摘要:
1972 143Unstable Intermediates. Part CV.l Radiolysis of Frozen Aqueous Solu-tions of Alkali-metal Halides : Electron Spin Resonance Spectra for MHand HalOH-By 1. S. Ginns and M. C. R. Symons," Department of Chemistry, The University, Leicester LE1 7RHExposure of glassy aqueous solutions of alkali-metal halides to 6oCo y-rays a t 77 K gave species identified by theire.s.r. spectra as He. MH+, OH, HalOH-, and Hal,-. The effect of gradual annealing is outlined and aspects of thedamage mechanism discussed. The fact that H * is preferentially trapped by the metal cations whilst OH i s trappedby the halide ions i s rationalised in terms of the electronic structures of the resulting radicals.THE main purpose of this and subsequent studies ofaqueous solutions is to investigate reactions between theinitial radiation products for water (et-, He, and *OH)and various substrates, particularly inorganic ions, usinge.s.r. and U.V.spectroscopy. Much attention has re-cently been given to optical studies of fluid, aqueoussolutions immediately following radiation pulses but,in general, it has not been possible to apply the moresearching technique of e.s.r. to probe the nature andstructure of the radical intermediates involved. Thiscan be done, in principle, by using aqueous glasses atsufficiently low temperatures together with subsequentcontrolled heating.In previous studies of this type, care has not alwaysbeen taken to ensure that glassy materials were studied.This is a particularly important factor for aqueous solu-tions because of the strong tendency for water to givepure ice crystals on freezing with consequent phaseseparation.If this occurs then, in general, radiationdamage results in e.s.r. spectra characteristic of theseparate phases, and are not directly relevant to thepresent study. It is often not possible to gauge theoccurrence of such phase separation by direct observ-ation, and we have, therefore, devised various tests todetect phase separations. When these were positive,minimum concentrations of certain additives wereincluded to assist in glass formation. These were chosen,as far as possible, so as to avoid unwanted side reactions.In the present study, we have confined our attentionto the alkali-metal halides. Of the species detected bye.s.r.spectroscopy, the hydrogen atom alkali-metaladducts MH+ have previously been detected in irradiatedbarium sulphate crystals doped with alkali-metal ions2The ion ClOH- was first detected in y-irradiated BaC4,-2H20,3 and BrOH- and IOH- have subsequently beenrep~rted.~ It is significant that none of these specieshas been detected in pulse-radiolysis studies, but kineticevidence for the involvement of ClOH- ions has beenadd~ced.~EXPERIMENTALThe alkali-metal halide salts were Reagent grade andwere used without further purification. Solutions of NaI,NaBr, KC1, and KF were prepared in water which had beendoubly distilled from alkaline permanganate. Solutionconcentrations ranging from ZM to saturation level were1 Part CIV, R.Catterall, I. Hurley, and M. C . R. Symons,2 M. B. D. Bloom, R. S. Eachus, and M. C. R. Svmons. T.preceding paper.prepared. In addition, aqueous solutions of the remainingpotassium, sodium, and lithium halide salts were studied.Samples were also prepared in D20 supplied by Koch-Light(99.7% D).Tests for GEassi~cation.-Samples containing low con-centrations of Mn2+ were frozen in the form of beads andtheir e.s.r. spectra were obtained at 77 K. When phaseseparation occurred, a very broad single-line spectrum forMn2+ was obtained. Samples which formed good glasses,e.g. aqueous solutions of LiI, LiBr, LiC1, and NaC1, gavewell resolved Mn2+ spectra at 77 K.Aqueous solutions were y-irradiated at 77 K using a6oCo Vickrad source (4 Mrad/h) to a total dose of 2 Mrad.All e.s.r.spectra were obtained on a Varian E.3 spectro-meter. Annealing experiments were carried out by allow-ing samples to warm for fixed times and then re-coolingthem to liquid-nitrogen temperature.FIGURE 1 E.s.r. spectra of a FOH- in y-irradiated aqueous KF;b FOD- from y-irradiated KF in D,O-extra features due toFOH- and D* are marked; c MI = -1 feature for the(Kf * D) centre formed in KF-D,O-(i) 0.32 mW, (ii) 5.0 mWRESULTS AND DISCUSSIONResults are summarised in the Tables 1 and 2, andtypical spectra, with an indication of our analyses, aregiven in Figures 1 4 . Only single lines in the normalR. C . Catton and M. C . R. Symons, J . Chem. SOC. ( A ) , 1969,446.I " I. Marov and M. C. R. Symons, J .Chem. SOC. ( A ) , 1971,201.M. Anbar and J. K. Thomas, J . Phys. Chem., 1964, 68, 3829. Chem. SOC. ( A ) , 1971, 833.144 J.C.S. DaltonTABLE 1parameters for the (Na+ - * * H), (K+ * - * D) centresHydrogen-deuterium hyperfinetensor (G) Alkali-metal hyperfine tensor (G)All A L A is0 All Al. A is0507-321.2 14.0 16.4 513.1 510.7 511.521.4 15.1 17.2 613-2 510.7 511.58.7 5.0 6.9 79.2 78.7 78.9(lBF) 62.0 (lBF) 24.6 522.54.0 500a This work. Ref. 2. Ref. 13.E.s.r.Radical &v(Na+ H) in HaOa 2.002(K+ - D) in D,Oa(K+ H) (BaSO,) Q(F- H) (CaFAC 2.002H atoms in gas phase(Na+ - H). (BaSO,) b2.00222.0012.0022.001RadicalFOH-*OH (1IOH-aZOH-81,- aI,-BrOH- aBrOH- 6Br,- aBr,- 8C12-fClOH- aClOH-fc1,- aTABLE 2E.s.r.data for radicals formed in y-irradiated aqueous solutions of alkali-metal halides at 77 Kg-Tensorgll gl2.1 2.0092-06 2.0071.98 2.131.98 2.131-93 2.181.975 2.1752.00 2-082-00 2-081.99 2.091.98 2.092.000 2.042.0027 2.0382.004 2.0172.0054 2.0174c AllA, ('OF) = f 7A, ('H) -0A , ('H) = 0 & 6517480418433(*1Br) 405426469470(35~1) 102( 3 6 ~ 1 ) 100("Cl) 59 [:g\) :;('H) 25.0Hyperfine tensor data (G)Al. AboA, (1eF) = 33 15.8A,('H) = -26 6.4A, (1°F) = f 7A, ('H) = -40A,('H) = -26A, ('H) = -4470 172 b*c70 155120 155120 19350 13750 14285 19485 195( 3 5 ~ 1 ) 10 40.8(35~1) 9 39.9(35~1) 16 9 p:\) fX.4 8.8('H) 24.6Spin densities2B ] u2s % u2p %17.2 ( A 1 > 0) 0.09 1.626.6 ( A 1 < 0) 0.04 2.53453102 6324026828327627661.261.25050.22.42.12.12.61.61-72.32.32.42.40-50.576.270.058.153.054.357.355.555.558.258.248.048.0a This work.b Values corrected for orbital magnetism when necessary as indicated in ref. 4. 0 Refs. 8, 9. d All is alwaysRef. 4. taken to be positive.3 Ref. 3.The sign of A 1 is ambiguous, and reasons for choosing the signs given have been discussed in ref. 4.-1 -2 -3FIGURE 2 E.s.r. spectra of a BrOH- formed in y-irradiated aqueous NaBr-the additional features are due to Br,- and *OH;The main parallel features are indicated in each b Br,- formed in same solution-the central features are obscured by *OH.spectru1972 145region for hydrogen atoms were detected in irradiatedglasses containing Cs+, and no hydrogen atoms were ob-tained from solutions containing ammonium ions.Thefeatures for MH+ ions formed in solutions prepared from90% D20 were narrower and better resolved. Those500 6 ,I 'I l l 1 lo v 1, -2 -3 -& - 5I I- 3 4 -512FIGURE 3 E.s.r. spectra of IOH- and I,- in y-irradiatedaqueous NaI. The parallel features are markedassociated with the M I = 5 1 lines for MD+ ions wereidentical with those for MH+. In general, the lowestpossible power level was used (ca. 0.3 mW).2 1 0 -1 - 2 -3quirement that the solids were glasses. Indeed, identifi-cation of this species would have been impossible had wenot known the regions in which strong features were tobe expected from previous studies of ClOH- in BaCl,,-In contrast, no features for F,- were ever detected,whilst those assigned to FOH- (Figure 1) were alwaysstrong.We do not understand why this difference wasfound.Annealing.-Hydroxyl radicals were lost first in allcases, but we were unable to detect any concomitant in-crease in the concentration of the HalOH- ions becauseof changes in line widths and, in some cases, partial lossof signal strength for these species also. The HalOH-and Hal,- ions were lost at roughly comparable rateswhen they could be studied together. However, theMH+ radicals were lost more readily than the ' normal 'trapped hydrogen atoms in all cases.Electronic Strwtzcre.-The HalOH- ions , in general,resemble the Vg centres in having their unpaired elec-trons in Q* orbitals.Attention has already been drawnto various trends in computed spin-densities for theseiso-structural species.3,4$6 However, our results for thenew radical FOH- are quite different, and the radical is,in many ways, better described as an hydroxyl radical2 ~ ~ 0 . 313200: 25025 G -HFIGURE 4 E.s.r. spectra of a C1,- from y-irradiated KCl in D,O -central features are obscured by *OD spectrum; b ClOH- fromKCl in ~M-KOH. The parallel features are indicated in each spectrumSpectra assigned to C12-, Br2-, and 1,- were all wellresolved (cj. Figures 2-3) and normal in that bothhalogen atoms were strictly equivalent, in contrast withresults sometimes obtained for these centres in crystallinehalidessSpectra for IOH- and BrOH- were better resolved thanthose previously rep~rted,~ and the parameters have beenslightly revised (Table 2).Spectra for ClOH- werealways largely masked by that for C1,- under conditionswhich were vaned as much as possible within the re-weakly interacting with a fluoride ion. A similar de-scription was proposed for the anion F02- which hasbeen identified in irradiated CaF, crystals.Considerable difficulty was experienced in interpretingthe powder e.s.r. spectra, and the interpretation used inFigure 1 is the one that gives the most internally con-sistent set of results, both for FOH- and FOD-. The&I. C. R. Symons, Advances in Chemistvy Series, 1968, 82, 1.H. Bill and R.Lacroix, J . Physique (Suppl.), 1967, 28, CH.138146 J.C.S. Daltonresults can be interpreted in terms of an FOH bondangle of ca. 90". This is consistent with the concept of aweak, long, a bond between oxygen and fluorine, theunpaired electron remaining primarily in an oxygen 29orbital perpendicular to the *OH bond. Thus the protoncoupling remains close to that for hydroxyl radicals inpure ice crystals 899 and the fluorine coupling correspondsto only ca. 2.5% delocalisation. In particular, the formof the anisotropy for the lH and 19F tensors agree withthis model, the AI1(l9F) features coming with the inter-mediate coupling for lH, etc.The form of the g-tensors for FOH- is as predicted,but it is surprising that gz (that is, the g-value along theO-H bond direction) is actually greater than that for' normal ' OH radicals in i ~ e .~ ~ ~ For the latter, orbitalmotion is largely quenched by asymmetric hydrogenbonding, (I), whilst for the former, the fluoride ion mustF' 0 0 ,;O-H ,;O-H(11 (II)also be considered. The effect of hydrogen bonding as in(I) is to constrain the filled P(n) orbital towards theprotons, leaving the unpaired electron as remote aspossible from these. The fluoride ion is surely expectedto enhance this trend, thus increasing the energy gapbetween the $(x) levels and decreasing Ag!l. In fact,the opposite occurs-this may be taken as evidence forweaker hydrogen bonding in (11), possibly because theparent water molecule hydrogen bonded to fluoride isless closely bound into the water network.The fact that the [MH+] can approximately equal oreven exceed that of ' normal ' trapped hydrogen atomsdespite the fact that [H,O] > [M+] strongly suggeststhat these units are energetically preferred.However,annealing studies show that they are destroyed atappreciably lower temperatures, as was found in earlierwork on these species in barium sulphate crystals.2Also ab ifinitio calculations lo suggest that the unit NaH+is dissociative. An alternative explanation is that thecations provide the best trapping sites because of theireffect on water structure. Normal ice is very closelyknit and, in fact, hydrogen atoms are not trapped thereinat 77 K. The interaction between cations such as Na+and water is weaker than that between water and water,lland also there is likely to be a somewhat disorderedregion of water surrounding the cations which canprovide cavities for the hydrogen atoms.It is signi-ficant, in this context, that Cs+ gave no noticeable* J. A. Brivati, M. C. R. Symons, D. J. A. Tinling, H. W.Wardale, and D. 0. Williams, Truns. Faraduy SOC., 1967, 63,2112.9 J. A. Brivati, M. C. R. Symons, D. J. A. Tinling, and D. 0.Williams, J . Chem. SOC. ( A ) , 1969, 719.10 T. A. Claxton and D. McWilliams, Trans. Furaduy SOL,1970, 66, 513.interaction with hydrogen atoms. This may be becausethe orbitals are too disparate to interact strongly, butwe recall that C0,- radicals interact with Cs+ just aseffectively as with the smaller cations.12 Alternatively,the organisation of water around the Cs+ ions may be lessconducive to trapping.We consider the fact that hydrogen atoms are pre-ferentially trapped at cation sites and hydroxyl radicalsat anion sites to be especially significant, The bondingsituations for these centres and the alternatives in whichH* and *OH are trapped a t anion and cation sites re-spectively are compared in Figure 5.The unpaired elec-tron in the anion centres is predominantly in a o*-orbitalwhilst that in the cation centres is in a a-orbital. Thus inthe former there is electron transfer from the anion tothe radical and in the latter, from the radical to thecation. This is, we suggest, the main reason for theobserved selectivity since hydroxyl radicals are definitelyelectrophilic whilst hydrogen atoms are not.(Unitssuch as NaOH+ could be present, but we would haveexpected to detect hyperfine coupling to 23Na.)We have previously drawn attention to the remarkablefact that although there is appreciable delocalisation ofthe unpaired electron onto the cation, the proton hyperfinecoupling for MH+ radicals is slightly greater than that forthe free atoms2 The most easily understandable ex-planation is that the radial extension of the 1s orbitalfor the hydrogen atom is strongly dependent upon theenvironment, and the hyperfine coupling is very sensitive(A)2P 4 2 J - 3 s 2 2 2 4 7 - 3 s F H Na F OH NaFIGURE 5 Bonding schemes for Ha a ; and .OH b where H. and.OH are trapped at cation (Na+) or anion (F-) sitesto small changes in the former and hence also in thelatter., The effect of the positive charge was thenpostulated to result in a small reduction in the radialextension and hence in a compensating increase inIt is interesting to compare our results with those forhydrogen atoms trapped in calcium fluoride ~rysta1s.l~Here also, A,(lH) is slightly greater than that for thefree atoms despite the fact that there is apparent de-localisation onto eight equivalent fluorine atoms.Theusual model for these centres is that the hydrogen atomsare trapped interstitially in sites that are surrounded byeight fluoride ions. If the hyperfine coupling (also givenin Table 1) to 19E' is taken to be positive, and allowance ismade for the direct dipole coupling from spin on hydro-11 R.N. Butler and M. C. R. Symons, Trans. Furuduy SOC.,1969, 65, 2569.l2 R. N. Butler and M. C. R. Symons, Truns. Furuduy SOC.,1969, 65, 945; J. H. Sharp and M. C. R. Symons, J . Chem. SOC.( A ) , 1970, 3075.13 J. L. Hall and R. T. Schumacher, Phys. Rev., 1962, 12'7,1892.Aiso(lH)1972gen (ca. 2 G), then the apparent spindensity on eachfluorine is ca. 2%, giving ca. 16% delocalisation in total.If interaction with each fluoride is as pictured in Figure5 (this might involve rapid migration via HF- units, forexample) then some electron transfer towards hydrogenmust occur and we would have predicted an orbitalexpansion and hence an exaggerated decrease in the lHhyperfine coupling.This may be outweighed by otherfactors, or just possibly, it may arise because the trappingsite is at a calcium ion rather than a cavity. In thatcase, coupling to 19F may be less direct, possibly involvingspin polarisation rather than a-delocalisation. It wouldbe interesting to search for coupling to %a in thesecentres.Mechanism of Formation.-Damage to water is thoughtto involve mainly electron loss (1) followed by electrontrapping (hydration) (Z), and reaction to give hydrogenatoms (3) and (4).H20 + y _+ H20+ + e- (1)(2)(3)(4)e- + cavity --+ et-H20f + H20 __t H30+ + OHe- + H30+ + H + H2014 B. G. Ershov and A. K. Pikaev, Radiation Res. Rev., 1969,2, 1.147Direct interaction with halide ions is also reasonable:Hal- + y + Hal + eHal + Hal- -+ Hal2-(5)(6)which may be followed byAll the paramagnetic species listed in (1)-(6) apart fromH20C and Hal were detected, if it is allowed that theintense blue-violet colours were indicative of et-.14 Tothese reactions we must now addMf + H+MH+ (7)and OH + Hal- -+ HalOH- (8)(or possibly, M+,H20 + e- + MH+ + OH-)Neither of these reactions (7) and (8) are normallyconsidered in liquid-phase studies: it may be that (7)is not of kinetic significance in view of the weakness ofthe interaction, but (8) certainly should be importantexcept for FOH-, when the bonding is again thought tobe weak.We thank the S.R.C. for a grant to I. S. G. and Mr. J. A.[1/1085 Received, June 29th, 19711Brivati for experimental assistance
ISSN:1477-9226
DOI:10.1039/DT9720000143
出版商:RSC
年代:1972
数据来源: RSC