spin energy diffusionbecomes the dominant mechanism.However within a distance bo equal to the radiusof the critical volume (Fp/6Hmax)% = (iiP/dH)%a, this spin energy diffusion process isslower because the protons inside this sphere have resonance frequencies which rapidlybecome different from each other as r + bo. Therefore an integration carried overa sphere, with 6, < r < b, leads to a spin energy diffusion process approximated bywhile for r > b, the spin energy diffusion process yields loAccording to eqn (3), when wiTfs % 1, TIP should be proportional to the square ofH,, as observed for H-Y, (fig. 2). Using a = 7.4A (see table l), bo = 21 A andT2(H-H) = 2.33 x s, we get T,, = 1.5 s instead of 1.8 s, found experimentally.According to eqn (4) and also for w$Tfs % 1, T,, should be proportional to thesquare root of Ho : this is observed for H-Y, (fig.2). It has been shown by Blum-berg l1 that the longitudinal relaxation for this case, is described by the followingrelationship :47t3hfz(t) = - N , C"t'3 re (5160 PROTON MOBILITY I N SOLIDSfor t < CkDS. To account for the observed Ti:, D must be about 8 x cm2 sriand then, for t < 14 s, MJt) should recover as a function of the square root of time.When t > 14 s the equilibrium value of the longitudinal magnetization should bereached exponentially. These predictions have been checked for H-Y, at 20°Cusing the values 1.42 x s for T2(H-H) and 18 A for bo. This T2(M-H) is veryclose to T,e obtained from the spin echo measurements (see fig.1).For the transverse relaxation time T;:, the experimental results in fig. 1 show thatthe line width does not depend on the iron content. The mathematical treatmentof these data yields essentially the same result as that obtained previously: themain contribution to the second moment is provided by the p2’Al interaction. Thesecond moment deduced from the low temperature T2p is 0.87 G2. This is close tothat obtained by Freude (0.85 G2).In the high-temperature region the spin-lattice relaxation TIB is determined by thecorrelation time z, of the proton motion. As shown in fig. 1, TlP is a function of theiron content and also depends upon the square of the resonance frequency.The model proposed earlier for the proton longitudinal relaxation, i.e.a protondiffusing within a network of oxygen atoms with a continuum of Paramagneticimpurities, thus appears well founded and, for a proton at a distance I’ from theparamagnetic impurity :Sinceeqn (7) may be integrated over a sphere element 4zr2drz, = r2/6DS4nCNFea2 1 Trp’ = ----5 cog 1’ 6 2 ,(9)where CI is the average jump distance and Z the distance of closest p-Fe approach.To obtain T~ in the absence of a minimum in the plot of T I , against temperature,more information on the motion must be derived from TZp. As shown experimentallyin fig. 1, T,, is independent both of the iron content and of the frequency. This stronglysuggests that when T2p becomes temperature dependent and tends to rise (Le.abovelo3 K/T e 2.1), the proton motion modulates the p-,’Al interaction, since at lowtemperature this interaction brings about the main contribution to the second moment.The onset of the T2p increase occurs sharply at lo3 K/T = 2.1 and is shown by thetransition from a gaussian to a lorentzian signal indicating that z, is equal tothe value of the “ rigid lattice ” T2,, i.e., 4 x From thisequality and from the activation energy derived from the slope of TlP it follows thatThis is valid for H-Y, and H-YI since both samples have the same activationenergy and rigid lattice T2p values.To fit the experimental TlP data as shown by the solid line in fig. 1, ct2/Z5 mustequal 2.41 x ~ m - ~ for H-Yp and 36.2 x ~ n 1 - ~ for H-YI. A variation of oneorder of magnitude in the a2/Z5 ratio is still acceptable in view of the uncertainty inihe distribution of Fe:+ ions in the lattice.The jump distance is evidently smaller than the supercage diameter and sinceZ(p-Fe) 2 Z(p-Al) = 2.2& from the cr2/Z5 ratios obtained experimentally it is con-cluded that a is -4.4 A.The proton diffusion coefiicient derived from eqn (10) and(8) is thens at this temperature.z, = 0.71 x exp(l9 kcal mol-l/RT) s. (10)D = 4.5 x exp(- 19 kcal mol-’/RT) cm2 s-IM. M. MESTDAGH, W. E. E . STONE AND 3. J . FRIPIAT 161Table 2 shows the values of D and of the proton jump frequency at different tempera-tures. A distance of 4 is of the order of magnitude expected for a proton jumpbetween two adjacent oxygen atoms.TABLE ~.-DISTRLBUTION OF THE MAGNETIC NUCLEI WITHIN THE LATTICE (u.c.= unit cell)nucleus HYI HYP27Al (cation per u.c.)* N ~ / c m - ~rAl-Al/A55Fe~ (cation per u.c.)541.15 x loz27.40.2(1 Fez-+ per u.c.)13.5 x lo1*42162.70 x lo2741 . 5 7 ~6.30.022 5(1 Fe:+ per 4.4 u.c.)1 . 5 0 ~ 10I83.21 x 1O2I8718'H per U.C. after outgassing at 300°C 45.5 652.3 x 10'' 3 . 2 7 ~ loz1- NH1g-I'H per U.C. after outgassing at 360°C 40.8NH1g-l 2.07 x 10'' -* A1 species are considered to occupy a volume (V, = 4.6 x cm3) of the lattice obtained bysubtracting the porous volume (V,) (soddite and supercages) from the total volume VT of the U.C.r ~ l - ~ l has been calculated assuming that the lattice is spread out as a " bidimensional " solid with atheoretical surface S = 1540 m2 g-l.TABLE 2.-PROTON JUMP FREQUENCY ( V ) AND DIFFUSION COEFFICIENT, ASSUMING AN AVERAGEJUMP DISTANCE a = 4.4 8,T I T D /cm2 s- 1 vls-10 2.9 x 0 .9 ~100 3.4 x 10-I" 1.1 x lo2200 7.7 x 2 . 4 ~ 104300 2.3 x 10-l' 8 . 2 ~ 105400 3.1 x 10-9 9.7x lo6The temperature at which TI should reach a minimum can be predicted from therelationship a = 6Dwg1. At 60 MHz it should be above 450°C, in the temperatureregion where dehydroxylation occurs.Since each oxygen tetrahedron always has one of its oxygen atoms in the hexagonalprism [see fig. 2, ref. (l)] linking the cubo-octahedra, and another in the supercage,the proton could jump from a position in the prism, to a position in the cubo-octa-hedron and in the supercage. The four oxygen atoms then have an equal probabilityof occupancy, in agreement with thc results of an ENDOR study of a y-irradiated€3-Y zeolite by Vedrine et al.In the temperature range in which the decationatedsieve is used as an acid catalyst (300-400°C) the turnover number of the proton onthe oxygen exposed in the supercage is thus rather high. By turnover number wemean the number of times a proton is associated with a given site.even for reaction in which the Bronsted surface acidity could still play a role. Thisis surprising since a fraction of the surface protons is lost. A possible explanationwould be that the proton motion parameters shown in table 2 are modified by thestructural changes arising from partial dehydroxylation.To investigate this possibility a comparison between a deep bed and a thin bed,Partial dehydroxylation of H-Y is known to enhance its catalytic properties,'1-162 PROTON MOBILITY IN SOLIDSpartially deaminated sample has been drawn in fig.3. Obviously the activation energyfor the deep bed sample, i.e. the slope of the variation of T i t with 1/T, is lower thanthe activation energy for the thin bed sample. This may be an important observationsince the higher the turnover number of the protons on an supercage oxygen atomsthe greater should be the catalytic activity of this proton.Thanks are due to Professor J. Jeener and his collaborators for many helpfuldiscussions and for permitting M. M. M. to use the facilities of his laboratory. The27Al relaxation measurements were kindly made available by Dr. McCarten ofBrucker (Karlsriihe) Co.M. M. Mestdagh, W. E. E. Stone and J. J. Fripiat, J. Phys. Clzenz., 1973, 76, 1220.E. G. Derouane, M. M. Mestdagh and L. Vielvoye, J. Catalysis, 1974, 33,169.B. D. McNicol and G. T. Pott, J. Catalysis, 1972, 25, 223.M. M. Mestdagh, W. E. E. Stone and J. J. Fripiat, J. Catalysis, 1975, 38, 358.L. Lerot, personal communication.G. T. Kerr, Adu. Chem. Series, 1973, 121, 219. ' D. Freude, D. Miiller and H. Schmiedel, Surface Sci., 1971, 25, 289.* D. Freude, D. Muller and H. Schmiedel, J. CoZIoid Interface Sci., 1971, 36, 320.A, Abragam, Principles of Nuclear Magnetism (Oxford University Press, 1961).1970).lo M. Goldman, Spin Temperature and Nuclear Magnetic Resonance in Solids (Clarendon, Oxford,'l W. E. Blumberg, Phys. Rev., 1959, 119, 79.l2 J. C. Vedrine, D. S. Leniart and J. S. Hyde, Ind. Chim. Belg., 1973, 38, 370.l3 W. D. Haag, h e r . Chem. SOC. 165th National Meeting, Petroleum Chem. Div., Paper 26,1973.(PAPER 4/2453

ISSN:0300-9599    

DOI:10.1039/F19767200154

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Quenching of the Luminescent State of theUranyl Ion (UOt') by Metal IonsEvidence for an Electron Transfer MechanismB Y EIUGH D. BURROWS,':' SEBAsTIiiO J. FORMOSiNHO, M. DA GRA~A MIGUELAND F. PiNTO COELHOChemical Laboratory, University of Coimbra, Coimbra, PortugalReceived 20th February, 1975The quenching of the luminescence of the uranyl ion by other metal ions has been studied in aque-ous solution. The quenching is shown to be a dynamic process, and the correlation of the logarithmof the quenching rate with the metal ion ionization potential suggests that intermolecular electrontransfer is the predominant mechanism. Evidence that this involves complete electron transfercomes from flash photolysis of solutions of UO:' and manganese(n), where a broad absorption (Amax= 505 nm) is observed which is assigned to Mn3+.Consideration of the energetics of the quenchingprocess suggests that in the quenching of uranyl by silver@, the products (UO: and AS"+) are pro-duced in their electronic ground states. Studies of the effect of temperature on the quenching suggestthat if an intermediate complex (exciplex) is involved in the quenching then this must involve onlyvery weak binding. With siiver(I), the quenching is sensitive to the ionic strength of the solution.Further studies suggest that the lifetime of the luminescent state of the uranyl ion in aqueous solutionvaries with both temperature and uranyl ion concentration.There has recently been considerable interest in the photochemistry of the uranyland photo-oxidation by U0;f has been demonstrated for a variety of organicand inorganic substrates.Both intra- and inter-molecular mechanisms have beendemonstrated,2+ and it is further suggested that the intermolecular mechan-ism can involve either hydrogen atom abstraction, or electron abstraction by theexcited UOi+ species4 In addition, uranyl ion can photosensitize certain pro-cesses by energy Whilst there is good evidence for the hydrogenabstraction mechani~m,~ evidence for intermolecular electron transfer is rather lessclear. Chibisov and co-workers flash photolysed solutions of uranyl ion with variousorganic substrates, such as phenols and aromatic amines, and observed absorptionswhich they assigned to uranium(v) and the cation radical of the organic molecule,confirming an electron transfer process.However, it is not clear how much of thisproceeds via an intermolecular route as UO;+ is known to complex with, for example,aniline and several other aromatic a m i n e ~ . ~ Matsushima lo has identified some kindof excited state electron transfer process in the quenching of uranyl luminescence byaromatic hydrocarbons by the correlation of the logarithm of the quenching ratewith the ionization potential of the hydrocarbon. As no permanent product isformed, Matsushima interpreted this behaviour in terms of formation of a non-fluorescent exciplex between the aromatic hydrocarbon and the uranyl ion.Our objectives in this study have been to obtain evidence for intermolecularelectron transfer to the excited uranyl ion.A number of studies '* 7 9 haveindicated that inorganic ions quench the uranyl luminescence. With anions, thisquenching may proceed via complex formation. However, with metal cations com-plex formation is mudi less likely, and an intermolecular mechanism is suggested.16164 LUMINESCENCE QUENCHINGWe have studied the quenching of uranyl luminescence by metal ions to attempt tofind evidence for the intermolecular electron transfer mechanism. During the pre-paration of this manuscript, Matsushima and co-workers published their resultson a closely related study of the quenching of uranyl luminescence by inorganic ions.Good evidence was obtained for intermolecular electron transfer in the quenchingby halide and pseudo-halide ions.With metal ions, intermolecular quenching byelectron transfer through a weak electron donor-acceptor interaction was suggested.Our study is complementary to theirs, and provides more definite information on themechanism of the quenching of uranyl ion luminescence by metal ions.EXPERIMENTALMATERIALSUranyl ion and the other metal ions were normally taken as their nitrate salts, andwere of the purest grade commercially available. Iron(@ nitrate was prepared from iron(@sulphate and barium nitrate.13 The absence of iron(m) in this solution was demonstratedby the absence of any red colour in the thiocyanate test.14 With cerium(m), there was aslight increase in the intensity of the uranyl luminescence in the presence of very purecerium(@ nitrate, possibly due to complexing.' Instead the quenching of the luminescenceof uranyl perchlorate by cerium(rrr) perchlorate was studied.Fluorescence enhancementwas also observed in solutions of uranyl ion with zinc@) nitrate or sulphate, and with lesspure grades of cobalt(@ nitrate. Solutions were prepared in singly distilled water with pHkept approximately constant (pH M 2.0-2.5) by addition of nitric acid. The intensity of theuranyl luminescence increased in more acidic solutions, possibly due to the presence ofdifferent uranyl nitrate complex species possessing different luminescent quantum yields.APPARATUS AND PROCEDURELuminescence spectra were run on a Perkin-Elmer MPF-3 spectrofluorimeter. Normallyspectra were run at room temperature, but for temperature studies the cell was thermostattedin a water-circulating cell block giving temperatures kO.5"C. Solutions were not normallydegassed as oxygen has little effect on the uranyl ion luminescence.16 The uranyl lumines-cence was excited at 406 nm and the emission studied at 510 nm.Nickel(@, iron(m) andcobalt(@, absorbed extensively at the excitation or emission wavelength and a correction wasmade to the quenching data.l Flash photolysis experiments were performed on an AppliedPhotophysics GD-20 ps apparatus with photoelectric detection.RESULTS AND DISCUSSIONQUENCHING OF URANYL LUMINESCENCE BY METAL IONSLuminescence spectra were obtained for aqueous solutions of uranyl nitrate (0.02niol dm-3) in the presence of varying concentrations of other metal nitrates.Potas-sium nitrate (up to 1 mol dm-3) and barium nitrate (up to 0.2 mol dm-3) had noeffect on either the intensity or form of the emission demonstrating that nitrate ionwas not a quencher. Quenching of the uranyl luminescence was observed by thenitrates of silver(I), mercury(1) (H&+), copper@), cobalt(Ir), lead@), iron(@, man-ganese@), iron(n1) and nickel(r1). The quenching of the luminescence of an aqueoussolution of uranyl perchlorate (0.02 mol dm-3) by cesium(r~r) perchlorate was alsostudied. In each case, the quenching followed good Stern-Volmer kinetics. Quench-ing is possible by either static (formation of ground state complexes) or dynamicprocesses. We favour dynamic quenching in all the above cases as the absorptionspectra of solutions of uranyl ion in the presence of the other metal ions were identicalto the sum of the spectra of the individual ions.Further, on flash photolysis ofuranyl sdphate (0.01 rnol dm-3) in concentrated sulphuric acid, the uranyl ion excite€I. BURROWS, S. FOKMOSINNO. M. MIGUEL AND 1:. COELHO 165state absorption (Lax 590 nm) decayed faster in the presence of Fez+ confirming adynamic quenching process with a bimolecular rate in this solvent 3.9 x lo6 dm3mol-I s-'. Whilst extrapolation of the quenching mechanism from concentratedsulphuric acid to water is not without risk, strong support for the same quenchingprocess in water comes from the work of Yokoyama et aL17 who have shown fromluminescence lifetime studies that the quenching of uranyl ion luminescence by thal-lium(1) in water is a dynamic process.Finally, the effect of temperature and ionicstrength on the quenching of uranyl luminescence by silver([) and copper(1i) (see later)are most reasonably interpreted in terms of a bimolecular quenching mechanism.From our Stern-Volmer quenching data, and the known lifetime of the uranylexcited state under these conditions (a lifetime of 1.25 p s is obtained for the excitedstate absorption following laser flash photolysis of 20 mmol dm-3 uranyl perchloratein water 18), bimolecular rate constants (k,) can be obtained for the quenching pro-cess. Stern-VoImer constants and values of k, are summarized in table 1. Manyof the quenchers in this table were also studied by Matsushima's group.' Whilstsilver(1) is found to be the most efficient quencher in both studies, the absolute magni-tudes of the quenching constants and the order of the efficiencies of the other quen-chers differ. We feel that this difference must stem mainly from the different experi-mental conditions employed.Thus, for example, decreasing the uranyl ion concen-tration increases the Stern-Volmer quenching constant (table 3). Further, variationsin pH markedly affect both the lifetime of the uranyl luminescence and the natureof the species present," and it is possible that under the more acidic conditionsemployed by Matsushima et aL7 quenching of protonated species becomes important.TABLE ~.-%"ERN-vOLMER CONSTANTS AND RATES OF QUENCHING OF URANYL ION EXCITEDSTATE BY VARIOUS METAL IONSion A"+&2+ce3+co2+CUZ'Fez +Fe3+Mn2+Ni 2+Pb2+TI+HgZ+z475658272926268025288281Ks.v.ldm3 mol- *4350~5 x 10-30.3812.13.36833G67.3 f12504.256.940ionizationkea/dmj mol-1 s 1 potentialb/V relevant energy levclc/cm-l3 .4 8 ~ 109G ~ X 1033 . 0 5 ~ los9.70 x lo62 . 6 9 ~ lo66 . 6 8 ~ lo8~ 5 . 4 0 ~ 1071 .oo x 1093 . 2 0 ~ 1075~ 109h3.41 x lo65.53x lo621.535.533.333.536.830.656.8 4T2g 18 5002633.735.16 ' A i g 1.5 40031.920.4ddd*TIO 20 OOO2TZg 12 6003rIg 20 000*AZu 16 OOO*Tig 12 600d4T10 18 9003T1g 13 500ddnumber ofunpairedelectrons00130 Calculated using a lifetime from ref.(18) of 1.25 ps for uranyl excited state ; b data from ref.(32) ; C data ref. (33) and (34) ; d no excited states below the lowest UOf+ excited state at 20 500cm-I ; e as perchlorate ; accuracy is lower in this case due to Fe3+ absorption at excitation andemission wavelengths ; calculated from oxidation potential of this ion, and oxidation and ionizationpotentials of Fez+ and Ag' ; h for sulphates in 1 mol dm-3 sulphuric acid from ref. (17).POSSIBLE MECHANISMS FOR THE QUENCHING PROCESSThree mechanisms have been suggested ' for the bimolecular quenching of excitedstates by metal ions : (i) electron transfer or charge transfer between the metal ionand the excited species ; (ii) electronic energy transfer from the excited species to themetal ion ; (iii) assisted intersystem crossing (spin exchange or heavy atom effect)166 LUMINESCENCE QUENCHINGThe luminescence level of the uranyl ion is possibly a triplet state 4* l9 ("a, andmechanism (iii) is consequently possible for the deactivation process.A heavy atomeffect is unlikely, as there is no obvious correlation of quenching ability with theatomic number (2) of the metal, so that if mechanism (iii) is operative it would beexpected to be induced by unpaired electrons. However, the most efficient quenchingis observed with T1+, Ag+ and Hg$+, all of which are diamagnetic, and with the otherions there is no correlation between the quenching efficiency and the number ofunpaired electrons.Electronic energy transfer has been observed from the excited uranyl ion toCO(CN)~-,~* C~fen):+,~ rare earth ions 4* ' and stilbenes.6 However, in the presentstudy, the most efficient quenching is observed with ions which do not possess elec-tronic excited states below the emitting level of UOi+ (20 500 cm-I).With the otherions an attempt was made to correlate the quenching data with the square root of theenergy gap between the emitting UOg+ state and the closest state of the other metalion by applying a tunnel effect theory 2o to the energy transfer mechanism. Only apoor correlation was observed suggesting strongly that mechanism (ii) is not thedominant mechanism here.The failure of mechanisms (ii) and (iii) in explaining the quenching suggests thatsome kind of electron transfer process [mechanism (i)] may be responsible.If thisis the case, we would expect the logarithm of the quenching rate to be linearly relatedto the overall free energy change in the reaction, or to the standard one electronoxidation potential (E") of the quenching metal ion.''* 21 Unfortunately, becauseof the instability of higher oxidation states, E" values are only known for a few of theions involved in this study. An alternative procedure which has been employed instudies of electron transfer quenching by organic species is to correlate the logarithmof the quenching rate with the gas phase ionization potential of the quencher.22This is reasonable for quenching by uncharged molecules in nonaqueous solvents,but may not be so reasonable for quenching by metal ions in water, where hydrationenergies are expected to make an important contribution to the overall energychange.1* 23 However, a theoretical treatment l of electron transfer from mono-,di-, and tri-valent ions to a divalent ion indicates that hydration only contributesslightly (<20 %) to the overall energy change, the major contribution coming fromthe ionization potential term. We thus feel that gaseous ionization potentials ofmetal ions are appropriate energy terms to use for the electron transfer quenchingof excited uranyl ion by metal ions. A plot of the logarithm of the quenching rateagainst the ionization potential of the ions studied gives a reasonable straight line(fig. 1) 24 suggesting strongly that some form of electron transfer is involved in thequenching process, in general agreement with the mechanism proposed by Mat-sushima et aL7 Iron(Ir1) and barium did not fit this plot at all.With iron(IIr), thereported rate constant is probably an upper limit, due to the difficulty in correctingthe observed quenching data for iron(II1) absorption at both emission and excitationwavelengths. In the study of Matsushima's group,7 iron(II1) was found to be a lessefficient quencher than either cobalt(I1) or manganese@). With barium, the ionizationpotential is thought either to be in error or not to be applicable to aqueous solution,las no reports exist in the literature for oxidation of this ion to Ba3+ in water. Iron@)and cerium(1u) give a very poor correlation in fig.1. The reason for this is not atpresent clear, although it is possible that different mechanisms are operating here.Trivalent rare earth metal ions,4* CO(CN)~-,~* ' and Cr(en):+ ' quench exciteduranyl ion by energy transfer, and it appears probable that with iron@) there is somecompetition between energy transfer and electron transfer as quenching mechanisms.The mechanism which predominates will depend both on the electronic energy gaH. BURROWS, S. FORMOSINHO, M. MIGUEL AND F . COELHO 167between the emitting level of the uranyl ion and the next lowest level of the quenchingion, and on the ionization potential of the quencher ion. The luminescent state ofcerium(Ir1) lies above the emitting level of the uranyl so that energy transferis unlikely in thislcase.The quenching by cerium(@ is less efficient than that expectedon the basis of the ionization potential. Whilst we are unable to offer any explanationfor this, it is known that the oxidation potential of the cerium(u1)-cerium(1v) coupleis very sensitive to changes in the degree of complexation128 and it is possible thatwe should correcqthe ionization potential of the free ion for complexation in this case.62 0 3 0 4 0ionization potential /VFIG. 1.-Correlation of logarithm of quenching rate and ionization potentials for quenching of uranylluminescence by: (1) Ag+; (2) Ce3+; (3) Co2+; (4) Cu2+; (5) Fe2+; (7) HgZ'; (8) Mn2+;(9) Ni2+ ; (10) Pb2+ ; (11) T1+.Two possible mechanisms exist for electron transfer quenching. Either the quench-ing occurs via formation of an excited state charge transfer complex (exciplex) whichsubsequently relaxes radiationlessly to the ground state of UO;+ and the ground orexcited state of the quencher (l), or there is complete electron transfer (possibly viaan intermediate exciplex) to give uranium(v) and the oxidized form of the other metalion (2)(UO;+)*+M"+ + (UO; .M("+l)+)* -+ UO:++Mnf (1)(U02+)*+Mn+ + UO,++M(n+l)+. (2)Flash photolysis allows us to distinguish between these possible mechanisms. Follow-ing flash photolysis of an aerated solution of manganese(@ nitrate (0.2 mol dm-3) anduranyl nitrate (0.02 mol dm-3) in water (PH 3.1) using light A > 300 nm, a broadabsorption, A,,, = 505 (f 10) nm, was observed (fig. 2). This decayed by apparentfirst order kinetics [k = 2.6 (k0.5) x lo2 s-l].An identical absorption was observedfollowing flash photolysis of an aqueous solution of the perchlorates under the sameconditions. No absorption at 500nm was observed following flash photolysis ofan aqueous solution of potassium nitrate (0.2moldm-3) and uranyl nitrate (0.02mol dm-3) under the same conditions, indicating that the 505 nm species cannot bedue to any nitrate derived radical. Assignment of the absorption can be made froma consideration of literature data. The most likely candidates are UO; and Mn3+.UO,' only has weak absorptions at 738 and 9401un.~~ However, rnanganese(1u16s LUMINESCENCE QUENCHINGpossesses a broad absorption in the visible region with maximum absorption varyingbetween 450 and 550 nm, depending on the degree of complexation,26* 27 and weassign the 505nm absorption to this species.Matsushima et a1.' have suggestedthat manganese@) might quench uranyl excited state by an energy transfer mechan-ism. If this were so, then formation of Mn3+ would be expected to proceed viaelectron expulsion from excited Mn2+. Flash photolysis of a solution of manganese-(zr) nitrate (0.2 mo1 ~ I r n - ~ ) in water using all of the light from the flash lamp did yielda very weak absorption at 500 nm (about 5-10 % of the intensity of the absorptionfrom flash photolysis of UO$+ and Mn2+). However: no such absorption was observedfollowing flash photolysis of an aqueous solution of manganese@) perchlorate underthe sameconditions, suggesting that a nitrate derived radical is responsible for the absorp-tion, and indicating that manganese@) does not photoionize in water under theseconditions. These results suggest very strongly that, in our experiments, manganese-(XI) ion quenches UOg+ excited state by the electron transfer mechanism (2).If theexcited uranyl ion can oxidize manganese@), it is probable that complete electrontransfer also occurs in the quenching of uranyl luminescence by the other more easilyoxidized ions. The existence of such an electron transfer is not unreasonable, asthe excited UO:+ species, with an electronic energy 20 500 cm-', is expected to be astronger oxidizing agent than ground state UO;+ (E" + 0.062 V 28). We can estimatea standard oxidation potential for the excited uranyl ion ' 9 29 using the E" value forground state UOz+, the UO$+ excitation energy, and the relevant energy levels of the1 0.04-. .---* -_I I I4 5 0 500 5s 0 6 0 0wavelengthlnmFIG.2.-End of microsecond pulse spectrum following flash photolysis of uranyl perchlorate (2 xUO: ion. Entropy contributions are ignored as they are expected to be smallcompared with the other terms. E" values of +2.5, 1.7, 0.9 and 0.5 V are estimatedfor the excited UO:+ ion, depending on whether UO,' is formed in the ground state,first, second or third excited state. The quenching studies suggest that excited UO;+will oxidize Agf, and as the standard oxidation potential for the Ag+/Ag2+ couple is1.93V,28 the E" value for the excited uranyl ion must be greater than this.Thissuggests very strongly lhat following quenching of excited UOi+ by Ag+, the productUO; and Ag2+ ions are both formed in their electronic ground states.mol dm-3) and manganese@) nitrate (0.2 mol drn-3) in water (PH 3.1).TEMPERATURE AND SALT EFFECTS ON THE QUENCHING PROCESSThe quenching of uranyl luminescence by Ag+ and Cu2+ was studied as a functionStern-Volmer quenching constants for this are summarised in of temperatureH. BURROWS, S . FORMOSINHO, M . MIGUEL AND F . COELHO 169table 2. Standard deviations of the slopes are rather high, but there seems to be adefinite slight decrease in the slope with increasing temperature. More marked,however, is the decrease in the intensity of the uranyl luminescence in the absence ofquencher. This decreases by a factor of three on going from 20 to 50°C.The mostprobable explanation is that the lifetime of the excited uranyl ion in water decreaseswith increasing temperature. A similar effect has been reported for the luminescencelifetime of U02+ in sulphuric acid, which decreases with increasing temperature above-200 K 3 0 If the lifetime of the excited uranyl ion decreases with increasing tem-perature, then the slight decrease observed in the Stern-Volmer quenching constantactually represents an increase in the quenching rate by a factor 2-3 over the range20-50°C for both Ag+ and Cu2+ as quenchers. The similarity of the temperatureeffects for the two quenchers is perhaps surprising in view of the difference in theirquenching abilities. The kinetic scheme proposed by Weller and Rehm 21 for theelectron transfer quenching of organic fluorescence contains the initial formation ofan encounter complex before the electron transfer step (3)k i k iF*+Q + F*.. . Q + F'-+Q'+.k - 1 k - 2(3)If such a complex is formed in the present case, it must be only weakly bound, evenwith Cuz+, to explain the similarity between the temperature effects on Ag+ and Cu2+quenching.TABLE 2.--EFFECT OF TEMPERATURE ON THE STERN-VOLMER CONSTANTS FOR THE QUENCHINGOF URANYL LUMINESCENCE a BY Ag+ AND Cu2fion temperature/'C Ke.v.b/dm3 mol-1 roc3850( & 860) 2.93810(f 800) 2.03740(&900) 1.43370(+ 420) 15.06(+ 1.17) 3.14.87(f 1.13) 1a for 0.02 mol dm-3 uranyl nitrate in water ; b least squares slope with standard deviation ;C intensity, relative to that at 50°C, of uranyl luminescence in absence of quencher.To attempt to obtain further information on the quenching process, the quenchingof uranyl luminescence by Ag+ was studied as a function of ionic strength.Quench-ing data are presented in table 3. The Stern-Volmer constant increases with increas-ing ionic strength, and, as the intensity of the uranyl luminescence is not affected byvariation in ionic strength, this must reflect a primary salt effect on the quenchingprocess, providing further support that quenching is a dynamic process. A Bronsted-Bjerrum plot 31 for the salt effects on the quenching at the three lowest ionic strengths(where the Bronsted-Bjerrum equation is valid 31) had a slope of+ 1.04 (k0.12).Itwas not possible to study the effect of ionic strength over a wider range because of theweakness of the uranyl ion luminescence at concentrations below that employed heremol dm-3). The positive slope for the Bronsted-Bjerrum plot clearly impliesreaction between two positively charged species. The experimental slope is less thantwo, which may reflect the charge distribution in the excited UO;+ species.35The slope of the Stern-Volmer plot for 1 mmol dm-3 uranyl nitrate solution wasgreater than that for 0.02 mol dm-3 uranyl nitrate solution, consequently the quench-ing was studied as a function of uranyl ion concentration, and data is reported intable 3. The Stern-Volmer slope decreases with increasing uranyl ion concentration170 LUMINESCENCE QUENCHINGAn increase in ionic strength has been shown to increase the quenching rate, and,consequently, this decrease in Stern-Volmer slope with increasing uranyl ion concen-tration must reflect a decrease in the lifetime of the uranyl excited state.It is possiblethat this decrease in emission lifetime results from the quenching of the uranylexcited state by an impurity in the uranyl nitrate. If there is an impurity whichquenches at a diffusion controlled rate it must be present at a concentration -0.2 %.An alternative explanation is that uranyl ion ground state quenches uranyl luminescencein the quenching reaction(UO;+)*+U0~+ 3 2uo;+. (4)TABLE 3.-QUENCHING OF THE LUMINESCENCE OF URANYL ION BY Ag+ a AS A FUNCTION OFIONIC STRENGTH AND URANYL CONCENTRATIONI UOpl /mol dm-3 p/mol dm-3 Ks."./dm3 mol-11 x 10-31 x 10-31 x 10-31 x 10-31 x 10-34~ 10-31 x2x3~ 10-38x5.5x0.250.51.2 x3x6x930010 80014 59016 16017 710916069804350a both as nitrates ; b ionic strength (p) varied by addition of potassium nitrate.A rate constant 2 x lo7 dm3 mol-l s-l can be estimated for reaction (4) from theobserved quenching data.Some support for the self quenching reaction (4) comesfrom experimental determination of the uranyl excited state lifetime in water usingflash photolysis and single photon counting, where there is a difference in the lifetimedetermined by the two techniques,l which, significantly, require different concentra-tion of uranyl ion.If self quenching is occurring, it may occur either by a heavyatom effect, or by excimer formation.We are grateful to Sra. D. M. C. S. Viais for carrying out some preliminaryexperiments, and to the Instituto de Alta Cultura (C.E.Q.N.R. project CQ-2) andServico Meteorol6gico Nacional for financial support.A preliminary account of this study has been presented to the Academia das Cihcias de Lisboa,H. D. Burrows, S. J. Formosinho, M, da G. Miguel and F. Pinto Coelho, Mem. Acud. Ci2nciusLisboa, in press.E. Rabinowitch and R. L. Belford, Spectroscopy and Photochemistry of Urunyl Compounds,[(Pergamon, Oxford, 1964).V. Balzani and V. Carassiti, Photochemistry of Coordination Compounds (Academic, London,1970).H.D. Burrows and T. J. Kemp, Chem. SOC. Rev., 1974,3,139.R. Matsushima, Chem. Letters (Japan), 1973, 11 5 .R. Matsushima and S. Sakuraba, Chem. Letters (Japan), 1973, 1077.G. Sergeeva, A. Chibisov, L. Levshin and A. Karyakin, Chem. Comm., 1974, 159.A. E. Comyns, Chem. Rev., 1960, 60,115.' R. Matsushima, H. Fujimori and S. Sakuraba, J.C.S. Faraday I, 1974, 70, 1702.lo R. Matsushima, J. Amer. Chem. SOC., 1972, 94, 6010. ' ' ref. (1) and references therein.M. Volmer and W. T. Mathis, Bull. SOC. chim. France, 1933,53,385.l3 Kirk-Othmer Encyclopedia of Chemical Technology (Interscienck, New York, 1967), vol. 12,p. 37H. BURROWS, S. FORMOSINHO, M. MIGUEL AND F . COELHO 171l4 A. I. Vogel, Quantitative Inorganic Atuzlysis (Longmans, 2nd edn., 1951), p.645.l5 D. D. Pant and H. B. Tripathi, J. Luminescence, 1974, 8, 492.l6 J. L. Kropp, J. Chem. Phys., 1967, 46, 843.l7 Y. Yokoyama, M. Moriyasu and S. Ikeda, J. Inorg. Nuclear Chem., 1974, 36, 385.R. J. Hill, T. J. Kemp, D. M. Allen and A. Cox, J.C.S. Faraday I, 1974,70,847.l9 J. T. Bell and R. E. Biggers, J. Mol. Specfr., 1968, 25, 312.2o S. J. Formosinho, J.C.S. Faraday 11, 1974,70,605.21 D. Rehm and A. Weller, Israel J. Chem., 1970, 8, 259.22 T. R. Evans, J. Amer. Chem. SOC., 1971,93,2081.D. A. Johnson, Some Thermodynamic Aspects of Inorganic Chemistry (Cambridge UniversityPress, 1968), pp. 126-133.24 In our preliminary account we plotted a function of the logarithm of the quenching ratecorrected for diffusion (log kdifflkdiff - k ~ ) 22 against the ionization potential, assuming adiffusion controlled rate constant (kdiff) of 10'' dm3 mol-' s-l. However, kdiff strictly varieswith ionic charge for reactions between ions [see, for example, ref. (31) pp. 158-1621 and weshould use different values for quenching by mono-, di and tri- valent ions. As most of thequenching rates in our study are very much less than diffusion-controlled, we have chosen toignore the diffusion correction to avoid this problem.25 J. T. Bell, H. A. Friedman and M. R. Billings, J. Inorg. Nuclear Chem., 1974, 36, 2653, andreferences therein.26 A. Y. Drummond and W. A. Waters, J, Chem. SOC., 1953,435.27 C. F. Wells, D. Mays and C. Barnes, J. Inorg. Nuclear Chem., 1968, 30, 1341.28 G. Charlot, D. Bezier and J. Courtot, Selected Constants, Oxydo-Reduction Potenfials (Per-29 see G. Navon and N. Sutin, Inorg. Chem., 1974,13,2159, for a related treatment of the oxida-30 V. N. Korobeinikova, V. P. Kazakov and Yu. N. Chuvilin, Doklady Phys. Chem., 1974, 213,31 see, for example, R. E. Weston, Jr. and H. A. Schwartz, Chemical Kinetics (Prentice-Hall,32 Handbook of Chemistry and Physics (Chemical Rubber Co., 52nd end., 1971-2), p. E-56.33 C. J. Ballhausen, Introduction to Ligand Field Theory (McGraw-Hill, New York, 1962), and34 C. K. Jorgensen, Absorption Spectra and Chemical Bonding in Complexes (Pergamon, Oxford,35 see, for example, ref. (4) for a discussion of charge distribution in the excited uranyl ion.36 R. W. Matthews and T. J. Sworski, J. Phys. Chem., 1975,79, 681.gamon, London, 1958).tion potential of excited Ru(bipy)g+.1078.Englewood-Cliffs, 1972), pp. 165-171.references therein.1962).(PAPER 5/366

ISSN:0300-9599    

DOI:10.1039/F19767200163

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Conduction and Relaxation of Cations in DehydratedPartially Copper(@-exchanged Synthetic FaujasitesBY ROBERT A. SCHOONHEYDT" AND FIRMIN VELGHECentrum voor Oppervlaktescheikunde en Colloidale Scheikunde,De Croylaan 42. B-3030 Heverlee, BelgiumReceived 5th March, 1975The electrical conductivity and dielectric relaxation of dehydrated synthetic faujasites X and Ywith various Cu2+-contents are measured at temperatures between ambient and 750 K and in thefrequency range 200-3 x lo6 Hz.The supercage cations are responsible for the conduction process. At exchange levels below11-14 Cu2+ ions per unit cell these are Na+ ions. At higher exchange levels a low temperature anda high temperature conduction process are evident, due to Na+ and Cu2+ ions, respectively, in thesupercages.The activation energies and entropies for Na+ conduction decrease with increasingCu2+ in the small cages : 74-32 kJ rnol-' and -20 to - 150 J mol-I K-l, respectively. The activa-tion energy and entropy for Cu2+ conduction are 120+ 10 kJ mol-l and -9+ 13 J mol-' K-',respectively. They both decrease significantly in the presence of protons.In the experimental temperature and frequency ranges 3 relaxations were observed. In orderof decreasing critical frequency at a given temperature they were assigned to (i) local migration ofsite 111' cations between two occupied sites 11, (ii) cationic jumps confined to the sodalite cages and(iii) electrode polarizations. A model explaining the variation of the intensities of relaxation (i) and(ii) with degree of Cu2+-exchange and with the temperature is proposed.Excess Cu is present in an unidentified hydroxylated form and does not intervene in the ionicprocesses studied.The replacement of Na+ by Ca2+ in the hexagonal prisms and sodalite cages ofsynthetic faujasites resulted in a significant decrease of the activation energy formigration of the supercage Na+ ions.'- Furthermore a cationic relaxation due tothese Ca2+ ions in the small cavities was observed.These measurements allowed adistinction to be made between exchangeable cations in the supercages and in thedense cages. For the latter no discrimination between exchangeable cations in thehexagonal prisms and in the cubo-octahedra was possible.The distribution of Cu2+ ions in dehydrated synthetic zeolites, types X and Y , isunusual as compared with other divalent cations, in that Cu2+ strongly prefers thesites I' in the sodalite ~avities.~'lO The question arises as to how far these differencesof partition of Ca2+ and Cu2+ among sites I and I' are reflected in ionic conductiondue to supercage cations.Moreover, if the differences in distribution of Ca2+ andCu2+ among sites I and I' are reflected in the ionic relaxation, a combination of electricconduction and dielectric relaxation measurements may be used to investigate quanti-tatively the distribution of exchangeable cations among different sites. Such measure-ments, when performed over an extended temperature range, will enable the investigatorto gain insight into the relative potential energies of exchangeable cations on their sites.EXPERIMENTALSamples were prepared in different ways and their cation contents axe given in table 1together with the symbols defining them used throughout the text.The number after thesymbols is the Cu-content in percent of the cation exchange capacity.1 7R. A. SCHOONHEYDT AND F . VELGHE 173The samples YCul4, YCu32, YCu42 and YCu68 were prepared at room temperatureby exchange of the Na+ forms with 0.1 N solutions of NaCl and CuC12 in a suitable ratioto obtain the desired exchange level. The solid/liquid ratio was 6 kg m-3 and the exchangetime 48 h. The samples YCu60 and XCu64 were prepared with 0.01 N CuClz solutionsunder the same conditions. The absence of NaCl decreases the pH of the exchange solutionand some protons are incorporated in the solid (table 1).The sample YCulOO was preparedby exchange at 363 K during 2 weeks with regular renewal of the 0.01 N CuCI2 solution.XCulOO was prepared under reflux conditions. In both cases an appreciable excess Cu waspresent. Finally, YCuLa was obtained by exchange with La3+ at room temperature for24 h, washing, drying in air at 773 K, followed by a CU’+ exchange with 0.1 N CuC12 atroom temperature for 24 h.TABLE EXCHANGEABLE CATIONS PER UNIT CELL AND THERMAL BEHAVIOUR OF THE ZEOLITEStemperatures/K ofsample IYCul4YCu32YCu42YCu60YCu68YCulOoYCuLaxcu64XCulOO\Tat ions1u.c. Cuzt ions1u.c. H+ ions1u.c.48.337.331.415.619.43.38.412.53.93.4 -8.8 -11.8 -16.8 5.521 .o -41.1 -6.6 -27.5 19.052.4 -water loss446,494443,503448,503443,493443,491438,508,608437,508383,443,523383,433,513,613crystallinity loss1193119311881183115395312351153988All the samples were characterized by their X-ray diffraction pattern.Water loss andthe thermal stability were recorded on a TG-DTG-DTA apparatus at a heating rate of0.25K s-l and in a N2-stream of 5.33 xThe equipment and pellet preparation procedure for the electric measurements weredescribed The only difference was the pressure applied to the pellet,1.9kO.1 x lo7 kg m-’. The pellet density was 70 % of the crystal density.The pellets were degassed in situ overnight and at room temperature down to a pressure< N m-’.The temperature was raised slowly over a 10 11 period to 400 K undercontinuous evacuation. 100 Torr O2 was introduced and the system kept overnight at400 K. O2 was renewed and the temperature raised to 723 K over 10 h. O2 was drivenoff overnight. Under continuous evacuation at 723 K the conduction and capacity weremeasured at 15 frequencies in the range 2 x lo2-3 x lo6 Hz. This was repeated afterequilibration at different temperatures in the range 295-723 K at intervals of -4OK.Reversibility was checked by working at both decreasing and increasing temperatures. Theequilibration time at each temperature was at least 8 h. Overnight and weekend equilibra-tion times gave identical results. After the experiments the crystallinity of the samples waschecked by comparison of the X-ray spectra with those obtained before the experiments.Spectral changes due to structure damage were never observed.m3 s-l.Table 1 summarizes the data.RESULTSAt every temperature d.c. and a.c. conduction were superposed. They wereseparated by extrapolation of the experimental conductivity values obtained in therange 200-3 x lo6 Hz at a given temperature to its static value us. This was possibleat every temperature except when a relaxation occurred in the low frequency range.In any case the measurements were completely reversible with respect to the tempera-ture and the frequency174 CU~+-MOBILITY IN ZEOLITESD.C. CONDUCTIVITYThe d.c. conductivity is strongly dependent on the Cu2+ content of the zeolitesand follows an Arrhenius law in its temperature dependence.Fig. 1 shows theseplots. The samples YCu14, YCu32, YCu42 and YCuLa are characterized by one,103 KITFIG. 1.-The d.c. conductivity as a function of the inverse of the absolute temperature. 0, YCuLa ; v, YCu42 ; A, YC~i60; 6, YCu68 ; V, YCu32 ; A, YCu14; ., YCulbO; 0, Xd1.164 ; 0,x c u 100.the others by two conduction mechanisms. The transition temperature between thetwo mechanisms decreases with increasing Cu2+ content. Fig. 2 gives the dependenceof the d.c. conductivity on the Cu2+ content at 724 K. This plot clearly shows thatthe high temperature (HT) conduction starts at 1412 Cu2+ ions per unit cell. Th175small circles in the lower right hand part of fig. 2 are taken from the low temperature(LT) conduction, extrapolated to 724 K. They suggest that this conduction mechan-ism is the same as for the samples with small Cu2+ content.Fig.3 and 4 show the dependences of the activation energies and the activationentropies of the LT conduction process on the Cu2+ or Na+ contents of the samples.R. A . SCHOONHEYDT AND F. VELGHECu2+ ions per unit cellFIG. 2.-Dependence of the d.c. conductivity on the Cu2+ content at 724 K. The lower set of circlesat the right-hand side are taken from the low temperature conductionI(fig. 1) extrapolated to 724 K.dCu2+ ions per unit cellFIG. 3.-The activation energies for the Na+ conductivity as a function of Cu2+ content. 0, Y typezeolites ; 0, X type zeolites176 cU2$-- M 0 B I LI T Y I N 2 E 0 t 1 T E SThe activation energy decreases linearly from that for YNa down to 32+ 5 kJ mok'at 14+2 Cuz+ ions per unit cell and remains constant at higher Cu contents.Theactivation energy for YCuLa is 41 kJ mol-I. The X type zeolites investigated fit thispicture too. The variation of the activation entropy is similar. Fig. 4 shows thatabove 11 +2 Cu2+ ions per unit cell (below 34+4 Naf ions per unit cell) AS remainsalmost constant at - 146+ 13 J mol-1 K-l.iNa+ ions per unit cellFIG. 4.-The activation entropy of the Na+ conductivity as a function of Naf content for Y typezeolites.Finally table 2 summarizes the thermodynamic parameters derived from the HTconduction process. It is remarkable that for YCu60 and YCulOO the activationenergy and entropy are significantly lower than for the other samples.The latterhave an average activation energy of 120 _+ 10 kJ mol-1 and average activation entropyof - 9 5 13 J mo1-1 K-l.TABLE 2.-ACTIVATION ENERGIES AND ENTROPIES FOR THE HT CONDUCTIONsample EjkJ mol-1rt10 AS/Jmol-1 K-lf13YCu60 99 - 61YCu68 I25 - 12YCulOO 84 - 51XCu64 120 f 2XCulOO 116 - 18DIELECTRIC RELAXATIONIn the experimental temperature and frequency ranges all the samples exceptXCulOO, XCu64 and YCu14 exhibit 2 relaxation processes. A typical example isshown in fig. 5, while fig. 6 gives the temperature dependence of relaxation I forYCu32. This relaxation appears in the spectrum around 400K. Its intensityincreases with temperature to attain a maximum value in the range 650-700 K.Thismaximum is in the tgd-notation 0.59 for YCul4 and l.OS&O.OS for all the otheR. A. SCHOONHEYDT AND F . VEtGHE 177samples. The corresponding value of E” is 9.7k 1 x 10-l’ F m-’. The critical fre-quencies at a given temperature follow the order of the conductivities, shown in fig. 2.Owing to the distribution of relaxation times (fig. 5) this cannot be translated directlyinto a number of contributing dipoles. Table 3 gives the activation energies for thisrelaxation. They are derived from the tg6-plots and checked on plots of the imaginarypart of the electric modulus against frequency at different temperatures. It may benoted from the table that the activation energies increase with increasing Cu2+ contentto reach a constant value for YCu42 and beyond.This value is within experimentalaccuracy equal to the activation energy for the WT conduction..3 I- --- 3 1 3.3 3.5d / E OFIG. 5.-Cole-Cole plot for the YCu68 sample. 0, 393 K ; 0, 414 K.1.2I0.7CQM Ya50.2 1log ( V l WFIG. 6.-tgS (= E”/E’) as a function of the logarithm of the frequency for YCu32. 0, 477 IS; a, 524 K ; a, 578 K; v, 649 K ; 0,713 K.At temperatures below 400 K all the zeolites, except YCu60 exhibit relaxation 11.Its intensity is about an order of magnitude smaller than that of relaxation I, butalso increases with increasing temperature. At a given temperature, its intensity isinversely proportional to the Cu2+ content (except for YCulOO). This is shown i178 CU~+-MOBILITY IN ZEOLITEStable 3 together with the corresponding activation energies where they could bedetermined. At a given temperature the critical frequencies increase with increasingCu2+-content. X-type zeolites also have relaxation 11, but due to its overlappingwith relaxation I it could not be characterized accurately.For a XCu zeolite with 16Cu2+ ions per unit cell Lohse et aZ.I2 reported a relaxation similar to our relaxation 11.The intensities were identical but the authors l2 did not report packing fractions.This makes comparison impossible. Their activation energy was 39 kJ mol-’. TheCole-Cole plots of fig. 5 give an idea of their relative intensities, their position on thefrequency scale at a given temperature and distribution of relaxation times.TABLE 3.-ACTIVATION ENERGIES FOR RELAXATION PROCESSES 1 AND IIE/kJ mol-l&-lOe“ at critical frequency ofsample I I1 relaxation I1 at 400 KYCul4YCu32YCu42YCu60YCu68YCulOoYCuLaxcu64XCulOO74.083113114100115133--0.3050.1700.1300.1 100.245----The samples YCul4 and XCu64 have a supplementary relaxation process (111) atthe low frequency side of relaxation I (fig.7). It appeared only at 650+20 K, butwas also characterized by a temperature dependent intensity and a large distributionof relaxation times. Its tg6-value at a given temperature was at least 2 times largerthan that of relaxation I. Finally XCulOO did not clearly show any relaxation effects.Probably they were hidden in the background of the high d.c.conductivity of thissample (fig. 1).1% ( V l H Z )FIG. 7.-tg& (= E”)IE’) as a function of the logarithm of the frequency for the YCul4 sample. A,728K; 0, 610K; 0, 560K.DISCUSSIONnumber and kind of exchangeable cations present.D.c. conduction as well as. the dielectric relaxation processes depend on theBoth phenomena are ionic as iR. A . SCHOONHEYDT A N D F. VELGHE 179the case for dehydrated Ca2+ zeolites and zeolites exchanged with monovalentcati0ns.l The distribution of exchangeable cations among the various exchangesites is used as the basis for the interpretation of the data. The reversibility of theelectric measurements shows that after a suitable O2 treatment no measurable reduc-tion of Cu2+ or structural changes take place under vacuum, i.e., in complete absenceof water.Nevertheless the lattice stability decreases with increasing Cu2+ content,as has already been observed by Bremer.14 The crystal structures of YCulOO andXCulOO samples particularly collapse at low temperatures. This may be due to thepresence of an appreciable number of protons in these Moreover, thepresence of a supplementary water loss peak at -610 K in the d.t.a. curves is strongevidence that the excess Cu is present as a hydroxylated phase. As this excess doesnot influence the ionic conduction and ionic relaxation, it must be present on theexternal surfaces of the microcrystals. Boreskov l6 found, for similar samples, acatalytic activity intermediate between that of CuO and a stoichiometric Cu2+ zeolite.IONIC CONDUCTIVITYThe presence of two conduction mechanisms above an exchange level of 42 %and one mechanism below that exchange level parallels the presence or absence ofCu2+ ions in the supercages, as determined by X-ray diffra~tion.~ Thus, below12-14 Cu2+ ions per unit cell all these ions are on sites I and 1’.Only one conductionmechanism is observed, due to Na+ ions in the supercages. Above that exchangelevel Cu2+ appears on sites 11. This gives rise to the HT conduction mechanism,whereas Na+ determines the LT conduction mechanism. Therefore these data con-firm our interpretation of conduction data for synthetic faujasites exchanged withCa2+ and monovalent cations in that only supercage cations are responsible for theionic conduction process.1* 2 p l3Na+ CONDUCTIVITYOwing to the preference of Cu2+ for sites in the small cavities, the positive chargedensity in these cages increases with respect to that of the dehydrated YNa zeolite.The effect is an increase of the potential energy of the Na+ ions on sites 11.As theactivation energy for migration is determined by the electrostatic attraction cation-lattice, l2 it decreases concomitantly. This conclusion agrees with that for Ca2+systems and fig. 8 compares the two. The effect of the variation of the small cagecharge density with respect to that of Na+ saturated zeolites is more pronounced forCu2+ than for Ca2+. This is in line with X-ray results which show that a Ca2+saturated zeolite contains 2.6 Ca2+ ions per unit cell on sites I’ and 14.2 on sites I,3whereas for Cu2+ these numbers are respectively 1 1.1 and 3.2.4 Thus, the facts thatthe distance 1-11 exceeds l’-II and that cations on I are shielded more effectively fromtheir environment by the 12 oxygen atoms of the hexagonal prism than cations on I’is reflected in the different slopes of the straight lines of fig.8.We conclude that the cation sites are not independent of each other. This hasalready been observed for the couple I, 1’,17 but is extended now to the sites in thesupercages.Fig. 8 allows an estimate to be made of the Cu2+ + Na+ content of the small cagesfor the samples YCu14, YCu32 and YCu42. The results are shown in table 4. Theinaccuracies both in the experimental activation energies (k 10 kJ) and in the numberof cations in the small cavities l8 make the results only qualitative.The movement of Na+ is not independent of the presence of other cations in thesupercages, as shown by the drastic diminution in the activation entropy (fig.4)t SO CU~+-MOBILITY IN ZEOLITESTheir motion is highly correlated with the presence of other ions and this may explainthe large negative AS values for Cu2+ and Ca2+ systems with divalent cations in thesupercages.cU2+ CONDUCTIVITYOn Y type zeolites different preparation methods for the samples give differentactivation energies for Cu2+ conduction. As the non-stoichiometric zeolites containan appreciable number of protons l 5 and excess Cu ions on the external surface, thedata can be rationalized in that the presence of protons, probably on O1 type oxygen(N-cM2+)/(N- CNa+)FIG.8.-The difference of the activation energy for Na+ conductivity in zeolites, partially exchangedwith divalent cations and Na+ saturated zeolites plotted against the relative change in negativecharge density in the supercages. N is the number of negative charges per unit cell. CMZ+ andCN~+ are respectively the positive charges in the sodalite units and hexagonal prisms, due to Cut+ orCaz+ and Na+. 0, Ca2+; A, Cu2+.TABLE 4.-cATION DISTRIBUTION IN DEHYDRATED, PARTIALLY cU2+ EXCHANGED ZEOLITEScations per unit cellsmall cages supercagessample cus+ Na+ cuz+ Na+ H’YCul4 3.4 18.1 - 19.1YCu32 8.8 9.9 - 25.6YCu42 11.8 4.3 - 27.2YCu60 1 1 - 1 4 - 3-6 1 5 .6 5.5YCU68 11-14 - 7 - 1 0 19.4XCu64 1 1 - 1 4 - 13.5-16.5 12.9 1 9 . 0atoms in the s~percages,~~ decrease the activation energy for migration of Cu2+ ions.We anticipate that protons and water molecules will have an enhancing effect on therate of reduction of CuZ+. The activation energy for Cu2+ conduction in X typezeolites resembles that in YCu68, although they also contain protons. We suggestthat the presence of Na+ or Cu2+ on sites 111’ near O1 type oxygen can explain thR. A . SCHOOMHEYDT AND F. VELGHE 181observed values. However, experiments on stoichiometric X-type zeolites are neededto confirm or reject this hypothesis.Independent of these aspects of the Cu2+ conductivity, it is striking that Cu2+ inthe supercages has a measurable mobility below 750 K, but Ca2+ does not.’.Atfirst sight, this is in contradiction with their ionic radii (0.72 A for Cu2+ and 0.99 Afor Ca2+ zO) and the electrostatic attraction between cation and site hypothesis.13However, the effect of increased charge density in the small cages and especially inthe sodalite units is more pronounced for Cu2+ than for Caz+.We therefore expect a higher potential energy for Cu2+ on sites 11 than for Ca2+and thus a greater mobility. Nevertheless, Freeman and Stamires 21 and Topchievaand co-workers 22 could measure a Ca2+ mobility in X and Y type zeolites with activa-tion energies in the range 88-107 kJ mol-l. The contradiction is only apparent.Indeed, the special pretreatments used by these authors lead to considerable ‘‘ deep-bed ” effects.lg* 23 Oxides of Ca2+ and A13+ are formed in the sodalite cages of the0form +Ca ’%a+ and 2+Al’ ‘A12+.The cation content of the sodaiite cagesincreases to values comparable to those found for Cu2+ systems. The result is thatCa2+ in the supercages acquires a measurable mobility.It seems that the mobility of divalent cations in the supercages is a sensitivefunction of the cation composition of the small cages.RELAXATIONSThe relaxation processes (1-111) have similar characteristics (temperature dependentintensities, broad distributions of relaxation times). This suggests that they have acommon origin. The only physical process which could cause these polarizations isthe migration of the exchangeable cations.The Cu2+-exchanged zeolites are heterogeneous on 3 levels : at the electrodes, atthe interfaces between the microcrystals and at the level of the cavities in the structure.When an ionic conductivity exists these heterogeneities give rise to electrode polariza-tions, Maxwell-Wagner effects and ionic relaxations.The fact that relaxation I11only occurs in the samples with the highest conductivity, YCul4 and XCu64, suggeststhat it is an electrode polarization or a Maxwell-Wagner effect. However, theMaxwell-Wagner theory predicts critical frequencies which fall outside the experi-mental frequency range on the high frequency side, according to the formulav = “4+(-)&:] 23-Q80 1-Q(go = 8.85 x 10-l2 F m-l, 85 and E; are the electrical permittivities of the zeolite andthe empty space between the microcrystals and Q is the volume fraction of zeolitein the pellet).As this expression gives critical frequencies for Maxwell-Wagner effectsin agreement with experiment 11 we think relaxation 111 is an electrode polarization.Relaxation I is assigned to the exchangeable cations in the sodalite cages. Thefollowing arguments favour this assignment.(i) The constant occupancy of the sites in the small cages for exchange levels of42 % and higher is reflected in a constant intensity of the relaxation process.(ii) This intensity is significantly higher than that for the corresponding Ca2+systems, in agreement with the different occupancy of sites I’ by these ions.24(iii) The activation energy for relaxation is equal to the activation energy forconduction by Cu2+ ions in agreement with the fact that only Cu2+ ions are in th182 CU~+-MOBILITY IN ZEOLITESsodalite cages.When Cu2+ and Na+ are present on sites 1’, the activation energy issomewhere between that of Na+ relaxation 24 and Cu2+ relaxation, depending on therelative numbers of the two ions.These activation energies indicate that a real migration of cations takes place.This is only possible in the sodalite cages and not in the hexagonal prisms. Moreover,one activation energy for systems with Na+ and Cu2+ implies a correlated motion,i.e., Na+ and Cu2+ do not move independently within a sodalite cage.This relaxation wasalso observed for Ca2+ systems but with much lower l2 This and thefact that its intensity is inversely proportional to the Cu2+ content suggests thatprocess I1 is due to Na+ on the interstitial sites 111’.Only a couple of Na+ ions perunit cell are involved. However, this number is definitively higher for Cu2+ than forCa2+ systems. Indeed, the higher Cuz+ charge density in the sodalite units increasesthe potential energy of site I1 cations. Again, the affinity of the cations for sites I1decreases and the probability of finding cations on interstitial sites 111‘ becomessubstantial according to Boltzman’s distribution law. Between two neighbouringsites I1 up to 6 potential energy minima can be vis~alised.~~ The motion of Na+over these minima between 2 occupied sites 11, acting as impenetrable boundaries, isresponsible for the ionic relaxation 11.Relaxation I1 is identical to that observed by Lohse et al.MODEL OF IONIC RELAXATIONS I N ZEOLITESThe ionic relaxation processes observed, 1-111, come about when the motion of thecations is hindered by impenetrable boundaries.If these boundaries are the blockinggold electrodes and the external surface of the microcrystals one speaks of electrodepolarization and Maxwell-Wagner effects respectively. If these boundaries are thesodalite cages or two neighbouring, occupied sites 11, one has the ionic relaxationsI and 11.With relaxation I the dipole is formed between the cation and the negativelycharged sites, to which it can jump. As there are more empty sites than occupiedsites and as the motion of the cations in the sodalite cages is correlated, it is impossibleto assign a dipole moment to a specific cation and a specific site.One can assigna dipole moment to the sodalite cage as a whole. Its magnitude and directionreflects the charge and the number of cations present. Every jump within thesodalite cage changes the direction of the dipole moment. In the presence ofan electric field a net migration of the cations in the field direction occurs, i.e., thedipole moments of the sodalite cages tend to align themselves. Thus, in a suitabletemperature and frequency range dipolar relaxation will occur.It can be seen qualitatively that, when there is no occupancy of the sodalite cagesor when they are fully occupied, there is no net dipole moment. Between theseextremes there will be a number of cations per sodalite cage which will give themaximum dipole moment.This explains the difference in intensity of relaxation Iin Cu2+ exchanged and Ca2+ exchanged zeolites. Moreover, the potential energy ofthe sites depends on the local environment. This environment is different €or differentsites but also for the same sites within one cavity. The result is that the local con-ductivity, i.e., the conductivity per site,26 changes from site to site, giving a distributionof relaxation times. Such an ionic motion in the sodalite cages implies that below acharacteristic temperature, i.e., the temperature of maximum intensity, the cations areconfined to this sodalite unit and do not move into neighbouring units. At highertemperatures the latter process becomes possible and the cations do not distinguishbetween sites in the supercages and in the sodalite cagesR. A .SCHOONHEYDT AND F. VELGHE 183This picture is also valid for relaxation 11. At low temperature the cations onsites I1 are immobile with respect to the cations on the interstitial sites 111’. Thus,the latter move over six slightly different potential wells 2 5 between two occupiedsites 11, resulting in a relaxation process with a distribution of relaxation times.CONCLUSIONBoth ionic conduction and ionic relaxation are produced by the migration ofcations. When the cation jumps are confined to the sodalite cages or, at low tempera-tures, to the interstitial sites, ionic relaxation processes occur. The activation energiesof these processes equal the activation energies for conduction.The heterogeneityamong the potential energies of these geometrically identical sites causes a distributionof relaxation times. The migration of the supercage cations produces ionic conduc-tion. The activation energy of conduction by Na+ ions is dependent on the cationcharge density in the small cages and especially in the sodalite units. All the Cu2+ions above the level 11-14 Cu2+/u.c. are located in the supercages and give ionicconduction with an activation energy of 120 kJ mol-l, decreasing in the presence ofprotons in Y type zeolites. Excess Cu is present as a hydroxylated species on theexternal surfaces of the microcrystals and does not intervene in the ionic processesstudied.W. De Wilde and R.A. Schoonheydt, Proc. 3rd Znt. ConJ Moleculur Sieues, ed. J. B. Uytter-hoeven (Leuven University Press, 1973), p. 129.R. A. Schoonheydt and W. De Wilde, J.C.S. Furaduy Z, 1974,70,2132.J. V. Smith, Adv. Chem. Series, 1971, 101, 171.P. GalIezot, Y . Ben Taarit and B. Imelik, J. Catalysis, 1972, 26, 295.N. N. Tikhomirova, I. V. Nikolaeva, V. V. Demkin, E. N. Rosolovskaya and K. V. Topchieva,J. Catalysis, 1973, 29, 500.I. D. Mikheikin, G. M. Zhidomirov and V. B. Kazanskii, Upspekhi Khim., 1972, 41, 909;V. B. Kazanskii and I. D. Mikheikin, Zzvest. Otdel. Khim. Nuuki, Bulgarian Akad. Nauk,1973, 6, 361. ’ J. Turkevich, Y. Ono and J. Soria, J. Catalysis, 1972, 25,44.* J. C. Vedrine, E. G. Derouane and Y. Ben Taarit, J. Phys. Chem., 1974,78, 531.N. M. Kuzmenko and V. I. Lygin, Proc. 3rd Int. Conf. Molecular Sieues, ed. J. B. Uytterhoeven(Leuven University Press, 1973), p. 347.A. Chapoton, Ph.D. Thesis (UniversitC des Sciences et Techniques de Lille, 1973).lo T. A. Egerton and F. S. Stone, J.C.S. Furuday Z, 1973, 69,22.l2 U. Lohse, H. Stach, M. Hollnagel and W. Schirmer, Munutsh., 1970, 12, 819.l3 F. J. Jansen and R. A. Schoonheydt, J.C.S. Faraduy I, 1973, 69, 1338.14H. Bremer, R. Schodel and F. Vogt, Z. Chem., 1972, 12, 423; H. Bremer, W. Morke, R.Schodel and F. Vogt, Adv. Chem. Series, 1973, 121, 249.l5 R. A. Schoonheydt, L. J. Vandamme, P. A. Jacobs and J. B. Utterhoeven, submitted toJ. Catalysisl6 G. K. Boreskov, N. N. Bobrov, N. G. Maksimov, V. F. Anufrienko, K. G. Ione and N. A.Shestakova, Dokludy Akad. Nauk S.S.S.R., 1971, 201, 887; G. K. Boreskov, Proc. 5th Int.Congr. Catalysis, ed. J. Hightower (North Holland, 1973), vol. 2, 981.l7 W. J. Mortier and H. J. Bosmans, J. Phys. Chem., 1971,75,3327 ; W. J. Mortier, H. J. Bosmansand J. B. Uytterhoeven, J. Phys. Chem., 1972, 76, 650.l 8 W. J. Mortier, Acta Cryst., 1973, A29,473.l9 P. A. Jacobs and J. B. Uytterhoeven, J.C.S. Furaday I, 1973,69, 373.2o L. Pauling, The Nature ofthe Chemical Bond (Cornell, Ithaca, 3rd edn., 1960), chap. 13, p. 514.21 D. C. Freeman Jr. and D. N. Stamires, J. Chem. Phys., 1961, 35, 799.22 K. V. Topchieva, I. F. Maskovskaya and V. Ya. Stetsenko, Russ. J. Phys. Chem., 1963, 37,23 G. T. Kerr, J. Catalysis, 1969, 15, 200; Adv. Chem. Series, 1973, 121, 219.24 U. Lohse, H. Stach and W. S c h e r , Z. phys. Chem. (Leigzig), 1973,254, 59.25 H. Lechert, Ber. Bunsenges. phys. Chem., 1973,77, 697.26 F. S. Howell, R. A. Bose, P. B. Macedo and C. T. Moynihan, J. Phys. Chem., 1974,78,639.1020.(PAPER 5/41

ISSN:0300-9599    

DOI:10.1039/F19767200172

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Aqueous Solutions containing Amino-acids and PeptidesPart 2.-Gibbs Function and Enthalpy Behaviour of the Systems Urea + Glycine,Urea + a-Alanine, Urea + a-Aminobutyric Acid and Urea + Glycylglycine at 298.15 KBY TERENCE H. LILLEY* AND R. P. SCOTT-/-Chemistry Department, The University, Sheffield S3 7HFReceived 14th April, 1975The osmotic coefficients and enthalpies of mixing of aqueous solutions of urea+ glycine, urea+a-alanine, urea + cr-aminobutyric acid and urea +glycylgIycine and the enthalpy of dilution of gly-cylglycine have been determined at 298.15 K. The experimental results are discussed in terms ofthe McMillan-Mayer theory of solutions.This is the second in a series of papers in which the thermodynamic propertiesof aqueous solutions containing aminoacids and some small peptides are investigated.Our long term objective is to obtain information on the mechanism of protein de-naturation which would seem to be a problem in non-covalent bonding interactions.The current objective is to gain some understanding of solute-solute interactions onsmaller and structurally less complex species in the hope that a greater knowledgeof these systems might lead eventually to an understanding of the protein problem.There is a considerable amount of information on aqueous solutions containingnon-electrolytes and electrolytes but thermodynamic data on aqueous solutionscontaining single and mixed non-electrolytes are relatively rare and the molecularbasis for the properties is little understood.Several authors have proposed modelsto explain the observed thermodynamic properties for a few systems in terms ofsolute-solute association 2* and solute with some success, but generallythe behaviour of such systems is not understood.A major advance in the under-standing of solutions containing a single non-electrolytic solute was made by Knight,Kozak and Kauzmann.’ They applied the exact theory of solutions proposed byMcMillan and Mayer to all of the systems at that time studied.In this paper an investigation of the osmotic and enthalpy behaviour of binarysolutions containing urea and the amino-acids, glycine, a-alanine and a-aminobutyricacid and the dipeptide glycylglycine is described. The results obtained are discussedin terms of the McMillan-Mayer approach.EXPERIMENTALISOPIESTIC APPARATUSThe apparatus used was basically that previously de~cribed.~ The aluminium block wasmodified so that better contact between it and the isopiestic vessels was obtained.CALORIMETEREnthalpy change measurements were performed using an L.K.B.10700-2 batch calori-meter. In the enthalpy of mixing experiments equimolal solutions of the two solutes undert present address : Chemical Defence Establishment, Porton Down, Salisbury.$ References are given in ref. (1) and in the following paper.18T. H . LILLEY AND R. P . SCOTT 185investigation were mixed in various mass ratios. The enthalpy of dilution measurementson glycylglycine were performed by mixing known masses of a 1 molal solution with knownmasses of water.All of the experiments were made at 298.15 K.MATERIALSThe method of purification of water has been given KCI was purified from theAnalaR reagent using the method of Pinching and Bates.'' The other materials used wereof the highest purity commercially available (all were stated by the manufacturers to be atleast 99.9 % pure) and were dried for several days over P205 before use.RESULTSThe molalities of the solutions in isopiestic equilibrium with aqueous MCl solutionsare given in table 1.The enthalpies of mixing of aqueous urea solutions with aqueousamino-acid solutions are given in table 2 and the enthalpies of dilution of aqueous 1mold glycylglycine are presented in table 3.TABLE 1 .-MOLALITIES OF SOLUTIONS IN ISOPIESTIC EQUILIBRIUMrn(KCl)*/mol kg-1 m/mol kg-1 Yt Y t m(KCI)* /mol kg-1 4% lmol kg- 1urea 4- glycine1.378 1 2.802 92.744 42.707 12.700 92.693 92.690 02.709 90.773 9 1.523 61.500 21.483 81.476 11.473 31.463 71.462 6urea + a-alanine0.913 0 1.618 21.652 81.668 71.682 11.689 71.716 11.737 60.639 6 1.147 41.158 21.166 81.173 81.181 01.190 81.199 300.200 50.394 00.492 90.627 S0.727 5100.200 50.394 00.492 90.627 80.727 5100.212 20.418 00.521 70.620 80.813 2100.208 90.412 00.515 80.611 90.809 21urea + a-aminobutyric acid0.865 3 1.478 41.513 21.543 51,569 21.593 81.614 31.628 91.647 20.550 4 0.959 70.982 00.994 31.001 71.008 S1.020 31.027 7urea + glycylglyci ne0.778 4 1.621 81.592 11.552 21.534 71.515 41.486 81.470 70.547 0 1.129 31.105 81.074 91.059 81.047 91.031 31.023 400.143 40.283 60.414 80.544 20.665 90.781 0100.212 80.422 10.516 G0.623 20.815 3100.188 50.388 30.489 80.590 60.795 0100.184 80.380 00.483 10.577 90.791 21* m(KC1) is the molality of a KC1 solution in equilibrium with the solutions investigated.-f y isthe molality fraction of urea in the solution = murca/m186 AQUEOUS SOLUTIONS OF AMINO-ACIDSTHERMODYNAMIC FORMALISMThe Gibbs function for a system containing solutes and 1 kg of solvent is given by(1) G = Gid+ Gexandi 1TABLE 2.-ENTHALPIES OF MTXING OF UREA AND AMINO-ACID SOLUTIONS AT 298.15 KY - A P l x / J kg-1 Y -AHmir/J kg-11 molal urea+ 1 molal glycine0.130 5 17.2 0.106 5 51.60.250 5 30.5 0.260 9 97.00.497 8 38.6 0.505 0 117.70.743 4 29.6 0.753 6 92.80.904 3 14.5 0.901 0 40.41 molal urea+ 1 molal a-aminobutyric acid2 molal urea+ 2 moIal glycine 1 molal urea+ 1 molal glycylglycine0.1 14 3 65.1 0.103 40.246 1 110.7 0.254 40.498 0 162.1 0.500 70.886 4 68.2 0.755 10.911 21 molal urea+ 1 molal a-alanine41.499.6114.489.138.00.102 10.270 30.507 60.757 40.894 642.181.8111.576.844.5TABLE 3.-ENTHALPIES OF DILUTION OF 1 MOLAL GLYCYLGLYCINE AT 298.15 Kmolality after dilution/ AHd""]J kg-1mol kg-10.102 3 86.40.237 6 177.20.481 7 227.00.768 2 152.40.902 9 71.2Ideality is defined on the molal scale for each solute species such that for any speciesThe corresponding definition of ideality for the solvent isThe excess specific Gibbs function of the solvent may be expressed l1pid = pP+RT In milme; rn + 0.(3)G:d = GP-RTm. (4)GrlRT = ~ ( 1 - 4 ) (5T. H . LILLEY AND R . P . SCOTT 187andIt is assumed that Gex may be expressed as a polynomial in the molalities of thesolute speciesConsequently using eqn (9, (6) and (7)If eqn (8) is applied to a binary solute mixture and it is assumed that the higherorder terms may be neglected, thenWhen yi = 1, y j = 0 and when yi = 0, y j = 1, so the expressions for the osmoticcoefficients of the single solutes areGexIRT = ZC,Xjgijm,mj + ZiZj&gijkmimjmk + higher order terms.(7)(8) 4 = 1 + mXiZjgijyiyj + 2m2&~j&gijkyiYjYk +higher order terms.4(g) = 1 + (gity: +&,jyiyj +gj,Y~)~~z+2(giiiy,3 + 3giij~i?yj + 3gijj~iUj” +gjjj$)*Z2* (9)#(i) = 1 + g i i ~ ~ + 2 g i i @ ~ ’ (10)4(j) = 1 +gjj~z+2gjj~z~*The coefficients in eqn (10) may be obtained from appropriate analysis of the singlesolute experimental osmotic coefficients. The cross terms gij, giij and gijj may beobtained by combination of osmotic coefficient data on binary solute mixtures withdata on single solute mixtures. Appropriate manipulation of eqn (9) and (10) giveswhere 4(i) now denotes the osmotic coefficient of a solution containing only solute iat molality mi.The excess enthalpy may be obtained from eqn (7)m+(ij)-mi$(i) - mj#(j) = 2gijmimj + @ i i jmi +gijjmj)mim j (11)a( GeX/RT)= ( a(l/T) >,,= hijmirnj+C hijkinirnjrnk+ higher order terms.(12)i j i j kIf a solution of a single solute i at molality m containing y kg of solvent is mixedwith a solution of a single solute j at the same molality but containing (1 - y ) kg ofsolvent, then using equation (12) the enthalpy of mixing can be shown to be given byThe enthalpy change on diluting a solution containing (1 -y) kg of solvent and asingle solute i of molality m, with y kg of pure solvent is-AHdiln*/R = m2y(l -y)(hit+m(y-2)hiit) +higher order terms. (14)Expressions for the entropy changes accompanying the above processes are easilyobtainable.AHmix/R = m2y( 1 -y)(2hi - hit - hjj) +higher order terms. (1 3)SOLUTIONS CONTAINING A SINGLE SOLUTEThe principal object of the present work was to obtain the terms in eqn (7) and (12)which contain information on the interaction of unlike solute species.Before thiscould be achieved it was necessary to have information on the terms which arise formlike-like solute interactions188 AQUEOUS SOLUTIONS OF AMINO-ACIDSEllerton et uf. l 2 have investigated the aqueous single solute systems glycine,a-aminobutyric acid and glycylglycine at 298.15 K using the isopiestic method. Theparameters presented by these workers were used in the present analysis of the binarysolute systems.The aqueous a-alanine system has been investigated by Smith and Smith l 3 butthe osmotic coefficients given by them are in error because of the incorrect valuesadopted for the osmotic coefficients of the sucrose reference solutions.We havereanalysed the osmotic coefficients for the sucrose +water system using the resultsobtained by Scatchard et a1.14 and by Robinson and Stokes.15 A slight differenceis apparent between the results obtained from these two sources but all of the resultswere combined and the osmotic coefficients fitted to an equation of the form ofeqn (10). It was necessary to include four ternis in the osmotic coefficient expansionbecause of the niolality range covered by the experiments. These parameters werethen used to re-analyse the osmotic coefficient data l 3 onthe a-alanine water systemJ . 0 C0.98' 0.960.940.92 ~rnufea /mol kg-'FIG. 1.-The osmotic coefficient of aqueous urea solutions at 298.15 K.0 Scatchard ef al.; (D Dunlop et a!.; this work; A Robinson and Bower.The aqueous urea system has been investigated at 298.15 K using the isopiestictechnique by several workers.14* 169 l7 The data of Scatchard et aZ.14 were re-analysed, since they had used incorrect values for the osmotic coefficients of the KClisopiestic reference solutions. The values given by Robinson and Stokes l 8 for thissalt were used in the re-analysis.Comparison of the various sets of results showedthem to be in fairly good agreement, although some discrepancies are evident. Weare uncertain which set or sets of results are the most reliable and so the osmoticcoefficients for the aqueous urea system, which were used in the analysis of the binarysolute systems, were obtained by interpolation from a large scale graph. The linedrawn through all of the experimental points was our assessment of the " best " lineT.H. LILLEY AND R. P. SCOTT 189The experimental results obtained by others (where necessary after re-analysis) andthe results obtained by us are shown in fig. 1, as is the '' best " line.The enthalpy of dilution data of Gucker et aZ.19* 2o for the aqueous glycine andaqueous urea systems at 298.15 K were used, together with their data analysis, toobtain the enthalpy coefficients for these systems. The results of Robinson et al.for the aqueous a-alanine 21 and a-aminobutyric acid 22 systems at 298.15 K werere-analysed, the enthalpy of dilution being regressed to an equation of the form ofeqn (14).The coefficients resulting from this analysis are given in table 5. Theenthalpy of dilution data given in table 3 for the aqueous glycylglycine system wereanalysed in a similar way to that described above and the coefficients for this systemare included in table 5.TABLE 4.-cOEFFICIENTS DEFINING THE EXCESS GIBBS FUNCTION FOR AQUEOUS SOLUTlONS AT298.15 K [SEE EQN (1111I lo3 gij/mol-i kg 1 0 3 gttj/mol-z kgz lo3 gt,,/mol--2 kgzgly cine urea -41.0k2.3 4.2+ 0.3 3.2k0.3a-alanine urea - 31.9k5.5 2.8f 1.3 5.2f 1.2gl ycylgl ycine urea -48.2k5.3 6.0+ 1.2 4.8+ 1.4a-aminobutyric acid urea - 14.0k4.5 -2.2k1.1 - 1.7k1.0SOLUTIONS CONTAINING TWO SOLUTESThe experimental osmotic coefficient results on the binary solute systems investi-gated together with the osmotic coefficients of the appropriate single solute solutionswere fitted, by a least squares procedure, to eqn (9).The results obtained are givenin table 4. There does not appear to be any simple relationship between the grjcoefficients and the corresponding gti and gij coefficients. The systems urea+a-aminobutyric acid and urea + glycylglycine have been investigated previously 9 24 usingthe isopiestic method. The methods used to analyse the experimental results weresimilar to that used here but because of the larger molality ranges in~estigated,~~. 24higher order terms in the equations analogous to eqn (1 l), were included. The agree-ment obtained for the pairwise unlike-unlike species interaction terms (g!!J> with thosegiven in table 4 is within experimental error.The problem of using solubility measure-ments 24* 25 to obtain gij coefficients has been discussed elsewhere.24TABLE 5 .-ENTHALPY COEFFICIENTS FOR AQUEOUS SOLUTIONS CONTAINING A SINGLE NON-ELECTROLYTE AT 298.15 Ksolute h"/K kg mol-1 hitt/K kgz mol-2glycine - 54.0 -a-aminobutyric acid 65.4k0.7 - 1.OkO.8gl ycylgl ycine - 144.0f9 23.04 6urea -43.0 2o --a-alanine 26.0k 0.3 - 0.41 0.4Eqn (13), without inclusion of higher order terms was found to be an adequaterepresentation of the enthalpy of mixing data presented in table 2. The average ofthe quantity AHmiX/nz2y(l -y) was found for each system and the unlike-unlike speciesenthalpy coefficient hij was obtained using eqn (13) and the like-like terms given intable 5.The values of hij are listed in table 7 and included in this table are the valuesobtained for the corresponding entropy coefficients s i j using the results in tables 4and 7190 AQUEOUS SOLUTIONS OF AMINO-ACIDSTABLE 6.-MCMILLAN-MAYER PARAMETERS FOR THE INTERACTION OF AMINOIACLDS WITHUREA AT 298.15 KsoluteLBd I [L&$--LB$(HS)I/i j cm3 mol-1 cm3 mol-1glycine urea 4a-alanine urea 45a-aminobutyric acid urea 91gl ycylgl ycine urea 24- 345- 371- 375-446TABLE 7.-cOEFFICIENTS DEFINING THE EXCESS ENTHALPY AND EXCESS ENTROPY FOR AQUEOUSSOLUTIONS AT 298.15 Ksolutei i --htj/K kg mol-1 --stj/kg mol-1gl ycine useaa-alanine ureaa-aminobutyric acid ureaglycylgl ycine urea5835191210.1540.0850.0500.358DISCUSSIONThe aim of the present work is to investigate intermolecular interactions of solutemolecules in aqueous solutions and it is important to know how these interactionsare manifested in the observed experimental properties.have developed an exact theory of solutions and havedescribed the thermodynamic properties of solutions in terms of certain integrals ofthe radial distribution functions. The osmotic pressure of a solution can be expressedin terms of the number density of solute species asMcMillan and Mayern/kT = pi +c BEp? +(&) 1 Bcpipj +higher order terms.(15)i C i jThe first term on the r.h.s. of eqn (15) represents Van’t Hoff’s law and the otherterms give rise to deviations from “ ideality ”. If we consider solutions containinga single solute i, eqn (15) becomesThe coefficient B i is given by 2 7n/kT = pi + B:p: +higher order terms.(16)[exp( -- Wii/kT) -- l} dr, drb. JvJy B; -- -(1/2V)Higher terms in eqii (16) are related to interactions between three or more solutemolecules in an infinitely dilute solution. In the present work we consider onlypairwise interactions. The coefficients gi obtained from the present data analysisare related to eqn (16) and (17) by 2 9 9 30gii = (2Bz - v?+ b,i)M,/2~? iwwhere bSi is related to solute-solvent interactions and is defined 27 asbSi = (I/v) J (expi- W,,/kT)- 1) dr, dr,. (19)v vIt has been shown 30 thaT. H. LILLEY AND R . P. SCOTT 191and consequently combination of eqn (1 8) and (20) givesgii = (2Bz-2~7+ kT~)M,/2v?.Thus to obtain information on the solute-solute interactions as expressed by eqn (17),from the experimental osmotic coefficients, requires information on the volumetricproperties of the solute and the solvent and the isothermal compressibility of thesolvent. The expression corresponding to unlike-unlike solute coefficients is 30Eqn (21) is similar to an equation given by Knight 31 except for the presence of thecompressibility term and Knight, Kozak and Kauzmann have tabulated thermo-dynamic data compiled from many sources for a variety of aqueous solutions contain-ing a single non-electrolyte.The results presented by these authors for which bothosmotic coefficients and partial molar volumes of the solute at infinite dilution 32existed were treated to obtain the gif terms.Using these Sir terms and the partialmolar volumes at infinite dilution 32 in conjunction with a value 33 of RTrc of 1.1 cm3mol-l for water at 298.15 K allowed the quantity LB; to be calculated. Fig. 2 showsa plot of LB; against the partial molar volume of the solute. There is a very real26I I1 f zoo 400@/crn, mol-FIG. 2.-PIot of LB,*, against Vd* for solutions of a single solute in water, Results at 298.15 K : 1,glycine ; 2, cc-alanine ; 3, ct-aminobutyric acid ; 4, x-aminobutyric acid ; 5, rx-aminoisobutyric acid ;6 , a-aminoisovaleric acid ; 7, 8-alanine ; 8, ,3-aminobutyric acid : 9, P-aminovaleric acid ; 10,y-aminobutyric acid : 11, c-aminocaproic acid ; 12, y-aminovaleric acid ; 13, glycylgIycine ; 14,glycylalanine ; 15, alanyiglycine ; 16, alanylalanine ; 17, triglycine ; 18, serine ; 19, proline ; 20,hydroxyproline ; 21, betaine ; 22, glucose ; 23, fructose ; 24, mannitol ; 25, sucrose ; 26, raffinose ;27, urea; 28, glycolamide; 29, lactamide; 30, glycerol.Results at approximately 273 K : 31,methanol ; 32, ethanol ; 33, propan-1-01 : 34, propan-2-01 ; 35, 2-methylpropan-1-01 ; 36,2-methyI-propan-2-01 ; 37, butan-1-01 ; 38, butan-2-01 ; 39, dioxan ; 40, methylamine ; 41, ethylamine ; 42,dimethylamine ; 43, diethylamine ; 44, urea ; 45, formamide ; 36, acetamide ; 47, urethane ; 48,acetone ; 49, formic acid ; 50, acetic acid ; 51, propionic acid ; 52, butyric acid.difficulty in trying to continue further since the information which is required is howthe intermolecular potential function varies with distance and angle for non-sphericalspecies.The experimentally determined properties only give this information underan integral and without further information an infinite number of models could befitted to each experimental result. There have recently been several discussions 34using a specific model but this model, not withstanding its plausibility, is not unique1 92 AQUEOUS SOLUTIONS OF AMINO-ACIDSnor has it been claimed to be so. The approach adopted here is simpler, ratherdifferent and essentially that used earlier.'If the solutes are represented by " hard-spheres " thenand if the approximation is made thatthen the line drawn in fig. 2 represents the values LB: take for this " hard-sphere "interaction [LR:(HS)].It is apparent from fig. 2 that subject to the assumptionsmade, all of the solute-solute interactions represented are attractive compared toLB:(HS). It has been shown 35* 36 that LB; may be calculated for " hard-core "interactions between molecules having ellipsoidal and " dumb-bell " symmetry andthat the differences between these and LB:(HS) are relatively small if the asymmetryof the species is not too great. The " hard-sphere " line drawn for the moleculesrepresented in fig. 2 is therefore a reasonable approximation. Knight et aL7 havetabulated the quantity [LB: -LBE(HS)] which characterises that part of the virialcoefficient not accounted for by " hard-sphere " interactions. In fig.3 this quantityvi = v yy;8/cm3 mol-1 ~ P l c m 3 mo1-1(4 (b)FIG. 3.--Plot of [LB;,--LB,*,(HS)] against V p for aqueous solutions at 298.15 K, of (a) alkan-1-01s(b) alkanoic acids.is plotted against the partial molar volume of the solute for the homologous seriesof the alkan-1-01s and the alkanoic acids and it is apparent that for both series anapproxiinately linear variation with zero intercept results. It was suggested ' thatthe increasing solute-solute interaction between species containing aliphatic sidechains, as the side chain is lengthened, is a manifestation of hydrophobic bonding 37. 38and head group-head group interactions were presumed to give a constant contribu-tion, within a given series, to [LB; -LBz(HS)]. This explanation seems reasonableif we coiisider a situation where the solute molecules are spherical and we assumethat the solute species interact through a " square-well " intermolecular potential offixed depth which extends over a given fixed distance froin the surface of the solute.Substitution of these assumptions into eqn (17) leads to the result that as the sizeof the solute increases so too does the attractive contribution to LB;.This argumentassumes a great deal about the intermolecular potentials but its conclusions are sup-ported by the fact that the two homologous series represented in fig. 3 have approxi-mately the same slope, suggesting that the contribution from the head groups havelittle erect on the observed interaction.Fig. 4 is an analogous plot to fig.3 for some ct-, p- and y-amino acids.' It iT. H. LILLEY AND R. P. SCOTT 193apparent that the values of LBE-LB:(HS) are similar within a given homologousseries and here it appears that the interactions are predominantly of a dipole-dipolenature and consequently head group effects overwhelm side chain interactions.Similar behaviour is observed for the dipeptides ’ and a similar explanation may beused.400r200-I 3L oO O L0 I00 uI 0 100i I- f I - 0 100@/cm3 mot-1(4 (6) (c)FIG. 4.-Plot of [LB:i-LB?i(HS)] against V? for aqueous solutions, at 298.15 K, of (a) a-, (6) /3- and(c) y-amino-acids.Using eqn (24) and the appropriate volumetric data 32 the values of LB; forurea-amino acid interactions were calculated and are given in table 8.The “ hard-sphere ” contribution for i-j interactions is 30B$(HS) = (n/6)(di + dj)3and if we continue to use our earlier assumption regarding the equalities of theintrinsic and partial molar volumes of solute species then B;(HS) is calculable. Infig. 5 we display the experimentally derived values of LB? against the partial molar5 0 0 ri E“I . I2 0 40 6 0 8 00V? /cm3 mol-1The solid line represents “ hard sphere ” interaction.FIG. 5.-Plot of LB; against V*e. The subscripts i and j refer to amino-acid and urea respectively.volume at infinite dilution of the amino-acid. The line represents the “hard -sphere ”term. Included in table 6 are values of [LB; -LB$(HS)J and as with systems contain-ing single solutes this quantity is always negative and corresponds to solute-soluteattractive forces.For the three a-amino acid +urea systems studied this attractiver-194 AQUEOUS SOLUTIONS OF AMINO-ACIDScontribution is the same, within experimental error, no side chain effects beingobserved. The constancy suggests a predominating interaction of the urea with thehead group of the amino-acids. The attraction is greater for the urea-glycylglycineinteraction and may be due to the greater dipole moment 39 of the dipeptide comparedto the amino-acids although a contribution from the peptide group might wellcontribute to the observed effect.TABLE &-THE VARIOUS TERMS IN EQN (25) FOR AQUEOUS SOLUTIONS CONTAINING UREA,GLYCINE AND UREA+ GLYCINEi i mol-1 kg K-1urea urea - 10.7 540 34 34 1.8glycine glycine - 24.7 637 28 28 1.8urea glycine - 10.5 702 34 28 1.8(dlnVse/dT) and K were taken from ref.(33), the temperature coefficients of the partial molarvolumes of the solutes were estimated from ref. (3), (42) and (43).T12t o9 0.?- 2 1 FIG. 6.-Enthalpy and entropy coefficients for the interaction of i and j species in water. Resultsat 298.15 K : 1, glycine + glycine ; 2, a-alanine + x-alanine ; 3, a-aminobutyric acid + a-amino-butyric acid ; 4, glycylglycine + glycylglycine ; 5, urea + urea ; 6, or-alanine + urea ; 7, glycine + urea ;8, a-aminobutyric acid + urea ; 9, glycylglycine + urea ; 10, or-aminovaleric acid + a-aminovalericacid ; 11, or-aminoisobutyric acid + a-aminoisobutyric acid ; 12, a-aminosiovaleric acid + a-amino-isovaleric acid ; 13, /3-alanine + /3-alanine ; 14, y-aminoisovaleric acid + y-aminoisovaleric acid ;1 5, z-aminocaproic acid + E-aminocaproic acid ; 16, glycolamide + glycolamide ; 17, glycerol + gly-cerol ; 18, sucrose+ sucrose.Results at approximately 273 K : 19, methanol + methanol ; 20,ethanol + ethanol ; 21, propan-1-01 + propan-1-01 ; 22, propan-2-01 + propan-2-01 ; 23, 2-methyl-propan-2-01 + 2-methylpropan-2-01 ; 24, butan-1-01 + butan-1-01 ; 25, butan-2-01 + butan-2-01.Combination of the definition for the enthalpy interaction coefficienT. H . LILLEY AND R. P. SCOTT 195with eqn (22) gives, assuming the temperature coefficient of the isothermal com-pressibility may be neglectedConsequently if one wishes to discuss the temperature variation of the solute-soluteintermolecular interaction the free energy parameter gi is required as are the expansi-bilities of the interacting solutes.There is not a great deal of experimental informa-tion on these quantities and as examples only the single solute aqueous systemscontaining urea and glycine and the binary solute system of urea+glycine in waterwill be discussed to get some estimate of the several terms in eqn (25). (Furtherdiscussion will be given el~ewhere.~~) In table 8 we present the various contributionsto the right-hand side of eqn (25) for these systems. It is apparent for these systemsthat the first term on the right-hand side is small compared to the second term andthat this latter term is dominated by the contribution from (dB$/dT).The relativelysmall contribution from the term containing grj means that entropy-enthalpy com-pensation 41 will occur. This is illustrated in fig. 6 where hi, is plotted against sijfor many systems including those just discussed, and it would seem that generally forsolutes interacting in water gi is small whereas its temperature coefficients are large.The reason for such compensation behaviour is not known although there have beenseveral discussions regarding its origin.41 We prefer not to speculate further on thissince it is apparent that before an acceptable explanation is possible a considerableamount of information is required on the intermolecular potential and its temperaturederivative, of solutes interacting in water.NOTATIONbB*d9GhHkLm423MPrRSintegral defining the interaction ofsolute with solventintegral defining the interactionbetween two solutesdiameter of a solute speciesGibbs function interaction coefficientGibbs function of a solution contain-ing 1 kg of solventen t ha1 p y interact ion coefficiententhalpy of a solution containing 1 kgof solventBoltzmann constantAvogadro constantmolalitytotal molality (sometimes called theTVWY2,471:PPAKtemperaturemolecular volumevolume or partial molar volumepotential of average forcesolute mole fractionosmotic coefficient, $ ( i j ) refers to theosmotic coefficient of a solution con-taining the solutes i andjosmotic pressuresolute number densityisothermal compressibility of solventchemical potentialchange in a propertySUPERSCRIPTS AND SUBSCRIPTSid idealex excess8 standard statei, j,k, solute speciess solventosmolality) = Ximimolecular weightpressuredistance separating two speciesgas constantentropy interaction coefficient =(h/ TI - 9We thank L.K.B.Instruments Ltd. for the loan of the calorimeter and one of us(R. P. S.) thanks the S.R.C. for the award of a Research Studentship196 AQUEOUS SOLUTIONS OF AMINO-ACIDSADDENDUMAfter this paper hzd been prepared Cassel and Wood 4 4 3 45 reported experimental dataon the enthalpies of mixing of urea+glycine aqueous systems at constant molality. Theresults agree, within experimental error, with those obtained here (see table 2).C.C. Briggs, T. H. Lilley, J. Rutherford and S . Woodhead, J. Solution Chem., 1974, 3, 649.J. A. Schellman, Compt. Rend. Trav. Lab. Carlsberg, 1955, 29, 223.R. H. Stokes, Austral. J. Chem., 1967, 20, 2087.R. H. Stokes and R. A. Robinson, J. Phys. Chem., 1966,70,2126.F. Franks, J. R. Ravenhill and D. S . Reid, J. Solution Chem., 1972, 1, 3.M. J. Tait, A. Suggett, F. Franks, S. Ablett and P. A. Quickenden, J. Solution Chem., 1972,1, 131.W. S. Knight, J. J. Kozak and W. Kauzmann, J. Chem. Phys., 1968,48, 675.W. G. McMillan and J. E. Mayer, J. Chem. Phys., 1945, 13,276.C. C. Briggs, R. Charlton and T. H. Lilley, J. Chem. Thermodynamics, 1973, 5, 445.H. L. Friedman, Ionic Solution Theory (Wiley, New York, 1962), pp.194-195.lo G. D. Pinching and R. G. Bates, J. Res. Nut. Bur. Stmd., 1949, 42,419.l 2 H. D. Ellerton, G. Renfelds, D. E. Mulcahy and P. J. Dunlop, J. Phys. Chem., 1964, 68, 398.l 3 E. R. B. Smith and P. K. Smith, J. Biol. Chem., 1937,121, 607.l4 G. Scatchard, W. J. Hamer and S. E. Wood, J. Amer. Chem. SOC., 1938, 60, 3061.l5 R. A. Robinson and R. H. Stokes, J. Phys. Chem., 1961,65,1954.l 6 H. D. Ellerton and P. J. Dunlop, J. Phys. Chem., 1966, 70, 1831.l7 V. E. Bower and R. A. Robinson, J. Phys. Chem., 1963, 67, 1524.R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworth, London, 1965), Appendix8.3.I 9 F. T. Gucker, H. B. Pickard and W. L. Ford, J. Amer. Chem. SOC., 1940, 62,2698.2o F. T. Gucker and H. B. Pickard. J. Amer. Chem. SOC., 1940,62, 1464.21 H. A. Benesi, L. S. Mason and A. L. Robinson, J. Amer. Chem. SOC., 1946,68,1755.22 L. S. Mason and A. L. Robinson, J. Amer. Chem. Soc., 1947, 69, 889.23 E. L. Cussler, J. Phys. Chem., 1967, 71, 901.24 H. Uedaira, Bull. Chem. SOC. Japan, 1972, 45, 3068.2 5 P. L. Whitney and C. Tanford, J. Biol. Chem., 1962, 237, PC1735.26 Y . Nozaki and C. Tanford, J. Biol. Chem., 1963, 238,4074.27 D. Stigter, J. Phys. Chem., 1960, 64, 118.2s D. Stigter, J. Phys. Chem., 1960, 64, 114.29 T. L. Hill, J. Chem. Phys., 1959, 30,93.30 J. E. Garrod and T. M. Herrington, J. Phys. Chem., 1969, 73, 1877.31 W. S. Knight, Thesis (Princeton, 1962).32 J. J. Kozak, Thesis (Princeton, 1965).33 G. S. Kell, J. Chem. Eng. Data, 1967, 12, 67.34 see e.g. H. L. Friedman, in Modern Aspects of Electrochemistry, ed. J. O'M. Bockris and B. E.Conway (Plenum, New York, 1971), vol. 6 ; H. L. Friedman, C. V. Krishnan and C. Jolicczur,Ann. N. Y. Acad. Sci., 1973, 204, 79 ; H. L. Friedman and C. V. Krishnan, J. Solution Chem.,1973,2,119.35 A. Isahara, J. Chem. Phys., 1950,18, 1446.36 A. Isahara, J. Chem. Phys., 1951,19, 397.37 W. Kauzmann, Advances in Protein Chemistry, ed. C. B. Afinson, M. L. Anson, K. Bailey and38 T. S. Sarma and J. C. Ahluwalia, Chem. SOC. Rev., 1973, 2,203.39 J. T. Edsall and J. Wyman, Biophysical Chemistry (Academic Press, New York, 1958), vol. I,40 T. H. Lilley, to be submitted for publication.41 See e.g. R. Lumry and S . Rajender, Biopolymers, 1970, 9, 1125.42 H. J. V. Tyrell and M. Kennerley, J. Chem. Soc. A , 1968, 2724.43 F. T. Gucker, W. L. Ford and C. E. Moser, J . Phys. Chem., 1939,43,153.44 R. B. Cassell and R. H. Wood, J. Phys. Chem., 1974, 78, 2465.45 R. H. Wood, personal communication.J. T. Edsall (Academic Press, New York, 1959), vol. 14.chap. 6.(PAPER 5/708

ISSN:0300-9599    

DOI:10.1039/F19767200184

出版商:RSC    

年代:1976

数据来源: RSC