摘要:

N. N. GREENWOOD P . G. PERKINS A N D D . H . WALL 59 GENERAL DISCUSSION Prof. V. I. Gol'danskii and Dr. E. F. Makarov (Moscow) (communicated):* For the establishment of analytical connections between the isomer shifts of Mossbauer spectral lines and the electronic structures of the compounds of tin which have been studied it is necessary to have calibration curves i.e. the " chemical difference " AX I $(O) 1 = 'c I $(O) I -'c I $(O) I Zms against the numbers of s p (and generally also d ) electrons in the valence shells. Authors of the earliest calculations l v 2 made use of an ionic model and assumed that the values of the chemical differences for tin compounds are determined simply by the number of 5s-electrons. At the present time two detailed methods of obtaining the requisite calibration curves have been developed (1) those based on the Fermi-Segre-Goudsmit evaluation of I $(O) I gS 3 9 4 with the use of empirically determined shielding parameters ; and A.J. F. Boyle D. S. Bunbury and C. Edwards Puoc. Physic. SOC. 1962,79,416. V. I. Gol'danskii G. M. Gorodinskii S. V. Karyagin L. A. Korytko L. M. Krizhanskii E. F. Makarov I. P. Suzdalev and V. V. Khrapov Dokl. Acad. Nauk S.S.S.R. 1962,147 127. E. Fermi and E. Segre 2. Physik 1933,82,729. S. A. Goudsmit Physic. Rev. 1933 43 636. V. I. Gol'danskii E. F. Makarov and R. A. Stukan J. Chern. Physics 1967,47,4048. * edited by R. H. Herber. 60 GENERAL DISCUSSION (2) those based on Hartree-Fock type calculations of electron densities at the nucleus for different valence shell configurations and various charge states of the Sn-ion.l Below is given a short description of the first of these two ways as well as the final results of both methods of calculations.The magnitude of the '' chemical difference " for tin-compounds which participate in the expressions for the isomer shift relative to a-Sn(sp3) can be written as where 1 $(O) 1 &,s is the density of one 5s-electron on the tin nucleus in a-Sn and b = a/(Zzft - 3a). Here a = shielding parameter of the 5s-electron by 5p electrons Zz$ = effective charge for the 5s electron in Sn3+ Po = relative change of the contribution of electron density on the nucleus from the ionic core when one 5s- electron is removed or added S,P = populations of 5s and 5p tin states. The expressions (I) are given in the following approximations (1) I $(O) 1 2cc Z,", (Fermi-Segre-Goudsmit approximation with Zeff = effective charge for 5s- electrons in any given charge state).(2) Shielding parameters of the 5s electron by another 5s- and Sp-electron are taken as equal as = ap = a (this follows from the analysis of terms according to Slater and Urusov 2). (3) The shielding of ionic core electrons by Sp-electrons can be neglected in comparison to such shielding by 5s- electrons (this follows from Hartree-Fock type calculations I). The value of Po can be estimated from different assumptions (1) from Hartree- Fock type calculations Po = 0.12 (2) from the Fermi-Segre-formula Po = 0-16; (such calculations are analogous to those performed for the mercurous ion Hg;+ in ref. (3)). (3) From the estimates of shifts of Kl -L1 -M1 terms for Sn Po> 0.2.In the following discussion the value Po = 0.17 has been used. The magnitude of b can be obtained from the following data (1) The ratios are determined from the values of I $(O) lig 1 $(O) I &I 1 $(O) I I2, 1 $(O) I S2, published in ref. (5) and obtained from the application of the Fermi-Segre formula to the HFS constants of optical spectra. Taking into account that in our approximation fccZ,Zf eff(neff = const.) we obtain S. L. Ruby G. M. Kalvius G. B. Beard and R. E. Snyder Physic. Rev. 1967,159,239. V. S. Urusov Zhur. Stukt. Khim. 1962 3,437. M. F. Crawford and A. L. Schawlov Phys. Rev. 1949,76,1310. 0. I. Sumbaev and A. V. Mesentzev Zhur. Eksp. Teor. Fiz. 1965 48,445. D. A. Schirley Rev. Mod. Physics 1964 36 339. GENERAL DISCUSSION 61 (2) The values of (Z,,,),, (Ze+ff)ca (Zz$)sn are taken from ref.(2 p. 60). Solving this system of equations which is overdefined we see that they are satisfied to good accuracy when 0*75a, a,50-8. Below it is assumed that a = a = a = 0.7 which gives (Z2f$) = 6-06.' Thus The a, a,-values are in good agreement with the data of ref. (2 p. 60). The absolute value of I $(O) I &3 can be obtained from the equation where do 1 dn - 1 + (- ATneff/T2Ao)' _ - and Tis the state term. we get Using the values of terms for 5s-and 6s-electron states in tin [ $(O) I &3 = 3.05 x 1026(2~ff/n~ff). (IV) (ze&,3 = (z:f;),n-3a = 3-96 ; neff = 3-52 (from ref. (7) or from the terms values- the result is almost the same). I $(o) 1 :p3 z 1.1 x ~ r n - ~ . This magnitude is in satisfactory agreement with the value of 1-2 x ~ m - ~ obtained from the extrapolation of data on HFS of optical spectra lo-this latter value of I $(O) I &3 = 1.2 x ~ m - ~ is used below.Thus (V) (AX I $(O) 1 :<1=1.2~ - 10266([1+0~17(3-P)]2S-1+(1-S)0~17) (AX I $(O) J :z 1 1.2 x lo2 { [ 1 + 0*17(4 - S - P)12 S - 1 + (1 - S)0-17) Fig. 1 represents the values of the " chemical differences " (relative a-Sn) as a function of S at different values of P. The solid lines are given in accordance with eqn. (V) (ref. (5 p. 59)) the dotted lines-according to the data of ref. (1 p. 60) obtained by Hartree-Fock type calculations for Sn. The ratios of differences at various numbers y1 of 5p- electrons obtained from the data of fig. 1 are given in fig. 2. One can see that the role of both s,s and s,p shielding manifests itself considerably more strongly for eqn.(V) than from Hartree-Fock type calculations (the effective values of 3,s and s,p shielding in the Hartree-Fock case are much less than from F. Bekker and S. Goudsmit Atornic Energy States (Mctiraw-Hiff New York and Londuii 1932). 62 GENERAL DISCUSSION Slater's or Urusov's treatment). The results represented in fig. 1 can be compared with experiment. For example the semi-empirical m.0.-method gives the following populations of the valence states of the Sn-ion in SnCl (5s) s N" 0.8 For these values of S and P one gets a " chemical difference " equal to 0.15 x ~ r n - ~ according to eqn. (V) and -0.12 x (5p) P = 2 ~ r n - ~ according to Hartree-Fock type FIG. 1 calculations. Using the experimental isomer shift for SnC1 which is negative (compared to a-Sn) one gets for the 23.8 keV excitation AR/R FZ -4.4 x from (V) and ARIR x + 5.5 x from Hartree-Fock type calculations with the above- mentioned populations.Thus even the sign of the AR/R(119Sn) ratio is opposite in this case for the two types of calculations (see also ref. (5 p. 64)). From a large body of experimental data (see e.g. ref. (3 p. 13)) it is more probable that P - 3 in tetrahalogenides. If we take the configuration So'*P3 for SnCl, both eqn. (V) and Hartree-Fock type calculations give the same value of the '' chemical difference " 0.2 x ~ m - ~ . This corresponds to AR/R(119Sn) = + 3.3 . which is in good agreement with the direct measurement of this value of ref. (4 p. 63). A definitive solution of this problem needs either the direct determination of the effective charge of tin in its compounds (at least in two of them) or the precise description of the electronic structure of tin in its two-valent compounds.V. I. Gol'danskii E. F. Makarov S . P. Ionov and G. V. Ionova unpublished. GENERAL DISCUSSION 63 For the evaluation of AR/R(ll9Sn) from the calibration curves it would have been necessary to get the isomer shift for some compound with a well-known population of 5s and 5p states of Sn. Unfortunately there are no strict criteria for the correct determination of electron populations of given valence states. However purely ionic models are unsatisfactory in the description of real chemical bonds in Sn- 4 Fermi- Segre- Goud smit Harkee -Fock 3 2 p a - * FIG. 2 compounds. Therefore the most appropriate procedure appears to be the use of isomer shifts for a set of compounds with similar structure and the extrapolation of these shifts to complete bond ionicity (i = 1).In this way based on different data for the ionicities of Sn-bonds in its halogen compounds SnHal 1-3 one can con- clude that the maximum value of the isomer shift (relative to a-Sn) for the 5s06p" configuraLon is (- 6-0 1- 0.6) mmlsec. Using the calculations of " chemical differ- ences " for 5s05p" configurations (as can be seen from fig. 1 these are similar for both methods of calculation) one obtains AR/R(l19Sn) z +(3-3f0.3) x in good agreement with the direct results of the Brookhaven group experiments (compari- son of 0-shell conversion coefficients in p-Sn and SnO,). In the ref. (1 p. 60) the ratio of AR/R(l19Sn) was determined on the assumption that the isomer shift (relative to a-Sn) for the 5s25p0 configuration is + 3 mm/sec.It can be seen from the Hartree-Fock graphs of fig. 1 that in this case AR/R = 1.2 x The authors of ref. (4 p. 64) have analyzed the isomer shifts of stannous (two-valent tin) compounds and concluded that the shift for the hypothetical compound 5s25p0 is equal to + 5.6 mm/sec which gives-for Hartree-Fock '' chemical differences " the value of AR/R(l 19Sn) = 2.2 x 'V. I. Gol'danskii G. M. Gorodinskii S . V. Karyagin L. A. Korytko L. M. Krizhanskii E. F. hlakarov I. P. Suzdalev and V. V. Khraphov Dokl. Acad. Nauk S.S.S.R. 1962,147 127. V. I. Gol'danskii E. F. Makarov and R. A. Stukan J. Chem. Physics 1967 47 4048. M. A. Whitehead and H. H. Yaffe Theor. Chim. Acta 1963 1 209. J. P. Bocquet Y.Y. Chu 0. C. Kistner M. L. Perlman and G. T. Emery Physic. Rev. Letters 1966 17 809. 64 GENERAL DISCUSSION As can be seen from fig. 1 the use of " chemical differences " for stannous com- pounds obtained by a semi-empirical method would give for AR/R(llgSn) a con- siderably smaller value = 1 x lo- in the last case. However the choice of popula- tion values corresponding to different isomer shifts for stannous compounds is indefinite due to the complexity of their structures. The calculation of orbital populations of the valence shells of tin in SnCl, per- formed on the basis of the semi-empirical m.0. method led to the result 5 ~ ~ ~ ~ ~ 5 P 1*795d0.08 (aSnCl4 = - 1.3 mmlsec) Now it is difficult to estimate the accuracy of these values but from general con- siderations one can judge that the accuracy of the calculations of the 5s population is better than the one for 5p and 5d-states.Taking the experimental value AR/R(119Sn) w 3.3 x 10-4,3 one finds that for SnCl with 5 ~ ~ " ~ one should accept the 5p3-configuration-and this result is the same for both types of " chemical difference " calibration curves. Thus the final solution of the question concerning the correctness of one or another method of calculation of " chemical differences " can be given only by an independent determination of the population of the valence shells of Sn in some compound with a known isomer shift. Considerable use can be made of the measurements of effective charges of Sn and particularly by detailed studies of stannous compounds since the distinction between various types of calculations of " chemical differences " for such compounds is much greater than for stannic compounds.Prof. J . F. Duncan (Victoria University of Wellington New Zealand) said In work in collaboration with Marianne Anderson Wellington Fulbright Scholar we con- sidered the reaction MX + M*Y = M*X+ MY where the asterisk indicates the excited state of the Mossbauer nucleus M in a chemical environment X or Y. This may be regarded either as an exchange of environment by a chemical process without any nuclear transitions or by exchange of the nuclear energy (by transfer of y-radiation) between the two atoms without any chemical process being involved. Using statistical mechanics and assuming the Einstein mass-energy relations it is possible to derive the free energy change AGZ of the above reaction for systems where the fundamental vibration frequencies are known.This free energy change refers to differences in MX and MY which involve all types of electron (i.e. both s(o) and non ~ ( n ) ) whereas the isomer shift 8 refers primarily to Mossbauer nucleus s- electrons. On the other hand we may set up a pseudo-equilibrium in which the distribution of excitation between MX M*X MY and M*Y is maintained by radia- tion exchange under conditions governed by the frequency of recoilless emission and absorption (i.e. by the Debye-Waller factor). This allows a quantity AG to be derived which we may regard as that contribution to AG," due to s-electrons only. The theoretical relations obtained axe AG," = AZ[G(U)YAV~ - G(u)~AvJ and AG,O = BZAV~/VX~V~~ V. I. Gol'danskii E.F. Makarov and R. A. Stukan J. Chem. Physics 1967,47,4048. V. I. Gol'danskii E. F. Makarov S. P. Ionov and G. V. Ionova unpublished. J. P. Bocquet Y. Y. Chu 0. C. Kistner M. L. Perlman and G. T. Emery Physic. Reu. Letters 1966 17 809. J. D. Donaldson and B. J. Senior J. Chern. SOC. A 1966 12 1796. I. B. Bersuker V. I. Gol'danskii and E. F. Makarov Zhrtr. Eksp. Teov. Fiz. 1965 49 699. GENERAL DISCUSSION 65 taken over all corresponding vibrations of X and Y. A and B are constants Av = vxI-vyi and u = hv/kT. AG should be directly dependent on 6 and (if the assumption of thermodynamic equilibrium is applicable) identical with it. The difference between AGE and AG must give a measure of the non-s (n) contribution to bonding. The table below gives some values for the evaluated quantities compared with 6 for one pair of compounds for which the full calculation can be done.compounds - AGE (mmlsec) - A6 (mmlsec) - AG; (mm/sec) SnC14-Sn14 0.033 1.0 0.410 Other examples could be quoted. Bearing in mind the experimental and theoretical uncertainties the values are reasonable. They suggest that an estimate of the G and rc contributions to bonding can be obtained by comparison of A6 or AG with AGZ. The close theoretical relation between these quantities can be seen from the tempera- ture dependence of both AGZ and AGZ which can be shown to be identical with that of A6 (via the Bodmer equation) when the system is in thermal equilibrium with its surroundings. In addition theory leads to the possibility of estimating 6 for an unknown compound by comparison with the vibrational frequencies of compounds for which 6 has previously been determined.It also leads to an understanding of the circumstances under which 6 values can be expected to be additive. Dr. A. J. Stone (University of Cambridge) said It seems to me that the calculated value of AR/R should be explicitly regarded as a semi-empirical number which may be used in conjunction with a Pople-Santry-Segal type of calculation but which has no wider application. It is unrealistic to expect a true value of AR/R from a method which is known to have considerable deficiencies even when applied to molecules containing only light atoms. Dr. P . G. Perkins (University of Newcastle upon Tyne) said In reply to Stone it is implicit in the text that the value of AR/R calculated by us is subject to the approxi- mations of the calculational methods.Prof. J . F. Duncan (Victoria University of Wellington N.Z.) said With reference to the paper by Greenwood et al. using the relations where A is the nuclear mass and B and C are evaluatable constants it is easy to show that only about 10-l3 of an electron need be annihilated for an energy equivalent to an isomer shift of 1 m.rn/sec. Although this is very small by comparison with the electron orbital occupancy numbers quoted in the paper it is about one-twentieth of that expected to be present in the nuclear volume. If such electron annihilation occurs it would be in direct contradiction with the primary assumption used in deriving their eqn. (l.l) viz. that the electron density within the nucleus is constant. One may therefore ask whether this is a realistic assumption whether electrons exist free within a nucleus and whether the isomer shift may not rather be due to electron-nuclear interactions occurring at the surface of the nucleus.If the answer to any of these questions is " yes " the validity of eqn. (1.1) is in doubt although it may still be a useful equation for handling experimental data. 3 66 GENERAL DISCUSSION Dr. M. Cordey Hayes (Birmingham University) said Probably one of the largest inaccuracies in the analysis of isomer shift data for heavy elements is that the direct contributions of p4 electrons are not included. According to relativistic theory the p+ electrons have a peak at the origin which is ( ~ 2 ) ~ times smaller than that of the s-electrons of the same principal quantum number ; u is the fine structure constant and 2 the nuclear charge.For tin and antimony this amaunts to approximately 10 %. The correlation of I19Sn isomer shifts for tin (IV) compounds with the occupation numbers of the 5s-orbital of tin calculated by SCMO methods by Greenwood et al. is a significant advance in the analysis of I19Sn Mossbauer data. But the conclusions obtained from an extension of their calculations to the inter- pretation of quadrupole splitting in compounds of the type R,SnX is not clear to me. They appear to conclude that the suggestions of Gibb and Greenwood are " confirmed " because there is a correlation between the calculated dn-pn bond order and the magnitude of the splitting for some of the compounds. Yet the p-imbalances calculated from cqn. (6.1) and given in table 3 could also be interpreted as giving some support for the interpretation of quadrupole splitting in terms of differences in the o-bond characteristics ; the dn-n bond order correlation then being regarded as a secondary effect and not responsible for the splitting.Dr. P. G. Perkins (University of Newcastle upon Tyne) said In reply to Cordey- Hayes it is clear from the data of table 3 of our paper that a high quadrupole splitting may be associated with a high d,-p bond order. I-Iowever we would not at this stage claim that this is the only interpretation of quadrupole splitting data. The effects of d occupation on p orbital imbalance would have to be examined more closely. We do possess the data necessary to decide whether the presence or absence of quadrupole splitting is due to a dG or dx mechanism and the next step will be to study this more closely.Dr. D. W. Davies (University of Birmingham) said I think that there are one or two general theoretical points that may be worth making. To the quantum chemist the variation in the electron density at the nucleus is the most interesting result of Mossbauer spectroscopy as magnetic hyperfine and electric quadrupole splitting can be obtained in other ways. To " confirm " either of these hypotheses the sign of the field gradient along the Sn-X bond direction should be measured-in back donation of electrons into the vacant orbitals of tin there is an excess of charge along the Sn-substituent bond direction whilst for mechanism based upon ionic charge transfer effects (but which can also include the effects of changes in hybridization) there is charge density missing along this bond direction and accordingly the sign of the e.f.g.is different in the two cases. We have heard so much about " s-electron " density that I must recall that I is not a good quantum number in molecules. There are no s- p- d- electrons and it is a severe approximation to use such a description even for many-electron atoms. If an orbital description of molecules is adequate for the qualitative discussion of chemical bonding (a doubtful assumption) then an orbital interpretation of the isomer shift may be useful. There is no need however to invoke the further assumption that an atomic orbital picture is adequate. In terms of molecular orbitals there is a clear distinction between molecules with and without a centre of symmetry.In octahedral or square planar complexes there will be orbitals with a node at the central atom T. C. Gibb and N. N. Greenwood ref. (16) in their paper. M. Cordey Hayes J. Inorg. Nucl. Chem. 1967 27 1177. GENERAL DISCUSSION 67 whereas in tetrahedral complexes there is no reason why any orbitals should have nodes at the central atom. I am not sure whether experimental results for isomer shifts support such a distinction. Perkins has presented the results of some molecular orbital calculations based on a Slater-type atomic orbital set. I agree with Stone’s comment that these calculations can hardly be regarded as giving reliable values for the electron density at the nucleus in tin complexes. The number obtained should be regarded as a semi-empirical quantity having some relation to the actual electron density a relation that might be different in different series of complexes.I view with grave suspicion the use of the Fermi-Segrk formula,2 derived over 30 years ago for a different purpose and any electron density value obtained from atomic orbitals depends critically on the choice of orbital exponent. The best choice for molecules is a highly controversial and for molecules containing heavy atoms a largely undiscussed question. TABLE 1 .-HARTREE-FOCK CALCULATIONS FOR ODD-ELECTRON ATOMS Percentage of “ correct ” spin density method Li B N Na P K RHF 72 0 0 76 -140 62 SPHF 97 4600 -5400 7500 * 320 86 -84 77 Proj. SPHF 81 1500 -1800 - ref. 3 4(4 4(b) 3 3 3 - - - - * different basis sets. For even-electron systems reliable theoretical values for the electron density at the nucleus are available for perhaps He H2 HeHf Be and LiH.I do not know if experimental Mossbauer spectroscopists can do anything with these. For odd- electron systems there is a considerable amount of evidence for the value of Hartree- Fock calculations on atoms in the interpretation of experimental hyperfine ~plittings.~ The simplest Hartree-Fock calculations however often give highly inaccurate values for the spin density at the nucleus as shown in the table of results for restricted (RHF) spin-polarized (SPHF) and projected spin-polarized (Proj. SPHF) Hartree-Fock calculations. Although the spin density is probably more difficult to calculate than the electron density I do not think the results for B N and P inspire much confidence. Dr. B. G. Perkins (University of Newcastle upon Tyne) said In reply to Davies the question with regard to calculations such as ours is an old one i.e.are all calcula- tions on large compounds or those containing heavy elements unjustified until the problems of small compounds have been completely solved or should one attempt to carry them out? I believe that because the problems of interpretation posed by techniques such as Mossbauer spectroscopy are with us now we cannot evade making some attempt to solve thcm. At the same time it must be realized that such attempts lack rigour. Input data for the calculations should be obtained as rigorously as possible and should not be arbitrarily varied. The justification for such calculations is that they explain the experimental phenomena under investigation and moreover suggest what the next step should be.N. N. Greenwood P. G. Perkins and D. H. Wall this Symposium. cf. L. L. Fodey Physic. Rev. 1958 111 1093. A. J. Freeman and R. E. Watson in Mugnetism ed. G. T. Rado and H. Suhl (Academic Press 1965) vol. IIA p. 167. N. Bessis H. Lefebvre-Brion and C. M. Moser Physic. Rev. (a) 1962 128 213 ; (b) 1961 124 1124. 68 GENERAL DISCUSSION Davies is correct in asserting that electrons in compounds may not be classified as s p or d. However in order to obviate the complexity of classifying wave functions from compouod to compound under the particular symmetry group concerned it is comenient to retain this nomenclature. At the same time it should be remembered that such a procedure is not strictly correct. The simplification attendant on classify- ing wave functions with respect to inversion only is an intermediate description but even this is not universally recognized.The best approach is to regard the ligands as having no other effect than to modify the s-orbital occupation of the free tin atom. Prof. J. Danon (Rio de Jcmeiro) said The Fermi-Segrk formula is valid only for unpaired s-electron density. Did you correct for mutual 5s electron screening as required for the comparison between 5s densities calculated from Fermi-SegrC and those given by molecular orbital populations ? Dr. P. G. Perkins (University of Newcastle upon Tyne) said In reply to Danon the quantity calculated by us is the total s electron density at the nucleus. There is no excess of one type of spin over another. The partial orbital occupation does lead to difficulty over electron screening as we pointed out in our paper.

ISSN:0430-0696    

DOI:10.1039/SF9670100059

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Isomeric Chemical Shifts of Mossbauer y-lines in Isoelectronic Compounds of Tin Antimony and Tellurium BY V. S. SHPINEL V. A. BRYUKHANOV V. KOTHEKAR B. Z. 1 0 ~ ~ a n d S. I. SEMENOV Institute of Nuclear Physics of the Moscow State University Moscow U.S.S.R. Received 7th August 1967 Isomeric cheniical shift (IS) of the Mossbauer y-line of Sn1I9 (23.8 keV) SbI2' (37-2 keV) Te12S (35-6 keV) have been studied in the isoelectronic complexes SnHa1g2 SbHalG' and TeHalZ2 (Hal = F Cl Br I). Some other compounds of antimony have been also measured. The IS-values for these compounds depend linearly on the electro-negativity differences of the Mossbauer atom and ligand. The relation between relative change of the mean-square charge nuclear radius for three y-transitions (&R/R)sb/(&R/R)s = - (5.5 f0.3) and ( & R / R ) T ~ / ( ~ R / R ) s ~ = + (1-1 f0.2) is obtained by comparing these dependences.In 37.2 keV and 35.6 keV y-transitions the calibration of the IS value associated with one 5s-electron was also carried out. We have obtained the recoilless fraction of the 37-2 keV y-quanta from Sn12102 source as equal to f = 0*3228:8$. From the measurements of the dependence of resonance effect on the Sb205 absorber thickness the value of the total internal conversion coefficient for the 37-2 keV y-transition is -10. The isomeric chemical shifts (IS) of the Mossbauer y-line in the isoelectronic compounds of the SnHal type displayed a simple correlation with the halogen e1ectronegativity.l Such correlation enables the IS values to be related to the nature of the chemical bonds in a more reliable way and is of essential importance for the analyses of the experimental data obtained using the nuclear y-resonance method.In order to derive the change in the charge radius between the ground and excited state 6R/R it is necessary to calculate the electron density on the nucleus of a multi- electron atom in a molecule or in a crystal. Such calculations cannot be carried out within a sufficiently high accuracy and the values of 6R/R obtained in such a way are likely to be uncertain. Additional useful information that is not connected with the absolute calculations of the electron density on the nucleus can be obtained from the IS investigations in isoelectronic compounds of elements having the same valent electron shells such as e.g. tin antimony and tellurium whose outer electron con- figurations are 5s2 5p2 5s2 5p3 and 5s25p4 respectively.One of the essential results of such investigation consists in obtaining the relative values of 6R/R for the nucleus of the atoms under consideration. In the present paper we compare the IS values of the Mossbauer y-line of Sn119 (2343 keV) Sb121 (37.2 keV) and Te125 (35.6 keV) in the isoelectronic complexes Me2SnHa16 MeSbHa16 and Me,TeHa16 where Hal = F,Cl,Br,I and Me is one of the following cations H+ Na+ (NH4)+ K+ Rb+ Cs+. In addition to these compounds the Me,Sn(OH)6 and MeSb(OH)6 compounds were investigated. We choose these complexes with the octahedral structure of the 02 symmetry group2 because despite differences in the valent electron configuration tin antimony and tellurium form equivalent bondings with six halogens in these compounds.It is usually considered (e.g. ref. (3)) that the chemical bonding of the central atom (Sn Sb Te) with the ligand is formed by sp3d2-hybridization. 69 70 ISOMERlC CHEMICAL SHIFTS EXPERIMENTAL AND RESULTS The IS measurement of the 23-8-keV y-line was made using the Snllgm isotope in BaSnO since this has the same probability of the Mtjssbauer effect as for SnO and the resonance emission line that is not broadened by the quadrupole interaction. The source of the 37.2-keV resonance radiation of SnlZ1 was metallic tin ( N 500 mg) enriched up to 98 % with Snlzo and irradiated with thermal neutrons in a reactor. The irradiated tin was subjected to chemical purification to remove radioactive Sb125 and TelZ5" that were activated in the reaction Sn124 (n y) Sn125-+Sn125-+Te12s".The sources in the /3-SnlZ1 form (electroplating on Au foil) and Sn12102 form were prepared using the purified tin. The fraction of recoilless resonance radiation for the Snl2lO2 source was determined as f = 0 . 3 2 ' ~ ~ ~ by investigating the dependence of the effect of magnitude on the Sb205 absorber thickness. This value is twice as large as that for the p-SnlZ1 source. At room temperature the value off for the Sn12102 source was about 0.16. This result is expected if one takes account of the resonance probability data for the 23.8-keV y-transition in Snl according to which thefvalue in SnO is larger than that in p-Sn (this fact can be explained by the effect of the optical branches of the phonon spectrum). On the assumption that the recoilless absorption probability in Sb205 is close to f in SnO the total internal conversion coefficient for the 37.2-keV y-transition was estimated to be about 10.During the investigation of the IS values for the 35.6-keV y-transition in Te125 the Telzsm isotope (in the ZnTe compound) and Sb12 were employed as sources of the resonance y-quanta. For SblZ5 the radioactive antimony Sb125 was extracted without carrier from a metallic tin sample enriched up to 80 "/o with Sn12" and irradiated in a reactor. In order to obtain a source displaying an unbroadened emission line Sb125 was diffused in the copper lattice using electroplating and annealing in an hydrogen atmosphere. The parameters of these sources were examined by observing the resonance absorption spectra in the ZnTe sample.It was found that the emission line width for both sources is close to its own width (r = 2.5+0.2mm/sec) and the IS value for the Sb125 source has been obtained as -0-1 +O-2 mm/sec with respect to ZnTe. The resonance absorption spectra were measured using the electrodynamic vibrator to provide a constant velocity. Measurements were carried out at liquid- nitrogen temperature. The standard scintillation spectrometer and the NaI(T1) crystal were used to record the 23-8-keV y-radiation of Sn119. The 37.2-keV y-line of Sb121 and the 35.6-keV y-line of Te125 were recorded by a NaI(T1) crystal (-0.3 mm thick) and by a proportional counter (mixture of xenon and isopentane) across the escape peak. Resonance absorption spectrum for all complexes of the SnHali2 type consisted of a single line whose width exhibited no observable broadening.This result is expected since tin in the SnHali2 complexes forms six equivalent bondings with six halogens. The absorption line in the samples of K,Sn(OH) and Na,Sn(OH) had a broadening of about 50 % that can be attributed to the non-cubic structure of these salts. The main portion is the IS measurements in the compounds of Sb was carried out at the liquid-nitrogen temperature using the SA2l0 source. The resonance absorption spectra have been observed for the following samples Sb205 Sb203 SbC15 NaSb(OH), NaSbF, HSbC16. xH20 InSb Sb. The absorbers were made from the antimony with its natural content of the Sb121 isotope. In all cases except Sb203 and Sb the absorption resonance spectra were found to be single lines without any observable broadening.The magnitude of the experimental effect for the antimony compounds was about 6-7 % or greater. SHPINEL BRYUKHANOV KOTHEKAR IOFA AND SEMENOV 71 The tellurium coniplex salts were made from tellurium enriched up to 60 % with TelZ5. The samples were examined using chemical and X-ray analyses. For all the tellrrrium complexes the resonance absorption line was a single peak without any considerable brcadening (re,, = 6-7 mm/sec). Since the TeFZ2 complex could not be made as a crystal salt we have prepared a solution of tellurium in concentrated HF acid which was frozen during measurements. In addition we investigated the resonance absorption spectra for solutions of Te in concentrated HCl and HBr acids. The corresponding spectra had the same form and the same IS value as those for TABLE 1.-VALUES OF THE ISOMER CHEMICAL SHIFT (Is) FOR COMPOUNDS OF TIN ANTIMONY AND TELLURIUM WITH RESPECT TO THE SOURCE SUCH AS snll”oz Sn12102 and Te125(C~) 1 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 RESPECTIVELY IS (6E) mm/sec 3 - 0.48 f0-05 0.50 f0.05 0.48 f0-05 0.45 f0.05 0.43 10.05 0-45 f0.05 0.80 f0.05 0-75 k0.05 1.35 f0.05 0.02 f0.05 0.03 f0-05 0-0 f0.3 1.4 f0-3 1.7 f0.3 2.0 f0.3 + 0-5 f0.2 - 3.0 f0.2 + 2.0 f0.2 - 3.5 f0.3 0.0 f0-2 - 10.4 f0.3 - 8.4 h0.3 - 11-2 f0.2 SE eV x 10-8 4 - 3.8 10.4 4-0 k0.4 3.8 50.4 3.6 10.4 3.4 f0.4 3.6 f0.4 6.4 i0-4 6.0 &0*4 10.6 f0.4 0.16 f0.4 0.24 f0.4 0.0 k3.5 16.6 *3*5 20.1 f 3 .5 23-7 13.5 + 6.2 f2.5 - 37.2 12.5 + 24.8 &2.5 -43.3 f3.7 0.0 f2-5 - 127.9 13.7 - 103.3 f2.5 - 137.7 f3.7 crystals of these salts within the experimental accuracy.This result indicates that the complex of the TeHali2 is formed in concentrated acids the chemical bonding in the outer co-ordination sphere affecting negligibly the IS values. A similar con- clusion about SdV has been drawn by one of the authors of ref. (4) in an investiga- tion of the resonance absorption spectra of Sn1Io in concentrated HCl solutions. The JS values for the compunds are given in the table. The IS values for the SnHal; * complexes containing the various cations in the outer co-ordination sphere have been measured for the dependence of the tin-ligand bonding length on the form of a cation.6 As seen from the table the IS values depend only upon the ligand and are independent of the cation within the experimental accuracy.The dependences of IS (6E) on the electronegativity difference AX (AX was taken from the literature 9 8 of the tin-ligand antimony-ligand and tellurium-ligand are linear in their middle section and similar to the dependence of IS in the compounds of the SnHal type. Indeed the IS value 6E must depend linearly on the bonding ionicity. Since the relationship between the bonding ionicity and AX is not linear 72 ISOMERIC CHEMICAL SHIFTS for pure covalent and pure ionic bondings,' the dependence observcd in our case must deviate from the linearity at the beginning and the end of the curve. DISCUSSION Comparison of the dependences obtained enables the following conclusions to be (i) The IS formula has the form drawn 6E = ~(Z)(C~R/R)A$~(O) (1) where K(2) is the atomic factor SR/R is the relative change of the nuclear charge radius at a resonance y-transition A$2(0) is the difference in the electron densities on the resonance nucleus in the source and absorber.The dependence of IS on AX had the same sign for the tin and tellurium complexes whereas the IS dependence for the antimony complexes had an inverse sign as seen from the figure. Since the tin antimony and tellurium in these compounds have equivalent chemical bondings we suppose that the electron density on the Sn119 Sn121 and TelZ5 nuclei is also changed in the same way with varying AX. Thus from eqn. (l) the sign of the change in the nuclear charge radius SR/R for the y-transitions in the Sn119 and Te125 nuclei must be the same while the sign for the Sn121 nucleus is the opposite. It has been found 941 that the 23.8-keV y-transition in Sn119 involves a value of 6R/R>0.Consequently the sign of 6RIR for the 35.6-keV y-transition in TelZ5 is also positive while the 37.2-keV y-transitions in Sb121 involves the value of SR/R <O. A decrease in the IS values with increasing ligand electronegativity shows also that the extent of withdrawing of s-electrons to the ligand is the main factor deter- mining the SE value in these complexes of tin and tellurium. This conclusion agrees with theoretical calculations made using the relativistic electron wave functions *sl from which it follows that the change in $2(0) arising from the screening of the internal s-electrons in the tin compounds is about 20 "/o of the t,h2(0) value corresponding to one 5-s electron. (ii) The dependence of IS on AX in fig.1 enables the SE values corresponding to unit electronegativity difference to be determined. We have obtained the following values 6E(Sn)* = = 8.8 x lo-' eV 6E(Sb)A = = 62.5 x lo-' eV and dE(Te)Ax = 1 = 15 X lo-* ev. Using these values one can obtain the relation of the SR/R values for the given y-transitions in Sn119 Sb121 and Te12'. Taking account of (1) we find The atomic factors including the relativistic corrections have been calculated using the expressions given in ref. (12) and were found to be K(Z), = 1.55 x eV cm3 K(& = 1.44 x eV cm3 K(Z)Te = 1.63 x eVcm3. In order to determine L\$2(sn)A x = 1 /d$2(Sb)A = and b$2(sn)A = /d$2(Te)A = ratios we used the non-relativistic single-electron wave functions l3 and took account of the relation : SHPINEL BRYUKHANOV KOTHEKAR IOFA A N D SEMENOV 73 ~ m - ~ $$,(Sb) = From the calculated values of $gS(O) such as $:,(Sn) = 1.1 x 1-34 x A$2(Sn)A,= l/At,62(Sb)Ax = 0.84; A$2(Sn)Ax l/At,62(Te)Ax= = 0.7.The relation between relative changes of the root-mean square and change of nuclear radius for three resonance y-transitions viz. 23-8 keV in Sn119 37.2 keV in SblZ1 and 35-6 keV in TelZ5 were found to be These results depend little on the extent of approximation when calculating the electron wave functions (only their ratios are important) and on the absolute values of bonding ionicity and the 6R/R ratio determined in such a way are more reliable ~ m - ~ and $;,(Te) = 1.5 x ~ m - ~ we found that (6R/R)S,/(6R/R)Sn = -(5*5&0*3); (6R/R)Te/(6R/R), = +(Is1 +_0*2). 0 0 ) 0 2 0 3 0 m 2 5 0 g izi 40 2 ,-I N 3 6 0 Lg 7 0 8 0 9 0 100 FIG.1.-The dependence of IS (SE) for the resonance y-transitions in Sn119 (23.8 keV) Sbl2I (37-2 keV) and Te125 (35.6 keV) on the electronegativity difference of AX. 1 2 3 are the complexes of the SnHalg2 type the SbHalgl type and the TeHalg2 type respectively. 4 are compounds of the SnHa14 type.' than the estimation of the 6R/R absolute value from the calculated $2(0) values. Assuming (6R/R)s to be + 3.3 x (a result obtained on the basis of the measure- ments l1 of internal conversion electron spectra for the 23.8 keV y-transition in p-Sn and Sn02) the (6R/R) values for the 3702keV y-transition in Sb121 and the 35.6 keV y-transition in Te125 are ( ~ I R ) = - 18.4 x 10-4 ; ( ~ R / B ) = +3.6 x 10-4. (iii) The comparison of the IS dependence obtained in the tin complexes and in the tetrahalogen compounds of the SnHal type indicates that the s-electron density on the Snl l9 nucleus in the octahedral compounds (5sp3d2-hybridization) is about 74 ISOMERIC CHEMICAL SHIFTS 15 % smaller than that for the tetrahedron compounds in which the chemical bondings of tin are formed by sp3-hybridization.This result can be explained in terms of the screening effect of the d-electrons in the sp3d2-hybridization bonding. An identical screening effect of d-electrons is also found in a comparison of the IS for the complex salt of HSbC16 . xH,O with that for the SbCl salt. The penta- chloride of antimony has the trigonal bipyramid structure the antimony forming five sp3d-hybridization bondings with the chloride.2 As seen from the table the IS value obtained in SbCl consists of about - 0.5 mm/sec with respect to HSbC16 .xH,O the s-electron density $2(0) being greater in the SbCl system. This result may be also attributed to the same screening effect of d-electrons the screening being less significant in the sp3d-hybridization than in the sp3d2-hybridization. (iv) The dependence of the IS value in the TeHalZ2 and SbHalil complexes on the electronegativity difference may be used for an independent estimation of the (6R/R) value if the extrapolation of the 6E values to the pure covalent bonding is carried out i.e. if the 6E value corresponding to the removal of one 5s-electron is determined. For this purpose the data giving the relation between the bonding ionicities and the AX values as given by Pauling l4 can be used.From these data chemical bonding with about a 50 % ionicity takes place at AX = 2. Using this result one finds that the removal of one 5s-electron in the octahedral tellurium compounds is associated with an IS value of 60 x eV. The inclusion of the screening effect of d-electrons (about 15 %) results in the IS value associated with one 5s-electron in the tellurium atom being 70 x lo-' eV. Using these values of K(Z)Te and $&(Te) and expression (1) we find that (6R/R), = f2.8 x This result is consistent with the value of (JRIR), obtained from the ratio of (6R/R)Te/(6R/R)sn since the value of (6R/R), was determined l 1 within 30 %. A similar calculation may be carried out for (6R/R)Sb. The figure shows the IS value for InSb in which the chemical sp3-bonding of antimony is considered to be practically c0va1ent.l~ Using the dependence of IS for Sb121 the IS value of 6E = 250 x eV corresponding to removal of one %-electron in the octahedral antimony compounds can be found.Taking account of the screening effect of d-electrons (about 15 %) we find the IS value associated with removal of one %electron in the antimony atom to be 6E = 290x eV. Using these values of K(Z)Sb and $gs(Sb) then (6R/R)Sb = 14.4 x This result shows no essential contradiction to the appropriate value determined from the ratio (6R/R),,/(6R/R), within experimental accuracy. (v) Among the tellurium compounds studied a reliable evaluation of the number of s-electrons has been made only for metallic tellurium,16 the value being 1.6. The IS value obtained by extrapolating the IS dependence to pure covalent bonding (AX = 0) which corresponds to one 5s-electron is octahedral complexes is consider- ably greater thaa the 6E value for metallic tellurium (7.5 x eV with respect to the Te125(Cu) source).This result may be attributed to an additional density of non- bonding pair of the 6s-electrons. The possibility of such a non-bonding pair in the TeHali2 type complexes was proposed by P a ~ l i n g . ~ The donor electrons of the outer cations appear to take part in such formation. If this assumption is true then the IS calibration that the additional density $zS(O) produced by one 6s-electron is about a half of that $zs(0) produced by one 5s-electron. Using the spectroscopic data available for a tellurium atom l7 and the Fermi- Segre l8 formula the value of $is(0) was estimated to be about twice as small as that obtained in the above way i.e.$is(0) 21 0.25$,2,(0). SHPINEL BRYUKHANOV KOTHEKAR IOFA AND SEMENOV 75 A. Yu. Aleksandrov N. N. Delyagin K. P. Mitrofanov L. S. Polak and V. S. Shpinel Zhur. Eksp. Teor. Fiz. 1962 43 1242. B. F. Ormont Structure of Inorganic Compounds (Gostekhizdat. M-L 1950). Semiconductor Substances. B. Z. Iofrt' K. P. Mitrofanov M. V. Plotnikova and S. Kopach Radiochem. 1964,4,419. J. D. Dannag N. Nowacki and J. Dannag Crystal Data Classification of Substances by Space Groups and their Identification from Cell Dimensions (New York 1954). G. Engel Kristal. 1935 90 341. M. Finemann J . Chem. Physics 1958 62 947. V. A. Bryukhanov N. N. Delyagin A. A. Opalenko and V . S. Shpinel Zhur. Eksp. Teor. Fiz. 1962 43 432. Aspects of Chemical Bonding ed. V. P. Iuse (M. 1960). ' L. Pauling The Nature of Chemical Bonds (Ithaca New York 1960). l o J. Lee and P. A. Flinn Physics Letters 1965 19 186. l 1 J. P. Bocquet I. I. Chu 0. C. Kistner M. L. Perlman and G. T. Emery Physic. Rev. Letters l2 D. A. Shirley Rev. Mod. Physics 1964 36 339. l3 F. Hermann and S . Skillman Atomic Structure Calculations (Prentice Hall Inc. New York l4 L. Pauling J . Chein. Physics 1952 56 361. l 5 R. W. G. Wyckoff Crystal Structure (Interscience Publ. Inc. New York 1948). l 6 C. E. Violet and R. Booth Phpic Rev. 1966,144,225. C. E. Violet R. Booth and F. Wooten l7 Atomic Energy Levels as derived from Optical Spectra ed. C. E. Moore (Nat. Bur. Stand. 1966 17 809. 1963). Physics Letters 1963 5 230. Circ. no. 467 1949). Z. Fermi-Segre Physik 1933 82 729.

ISSN:0430-0696    

DOI:10.1039/SF9670100069

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

SHPINEL BRYUKHANOV KOTHEKAR IOFA AND SEMENOV 75 GENERAL DISCUSSION Prof. N. N. Greenwood (University of Newcastle upon Tyne) said With regard to to the paper by Shpinel et al. TeXg- is not isoelectronic with SnXg- and SbX ; it has two mere electrons. This will affect the numerical details of the analysis but probably not the sign of 6R/R. Table 1 refers to TeFg-; what is the evidence that this species is actually present in solution? We have tried unsuccessfully by numerous methods to prepare salts of this anion but invariably salts of TeF; are obtained. As the geometry of this ion is that of a square-based pyramid with a pendant lone pair of electrons,l it is perhaps surprising that no quadrupole splitting was observed if this were the species in solution. TeF itself can readily be prepared and is isoelectronic with SnFg- and SbF .Was its Mossbauer spectrum recorded ? It would be interesting to know the effect on the spectrum. of adding the two electrons to give TeFg-; this latter ion is isoelectronic with XeF which is now considered to be an irregular octahedron. Is it possible to estimate the degree of distortion which could be tolerated in TeFg- before a quadrupole split spectrum would be observed rather than the single line as reported? Dr. R. V. Parish (University of Manchester Inst. of Science and Technology) said Shpincl has suggested that the difference in isomer shifts for SnX and SnXg- is due to the shielding effect of the5d-electrons donated to thetin in theoctahedral complexes. Has he considered the effects of the change in bond length? The Sn-X distance in SnX2- is considerably larger than that in SnX and this effect alone would probably decrease the s-electron density.Prof. V. S. Shpinel (Moscow State University) said In reply to Greenwood the Since N. N. Greenwood A. C. Sarma and B. P. Straughan J. Chem. SOC. A 1966 1446. terminology " isoelectronic " may be replaced by " isostructural " compounds. 76 GENERAL DISCUSSION in all the three sets of compounds SnXz- TeXz- SbX; octahedral bonds are formed on the basis of 5sp3d2 hybridization the lone pair of electrons did not seem to have any influence on the change of isomeric shift with electronegativity. And as only the slopes are used in the calculation the value of dR/R was not much influenced by this lone pair. There is no direct evidence that TeFi- was formed. In fact since F is a highly electronegative atom isomeric shift in the two cases TeF; and TeF; - would be very close.Small quadrupole splitting which can be present in the former case cannot be observed because of the large line width of the resonance line. The fact that the salt TeFg- is not formed does not mean that the TeF2- ion cannot be formed. It may be that because of the high solubility it is difficult to separate it from the solution. Possible formation of the compound Me,TeF6 was mentioned by Urch. We could not measure the Mossbauer spectrum of TeF6. But we recorded the spectrum for Te(OH)6 which is isoelectronic with Sn(OH)Z-. The effect of adding a lone pair of electrons can be seen. The observed isomeric shift was equal to -0.5 mmlsec I.S. for Te(OH);- which can be predicted from our curve is equal to 0-65 mm/sec.This difference of 1.2 mm/sec can be attributed to the lone pair of electrons. It seems probable that in the compounds TeX2- the lone pair of electrons gives some constant addition to the s-electron density in the same way as the electrons of the inner closed shells. In reply to Parish since the occupation number of the 5s electrons in the two sets of compounds SnX4 and SnXi- (possessing valence-electron hybridization sp3 and sp3d2) is proportional to the electronegativity difference between tin-ligand the decrease of s-electron density at the tin nucleus in the latter case can be taken either as a change of electroncgativity of the tin atom in the two cases or as being due to the change of density caused by screening by two d-electrons. In both cases this decrease can be related to two extra d-electrons. We examined a number of compounds with different cations in which the Sn-Hal bond-length was different. Within the limits of experi- mental errors no effect on isomeric shift was observed.

ISSN:0430-0696    

DOI:10.1039/SF9670100075

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Isomer Shifts in Neptunium Compounds* BY J. A. STONE AND W. L. PILLINGER Savannah River Laboratory E. I. du Pont de Nemours and Co. Aiken South Carolina 29801 Received 1st September 1967 Characteristic isomer shift ranges have been measured for several valence states of neptunium using the Mossbauer resonance of the 59.54-keV gamma ray of 237Np. Isomer shifts were observed (a) for various neptunium compounds as absorbers with an 241Am-Th alloy as a single-line source and (b) for the charge states produced by alpha decay of various 241Am compounds as sources with a single-line NpOz absorber. Approximate shifts for the 0 +3 f-4 +5 and $6 valences of neptunium are respectively 0.0 +4.1 $0-2 - 1.5 and -3-5 cm/sec with respect to Np02. These data show that the fractional change in nuclear deformation is negative and that the excited- and ground-state quadrupole moments are equal to within 1%.The unusually large isomer shifts observed spanning more than 2000 natural line-widths offer an extremely sensitive method for investigating chemical bonding in neptunium-containing systems. The Mossbauer resonance in 237Np is a potentially useful tool for chemical studies in the actinide elements. The resonance has now been characterized considerably and found to have favourable nuclear properties. Several nonchemical aspects of the resonance have been investigated. The 59.54-keV gamma ray in 237Np can give the Mossbauer effect after either beta decay of 237U (half-life 6.75 days) or alpha decay of 241Am (half-life 458 y). l* From the Mossbauer spectra of materials that exhibit resolved hyperfine structure nuclear moment ratios between the 59.54-keV level and the ground state of 237Np were mea~ured.~ With these ratios complex spectra in which both magnetic and quadrupole splitting are present may be inter- preted.A convenient source of recoilless 59-54-keV gamma rays is a dilute solution of 241Am metal in thorium metal which emits a relatively narrow unsplit line with good effi~iency.~ By analogy with other systems such as 57Fe and ll9Sn isomer shifts in the neptu- nium system are expected to yield much chemical information. The most pronounced effect of chemical environment on isomer shifts is that of different oxidation states ; therefore we have investigated the effect of neptunium oxidation states on isomer shifts in the 237Np Mossbauer resonance. Two experimental approaches were used (a) representative neptunium compounds in each oxidation state were studied as absorbers and (b) different charge states of neptunium recoil atoms were observed following alpha decay of americium compounds as sources.Hyperfine structure parameters also are given for the neptunium compounds. The relationship of the quantities measured to the fractional change in nuclear deformation and to the isomer shift calibration is discussed. EXPERIMENTAL Neptunium metal was obtained from the Target Center Oak Ridge National Laboratory in the form of a rolled foil (500 mg/cm2). A specimen of NpC was obtained from Argonne National Laboratory. Sixteen neptunium compounds were prepared by standard *The information contained in this article was developed during the course of work under Contract AT(O7-2)-1 with the U.S.Atomic Energy Commission. 77 78 ISOMER SHIFTS IN Np COMPOUNDS methods ;6-13 each material was examined by X-ray powder patterns spectrophotometry and/or chemical analyses. A standard NpO absorber (140 mg/cm2) was used to study the 241Am sources. The neptunium absorbers were studied with a source prepared at Argonne National Laboratory by arc-melting 3 mg of americium with 98 mg of thorium ; the melted button was flattened annealed at 940."C for 20 h mounted in plastic and then sealed in a copper can.14 A 14 mg source of 241Am02 was prepared by igniting the oxalate at 800°C. A 16 mg source of 241AmF3 was prepared by precipitating AmF3 from an aqueous Am3+ solution with 2M HF followed by drying it in air at 85°C. Velocity spectra were taken with a standard loudspeaker-type Mossbauer spectrometer using a constant-acceleration drive (triangular velocity waveform) and employing a 400- channel analyzer operated in the multiscaler mode.The gamma ray detector was a NaI(T1) scintillation counter. Experiments at 4.2"K with liquid helium used a cryogenic system which has been described previously.2 RESULTS The shapes of typical 237Np Mossbauer spectra are illustrated in fig. 1 and 2 for different degrees of hyperhe splitting. Velocity spectra are relatively simple for an unsplit line (fig. la) pure quadrupole splitting (fig. lb) and pure magnetic splitting -2 0 +2 +4 0 +2 +4 +6 4 i I I I I I I I 1 I I 1 veIocity cm/sec FIG. 1.-Mossbauer spectra of the 59.54-keV y-ray in 237Np at 4~2°K. (a) NpOz ; (b) NpBr3 ; (c) NP&.(fig. l c ) ; the isomer shifts are given by the centroids of the patterns. For combined magnetic and quadrupole splitting the analysis is more difficult because the spectra are more complex. Moderately large quadrupole splitting combined with much larger magnetic splitting is shown in fig. 2a; fig. 2b shows magnetic and quadrupole splitting of similar magnitudes. Isomer shifts and hyperfine-structure parameters at 4-2"K are given in table 1 for eighteen neptunium compounds. The oxidation states represented are NpO NpIII NpIV Npv and Npv' ; for the latter two only data for compounds containing the oxygenated ions NpO; and NpO;+ are given. For neptunium metal and for Np,08 the velocity spectra consist of two superimposed hyperfine patterns. J . A . STONE AND W.L . PILLINGER 79 Isomer shifts relative intensities and linewidths for three 241Am sources at 4.2"K are given in table 2. For direct comparison with the neptunium absorber data source isomer shifts are quoted as the resonance Doppler velocities with their signs reversed. The lines have been assigned previously to different charge states of the TABLE 1 .-HYPERFINE-STRUCTURE PARAMETERS AND ISOMER SHIFTS OF NEPTUNIUM COMPOUNDS compound ref.* Np Metal (A) 6 NpAl2 7 NPA14 7 NPC 5 (B) 9 9 6 AT 4-2"K gO/lnHeff cmlsec - - 5.34 f0.05 4.90 10.05 8-46 rt0.10 6.00 f0.60 - - - 4.74 30.05 7.46 10.05 10.4 fO.1 - 9-30 i-0.05 9.81 30.05 10.1 f0-1 9.84 f0.10 4.81 rt0-05 5-29 f0-05 - cm/sec 2.23 +0*04 0.70 f0.01 -0*47&0*10 - - (1.04 &0*03)f 0.49 f0.05 0.55 zt0.03 - - 0.42 40.05 - 0.30 f0.05 -0.72 &0.10 2-58 f0.03 + 1-29 10.10 + 2-55 f0.05 + 2.30 f0.10 + 2.25 k0.25 +5*96+0.15 + 6-19 f0.15 4.04 +0*10 4 cmlsec -0.13 f0.01 + 0-1 3 A0.01 + 0.57 f0.05 f l .4 f0.1 -1.2 5 4 . 2 (3.5 fO-l)f +4*1 f0.1 +4-1 f 0 - 1 0.00 + 0-22 f0.05 +0*33 f0.13 $0.13 f0.05 -2.4 f0.1 - 1.55 10.07 -0.60 f0.10 -1.4 f0.1 -1.8 -40.1 -3.4 k0.1 -3.2 50.1 -3.9 f0.1 0 reference to method of preparation h magnetic splitting ; 1 cm/sec = 0.562 x lo6 G (assuming p = 2-8 n.m. for 237Np) C quadrupole splitting ; 1 cm/sec = 480.2 Mc ; only experimentally determined signs arc given d isomer shift with respect to NpO e from ref. (3) fcomplex spectrum at 4-2°K; data given are for 77°K 4 from ref. (1). TABLE 2.-MOSSBAUER PARAMETERS FOR 241Am SOURCES AT 4-2°K source intensity state cm/sec cm/sec relative charge linewidth 4 Am-Th (A) S 0.35 f0.01 -0.26f0.01 (B) W 0.36 f0.02 + 1.71 f0.03 Am02 (A) S + 3 0.41 f0.07 +4*1 i 0 .I (B) m +4 0.96 f0.14 -0.16 f0.05 (C> m + 5 0.96 40.14 -2.7 50.1 -3.8 10.1 (D) W + 6(?) 0.54 f0.14 AmF3 S + 3 0.82 f0.03 +4-4 f0.1 s = strong; m = medium; w = weak. 80 ISOMER SHIFTS XN Np COMPOUNDS neptunium recoil atoms on the basis of chemical arguments.16 Together the data from source and absorber experiments offer strong evidence for the existence of characteristic isomer shift ranges for the different neptunium oxidation states. DISCUSSION Isomer shift measurements in 237Np are summarized in fig. 3 which shows the ranges of isomer shifts for each oxidation state of neptunium. Isomer shifts are large in the 237Np system ranging from - 3.9 to + 4.4 cm/sec in the materials studied and t6 +s +d 5 1’.5 f2 5f3 5f4 0 Sf46d Ys2 -5 0 +5 velocity crn/sec FIG. 3.-Isomer shifts in Np compounds with respect to NpOz. they fall into well-defined bands for each oxidation state. The mean energy of these isomer-shift bands increases monotonically with decreasing oxidation number from Npvx through Np”’. The isomer shift 6 between two absorbers a and b is 6 = (4nZe2r2c/5E,)S’(Z)[D - Db](Ar/r) (1) J . A . STONE AND W . L. PILLINGER 81 where in the first term for 237Np (2 = 93) the nuclear radius r IS (1.20~ (237)* cm and E = 59.54 keV. The relativity factor S'(93) is given by Shirley l7 as 13.6; the quantity D denotes the density of s-electrons at the nucleus I $(O) 1'. The last term in eqn. (l) Ar/r is the fractional difference in nuclear radii between the excited and ground states.For a highly deformed nucleus such as 237Np eqn. (1) is modified to include the fractional change in nuclear deformation. An axially sym- metric nuclear surface may be described by an expansion in spherical harmonics as where the coefficient P is the deformation parameter.18 If the isomer shift is due only to changes in deformation it has been shown l9 that Ar/r should be replaced by (5P2/4;n)(AP/P) so that r = ro[l+PY~(O,@)+ . . .] (2) 6 = [(ze2r2C/E~)S'(Z)][D,-Db1P2(ABjB) (3) The first term in brackets is evaluated as 5.06 x The valence electronic configuration of a neutral neptunium atom is 5f"6d7s2 and in going from Npvl through Np*I1 the electronic configurations presumably change from 5f1 to 5f4. The 5felectrons influence the s-electron density at the nucleus and thus the isomer shift only indirectly by their shielding effect on the closed-shell 6s electrons.Therefore the addition of a 5felectron will increase the shielding and give a decreased density of 6s electrons at the nucleus. For example in the reduction of NpIV to Np''' the quantity [D(5f4) - D(5f3)] in eqn. (3) is negatiue. However from fig. 3 the isomer shift increases with each addition of a 5felectron in going from Npvl to Np"'. Therefore it is concluded that AP/P is negative i.e. the 59.54-keV level of 237Np is less deformed than the ground state. The intrinsic quadrupole moment of 237Np in the ground state has been measured 2o by Coulomb excitation as 9.0 x Then the isomer shift for 237Np is given by A quantitative estimate of Ap/p requires knowledge of s-electron densities in neptu- nium compounds ; unfortunately these are not yet available either experimentally or theoretically.Therefore an absolute calibration of the isomer shift in 237Np will not be given here. AP/P can be crudely estimated by (a) noting that on the average addition of a Sfelectron to neptunium changes the isomer shift by about +2.7 cm/sec and (b) assuming that [D - Db] is about the same as for the addition of a 4felectron to europium for which a calibration has been made. Brix et al. 21 give [D(Eu3+)- D(Eu2+)] = 1.9 x loz6 ~ m - ~ from which we calculate AP/P = -0.006 for 237Np. It seems likely that the electron-density changes for the heavier neptunium will be greater than for europium so that this result should be a lower limit The ratio of the quadrupole moments K between the 59.54-keV level aiid the ground state of 237Np can be expressed as cm4/sec for 237Np.cm2 from which the value of p is 0.21. 6 = (2-26 x lo- 24 cm4/sec)[D - Db](A/?/P). (4) - 0.006 < Ap/p < 0. (5) K = 1 +(A/?/p)[l+(5/47~)*P+ . . .I. (6) From hyperfine structure in 237Np Mossbauer spectra we found that K = 1-0+0.1 ;3 from the present work with isomer shifts we conclude that 0.99 < K < 1.0. (7) The range of isomer shifts observed in neptunium compounds (about 8 cm/sec) is The spans of the largest yet found spanning more than 2000 natural linewidths 82 ISOMER SHIFTS I N Np COMPOUNDS isomer shifts in 57Fe and 19Sn are only about 30 and 12r respectively and significant chemical information has been obtained 22 with the span on only 2r in 12’I.Thus the 237Np isomer shift should be markedly sensitive to chemical environment. Indeed the sensitivity of the isomer shift to slight variations in local environment may partially account for the breadth of 237Np resonance lines (40 to 1OOro). Even without sharper lines however isomer shifts in 237Np promise to be an important tool in the study of chemical systems in the actinides. In addition to the isomer shifts table 1 gives hyperfine structure parameters for the neptunium compounds studied. Although these will not be discussed in detail definite correlations between the quadrupole coupling constants and the chemical state are present. Also the internal magnetic fields range from 3 to 6 MG (assuming the ground-state magnetic moment of 237Np is 2-8 n.m.). Further work is planned to study these effects.In summary Mossbauer spectra of several neptunium compounds in different valence states have been obtained. The isomer shifts observed are both large and strongly correlated with the valence states. The fractional change in nuclear deforma- tion is negative and the quadrupole moments of the 59.54-keV level and the ground state of 237Np are equal to within 1 % with the excited-state moment being smaller. The 23 ’Np resonance should greatly increase the understanding of chemical bonding in the heavy elements. We thank Dr. D. G. Karraker Dr. M. C . Thompson and Dr. G. A. Burney for preparing several of the neptunium compounds and for valuable discussions and Mr. M. H. Goosey for the design and construction of electronicportions of the spectrometer. We are grateful to Dr.M. B. Brodsky and Dr. D. J. Lam of Argonne National Labora- tory for the loan of the Am-Th source and NpC absorber. J. A. Stone and W. L. Pillinger Physic. Rev. Letters 1964 13 200. J. A. Stone in Applications of the Mossbauer Effect in Chemistry and Solid-State Physics Technical Report Series No. 50 (International Atomic Energy Agency Vienna 1966) p. 179. J. A. Stone and W. L. Pillinger Physic. Rev. 1968 165 1319. S. L. Ruby and G. M. Kalvius private communication. J. W. Ross and D. J. Lam J. Appl. Physics 1967 38 1451. B. B. Cunningham and J. C. Hindman The Actinide Elements Natl. Nuclear Energy Ser. Div. IV. 1954,14A chap. 12. 0. J. C. Runnalls J. Metals 5 AIME Trans. 1953 197 1460. J. P. Bibler and D. G. Karraker Inorg. Chem. 1968 7 982. S. Fried and N. R. Davidson J.Amer. Chem. SOC. 1948,70 3539. lo J. J. Katz and D. M. Gruen J. Amer. Chem. Soc. 1949,71,2106. l1 T. K. Keenan and F. H. Kruse Inorg. Chem. 1964,3 1231. l2 G. Gibson D. M. Gruen and J. J. Katz J. Amer. Chem. SOC. 1952,74,2103. l 3 G. H. Dieke and A. B. F. Duncan Spectroscopic Properties of Uranium Compounds Natl. Nuclear Energy Ser. Div. 111 1949 2 139. l4 M. B. Brodsky private communication. l5 E. Kankeleit Rev. Sci. Instr. 1964 35 194. l6 J. A. Stone and W. L. Pillinger Bull. Amer. Physic. Soc. 1966 11 809 ; to be published. l7 D. A. Shirley Rev. Mod. Physics 1964 36 339. l8 A. Bohr Kgl. Danske Vid. Selsk. Mat.-fys. Medd. 1952 26 no. 14. l9 L. Wilets D. L. Hill and K. W. Ford Physic. Rev. 1953 91 1488. 2o J. 0. Newton Nucl. Physics 1958 5 218. 21 P. Brix S. Hiifner P. Kienle and D. Quitmann Physics Letters 1964 13 14Q 22 G. J. Perlow and M. R. Perlow J. Chem. Physics 1966,45 2193.

ISSN:0430-0696    

DOI:10.1039/SF9670100077

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

GENERAL DISCUSSION Prof. N. N. Greenwood (Newcastle upon Tyne) said Is some of the line broadening in 237Np reported by Stone due to nonstoichiometry and differing site symmetries in the americium source compound? Would it be possible to choose a soluble stoichio- metric ionic coordination compound of americium and use a frozen solution as a source to achieve narrower lines? Dr. J. A. Stone (du Pont de Nemours & Co. S. Carolina) said Factors to be considered in using a frozen-solution 241Am source are (a) the effect of recoil after alpha decay on the bonds in the molecule and (b) the possible reduction of recoilless fraction in such a system. The use of a frozen solution as an absorber to achieve line narrowing is known particularly in lS1Eu systems. However so far as I know the use of a frozen solution as a source has not been reported; the question of line- narrowing in this case could only be answered by experiment.Prof. J. Danon (Rio de Janeiro) said Why was 241Am chosen as source for the Mossbauer effect with 237Np instead of 237U? With the latter one would expect a Mossbauer nucleus in a more “ normal ” situation since it is formed by beta-decay instead of the much more energetic alpha-decay of 241Am. The 4f electrons in the lanthanides have less expanded radial function than the 5felectrons in the actinides. Is it correct to expect a larger inner shielding effect on the closed s-shells by the 4f electron of the lanthanides than by the 5felectrons of the actinides ? Prof. N. N. Greenwood (Newcastle upon Tyne) said In reply to Danon’s comment on Stone’s paper chemical shift depends on the magnitude of both St+b(0)2 and 6R/R (or Sp/p) ; it is this second term which is much larger for 237Np than for lslEu or 3Eu.Prof. J. F. Duncan (Victoria University of Wellington N.Z.) said Does the evidence presented by Stone of isomer shift values for different oxidation states of neptunium provide any confirmation of the view that this element should be con- sidered as a member of the actinide series of elements only in low oxidation states? From the regularity of the isomer shift with number of d electrons (fig. 3) it appears that a change from 6d to 5foccupancy does not take place. If this conclusion could be unequivocally drawn it would be important as it is contrary to the currently accepted view in this region of the actinide series and implies that neptunium is not a member of the series.Secondly in some cases (e.g. the iron phosphides) it has been observed that the nuclear hyperfine field is proportional to the observed magnetic moment as con- ventionally measured by the Gouy method and follows the same temperature depend- ence. If one could establish under what conditions this is found to be the case a new method of determining the spin types of paramagnetic ions of Mijc;sbauer nuclei would be possible. In view of the importance of this proposed technique to general inorganic chemistry has Stone any knowledge of hyperfine fields being compared with magnetic moments in systems of his type or others and what results were obtained ? Dr. J. A. Stone (du Pont de Nemours & Co. S. Carolina) said In reply to Duncan we do not think that the isomer shifts for the different valence states of neptunium 83 84 GENERAL DlSCUSSION offer any evidence to contradict the actinide hypothesis.Qualitatively the effect of 6d electrons on isomer shifts should be similar to that of 5felectrons i.e. both would be expected to shield the closed-shell 6s electrons from the nucleus. From the data available however it is not possible to distinguish between configurations involving 5f electrons and those involving 6d electrons. Few data on the magnetic properties of neptunium compounds are available and indeed the Mossbauer technique may provide the first magnetic information for many of these materials. For the compounds listed in our paper the temperature depend- ence of the hyperline fields has not been measured although studies of this type are planned.For one compound NpO, there is sufficient information to establish a degree of correlation between the hyperhe field and the magnetic moment. Heat- capacity and magnetic susceptibility measurements indicated that NpO undergoes a magnetic transition at 25°K ; however no evidence for this transition was found in neutron-diffraction measurements 3* or the early Mossbauer data.5 The discrepancy has been resolved by the Argonne group who performed careful measurements of the temperature dependence of the Mossbauer spectrum ; discontinuities in the linewidth and the isomer shift were noted at the magnetic transition temperature. Here the very small magnetic moment of the neptunium atoms is accompanied by a correspondingly small hyperhe field (0.05 x lo6 gauss).This is in contrast to the fields we find in most other compounds of neptunium which are greater than 3 x lo6 gauss. Dr. P. Gatlich (Technische Hochschule Darmstadt) said Stone has looked at the chemical consequences associated with the a-decay of 241Am. The results found are indicative of different oxidation states of 237Np in the source due to the recoil effect. Has he also looked at the chemical effects caused by the p-decay of 237U and if so how do those compare with the chemical effects in case of the a-decay of 241Am? One would expect different findings since the recoil energies involved in the two nuclear transformations have quite different magnitudes. Stone also mentioned some disadvantages in the use of 237U as a source. Under certain circumstances however 237U might be the better choice from the econo- mical point of view.At the Technische Hochschule Darmstadt we have a linear accelerator where we can carry out y,n-reactions on 238U giving 237U. So we would not have to buy the rather expensive 241Am. Does Stone think that the disadvantages of 237U as a Mossbauer source are such that he would not recommend making use of the linear accelerator to obtain 237U? Dr. J. A. Stone (du Pont de Nemours & Co. S. Carolina) said In reply to Giitlich our earlier work with 237U was aimed primarily toward finding a suitable single-line source and to this end nearly all the sources studied were in host lattices of NpO, which was successful or Tho2 which was not. No attempt has been made to prepare 237U sources in a variety of chemical environments in order to study recoil effects following p-decay.Such a study would be of considerable interest for radiation chemistry. E. F. Westrum Jr. J. B. Hatcher and D. W. Osborne J. Chem. Physics 1953 21,419. J. W. Ross and D. J. Lam J. Appl. Physics 1967,38 1451. D. E. Cox and B. C. Frazer J. Physics Chem. Solids 1967 28 1649. L. Heaton M. H. Mueller and J. M. Williams J. Physics Chem. Solids 1967 28 1651. J. A. Stone and W. L. Pillinger Physic. Rev. Letters 1964 13 200. B. D. Dunlap G. M. Kalvius S. L. Ruby M. B. Brodsky and D. Cohen Con$ Nuclear Hyper- fine Interactions (Asilomar California 1967) to be published. GENERAL DISCUSSION 85 The principal disadvantages of 237U as a Mossbauer source are (i) because of the short half-life periodic renewal of the source is required and (ii) the gamma-ray spectrum is more complex than that of 241Am.We thus chose to use 241Am sources as a matter of convenience. However the disadvantages of 237U sources are not so serious as to eliminate them from consideration and under some circumstances they might be the better choice. One advantage to be expected from 237U sources is that in general they should exhibit larger recoilless fractions than the 241Am sources. It would be interesting to see if useful Mossbauer sources could be made with the (y,n) reaction; but sources made in this way would have low specific activity since the 237U activity could not be chemically separated from the 238U target. Our knowledge of sources for 237Np Mossbauer spectroscopy is still rudimentary ; many lines of investigation are available which may lead to improved sources.Dr. G. M. Bancroft (University of Cambridge) said I have a number of points to to raise on the paper by Stone et al. relating to the practical nature of Np Mossbauer work. Is theffactor known? It appears that iffis about 0.1 a good fraction of the broadening observed could be due to thickness effects. Is this thickness needed for a good absorption? This amount of Am would appear to be rather difficult to handle in a normal radiochemical laboratory. Dr. J. A. Stone (du Pont de Nemours & Co. South Carolina) said In reply to Bancroft the measurement of recoilless fractions for the 237Np resonance presents a number of difficulties and thus far no determination of these quantities have been made. However we have done experiments with NpO absorbers spanning the range of thickness from 7 to 300mg/cm2; we found essentially no change in line- width over this range and the areas of the resonances were proportional to thickness. We therefore believe that there are important causes of line-broadening present in addition to thickness effects. Because the resonances are weak in many of the Np compounds studied thick absorbers were used to enhance the absorption ; resonances with only 0.1 % change in count-rate are typical. In amounts up to 200 mg 237Np can be handled safely on a radiobench ; however 241Am sources should be prepared in a glove box. A properly encapsulated 241Am source should present no unusual problems in handling.

ISSN:0430-0696    

DOI:10.1039/SF9670100083

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Mgssbauer Spectroscopy of Stereochemically Non-Rigid and Related Organometallic Molecules BY ROLFE H. HERBER School of Chemistry Rutgers-The State University New Brunswick New Jersey 08903 U.S.A. Received 8th September 1967 The Mossbauer parameters for cyclo-octatetraene iron tricarbonyl [COTFe(CO)3] and a number of related iron-tricarbonyl complexes have been determined both for the neat solids and for frozen solutions using solvents which set to a " glassy " matrix. The quadrupole splitting parameter is found to be diagnostic with respect to distinguishing between the ligand symmetry of the hydrocarbon framework around the Fe(CO)3 moiety. The Mossbauer data for COTFe(C0)3 in conjunction with new low temperature n.m.r. data on a number of iron-tricarbonyl complexes indicate that the cyclo-octatetraene ring acts as a 1,3-diene in solution at low temperatures in consonance with the bonding information derived from X-ray diffraction data on the neat solid.Moreover the usual isomer shift-quadrupole splitting correlation diagram for these absorbers shows that the Mossbauer measurements can serve as a useful bridge relating structural information derived from measurements on solids and those requiring liquid or solutions samples. Detailed measurements of both the s7Fe and 'I9Sn resonances in [x-C5H5Fe(CO),I2SnCl2 and X - C ~ H ~ F ~ ( C O ) ~ S ~ C ~ ~ in glassy matrices as well as the neat solids show that low lying optical modes which are normally observed in the far infra-red (10-200 cm-') region of the spectrum can be studied by Mossbauer methods. That such optical modes contribute to the Mossbauer-Lamb recoil-free fraction is deduced from the non- constancy of the normalized area ratio for the two resonance spectra in different matrices in con- sonance with a simple Debye model for the lattice dynamics of the Mossbauer atom.The most reliable information concerning the structure of chemical compounds is obtained from X-ray diffraction data on single crystal samples and with the availablity of computer linked diffractometers such data have become increasingly available. However X-ray diffraction data are not always appropriate in solving structural problems due to an intrinsic interest in non-solid or amorphous materials the diffi- culties of growing single crystals of adequate size the lack of resolution and indeter- minacy of proton positions and it is for these reasons inter alia that the structural information is frequently derived from other spectroscopic techniques principally infra-red nuclear magnetic resonance and Raman data.Such techniques invariably depend on a comparison of experimental spectra to those obtained for model com- pounds which have in turn been examined by X-ray methods. Since X-ray diffraction gives structural information on the solid state modification of the compound while infra-red n.m.r. and Raman spectroscopic data are generally obtained on solution samples the reconciliation of the structural information by the two types of studies assumes a conformational integrity of the molecule in the two states. A number of examples involving molecular conformation have become the subject of speculation since the interpretation of the X-ray data and the n.m.r.(infra-red) data seemed to be at variance with each other. Mossbauer spectroscopy provides an effective bridge between solid state and solu- tion methods since it is possible to determine the resonance parameters of the neat solid as well as of frozen solution samples of the same molecular species and thus to obtain information concerning conformational integrity from such data. The present study is an investigation of the solution conformation of cyclo-octatetraene iron tri- carbonyl COTFe(C0)3 and related metal-organic compounds. 86 R. H. HERBER 87 EXPERIMENTAL The iron-organic compounds were prepared according to literature meth0ds.l The solvents in the solution studies-methyltetrahydrofuran (MTF) nitrobenzene and n-octane- were freshly distilled prior to use.EPA (16 "/o v/v ethanol 42 "/o v/v i-propanol and diethyl ether) was freshly prepared from spectroscopic-grade reagents. The Mossbauer measurements were carried out using a constant acceleration drive in conjunction with a 400 channel analyzer operated in the multiscaler mode and have previously been described.2 Velocity calibrations were based on natural metallic iron foil Mossbauer spectra making use of the magnetic hyperfine splitting data of Hanna et d3 Isomer shifts are reported with reference to the centroid of the Mossbauer spectrum obtained using an N.B.S. standard sodium nitroprusside absorber at 294°K in conjunction with the j7Co- in-palladium source used for the measurements here reported. '' Frozen solution " samples were prepared by dissolving -50 mg of the sample compound in 2.0 ml of solvent and then transferring this material to a liquid sample cell by means of a hypodermic syringe injection system.The sample cell which consisted of a solution volume of about 40 mil thickness bounded by two 3 mil thick aluminum windows was immersed immediately after filling into liquid nitrogen and maintained at this temperature until mounted in the experimental dewar for measurement. All spectra were obtained with the samples at liquid-nitrogen temperature using the source at room temperature (294 &2"K). RESULTS Fig. 1 shows the spectrum of solid COTFe(CO) and consists of a well-resolved doublet for which the isomer shift and quadrupole splitting parameters are calculated by the Typical Mossbauer spectra for the compounds are shown in fig.1 2 and 3. e I l l t l l l l l l l -2.0 -I.o 0 I.0 20 30 isomer shift mm sec-I with respect to SNP at 294°K. FIG. 1 .- Mossbauer spectrum of C8H8Fe(C0)3 at liquid-nitrogen temperature. The isomer shift is usual method-of-chords treatment.4 These parameters are summarized in table 1 and are in reasonably good agreement with the values reported earlier.5 The asym- metry in the intensity of the quadrupole doublet arises from the accidental preferen- tial crystal orientation of the sample (which consists of long red-black needles or platelets) and has no particular relevance to the present study. A sample of 88 ORGANOMETALLIC MOLECULES COTFe(CO) ground with powdered quartz prior to mounting in the sample holder showed essentially equal intensities for the two peaks leading to the conclusion that the Gol’danskii-Karyagin ti asymmetry is very small or absent in this compound.Fig. 2 shows a spectrum typical of the “ frozen solution ” samples ; it consists of an asymmetric resonance pattern. This resonance curve can be resolved into a - 2.0 - 1.0 0 +I*O +2*0 +3*0 Doppler velocity mm/sec lower data curve represents the “ singlet ” absorption (see text). M. C. A. address FIG. 2.-Mossbauer spectrum of C8H8Fe(C0)3 in nitrobenzene at liquid-nitrogen temperature. The 60 70 80 90 100 110 120 130 140 150 24.0 23-5 ct 23.0 x 22.5 E 22.0 r a 0,45mm/sec c -ri. Y I 2 0 l.S.=+0.265 f0.060 mm/sec Q.S.. 0*618*0+060mm/sec d 35.0 8 34.0 1.5.*+0*278 f0~060mmlsec 33.0 Q.S. 0 6 0 6 *0~060mm/sec P 5 0.43mmlsec EPA Solution 32.0 -2.0 -1.0 0 +I.O +2.0 $ 3 0 Doppler velocity mm/sec is that of the neat solid the lower that of the EPA solution.FIG. 3.-Mossbauer spectrum of C4H6Fe(CO) at liquid-nitrogen temperature. The upper spectrum R . H. HERBER 89 doublet which is nearly identical with that observed for the neat solid and a broad absorption peak which may be a true singlet or a poorly resolved doublet of small quadrupole splitting. Under the foregoing assumption the isomer shift and quad- rupole splitting parameters have been determined for the resolved doublet and these values are included in the data summarized in table 1 and in the correlation diagram given in fig. 4. TABLE 1 .-M~SSBAUER PARAMETERS FOR RFe(C0)3 COMPOUNDS AT 78°K R isomer quadrupole solvent shift b splitting mmlsec mm/sec cyclo-octatetraene I +0.31 1 -24 EPA + 0.34 1.16 C6H5N02 + 0.36 1.21 n-CsH1s + 0.36 1.23 MTF + 0.35 1 -20 1,s-dicarboinethoxy COT CsH 6(C02CH3)2 - + 0-33 1 -27 COT lactone +0*33 1 *25 MTF + 0-33 1.17 norbornadiene - + 0.29 2.1 5 cyclo-octa-l,5-diene - + 0.23 1 -83 CsH 23[Fe(CO) 31 2 - + 0.24 1-32 cycle but adiene - 3.0.28 1 -54 EPA + 0.26 1 *55 a EPA = 16/42/42 % v/v ethanol+ 1-propanolS diethyl ether ; MTF = methyltetrahydrofuran.b from Na2[Fe(CN),(NO)] . 2Hz0 at 294+ 1 "K. c all values reported to f0-03 mmlsec. - 1 I I i 1 I I I 060 1.00 1.40 1.80 2.20 Q.S. mmlsec. FIG. 4.-Correlation diagram for the absorbers discussed in the text. Fig. 3 shows the resonance curve obtained for trimethylene methane iron tricar- bony1 both as the neat solid and in frozen EPA solution and shows no evidence for the type of " singlet " absorption noted in the frozen solution spectra of COTFe(CO), C4H4Fe(C0)3 and C8H,(COOCH,)Fe(CO),.The nature of this '' singlet " absorp- tion is not clearly understood although the following facts appear reasonably well established. (a) The " singlet '' absorption accounts for 15-20% of the total area 90 ORGANOMETALLIC MOLECULES under the resonance curve and its intensity does not appear to be related to the basicity (or electron-donating property) of the solvent since it is observed with nearly equal relative intensity in COTFe(CO) solutions in EPA nitrobenzene and n-octane. (b) The centre of the " singlet " absorption peak lies in the range +0.48+0-03 mm/sec with respect to the sodium nitroprusside standard. This isomer shift (I.S.) is more positive than that of the parent compound but well within the range of other metal organic absorbers especially those containing a n-cyclopentadienyl ligand (e.g.[n(C,H,)Fe(CO),] has an I.S. of +Om48 mm/sec). (c) The solid recovered from the solution sample from which the solvent has been removed by vacuum (room- temperature) distillation shows a resonance spectrum consisting only of the doublet resonance pattern. From this is may be inferred that the mechanism which gives rise to the species responsible for the " singlet " absorption in the frozen solution samples is completely reversible at ambient temperature. DISCUSSION Cyclo-octatetraene iron tricarbonyl 7-9 has a room-temperature proton magnetic resonance spectrum which consists of a single sharp line. From this result a structure consisting of an Fe(CO) group bonded symmetrically to an essentially planar CSHs moiety in an open-faced sandwich configuration was proposed.'* ' 9 lo However an X-ray crystallographic study l1 showed that in the solid state the Fe(CO) group was bonded to only four of the eight carbons of the ring and moreover that the hydro- carbon fragment consisted of two planar segments having a dihedral angle of 42".On the basis of this information the n.m.r. data were accounted for by postulating a stereochemical non-rigidity which operated in such a way that averaging of all of the protons occurred on a time scale rapid compared to the typical n.m.r. scanning rate (60 Mc/sec).' This idea was substantiated by subsequent low-temperature n.m.r. observations l2-I4 which showed that as the temperature was lowered the sharp singlet gradually broadened and ultimately (at about - 155°C) reappeared as two well- separated still asymmetric resonances.The two resonances are clearly due to two sets of protons those belonging to the four carbon atoms involved in coordination to the metal and those belonging to the four uncoordinated carbon atoms. However starting from this assignment three different structures have been proposed for the species present in solution and these will be referred to as the 1,3-biplanar,12 the 1,3-tub l4 and the 1,5-tub l3 configurations respectively. have permitted a resolution of the structural controversy concerning the configurational integrity of COTFe(CO) in the solid and solution states. From the data summarized in table 1 and fig. 4 it is evident that the quadrupole splitting parameter for CSM8Fe(C0)3 solid and for the frozen solutions in EPA MTF nitrobenzene and n-octane are all identical within the experimental error of 40.03 mmlsec.Moreover this splitting is the same as that observed for solid ~&,(COOCH,),Fe(CO) and for C8H6(C02CH2)Fe(C0)3 both as the solid and in frozen MTF solution. The proton n.m.r. spectrum of the diester consists of three sets of chemically non-equivalent equally intense ring proton resonances centred at T = 2.72,4.91 and 5-91 and leads unambiguously to a 1,3-diene structure assignment. The spectrum of the lactone although more complex than that of the diester due to the presence of two chemically non-equivalent but barely resolvable protons in the lactone ring again leads to a consistent assignment of the Fe(CO) moiety being bonded to a 1,3-diene hydrocarbon fragment.Moreover the n.m.r. spectra of the diester and lactone recorded over a temperature range of -60" to + 120°C do not show any The present Mossbauer data together with additional synthetic and n.m.r. studies,l R. H . HERBER 91 appreciable change in appearance in contrast to the observations made on the stereochemically non-rigid cyclo-octatetraene compound. In contrast to this essential constancy of the Q.S. parameter for the absorbers referred to above it is of interest to consider two other compounds in which the Fe(CO) group is bonded to a cyclic carbon framework but of different configuration. The cyclo-octadiene Fe(CO) compound CsHl 2Fe(CQ)3 is known from n.m.r. evidence to have a 1,5-diene configuration.16 The Q.S.value obtained for this solid is 1.83 mm/sec and this 50% increase in the magnitude of the hyperfine interaction leaves little doubt concerning the sensitivity of this Mossbauer parameter with respect to conformational change. In this context the largest Q.S. which has been observed in this series of compounds is that for norbornadienyl irontricarbonyl C7HsFe(CO) in which the metal atom is bonded to a non-alternant cyclic diolefin in which the two double bonds are in a 1,4 position relative to each other. (a) There is no major conformational change reflected in the Mossbauer data between the solid COTFe(CO) sample and the various frozen solution matrices which have been examined. (b) The frozen solution Mossbauer spectra do show evidence for the (apparently reversible) formation of a new chemical species which accounts for 15- 20 "/o of the iron containing material.Although the nature of these species is not well understood its formation does not appear to be sensitively related to the basicity of the solvent. Consequently any differences between the infra-red spectra of the solid and of the solutions1 must be interpreted with caution with respect to the structural in- tegrity of the parent compound ; * (c) From the Q.S. systematics for the COTFe(CO) systems which have been studied in the present investigation and the parameters for iron-tricarbonyl diolefin compounds of known configuration of the three structures which have been proposed for the low-temperature solution configuration only the 1,3-biplanar model 12* l7 is consistent with the present results.Finally a qualitative rationalization of the relative magnitudes in the Q.S. para- meter in 1,3- and 1,5-diene complexes of iron tricarbonyl can be based on the assump- tion that the major contribution to the electric field gradient tensor arises from the ligand symmetry (rather than from the symmetry of non-bonding metal atom orbitals). Dickens and Lipscomb l1 have suggested that the bonding in solid CQTFe(CO) can be described as giving rise to a quasi-octahedral configuration around the iron atom with two CT bonds to the two " interior " carbon atoms of the CsHs ring and ,u bond to the n-electron system localized between the two " outside " carbon atoms of the ring and the remaining three bonds being to the three carbonyl ligands. Assuming a nearly equivalent bonding interaction to the ligands this symmetry is expected to lead to a relatively small electric field gradient which would vanish if the six ligands were identical and located at regular (undistorted) octahedral positions.Per contra in the 1,5-diene configuration the bonding consists of three metal-carbonyl bonds and two metal-olefin rc bonds to give a quasi 5-coordinate structure. Even in the limiting case of 5 identical ligands with a regular trigonal-bipyramid structure such a configuration The above results lead to the following interpretation of the available data. *Note added in proof Subsequent measurements have shown that a significant contribution to the " singlet " absorption in the frozen solution spectra arises from the presence of a small amount of iron impurity in the aluminum foil used to shield the photomultiplier tube and to provide good thermal contact between the sample and the cold finger of the dewar.The reason this absorption (which amounts to a resonance effect of 4 - 8 %) is observed only in the frozen solution spectra is due to the magnification of the vertical (counts per channel or percent transmission) scale in these runs due to the small resonance effects in the sample itself. The 57Fe resonance observed for the soild absorbers is usually -10 % and the impurity resonance is masked by these large effects whereas in the solution spectra in which the resonance absorption is usually -1 % the iron contamination is easily seen (and corrected for). The dashed lines in fig. 2 and 6 reflect this correction. 92 ORGANOMETALLIC MOLECULES leads to large Q.S.values relative to the 6-coordinate case. (Compare e.g. Fe(CO) and Fe2(C0)9 for which the hyperfine splittings are 2-6 and 0.5 mm/sec respectively.) On this basis the increase in this parameter from N 1.2 mm/sec for COTFe(CO) to 1.83 mm/sec for CsH12Fe(C0)3 is consistent with the formulation of the former as a 1,3-diene and the latter as a 1,5-diene complex. OPTICAL MODE DEPENDENCE OF MOSSBAUER f VALUES an analytical expres- sion for the relationship between the recoil-free fraction (Mossbauer-Lamb factor) f and the Debye temperature of the solid 6 From the Debye theory of solids it is possible to derive and 0 is the lattice (Debye) tempera- in which ER is the free-atom recoil energy - 2mc2 ture of the matrix. E? In the low-temperature limit the integral goes to n2/6 so that 19* 2o Mossbauer measurements extended to very low temperatures relative to 8 have confirmed the temperature independence offin the limit and such data can be used to obtain information related to lattice dynamics of perfect crystals and related solid st ate parameters.At higher temperatures? the integral in eqn. (1) approaches 6/T so that Eqn. (5) can be used to determine 8 for a given lattice from temperature dependence measurements on f (or more commonly from measurements of the temperature dependence of the areas under the resonance curves 21) since d In f/dT = -6E,/k6' so that For a real system in which the Mossbauer atom does not fulfil the requirements of the Debye model the temperature so calculated should be identified as a " Mossbauer temperature " 21 and will be referred to by the symbol OM in the following discussion.The Mossbauer temperature can be related to a vibrational frequency 22 by the Einstein relationship v = kO,/h. (8) R. H. HERBER 93 This frequency is related to the Mossbauer-Lamb factor fnot only through the value of OM but also in terms of the mean-square amplitude of vibration (x2) since f = exp - [4n2(x2>/A2] (9) in which 2 is the equivalent wavelength of the Mossbauer gamma ray. If the vibra- tional amplitude of the Mossbauer atom arises only from vibrational modes of the whole crystal lattice assembly the excitation is considered to be acoustic and can be related (at least in principle) to the bulk properties of the solid. On the other hand we may speculate on what kind of local (optical) vibrational modes are needed to account for the temperature dependence of the recoil-free fraction.From eqn. (8) it is seen that if the lowest lying optical mode which involves motion of the Mossbauer atom occurs at 35 cm-l O M N 50°K. Similarly if the lowest lying optical modes are at 70 100 or 200 cm-l the corresponding Mossbauer tempera- tures are 100 144 and 287"K respectively. That such optical modes do indeed contribute to the temperature dependence off (although they do not necessarily account for the complete dependence 21) is shown by the following argument. We consider a molecular solid which contains two distinct Mossbauer atoms A and B which are linked to each other (not necessarily directly) through normal chemical bonds. The ratio of the recoil-free fractions for the two atoms is given from eqn.(5) in the high-temperature limit by which for T/O21 reduces to If only acoustic modes contribute to the lattice temperature then the lattice tempera- ture will be the same at A sites and at B sites i.e. OA = & so that (11) becomes (12) i.e. the ratio of the recoil-free fractions is independent of the lattice temperature and is given by the ratio of the free atom recoil energies. On the other hand if local (optical) vibrational modes make a significant contribution to the mean square amplitude of vibration then OA # 8 and hence lnfA/lnfB will depend on the nature of the matrix in which the Mossbauer atoms are embedded. To explore the significance of optical modes in the Mossbauer study of organo- metallic compounds the compound 2 3 [n-C5HSFe(CO)2]2SnC12 (I) has been examined in detail since it is a readily prepared air-stable material for which both the 57Fe and ll%n resonances can be determined with considerable preci~ion,~ and for which a detailed X-ray crystallographic study has been published.24 In addition the related compound 23 ~ T - C ~ H F ~ ( C O ) ~ S ~ C ~ (11) has also been studied.In f J n f B = ERAIERB EXPERIMENTAL The Mossbauer methodology was identical to that referred to in the first part of this paper. Samples of the absorber in the paraffin and polymethylmethacrylate matrices were prepared by orthodox methods. DISCUSSION The Mossbauer spectrum of the neat solid absorber (I) is shown in fig. 5 in which the upper curve pertains to the 57Fe resonance and the lower curve to the ll'Sn reson- ance. The isomer shift and quadrupole splitting parameters extracted from these I I I i I I I I I 1 - - 99- - Fe - POlY.98 - - FIG. I 100- Sn - poly. 99- - I I I 1 I I I I I I I I . I I I I I I I I 1 60 100 140 I80 220 analyzer address 5.-Mossbauer spectrum of [ x - C ~ H ~ F ~ ( C O ) ~ ] ~ S ~ C ~ ~ at liquid nitrogen temperature. The 40 80 120 160 200 analyzer address FIG. 6.-Mossbauer spectra of [ X - C ~ H ~ F ~ ( C O ) ~ J ~ S ~ C ~ ~ in polymethyl methacrylate at liquid- nitrogen temperature. The upper curve is the "Fe resonance the Iower curve the lr9Sn resonance. for (I) in a polymethylmethacrylate (Lucite) matrix are shown in fig. 6 from which it is seen that while the 57Fe resonance can still be analyzed in terms of the hyperfine parameters the 13Sn resonance effect is barely discernible (although - lo6 c/channel have been accumulated in this experiment).Using the same data analysis method as above the normalized area ratio for (I) in the polymethylmethacrylate matrix is A(Sn)/A(Fe) = 0.60. This decrease by a factor of -60% in the area ratio suggests in consonance with eqn. (ll) that 8Fe # OSn presumably due to the contribution by local (optical) modes to the vibrational excitation of the two Mossbauer atoms. R . H . HERBER 95 TABLE 2.--MOSSBAUER PARAMETERS FOR (Fc Sn) METAL ORGANIC COMPOUNDS AT 80°K isomer shift quadrupole splitting compound matrix a resonance inrnlsec * mm/sec [n-Cs€fsFe(CO)21~SnCI 2 neat neat P P MTHF neat neat P P MTHF MTHF Fe Sn Fe Sn Fe Fe Sn Fe Sn Fe Sn -1-0'359 f O * O l O +1-95 f0-02 +0.351 f0.010 +1*96 ZO-02 +1.31 10*02> 4- 0.360 &O*O 10 -to401 f0-010 +1*74 10.02 +0.381+0.010 +1*66 3=0.02 +0*394&0*010 -1-1.54 *0*05 1.676 1 0.0 10 2 2 8 i0.02 1-74 f0.05 2.25 &0*07 2.29 f0.02) 1.68 t0.01 1.865f0-010 1.77 +0.02 i.779 Zo.oio 1.78 f0.02 1 -764&0*010 1.64 10-05 p = polyrnethylmethacrylate MTHF = methyl hydrofuran.Fe isomer shifts with respect io sodium nitroprusside dihydrate Sn iso:ner shifts with respect to SnOz. This doublet tin resonance has been discussed eisewhere (R. Ef. Herber and Y. Goscinny Inorg. Chem. in press) Similar data for compound (11) in the neat solid and in a frozen methyl tetrahydro- furan glass are shown in fig. 7 and the numerical data are summarized in table 7. (It was not possible under the ex- pcrimental conditions employed to observe a tin resonance in frozen methyltetrahydrofuran solutions of compound (I).The polymethyl- methacrylate matrix was chosen to provide a structureless (glassy) matrix for the solute with a higher lattice temperature than obtains in the furan system.) From the reso- nance data for (11) the normalized area ratio for the solid in A(Sn)/ A(Fe) = 5.91 while that for the frozen solution is A(Sn)/A(Fe) = 2.67 and these results again suggest the participation of local vibrational modes in the excitation of the resonant atoms in these solids. These results complement the earlier measurements on stannic iodide 21 in which the double reso- nance data (involving 'leSn and 1291) were determined over a tem- perature range of 85-220°K and showed that the lattice tempera- tures at the metal atom sites and the halogen atom sites were not identical.From the present data on the two metal-organic com- pounds which have been studied it is inferred that a major contri- bution to this non-equivalence of the lattice temperature at the two sites derives from participation of optical modes in the Mossbauer atom vibrational excitation. .;i 20 60 I00 1-40 I80 analyzer address FIG. 7.-Mossbauer spectra of x-C5H,Fe(C0)2SnC13 at liquid-nitrogen temperature. The upper two curves pertain to the neat solid the lower two curves to the frozen methyltetrahydrofuran solutions. 96 ORGANOMETALLIC MOLECULES Extension of these studies to determine the temperature dependence of the recoil- free fraction in a variety of glassy matrices should serve to elucidate the role played by low-lying optical modes in determining the magnitude of the resonance effect.Such data in turn have applications to the determination of the conformations of molecular species (in which the optical modes are dependent on structural paraneters) and the integrity of conformation in the solid and solution states as discussed in part 1 of this paper. The author is indebted to Professors S . J. Lippard and R. Breslow and Mr. R. Grubbs for the collaboration on the COTFe(CO) conformational problem discussed in part 1 and to Prof. 6. Emerson for several of the samples used in this work. Thanks are also due to Mr. A. Hoffman for the synthesis of several of the organo- metallic compounds and to Mr. Y. Goscinny for assistance with Mossbauer measure- ments. This research has been supported in part by the U.S. Atomic Energy Commis- sion and this paper constitutes document NYO-2472-48.In addition support by the Petroleum Research Fund of the American Chemical Society is similarly gratefully acknowledged. R. Grubbs R. Breslow R. H. Herber and S . J. Lippard J. Amer. Chem. Soc. in press. The author is indebted to Prof. G. Emerson for the samples of cyclo-octa-1,5-diene iron tricarbonyl cyclobutadienyl iron tricarbonyl and trimethylene methane iron tricarbonyl used in this study. ' G. K. Wertheim Mossbauer Efect Principles and Applications (Academic Press New York 1964). G. K. Wertheim and R. L. Cohen Applications of the Mossbauer Effect in Chemistry and Solid State Physics Tech. Repts. Series no. 50 (Int. Atomic Energy Agency Vienna 1966). S. S. Hanna et al. Physic. Rev. Letters 1960,4 177. R. S. Preston S. S. Hanna and J. Heberle Physic.Rev. 1962 128 2207. H. Brafnian M. Greenshpan and R. H. Herber Nuclear Instruments and Methods 1966,42,245. R. H. Herber Chemical Aspects of Mossbauer Spectroscopy in Progress in Inorganic Chemistry vol. 8 F. A. Cotton ed. (Interscience New York 1967). G. K. Wertheim and R. H. Herber J . Amer. Chem. Sac. 1962 $4 2274. V. I. Gol'danskii E. F. Makarov and V. V. Khrapov J. Expt. Theor. Physics Acad. Sci. U.S.S.R. 1963 44 752. T. A. Manuel and F. G. A. Stone Proc. Chem. Soc. 1959,90; J. Arner. Chem. SOC. 1960 82 336. A. Nakamura and N. Hagihara Bull. Chem. Soc. Japan 1959,32 880. * M. D. Rausch and G. N. Schrautzer Chem. and Znd. 1959,957. lo F. A. Cotton J . Chem. SOC. 1960,400. l1 B. Dickens and W. N. Lipscomb J. Amer. Chem. Soc. 1961,83,489. l2 C. G. Kreiter A. Maasbol F. A.L. Anet H. D. Kaesz and S . Winstein J . Amer. Chem. Soc. l 3 F. A. Cotton A. Davison and J. W. Faller J. Amer. Chem. SOC. 1966,88,4507. l4 C. E. Keller B. A. Shoulders and R. Pettit J . Arner. Chem. SOC. 1966,88,4760. lS These studies have been carried out in collaboration with R. Grubbs R. Breslow and S. J. l6 R. Pettit J. Amer. Chem. Soc. 1959 81 1266. l7 F. A. L. Anet H. D. Kaesz A. Maasbol and S. Winstein J. Amer. Chem. Soc. 1967,89,2489. l8 R. L. Mossbauer and W. H. Wiedemann 2. Physik 1960 159,33. l9 A. H. Muir Jr. Tables and Graphs for Computing Debye- Waller Factors in Mossbauer Effect Studies AI-6699 (Atomics International California 1962). 2o R. H. Nussbaum Interpretation and Applications of Mossbauer Fraction Experiments in Mossbauer Effect Methodology vol. 2 I. J. Gruverman ed. (Plenum Press New York 1966). 21 S. Bukshpan and R. H. Herber J. Chem. Physics 1967,46 3375. 22 For a detailed discussion see G. K. Wertheim Mossbauer Effect Principles and Applications 23 F. Bonati and G. Wilkinson J. Chem. SOC. 1964 79. 24 J. E. O'Connor and E. R. Corey Inorg. Chem. 1967 6,968. 1966,88,3444. Lippard of Columbia University and are discussed in detail in ref. (1). F. A. L. Anet J. Amer. Chem. Soc. 1967 89 2491. (Academic Press New York 1964) chap. IV.

ISSN:0430-0696    

DOI:10.1039/SF9670100086

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

GENERAL DISCUSSION Prof. R. H. Herber (Rutgers-The State Utiiversity) said Since completion of the paper for this Symposium some additional data on the compounds [(n-C5H5)Fe (C0),l2 SnCI (I) and n-CSHS Fe(CO) SnCl (11) have been obtained by Y. Goscimy and myself. This work was prompted by the unusual crystal structure of (I) which was reported by O'Connor and C0rey.l In the solid the Sn-Fe bond length of 2.492 A is shorter than that reported for any other tin-iron metal-metal bond while the Sn-Cl bond length of 2.43A is 0.12A longer than it is in SnCl and 0-09A longer than it is in (CH,),SnCl. The main question which we attempted to resolve was whether these bond lengths as well as the considerable deviation of the C1-Sn- C1 bond angle (94.1") and the Fe-Sn-Fe bond angle (128.6") from tetrahedral values was due to the stacking of the molecules in the crystal or a reflection of the unusual bonding in this compound.Mossbauer spectroscopy may be used to eluci- date uniquely this question since it is possible to examine both the neat crystalline solid as well as frozen glassy matrix solutions as I have already pointed out in my paper. Extensive measurenients on (I) and (11) in glassy matrices of poly-methylmetha- crylate (Lucite perspex) and/or methyl-tetrahydrofuran have shown that the para- meters for the Fe resonances in either compound change little in going from the neat solid to the solution matrix nor does the Sn resoiiance in (11) show any major differ- ence in the isomer shift or quadrupole splitting parameters in the two media. The Sn resonance in (I) however undergoes significant broadening which from high statistical accuracy data can be decomposed into a pair of doublets of almost the same isomer shift but with a difference of the Q.S.parameter of - 1-30 mrnlsec at 93°K. These results are interpreted in terms of a cis-trans isomeric pair which is possible for (1) but not for ([I). Such isomerism has been suggested for n-CSHs Mo(CO)~- 7tC3Hs by King for n-C,H Fe(CO) Si(CH3)C12. In addition a conformational change on dissolution of crystalline [n-CsH5 Fe(C0),l2 GeCI has been suggested by Bush and Woodward on the basis of their X-ray diffraction data and the infra-red data reported by Stone et Thus Mossbauer spectroscopy can elucidate subtleties of conformational change in organo-metallic compounds as suggested in part 1 of this paper.The detailed results of our study of the tin derivatives of [7r-CsHs Fe(CO),] will be reported elsewhere. and by Jetz and Graham Dr. P. Gatlich (Techizische Hochschule Darmstadt) said I should like to ask Herber a question concerning the technique of preparing the frozen solutions. The point that worries me is the rate of freezing. If on the one hand he immersed an absorber cell containing the solution under consideration into liquid nitrogen the solution will be allowed to freeze rather slowly. But if on the other hand he sprayed the solution (in the form of microdrops) into an indifferent organic solvent at low temperature ( e g cis-2-pentene m.p. - 180') the solution will freeze much more rapidly. (For the Mossbauer experiment the liquid solvent has to be separated from 4 J.E. O'Connor and E. R. Corey Znorg. Chem. 1967 6,968. King Znorg. Chem. 1966 5 2242. W. Jetz and W. A. G. Graham J. Amer. Chem. SOC. 1967,89,2773. M. A. Bush and P. Woodward J. Chem. SOC. A 1967 1833. F. G. A. Stone et a!. J. Chem. SOC. A 1966 1130. 97 98 GENERAL DISCUSSION the frozen solution.) The different freezing rates will cause considerable differences in the concentration of the solution since in a slow freezing process the pure solvent will solidify first. I would expect that this effect would show up in the Mossbauer spectra. Did he investigate the effect of freezing rate in detail? Prof. R. H. Herber (Rutgers-The State University) said The question asked by Gutlich is certainly an appropriate one and is one to which we have given much thought. Clearly if one is to use Mossbauer spectroscopy as a tool for differentiating between crystalline solids and structures of molecules found in frozen solutions one must establish clearly that one has indeed a homogeneous solution for examination.We have done a number of experiments to see whether there was any concentration dependence of the Mossbauer parameters that we observe and have always found a negative answer to this. Moreover the asymmetry in the quadrupole doublets which one observes for a number of the compounds that we have investigated can provide a reasonably good index of the question of homogeneous dispersion of the solute in a solvent. To cite a specific example in bis(cyclopentadieny1 iron dicarbonyl)SnCl we have observed a considerable asymmetry in the quadrupole doublet of the iron resonance when we pack the neat solid into an absorber holder.Subsequent experi- ments in which we have ground the absorber material with fine quartz powder to insure a random orientation of microcrystallites has confirmed the fact that this asymmetry arises out of crystal orientation and not from a Gol’danskii-Karyagin effect. When we examine the frozen solution spectra of this material we observe an equal intensity in the two peaks corresponding to the iron quadrupole doublet reson- ance. Since it is reasonable to assume that if we were dealing with small crystals which have nucleated from solution and that these crystals would grow preferentially on the window of our sample cell one would observe the effects of a preferential crystal orientation. The lack of asymmetry in the quadrupole doublet can be used as a reasonable index of having obtained a homogeneous solution.Furthermore when we prepared solutions in polymethyl methacrylate we have looked carefully at the discs which we obtain when the solvent has been allowed to evaporate off at room temperature. We observe no microcrystallites in this material even under high magnification. Again this is only indirect evidence but there is nothing in our data to suggest that we are not dealing with a homogefieous dispersion. Finally I mention briefly the experimental details of making-up our solutions in liquid solvents such as methyl tetrahydrofuran. The spectra on these solution are run in a cell which has a thickness in the direction of the optical axis of about 1 mm and a diameter of about 23 cm and a window made of thin high-grade purity aluminum foil.We filled these cells with the solution prepared at room temperature by the use of a hypodermic syringe and an inlet system constructed of a hypodermic needle which leads directly to the interior of the cell. Once the cell has been filled it is plunged immediately into liquid nitrogen and experiments with samples of similar size have convinced us that we obtain a glassy solvent matrix of very high viscosity just before actual solidification occurs. It is doubtful whether under these canditioiis there is sufficient time for the migra- tion to and coalescence of the solute on the cold surface of the window and crystal formation before the viscosity becomes so high that migration through the medium becomes an extremely slow process. In summary all of the indirect evidence points to the fact that we are observing homogeneous solutions.Clearly one could check this point further by reflectance x-ray diffraction experiments and precision high resolution infra-red scattering experiments. GENERAL DISCUSSION 99 Prof. N. N. Greenwood (Newcastle upon Tyne) said With regard to Herber’s paper is there any independent evidence for the existence of rotational isomers of [CpFe(CO),],SnCl, for example from n.m.r. or infrared spectroscopy? Prof. R. €3. Herber (Rutgers-The State University) said In reply to Greenwood there is such evidence although we were not aware of it at the time of the Mossbauer measurements. In their cry st a1 structure determination of { x-C H Fe( CO) 2 } GeC12 Bush and Woodward comment that in the solid the molecular symmetry is C2 while solution infra-red data in the carbonyl region shows that the molecular symmetry is C1 suggesting a conformational change on dissolution.That a similar situation obtains for {(x-CSH5)Fe(C0)2)2SnC12 may be inferred from the paper by O’Connor and C ~ r e y . ~ Finally the comparison of the solid and solution Mossbauer spectra for (IT(C5H,)Fe(C0)2)2SnC12 and n(C2H,)Fe(CO),SnC1 in which the iron resonance parameters remain essentially unchanged while the tin resonance in the dichloro (but not in the trichloro) compound shows the presence of two doublets is most readily understood on the basis of the existence of a pair of rotational isomers although clearly this is not the only tenable explanation. Prof. J. F. Duncan (Victoria University of Wellington N.Z.) said Did Herber use the glass-forming technique he describes for obtaining hyperfine fields with para- magnetic compounds? This technique is a possible general method for determining spin types of such compounds by measuring the nuclear hyperfine field provided the glass employed and the temperature of measurement are such as to make the field observable.Prof. R. H. Herber (Rutgers-The State University) said We have not applied the glassy-matrix technique to the study of paramagnetic solids although this is a potentially powerful technique as Duncan suggests. All of the compounds reported in the present work are diamagnetic and hence do not show the magnetic hyperfine effects associated with paramagnetic solids. Dr. T. C. Gibb (University of Newcastle upon Tyue) said Since eqn.(10) and (11) of Herber’s paper are only valid in this application if 8 < 80”K has he any information as to the probable values of 8 A and OB in the compounds? If OA and 8 B are essentially derived from an optical (i.e. Einstein) mode model might not 0;/8; be nearly independent of the lattice? If OA and 8 change in constant ratio the value of In fA/ lnf will still be a constant. I would like to ask you how the normalized area ratios were obtained. The peak area can be described in the context of a resonance integral such that A = 1 yam - / r e v (-~(~)~)W(adE)d& where W(E) specifies the source emission line and a(E)n is a function of the Doppler shift E and the absorber thickness tA and recoil-free fraction fA. The form of A is appropriate to both l19Sn and 57Fe if suitable parameters are used.However since a particular Fe-Sn compound has a fixed ratio of iron to tin the curve shape for A with variations in o(E) will not be identical for both elements. In other words the area ratio of the two resonant spectra will be a function of the absorber thickncss and recoil-free fractions. This hypothesis has been tested by evaluating the appropriate M. A. Bush and P. Woodward J. Chern. SOC. A 1967 1833. N. Flitcroft D. A. Harbourne I. Paul P. M. Tucker and F. G. A. Stone J . Chem. SOC. A 1966 1130. J. E. O’Connor and E. R. Corey Inorg. Chem. 1967 6 968. 4* 100 GENERAL DISCUSSION integrals using a computer programme devised for some work on Fe,(CO),. Some sample data sets are as follows (tA is in mg/cm2 of the natural element) f A tA area ratio fA ‘A area ratio varying factor Fe 0-7 Sn 0-7 Fe 0-2 Sn 0.2 0.44 0.53 0.7 0.7 0-2 0.2 0.52 0.43 tA tA 1 3 0.2 1 1 3 0 .1 7 f A ;:} 0.32 ;} 0.27 fA,tA 0.2 0.2 0-2 0.2 Thus the area ratio can alter by at least up to 20 % if the thickness of the sample in the two matrices is different or if the recoil-free fractions have altered such that lnfA/ lnf = constant. Although Herber may well have corrected for these points in his calculations they should be emphasized as a caution to others. These corrections are in addition to the usual caunter background corrections. Fe 0.7 Sn 0.7 Fe 0.7 Sn 0.7 Prof. R. H. Herber (Rutgers-The State University) said In reply to Gibb’s well taken points (i) except for a few of the compounds which we have studied (e.g. C8H8Fe(C0)3 ( Z - C ~ H ~ ) ~ Fe etc.) we do not know 8 for these compounds- and I am not certain that I understand what a Debye temperature for such a compound means (as I pointed out in my paper).To obtain an optical eE one needs far infra-red data and these unfortunately are not generally available. However from the resonance effect magnitudes one might estimate that OM- 100°K. Using Muir’s tables one finds that at 110°K the high temperature limit (which is used in eqn. (10) and (11) of my paper) is in error by only 10 % and at room temperature by about 4 % so that these equations-which are valid only in the high temperature limit- give a reasonably good approximation to the ratio f&Kie. (ii) If only acoustic modes contribute to d lnf/dT the area ratio should be a con- stant for the two Mossbauer atoms and independent of the lattice; if both acoustic and optical modes contribute this ratio will not be constant but will depend on the effective lattice temperature of the matrix.I did not specifically consider the possi- bility that only optical modes contribute to d lnf/dTin the kinds of matrices which we have used although such a situation is-at least in theory-certainly possible. As Gibb correctly points out under such conditions one may in fact observe a constant area ratio. However I am not familiar with any polyatomic non-cubic non-isotropic solids which would show a Debye spectrum with no appreciable acoustic modes being populated at low temperatures. If such lattices exist an experimental verification of this point might be worthwhile. (iii) In the samples which we used for the solution studies t(n) is - 0.2 if one makes the assumption that f’ at liquid-nitrogen temperature is - 0.5.Therefore the “ thin- absorber ” approximation (i.e. that the area under the resonance curve is a linear function of the number n of absorber atoms) is valid in the experiments which we have described. Finally the ‘‘ normalization ” which is referred to in the discussion section of my paper pertains to the method of integration used to obtain the areas under the resonance curve i.e. the data are plotted in the usual manner (counts per channel against Doppler velocity) the absorption maxima are cut out and weighed and the normalization corrects for different scales used for the ordinate in different A. H. Muir Jr. Tables and Graphs for Computing Debye- Waller Factors in Mossbauer Efect Studies AI-6699.GENERAL DISCUSSION 101 experiments. If we always counted to the same number of counts per channel (at u = GO) then this normalization would not be needed (provided the graph paper was of consistent thickness !). However we find it more accurate and convenient to do it the way I have described. Mr. M. G. Clark (University of Cambridge) said Assuming that the recoil-free fraction f is given by the well-known relationship where is the wavelength of the Mossbauer radiatiun then the natural measure of the anisotropy of the recoil-free fraction is a quantity a of the farm For example for an axially symmetric site It follows that the ‘I9Sn Mossbauer resonance is (23.9/14~4)~ = 2.75 times more sensitive than the 57Fe resonance to a given anisotropy in the mean square displace- ment.With this in mind it is interesting to note that in the spectra obtained by Herber for n-C,H,Fe(CO),SnCl the two lines of the l19Sn quadrupole doublet have markedly different areas (which could be interpreted as arising from anisotropy of the recoil-free fraction) whereas there is little ar no difference between the areas of the two lines of the 57Fe doublet (see fig. 7 of Herber’s paper). In fact one may make the general prediction that the higher the energy of the Mossbauer transition the easier it is to observe anisotropy of the recoil-free fraction. f = exp [ -4n2(x2)/A2] a = (anisotropy of mean square displacement)/A2. a = 4n2Kx2) 11 - (X”.L1/A2 = -In (41 Ifi). Dr. M. Cordey Hayes (Birmingham University) said Resonance detectors for 57Fe and 19Sn are also used at Birmingham.*3 The 19Sn detector which consists of enriched SnO in a plastic scintillator mounted on a cooled EM1 9524 photo multiplier tube gives with a SnO source a resonance effect of approximately 300 %.The counter has been used to study lgmSnF2 in very low concentrations (parts per million) in dental enamel. Dr. P. K. Gallagher (Bell Telephone Lab. N.J.) said I would ask Herber does the calculation of the pressure imposed upon an iron(II1) ion in a cobalt(II1) lattice assume that there is no adjustment by the crystal lattice? If so is this unrealistic and in fact really representative of an upper limit which is probably not closely approached? A more appropriate experiment might be to investigate a typical oxide lattice e.g. LaCoO in which a small percentage of the Co is replaced by s7C0 and comparing this with a similar sample containing an equivalent amount of ,‘Fe.Under these circumstances it should be possible to determine whether the isomer shifts (associated with Fe(I1)) observed are the result of the compressive forces on Fe(II1) suggested by Herber by or during the radioactive decay process. Prof. R. H. Herber (Rutgers-The State University) said The pertinent point raised by Gallagher has already been touched on by Greenwood. It is clear that an G. K. Wertheim Mossbauer Efect Principles and Applications (Academic Press Inc. New York 1964) p. 38. J. E. O’Connor Proc. 3rd Mossbauer Con$ Cornell 1963. F. A. Deeney Ph. D. Thesis (University of Birmingham). 2P. A. Flinn S. L. Ruby and W. L. Kehl Science 1964 143 1434. 102 GENERAL DlSCUSSlON adiabatic compression of the type envisioned in my preceding remarks can be expected to be propagated through the crystal with the velocity of a sound wave (- 5 x lo3 m/sec) leading to the characteristic time cited by Greenwood.However it is not at all clear without a detailed analysis of the nature of the restoring forces acting on a charged cation sitting in an ionic lattice site to what extent a local readjustment of the positions of the constituent ions will reduce the " internal pressure " to - 1 atm. For example one might expect local heating effects to occur as a result of the cascade process and its consequences and this may well reduce the " stiffness " of the lattice to the point where the pressure effect has vanished in a time short compared to lO-'sec. In this context I would agree with Gallagher that one is talking about an upper limit to such effects.Concerning the second of Greenwood's points I can only comment that the difference in polarizability between Fe3+ and Co3+ has been completely ignored in out qualitative conception of the problem. Such refinements and sophistications must be considered in any detailed treatment of these phenomena. Finally concerning the third of Greenwood's comments it was this readily observed difference in the quadrupole splitting associated with high spin Fe2+ and Fe3+ in analogous non-cubic lattice sites that prompted Drickamer et al. to identify the high pressure form with Fez+. The much broader resonance peaks observed by Sano in the cobalt cyanide source study may be due either to the presence of a range of " intern21 pressures " or to a range of distortions from cubic symmetry or to both or to other effects which have not been considered in this discussion. It must however be clear that there is still ample room for more definitive experi- ments to be done which might serve to clarify some of the points brought up by Greenwood and Gallagher.

ISSN:0430-0696    

DOI:10.1039/SF9670100097

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Migration of Cations in Two Solid-state Reactions B Y J. F. DUNCAN K. J. D. MACKENZIE AND D. J. STEWART Chemistry Dept. Victoria University of Wellington New Zealand Received 21st August 1967 The quantitative use of Mossbauer spectroscopy in studying reaction kinetics is illustrated by the results obtained on the following two reactions A the rearrangement of kaolinite to form mullite and cristobalite at 1,100" ; B the reaction between iron(II1) oxide and zinc oxide to give zinc ferrite spinel (ZnFe20,) at 800". Both absorber and source methods have been used to identify the sites occupied by the cations at various stages during the reaction and to determine the rate at which the Mossbauer atoms move from one site to another. By comparing the Mossbauer results with X-ray spectroscopic data the following conclusions arise.in the first reaction (i) impurity ("Co) cations become associated with lattice sites at temperatures well below dehydration temperature ; (ii) the cations are not firmly bound in material which can be well characterized as mullite by X-ray methods. This suggests that the cations take longer to attain their final positions than may be inferred from X-ray data. In both reactions A and €3 the cations are still mobile after the oxygen lattice is established. We have applied Mossbauer infra-red X-ray and mass-spectroscopy and also conductance and photo-micrograph techniques to the study of the two reactions. This allows the results obtained by Mossbauer methods to be interpreted more definitively to give information of an unusual kind. The results obtained using conventional techniques on these two reactions leads to the following conclusions.THE MULLITE REACTION. The reaction sequence may be summarized as 5 50'C A1203. 2Si02. 2H20-+A1203. 2SiO2+2H2O (1) kaolin i te 6 5 0°C metakaolinite 980°C 2(Al,03 . 2Si02)--+Si3A14012 + SO2 metakaoli nite AI-Si spinel 1 1 O O T 3(Si3A1,012)-+2(3A1203 . 2Si02) + 5sio2. (3) Al-Si spinel mullite cristobalite On dehydration of kaolinite at 550" a phase which gives weak X-ray patterns (metakaolinite) results ; heating this to 980" gives rise to the spinel phase (also with weak X-ray lines). At 1100" the X-ray doublet characteristic of mullite is observed which we have used for most of our X-ray kinetic w0rk.I The structure of mullite is known from X-ray crystallography to consist of aluminium-oxygen octahedra cross-linked by silicon-oxygen and aluminium-oxygen tetrahedra.Some of the silicon positions are substituted by aluminium and there are oxygen vacancies in the structure. Reactions (1) and (3) are markedly influenced by the presence of trace amounts of ~ a t i o n . ~ 103 1 04 TWO SOLID-STATE REACTIONS From the experimental evidence we have obtained using the techniques listed above,l the reaction at 1100" appears to take place by a series of steps involving loosening and rearrangement of the oxygen lattice along with migration and re- arrangement of the cations which may or may not occur simultaneously with the oxygen lattice rearrangement. THE FORMATION OF ZINC FERRim SPINEL. This reaction may be written ZnO + Fe,O = ZnFe204. (4) Previous work suggests that the reaction proceeds by the following steps (i) vapour phase transport of ZnO to the surface of the Fe20, (ii) diffusion of zinc and iron cations throughout the continuous oxygen lattice obtained by rearrangement of the oxygen ions.The X-ray spectrum of the product begins to appear as soon as those of the reactants begin to decrease in intensity. No other lines are present showing that there is no intermediate crystalline form present in significant amounts. There is some confusion in the literature about the nature of the transition step in this reaction but our work using X-ray techniques suggests strongly that the rate determining step is the diffusion of iron cations throughout the lattice. EXPERIMENTAL In this paper we describe only the techniques related to Mossbauer spectroscopy.Spectra were:determinediat room temperature using polycrystalline samples of the reaction material. In the kaolinite work the absorbers were 4 mm thickx 14 mm diam. ; in the spinel work they were 10-30 mg/cm2 thick. Kaolinite samples were also doped with 57C0 by ion exchange methods as described e1sewhere.l Small samples of this material were placed in platinum foil discs on ceramic planchets and used as source material. ANALYTICAL TECHNIQUE In the spinel reaction Mossbauer methods were used as a precise analysis for the pro- portion of iron ions in different sites during the reaction. Careful consideration of the best conditions under which definitive Mossbauer spectra could be obtained was necessary. It is difficult to determine Mossbauer absorption with accuracy if Compton scattering is large.The relative contributions of these two methods of absorption were studied by varying the thickness of the absorber. Different weights of sample were used in absorbers of standard diameter to obtain absorbers of different thickness. From the spectra obtained using these samples the relative contributions of Mossbauer and Coinpton absorption were estimated using the equations derived in the next section. Such considerations enabled the best conditions for quantitative analysis by Mossbauer methods to be established. A typical spectrum for analysis of ZnFe204 and Fe203 is shown in fig. 1 where the magnetic hyperfine peaks due to Fe203 are superimposed on a doublet close to zero velocity due to ZnFe204. To obtain the proportion of ZnFe204 Fe2O3 calibration curves were prepared using samples of known composition made by mixing pure samples of the com- ponents in appropriate proportions.If it is assumed that the widths r+ at half maximum of all peaks are the same the relative proportions of the components can be estimated from the peak intensities. With ZnFe204 the doublet was not resolved using the velocity range necessary to include the magnetic hyperfine spectrum of Fe203. The intensity of the doublet peak was therefore used as a direct measure of the amount of ZnFe204 present. For Fe203 two estimates were made. In the first the mean of the intensities of the two outer lines was used; in the second the mean of the intensitites of the two next inner lines were used. The two peaks of the magnetic hyperfine spectrum closest to zero were not completely J .F. D U N C A N K . J . D. MACKENZIE A N D D. J . STEWART 105 resolved from the ZnFezOI doublet and therefore were not used. The intensities were related to the spectra obtained with 100 % pure material in each case so that no assumption is made concerning the identity of the structure factors for the various components. However these methods of estimating the amounts of material present make the assumption that the intensities are unaffected by adventitious Compton absorption. P I 1 I -0.00 -0.40 0 0'40 0.80 1.20 velocity (cmisec) FIG. 1.-A typical Mossbauer spectrum of a mixture of ZnO Fe203 (A) and ZnFe204 (B). Inset peak due to ZnFe204 enlarged to show quadrupole splitting. The isomer shift is quoted relative to natural iron and has been determined for all the spectra in this paper using a sodium nitroprusside standard.The error bars indicate &G. MATHEMATICAL TREATMENT OF ABSORPTION MEASUREMENTS We make the following assumptions (a) that the incident radiation contains radiation of approximately resonant energy of incident intensity I; and radiation of non-resonant energy of incident intensity ZB. This background radiation could be due to other nuclear transitions such as the 123 and 137 KeV radiation with 57Fe ; (b) that the non-resonant radiation is unattenuated in the Mossbauer absorber ; (c) that the radiation which is approximately of resonant energy can be attenuated by two absorption processes; first due to resonant absorption with an absorption coefficient pm; and secondly due to non-resonant absorption (such as Compton effect) with an absorption coefficient p,.The latter will obtain for any radiation of this energy appearing in the energy window of the single channel analyzer ; the former will apply only to that radiation which is in fact resonant ; (d) that the absorption of both types of radiation specified in (c) is strictly exponential with thickness d. If lo is the total initial radiation intensity (see fig. 2) then I0 = I ; + I& 106 T W O SOLID-STATE REACTIONS L= I. + 1 velocity FIG. 2.-Diagrammatic illustration of significant intensities in a Mossbauer spectrum. absorber thickness (mg natural iron/cm2) FIG. 3.-Plot of R and log R against d. A ZnFe20 ; B Fe203 in a mixture of the two components. J . F. DUNCAN K . J . D. MACKENZIE AND D . J . STEWART 107 We are interested in the ratio of the resonance intensity ( I - 12) to the non-resonance intensity I experimentally observed (see fig.2). For this When d is small and IB4Zi' (7) leads to R = ~ m d i.e. R is linear with d as experimentally observed (see fig. 3); also log R tends to minus infinity. When d is large (and I,+Ii') (7) leads to i.e. the slope of R against d is negative and equal to -pc(If/lB) exp ( -pcd) and the plot is not linear. However which indicates that log R is linear with d for large dand the intercept on this ordinate is In (IT/IB) (see fig. 3). The condition for a critical value in the plot of R against d is Table 1 gives some of the parameters evaluated from fig. 3 for an absorber con- taining a mixture of ZnO Fe20 and ZnFe,O,. It is thus clear that even in the absence of line broadening due to scattering (as assumed here) the quality of the spectrum obtained depends on the thickness of the absorber and the care with which non-resonant radiation is eliminated.However the experimentally determined TABLE I .-PARAMETERS EVALUATED FROM FIG. 3 dm,,. (mglcm*) 911B calc. from eqn. (11) expt. 0.351 0.0121 1-56 19.3 22-6 p c for A1 is 0.01 35 mg/cm2. h n I'C (cm2jmg) (crn2img) ratio of ZnFe,O Fe,03 did not show any variation with thickness outside experi- mental error (fig. 4) although this latter uncertainty showed some variation with thickness. We therefore recommend that intensity measurement be conducted close to the maximum of fig. 3 for better accuracy. 108 TWO SOLID-STATE REACTIONS I 5 I I I0 2 0 3 0 absorber thickness (mg natural iron/cm2) FIG.4.-PIot of experimentally determined percentage of ZnFe204 in a reaction mixture for different thickness of absorber using the peaks referred to in the text. 0 from outer peaks; a from next inner peaks of Fe203 spectrum. SIGNiFICANCE OF MOSSBAUER SPECTRA TO THE REACTIONS The primary data obtainable from all Mossbauer spectra give direct information about changes in site occupancy and symmetry during different stages of the reaction. However quenched samples are invariably used in such work from which one infers the reaction conditions at the temperature of reaction. This is legitimate provided the quenching does not significantly alter the type of site occupied. That meaningful kinetic results are obtained in such systems is reasonable evidence that the quenching procedure is reproducible.However there is no doubt that because of atomic vibrations and possible rearrangement of ions during quenching the sites indicated from measurement at low temperatures are only approximations to the conditions obtained at the temperature of the reaction. KAOLINITE REACTION The kaolinite we used was known to contain cationic impurities. The most important are iron and titanium which are believed to occupy the trivalent aluminium lattice sites. Since 57Fe is a Mossbauer nucleus it became possible to study the migration rate of this impurity cation as the reaction proceeded. We first discuss the results obtained from the absorption technique. KAOLINITE AS AN ABSORBER. The results obtained are shown in fig. 5 and table 2. From these data the following conclusions emerge. (a) There is no change in the valence state of any of the iron atoms which remain high spin iron (111) throughout the whole course of the reaction.(b) Since the line widths obtained on the original material are small the iron atoms are firmly bound in lattice sites. J . F. DUNCAN K . Y . D. MACKENZIE AND D. J . STEWART 109 (c) As dehydration proceeds by heating to 650" the spectrum becomes much more diffuse indicating a loosening of the iron atoms in their sites. This result should be compared with that obtained using kaolinite sources discussed below. (d) This diffuseness continues as the temperature is raised even up to 1100" (see fig. 5). X-ray identification of the mullite doublet at 26"-28 (3.42A) using copper 0 - I I -0.30 -0.20 -0.10 0 0.10 0.20 0 ' 3 0 velocity (cm/sec) FIG.5.-Mijssbauer spectra with a palladium source of kaolinite absorbers after heating for about 30 min as follows A 20"; B 650"; C 980"; D 1100"; E 1280". The error bars indicate fa. TABLE 2.-h!%SSBAUER PARAMETERS FOR KAOWITE ABSORBERS. THE ISOMER SHIFTS HAVE BEEN OBTAINED BY CALIBRATION WITH SODIUM NITROPRUSSIDE FOR WHICH A VALUE OF 6 = - 0 . 0 2 8 cmlsec HAS BEEN ASSUMED temp. ("C) a(cm/sec) assignment unfired + 0.03 Fe3+ high-energy site 650 (-0-12 + 0.12) Fe3+ in many sites loosely bound 980 + 0.03 Fe3+ in fewer sites tightly bound 1100 + 0.05 Fe3+ in fewer sites tightly bound 1280 +o'ol} 3.0.09 Fe3+ in two sites tightly bound K radiation indicates that at 1100" over 20 % of the reaction mixture is converted to mullite in 4 h. Even when this has been completed the MiSssbauer spectrum is still diffuse indicating that the iron atoms are still loosely bound in the lattice in spite of the fact that X-ray evidence suggests that clearly distinguishable interlayer spacings are present.(e) Not until the reaction mixture is heated to 1280" do the peaks of the Mossbauer spectrum become sharp and even then the line widths are still greater than that obtained from the unheated materials. This thermal treatment is much more than enough 5 110 TWO SOLID-STATE REACTIONS essentially to convert the kaolinite to mullite (as judged from X-ray work). Thus the X-ray investigation of this reaction provides only limited evidence about the positioning of the atoms within the lattice. Even though mullite has been formed the properties thereof can be varied considerably by further heat treatment so that the cations become progressively more firmly bound.X-ray evidence merely confirms that crystallinity improves on further heating.' (f) The Mossbauer spectrum obtained after heating at 1280" does however reveal the presence of more than one cationic site (see fig. 5). The isomer shifts of both resonances clearly indicate Fe3+ atoms in more than one site. This is reasonable if isomorphous replacement of the aluminium by the iron occurs because the former is known to be present in both the tetrahedral and octahedral sites in mullite. KAOLINITE SOURCES. Whereas the natural material used for the absorber work is likely to contain most of the iron in non-exchangeable positions the material used in the kaolinite source work necessarily involves readily exchangeable cations so that the same results may not be obtained.The results of the source work are shown in table 3 and fig. 6 from which the following features emerge. (a) After the kaolinite has been treated with 57C02+ cations the Mossbauer spectrum exhibits two broad peaks indicating many sites of different energy. The isomer shift however is in accordance with expectations for high spin iron (11). (6) When the material is heated to 650" (at which stage it is known from X-ray methods that metakaolinite is formed) a spectrum with five distinct peaks is obtained. The resolution of these peaks is not easy but we have tentatively assigned them as follows. (c) Peaks X are assigned to a quadrupole split Fe2+ in one of two aluminium sites whilst peaks Y are also high spin Fe2+ in another aluminium site.Both of these assignments lead to quadrupole interactions and isomer shifts characteristic of high spin Fe2+ (see table 3). Peak Z falls characteristically in the high spin Fe3+ region. This could arise by oxidation of 57C02+ ions in some of the sites. The asymmetry of the right hand peak Y in spectrum. E and F could also arise from Co3+ due to oxidation but we believe there is a contribution from an Auger cascade arising from the nuclear process in some of the decays. Although there is some change in detail the Mossbauer spectrum does not alter significantly from spectrum E with further heating at 1100" even after extensive mullite formation. This confirms the results obtained with the kaolinite absorber discussed in the previous section. (d) After heating to 1100" for 12 h conversion to mullite is virtually complete (as shown by X-rays).Nevertheless the Mossbauer spectrum (see fig. 6 1100" 17.5 h) now becomes much more diffuse indicating that the cations are still loosely bound. (e) As the material is heated (e.g. at 1280" for 17 h) the resonance intensity progressively becomes larger (see fig. 6) although the peaks are still broad. Inspection reveals that Fe2+ still persists (X). An Fe3+ peak (Y) is obtained but this is not split. Both the single Fe2+ quadrupole and the unsplit Fe3+ peak suggest that there is only one type of site in each case present. This conclusion appears to be at variance with the kaolinite absorber work. However in that work one studies the movement of iron atoms initially present in the kaolinite before the material is heated.In source work cobalt atoms are artificially introduced into the kaolinite and therefore no evidence is obtained concerning the movement of iron atoms which are already present. It seems likely therefore that some of the iron (111) atoms originally present are in highly bound sites which are not easily replaceable even after heating at 1280O. J . F. DUNCAN K . J . D . MACKENZIE AND D. J . STEWART 111 TABLE 3.-haOSSBAUER PARAMETERS FOR KAOLINITE SOURCES DOPED WITH 57Co/57Fe. THE ISOMER SHIFTS HAVE BEEN OBTAINED BY CALIBRATION WITH SODIUM NITROPRUSSIDE FOR WHICH A VALUE OF 6 = -0.028 cm/sec HAS BEEN ASSUMED firing time temp. ("C) (min) 110 10 650 30 980 90 1100 80 600 1280 1050 1020 0 .- U 3 2 c C peaks X X Y Z X Y z X Y z X Y Z X Y X Y 6 (cmisec) +0.13 +0.12 + 0.08 + 0.08 +on12 + 0.08 + 0.05 +Om13 + 0.09 + 0.10 +0.12 + 0.08 + 0.08 + 0.08 + 0.05 +0.11 + 0.04 AEQ + 0.20 + 0.24 + 0-23 f0.17 +0.17 +0-17 +0.19 -1-0.16 +0.16 + 0.23 + 0.26 tcmlsec) - - - - assignment Fe2f quadrupole many Fe2+ quadrupole Fe2+ quadrupole Fe3+ Fe2+ quadrupole Fe2+ quadrupole Fe3+ Fe2+ quadrupole Fe2+ quadrupole Fe3+ Fe2+ quadrupole Fe2+ quadrupole Fe3+ Fe2+ quadrupole Fe3f Fe2+ quadrupole Fe3+ low energy sites 10 0 .3 0 0.20 0.10 0 -0.10 -0.26 velocity (cm jsec) FIG. 6.-Mossbauer spectra for 7Co/5 7Fe-doped kaolinite sources with a stainless steel absorber after heating as follows. A 200" for 0.5 h ; By 650" for 0.5 h ; C 980" for 0.5 h ; D 1100" for 10 h; E 1100" for 17-5 h ; F 1280" for 17 h. The error bars indicate f Q. I12 TWO SOLID-STATE REACTIONS For the easily exchangeable positions in source work the line broadening obtained even after the final treatment at 1280" can be interpreted as due to a large number of sites of similar but different energies.We do not believe that the Fe2+ ions and the Fe3+ ions in these positions are sited differently or that Fe2+ cannot be oxidized to Fe3+ easily at 1280". The presence of the Fe2+ quadrupole resonances in all the spectra in fig. 6 can be understood in that the spectrum reflects the siting of 57C02+ ions in a quenched matrix after the heat treatment. We would certainly expect Fe2+ resonances to be present (since Co2+ would not readily be oxidized) as well as Fe3+ resonances (either due to partial oxidation of Co2+ or to an Auger cascade process). By contrast with the previous reaction ZnFe204 forms without the appearance of any intermediate phase.The Mossbauer results shown in fig. 7 may be interpreted in terms of the structures of Fe203 and ZnFe204. The latter is a normal spinel SPINEL REACTION velocity (cmlsec) FIG. 7.-Change in Mossbauer spectrum of an equimolar mixture of ZnO and Fe203 on heating with the iron cations in octahedral sites which the small quadrupole splitting shows to be slightly distorted (see fig. 1). The iron atoms in Fe203 are present in one type of octahedral site only as indicated clearly by the single magnetic hyperfine spectrum obtained for the pure material. The results are illustrated in fig. 7 and 8 from which the following features are evident. There is considerable variation of line width as the reaction proceeds (see fig. 7).If the % conversion determined from either the integrated Mossbauer line intensity or the Mossbauer peak intensity is compared with that determined independently by X-ray methods a close correspondence is obtained in the latter stages of reaction and also at higher temperatures (see fig. 8). This shows that the two techniques relate to a similar feature in the reaction the iron cation siting. The line broadening obtained in the Mossbauer work shows however that the iron at 760" for the fullowing times. A 3 min ; B 6 min ; C 9 min. J . F . DUNCAN K . J . D. MACKENZJE AND D . J . STEWART 113 cations are not initially occupying sites of identical energy but they become more similar as the reaction proceeds because the line width decreases to its natural value. In all cases the Mossbauer estimate of the percentage of ZnFe,O with repro- ducible site energy formed in the early stages of the reaction is lower than that obtained by X-ray methods although they become identical after a prolonged time or at higher A' time (min) 760" 720" 690" FIG.&-Percentage ZnFe204 in reaction mixture estimated from integrated Mossbauer line intensity 0 Mossbauer peak intensity a and X-ray scattering 0 plotted against time of reaction. reaction temperatures. This last is not an artefact because the two estimates were calculated by independent methods and no assumption was made that a particular percentage would be obtained at infinite time or that the Mossbauer and X-ray results should be identical at infinite time. Quantitative addition of the intensities TABLE 4.-hTEGRATED INTENSITY OF MOSSBAUER ABSORPTiON FOR EQUIMOLAR SAMPLES OF ZnO AND a-Fez03 HEATED AT 700" FOR VARIOUS TIMES time of heating integrated intensity 30 45-6 90 48.1 180 55.1 48.2 240 50.0 The error in these figures is f10 %.for ZnFe,O and for Fe,03 did not vary outside experimental error (see table 4). The results therefore confirm that the observations made at small times are real. There is however some evidence in the Mossbauer work that a small fraction of the iron atoms may temporarily enter a second site during part of the reaction as seen by the shoulders evident in some of the spectra (see fig. 7). This second site however (min) (arbitrary units) 114 TWO SOLID-STATE REACTIONS has properties (including the magnetic hyperfine splitting) which are closely similar to those of the fully formed spinel and implies minor changes in the cation environ- ment during the reaction.The conclusion is therefore that the rate at which the oxygen lattice spacings achieve a value characteristic of the spinel phase is faster than that at which the iron atoms enter their final sites. This conclusion implies that the zinc ions are achieving their final positions more rapidly than the iron ions (since cations rather than oxygen atoms are the principal scattering agents of X-rays). It also accounts for the obser- vation that ZnFe,O can be prepared in a form which has ferromagnetic behaviour by rapid quenching of the spinel from a high temperature which treatment causes zinc ions but not iron ions to be displaced from their normal lattice positions. COMPARISON OF RESULTS OBTAINED BY X-RAY AND MOSSBAUER METHODS Although these two reactions are different one being the reaction of two different solids and the other an inter-crystalline rearrangement of material which is initially homogeneous they have one common feature that the results obtained by Mossbauer spectroscopy and X-ray crystallography appear to be discordant.One can rely on X-ray crystallographic methods for interpreting the way the structure of the solid changes only if the interplanar spacings used are significant to the reactions. In both reactions the X-ray peaks used for following the kinetics were related to the spacing between the oxygen layers in which the cations are embedded (e.g. the (220) plane for ZnFe,O, (110) for ZnO and (104) for a-Fe,O ; and the (120) and (210) planes for the mullite doublet).Use of such lines thus indicates that the oxygen lattice has formed with the interlayer spacing expected and the cations are present in these layers. It does not necessarily follow that the cations have obtained sites which are uniformly distributed or of sufliciently high energy to be regarded as immobile. Indeed the Mossbauer spectra show clearly that this is not the case in both reactions (see fig. 6 and 7). Before one can be confident that such a discrepancy is ascribable to causes related to the reaction experimental variables and errors must be excluded. Perhaps the most important of these is the possibility that the X-ray intensity might not be pro- portional to the amount of a particular component present due to changes in the structure factors owing to different ion-ion interactions as the reaction proceeds.The structure factors of the metal ions (zinc and iron in the second reaction and aluminium and silicon in the first reaction) are not greatly different although they are considerably different from that of oxygen. Further in the spinel reaction the calibration technique involved both the two reactions and the product which must necessarily allow for any errors due to this cause. In the kaolinite reaction there was no quantitative comparison between the results obtained by X-ray and Mossbauer techniques and the conclusions drawn from the two techniques are independent. CONCLUSIONS We conclude that Mossbauer spectroscopy can be used to study the mechanisms of solid state reactions in a more sophisticated way than merely following the migra- tion rate of a Mossbauer nucleus from one chemical state to another.In this study the comparison of the Mossbauer technique with X-ray methods has revealed details which were unsuspected viz. (a) The structure and properties of solids are dependent upon the treatment they have received. This is not a new conclusion but if reliance had been placed solely on X-ray techniques one would have erroneously concluded that no further reaction J. F. DUNCAN K. J. D. MACKENZIE AND D. J . STEWART 115 occurs in the spinel case after heating for instance to 700" for 3 h and in the kaolinite case after heating to 1100" for 4 h. (b) The kaolinite reaction proceeds through a phase in which the impurity cations are in well-characterized sites. The variable crystallinity of mullite is due to varia- tions in cation distribution.The properties of mullite can be considerably changed even after long heat treatments at high temperature due to changes in thedegreeof cationic ordering. (c) In the spinel reaction the results imply that the zinc ions diffuse into their sites more rapidly than the iron(II1) ions. One of the authors (K. McK.) thanks the University Grants Committee of New Zealand for the tenure of a Post Graduate Scholarship. Another author (D. S.) thanks Victoria University of Wellington for the tenure of a Post Graduate Research Scholarship. This work is supported in New Zealand by the United States Air Force Office of scientific research (Grants 27/63 ; 26/65 ; and 1236/67 principal investigator J. F. D.). J. F. Duncan P. K. Foster and K. J. D. MacKenzie forthcoming publication. C . W. Burnham Yearbook of the Carnegie Institute 1964 63 223. A. I. Avgustnik A. F. Nazarenko and V. A. Sviridenko 2. Priklad. Khim. 1954 27 782. J. F. Duilcan and D. J. Stewart Trans. Faraduy Soc. 1967,63,1031. (a) E. J. W. Verwey and E. L. Heilmann J. Chem. Physics 1947,15 174. (b) G. E. Bacon Neutron Diflt-action (Cambridge University Press 1963). J. Smit and H. P. J. Wijn Ferrites (Philips Technical Library 1959). p. 94. ' S. T. Lundin Studies on Triuxiul Whiteware Bodies (Almquist and Wiksell Stockholm 1959,

ISSN:0430-0696    

DOI:10.1039/SF9670100103

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

J . F. DUNCAN K . J . D. MACKENZIE AND D. J . STEWART 115 GENERAL DISCUSSION Dr. G. M. Bancroft (University of Cambridge) said A number of assignments in the spectra in the paper by Duncan et al. seem to be incorrect. From our work on silicate minerals and from other work we can conclude that the isomer shift for six coordinate Fe3+ invariably falls between 0.3 and 0.6 mm/sec (relative to Fe) and that the isomer shift for six coordinate Fe2+ invariably falls between 1.0 and 1-2 mm/sec. Tetrahedral species give shifts somewhat lower. Since we have to go by previous experience in assigning peaks a number of conclusions are doubtful. First Fe3+ does not remain as such throughout the kaolinite reaction; spectrum B fig. 5 is more indicative of Fe4+ than Fe3+. Secondly spectrum E is most likely a quadrupole doublet.Thirdly the value of for spectrum A appears to be incorrect in table 2. Similarly assignments for the source work are doubtful. Peaks x y and z do not fall in the characteristic regions for Fe2+ and Fe3+. Prof. J. F. Duncan (Victoria University of Wellington N.Z.) said The unequivocal assignment of the peaks of these spectra is not easy so that we were not greatly con- cerned with the validity of them. It is fair to say that the isomer shifts for Fe3+ and Fez+ are normally within the ranges specified by Bancroft but one should remember that there are examples of compounds with unusual isomer shifts. The values quoted by us are not unreasonable for the assignment given (except for the original typo- graphical error in table 2 spectrum A). However the alternatives suggested by Bancroft and Gibb can be summarized as follows : 116 GENERAL DISCUSSION (a) Assignment XYYX instead of XYXY in fig.6 leads to many satisfactory assignments but we usually considered these less satisfactory. Thus very high quadrupole energies (approx. 0.27 cmlsec) would often have to be combined with isomer shifts which would be rather low for Fe2+ (approx. 0-08 cm/sec) (e.g. the two external peaks of spectrum B fig. 6). This was the primary reason why we chose the alternative assignment. It seems that neither choice is entirely satisfactory and further evidence is needed for an unequivocal assignment. (b) Bancroft's suggestion that spectrum E fig. 5 is a quadrupole interaction is possible but would disagree with the general picture of cation replacement built up for these structures from other evidence.It would however still be Fe3+ with = 0.05 cm/sec AE = 0.08. I find it difficult therefore to understand why Bancroft asserts that Fe3+ does not remain as such throughout the reaction. The results can certainly be interpreted this way and there is no definite evidence from fig. 5 which shows this interpretation to be in error. The source work does not as I understand Bancroft's comments enter into this argument since it refers to a different situation (c) I have made reference elsewhere to Fe4+ which I believe to be an unnecessary postulate in many cases. If we consider this as a formal definition of oxidation state (without referenee to the atomic change) then we imply that there is energy change associated with a process which we may formally write e-+ Fe4+(s) = Fe3+(s) and that this energy change is related to some experimentally measurable quantity.If this were not the case the concept would be pragmatically useless. In the present case we measure changes in nuclear excitation level which are only tenuously related to the chemistry of the system and which can be adequately explained by changes in electron orbital occupancy. It therefore seems to me that in this field (and indeed in almost all others where unusual oxidation states are invoked) that we are defining a certain experimental observation (i.e. a certain value of 6) as indicative of Fe4+ rather than the other way round. We can say the same about Fe2+ Fe3+ Fe" and Fe"' but there we have good experimental observations as the basis for our classifica- tion (e.g.Fe2+ is green Fe3+ brown). With Fe4+ I can see no experimental observa- tion which makes it compelling to introduce it for classifying chemical behaviour and I therefore consider its use unnecessarily confusing. If in addition one asks what special electronic structure is characterized by compounds formally designed as Fe4+ and how these differ from Fe2+ and Fe3+ then one gets into difficulties because of the complicating effects of coupling between adjacent species defects and other features. In addition the lifetime of any real species approximating to Fe4+ in ionic solids is likely to be extremely short. I therefore find it hard to agree that any spectrum is more indicative of Fe4+ than Fe3+. If this statement is to be taken formally it is meaningless because it is reiterative of a definition.If it is to be taken literally it is demonstrably impossible to accept from other evidence. Prof. N. N. Greenwood (Newcastle upon Tyne) (communicated) In Duncan's paper was the curve fitting done visually or by computer and how closely was the assumption of uniform half-widths for all lines borne out experimentally ? Prof. J. F. Duncan (Victoria University of Wellington N.Z.) said In reply to Greenwood the curve fitting of the results quoted in our paper was done about two years ago by visual methods. We are now equipped to fit the data by computer methods and have several times used this technique for analyzing complex magnetic hyperfine spectra. The line widths of the spectra of the mixtures used for calibration GENERAL DISCUSSION 117 (fig. 3) were always not more than 5 % greater than natural.The line widths of spectra determined for kinetic purposes were frequently much larger than this which feature is discussed in extenso in the paper. It did vary with the temperature and time of heating and closely approached natural line width for the spinel reaction after about 75% completion of reaction as measured by the X-ray technique. 0 10 2 0 tA (mg/cm2 of iron) F I ~ . 1. Dr. T. C. Gibb (University of NewcastZe upon Tyne) said The method of optimizing ing absorber thickness treated mathematically in eqn. (5)-(12) in the paper by Duncan et al. is an approximation to a more rigorous treatment worked out some time ago but not widely published.' Treating the resonant absorption process as an exponen- tial term is not strictly accurate.Most of the necessary equations have been given by Margulies and Eh.rmaa2 For a zero thickness source and a uniform absorber with Compton scattering coefficient pA we can express the transmission at the resonance maximum as (1) p(0) = exp (-pAtA)[(1 -fs) +fs exp (- 3TA) JO(3iTA)I* fs is the absorber recoil-free fraction T A is an effective absorber thickness and Jo is a Bessel function. pA can be expressed as a scattering coefficient per unit mass thickness tA of the resonant element and will therefore vary from absorber to absorber. Eqn. (1) is a slight modification of the expression by Margulies and Ehrman because the non-resonant attenuation term is re-introduced. The chief object is to maximize the change in transmission which can be achieved due to resonant absorption and this can be represented as (2) A computer programme was written to evaluate this expression for appropriate values of the parameters and a specimen set of curves is shown in the diagram for 57Fe with four values of fA and an eflectiue mass absorption coefficient of 0.10.The Ap(0) = exp (-pAtA)f[l -exP( - %TA)JO(3iTA)I. 118 GENERAL DISCUSSION position of the maximum is not easy to estimate by taking the limiting slopes. For- tunately the optimum thickness is not very dependent on the source and absorber recoil-free fractions. The experimental points of fig. 3 in Duncan’s paper agree well with the shape of the theoretical function (fig. 1). Unfortunately application of these curves to 57Fe work is not always successful. In cases where optimization is essential viz. where the effective pA value is large because of the presence of heavy elements in the compounds there is a considerable increase in the low-energy scattering background from the 129-keV gamma which we cannot easily account for.Neither have we included the effects of hyperhe effects which alter the TA values. Duncan states that the broad lines in some of the spectra are due to “ a loosening of the iron atoms in their sites ”. Surely it is better to say that the iron atom is “ bound ” strongly in its site but that the site environment is not unique while the ordering processes are still taking place. Variations in the nearest neighbour cations would cause line broadening. In fig. 6 of Duncan’s paper is there any reason to prefer the XY-YX allocation of peaks rather than XY-XY? Ferrous cations usually show more sensitivity in the quadrupole splitting than the chemical shift.Prof. J. F. Duncan (Victoria University of Wellington N.Z.) said In reply to Gibb there are two features related to the shape of the spectrum which provide evidence about the bonding of the Mossbauer atom (a) line broadening gives information about the energies of the sites in which the Mossbauer atoms find themselves. If this were the only feature Gibb’s comments would be entirely justified-although we referred to a diffuse spectrum rather than line broadening. (b) Lf a Mossbauer resonance is to be observable the y-ray must be emitted with no energy loss by recoil and this requires that the Mossbauer atom be in a site of high energy. The intensity of resonance thus gives direct information about the “ stiffness ” with which the Mossbauer atom is held in its site. We measure the absorption intensity to provide information about the rates of reaction which were compaed with those determined with X-ray techniques. It is the discrepancy between these two techniques which provides evidence that the iron atoms become loosened in their sites! Any line broadening will also indicate changes in the binding energies of those atoms which are sufficiently strongly bound to allow emission or absorption of recoilless y-rays and the Mossbauer spectrum to be observable. T. C. Gibb Ph.D. Thesis (Newcastle upon Tyne 1966). * S. Margulies and J. R. Ehrman Nuclear Instr. Methods 1961 12 13 I.

ISSN:0430-0696    

DOI:10.1039/SF9670100115

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Application of the Mossbauer Effect in Tool for Chemical Structures and Investigation 1291 as a Spectroscopic Lattice Dynamics BY M. PASTERNAK~ Nuclear Physics Dept. Soreq Nuclear Research Center Yavne Israel Received 4th September 1967 This paper deals with the interpretation of the Mossbauer effect (ME) of the 28 keV level in and the analysis of the deduced constants in terms of chemical structure and lattice dynamics. The calculation of the magnitude and sign of e’qQ the r ) parameter and the isomer shift from the quad- rupole splitting spectrum is described. These parameters are directly related to the p electron population in iodine. The effect of the behaviour of the anisotropic recoilless fraction f on the various Am transition intensities is demonstrated. Illustrations of the application of the ME in 1129 for the structural determination of some tellurium compounds are given.Cr13 is cited as an example of a magnetic iodine compound and its spectra above and below the Curie temperature are analyzed. Finally possible trends for future studies of the ME in are outlined. The Mossbauer effect has been observed in about 32 isotopes yet implementation of this spectroscopic method for physico-chemical investigation has been confined chiefly to FeS7 and Sn119 compounds. This is because the latter isotopes have convenient nuclear properties i.e. good energy resolution and strength of effect. Another isotope with nuclear properties suitable for spectroscopic application which has not yet been widely exploited is P9. The Mossbauer effect of the 27.8 keV line in 1129 was observed by Jha et aZ.,I but utilization of this nucleus for spectroscopic purposes has been primarily carried out by the Illinois and the Soreq groups.Hafemeister et aL2- measured the nuclear properties relevant to the ME and applied the ME to studies of lattice dynamics and structural properties of alkali iodide and some of the oxy-iodine compounds. These studies and those carried by our group on iodine molecular crystals show the potentiality of the ME in 112’ for obtaining diverse information concerning the chemical structure and lattice dynamics not always attainable with other Mossbauer nuclei. Iodine is the only halogen with which the ME can be applied. Iodine compounds have been extensively studied by nuclear quadrupole resonance (n.q.r.) method by which accurate quadrupole coupling e2qQ and asymmetry parameter q constants can be mea~ured.~ However due to the high frequencies involved in iodine compounds (500- 100 Mcfsec) this technique is experimentally troublesome.On the other hand the ME in IIZ9 not only provides such constants but also the sign of the electric field gradient (e.f.g.) the isomer shift (is.) and lattice dynamics information. Mossbauer effect studies were also performed with 112’ (E = 57.6 keV),6 but due to its nuclear properties it is inferior in comparison to It is the aim of this work to outline the main features of the ME in 1129 to consider the experimental problems encountered due to the characteristics of this nucleus and to present the important parameters deduced from the Mossbauer spectrum with which structural and lattice dynamics information can be obtained not only for compounds of iodine but also of tellurium the 1129 nuclear precursor.* present address Natuurkunding Laboratorium der Rijks-Universiteit Groningen Holland. 119 for spectroscopic purposes. 120 APPLICATION OF THE MOSSBAUER EFFECT I N 1129 EXPERIMENTAL For the ME in 1129 there are additional problems resulting from the Te129-+1129 nuclear system. The decay scheme of Te129-1129 is shown in fig. 1 and the relevant properties of the 27.8 keV level pertinent to the Mossbauer effect are given in table 1. 33d 70m 482 8- 1450 keV 02% 275 28 39 0.7 x I O - ~ sec 0 FIG. 1.-The Te129-Ii29 decay scheme. TABLE NUCLEAR PROPERTIES OF THE 28 keV LEVEL OF PERTINENT TO THE M~SSBAURR EFFECT 27.78 f0.05 keV Q(7/2) = -0-55 cm2 E,= { 27.72 f0.06 keV a = 5.3rt0.3 Qex/Q = 1.231 f0.001 T+ = 16.8 f0.2 lo-’ sec I? = 0.278 lo-’ eV ; 0-029 cmlsec ; 6.7 Mc/sec p(7/2) = +2-617 n.m Ekm = 27.4 keV peX/pg = 1-07 f0.07 - 3.74 cm2 AR/R = +6.10-5 221ex+ 1 1 21,+1 l + a 0 0 = 2n12 - - - ER = 2.97 eV; 34°K somcEs.-For the investigation of absorber properties one requires a source with a nonsplit line.This requirement is fulfilled by using a tellurium compound with cubic symmetry around the Te atom. Most of the tellurides have this property and a widely used source is ZnTe.3 As seen from fig. 1 the 28 keV line (27.78 f0.05 keV,7 27-72 f0.06 keV *) can be populated either from the metastable decay (T+ = 33d) or from the ground state of Te12g(T+ = 70 m). The second mode of decay is preferable because higher specific activities can be obtained and less background is present because more than 50% of the Te129m decay product are K X-rays.(Ek(Te) = 27.2 keV). The Te129 source is activated by (n,y) irradiation in the reactor. Due to the relatively short half-life the telluride consti- tuents must not produce any competing irradiation which may add to the background. Therefore the enriched isotope Z P is used in the ZnTe source. The recoilless fractionfof ZnTe at 100°K is 0.23 3~O.01.~ M. PASTERNAK 121 ABSORBERS.-The synthesis of I' 29 absorbers requires special treatment because this isotope is not found in nature and because it is radioactive (T+ = 1 . 6 ~ lo7 y). Since most iodine compounds are produced from elemental iodine one has to transform the iodide which is the commercial form obtained from Oak Ridge to 12 either by decomposing Pd12 at 350°C or by oxidizing the iodide and then extracting I2 with an inert solvent such as pure ether or methylene chloride.The Xe K X-rays produced from the internal converted level in Xel 29 (see decay scheme) can be absorbed with a critical absorber such as Sn or In. ANALYSIS OF THE ABSORPTION SPECTRUM OF 1129 Most of the iodine compounds except the monoiodides and some of the oxy- iodides reveal a quadrupole splitting spectrum. Since iodine is diamagnetic there have been only a few instances where a magnetic splitting spectrum has been analyzed. Ferromagnetic CrI lo is one example. Magnetic splitting achieved by an external magnetic field has been observed l1 in K1129 as well as in atomic iodine (produced by beta decay of Te129m) embedded in an iron foil.12 QUADRUPOLE SPLITTING SPECTRUM As a result of the interaction of the electric field gradient with the nuclear quad- rupole moment the nuclear ground state ( I = 7/2) and the excited state (I* = 5/2) split into 4 and 3 sublevels respectively.Due to the nature of the nuclear dipole transition only Am = 0 and Am = +1 transitions are allowed and therefore eight transitions occur. The positions of the energy lines may be determined by the relation where A = e2qQ,/4 R = Qex/Qs = 1*231+0401,3~ 6 is the isomer shift and f(I,m,q) is related to the spin Hamiltonian eigenvalues. The values of f(I,m,q) were computed by Bersohn l2 in terms of a series of even powers of y. We have calculated the line positions (see table 2) for y2 d0.2. Values off(l,m,y) were numerically cal- TABLE 2.-ENERGY POSITIONS Eg FOR y2 < 0.2 lint type of transition 5/2+3/2 5 /2+ 5 /2 3/2-+3/2 3/2+5/2 1 /2-+ 1 /2 5/2+7/2 3/2+ 112 1 /2+3/2 Am energy position A( 1 -654 - 0.392 q ') + 6 A(1.087- 0.051 q2) + 6 A(0.230+0.035 y2)+6 A(0.469f0.964 q2)+6 A(0.178-0*091 q2)+6 A( - 0.389 + 0.250 y 2 ) + 6 A( - 0.269 + 0.1 59 q2) + 6 A(-0*555-0.897 y2)+6 culated by Cohen l3 for values of y from 0 to 1.Using these values we have cal- culated the energy levels at intervals of 0.1 q which are given in table 3 and graphically shown in fig. 2. Except for lines 4 and 8 the other transitions are only slightly y dependent. In fig. 3 the schematic quadrupole splitting spectrum of an absorber is shown. The lines may be identified from their energy positions and relative intens- ities which are proportional to the square of the Clebsh-Gordan coefficients (C.G.)2 (see last line in table 3).The values of e2qQ and the isomer shift are found from any of the lines (except 4 and 8) by the least-squares method which yields accurate values. 122 APPLICATION OF THE MOSSBAUER EFFECT IN Ii2' r i l ' l ' l ' l i l ' l ' l ' l - 0 8 EG (in units of e2qQ/4) FIG. 2.-Dependence of the energy positions Eu on r]. Am-] Am=O - 1.2 ___) - I - V 0 + V FIG. 3.-Schematic description of an I"' quadrupole splitting spectrum. This spectrum corresponds to a negative e2qQ and y = 6 = 0. The horizontal arrows indicate the influence of 7 on the line positions. For an anisotropicf((x2) 8)(x2)L)Am = 0 and 1/2+1/2 transitions will be enhanced. TABLE 3.-ENERGY POSITIONS Eg OF THE EIGHT TRANSITIONS IN THE QUADRUPOLE SPLIT CLEBSH-GORDAN COEFFICIENTS (q = 0) ARE GIVEN IN THE LAST LINE SPECTRUM OF 1129 FOR VALUES OF FROM o To 1 AT INTERVALS OF 0.1.THE SQUARE OF THE transition tl 0 0.1 0-2 0.3 0.4 0-5 0.6 0.7 0.8 0.9 1.0 1 C.G 5/2-+3/2 1 a654 1 -655 1 -645 1.631 1.616 1.602 1-589 1.579 1-573 1-571 1 a572 1 ' 1 5/2+5/2 2 1.087 1 -087 1.085 1 -082 1 -079 1 -074 1 -068 1 -060 1 -052 1 -043 1 -032 6 5/2+7/2 3 0.230 0.23 1 0.232 0-23 3 0-23 6 0.239 0.243 0.247 0.253 0.260 0.267 21 3/2+1/2 0.469 0.478 0.505 0.546 0.600 0.662 0.729 0.803 0.877 0.955 1.035 3 4 3/2+3/2 5 0-1 78 0.182 0.1 80 0.1 80 0-1 83 0.190 0.192 0.212 0.228 0-248 0.270 10 312-412 6 - 0.389 - 0.386 - 0.380 - 0.369 - 0.354 - 0.338 - 0.323 - 0.307 - 0.293 - 0.280 - 0.270 15 1/2-+1/2 7 - 0.269 - 0.268 - 0-264 -0.261 - 0.256 - 0.255 - 0.254 - 0.255 - 0.25 8 - 0.263 - 0.267 18 1/2-+3/2 8 - 0.555 - 0.564 - 0.589 - 0.627 - 0.673 - 0.727 - 0.785 - 0.846 - 0.907 - 0.970 - 1.032 10 M.PASTERNAK 123 To find y one makes use of transitions 4 and 8 (see table 2 or 3). The influence of y on line positions 4 and 8 can be seen from l4 the spectra of TeO and Te(N03)4 sources against the nonsplit absorber C U I ' ~ ~ as shown in fig. 4. I I I I 7 6 35 - I I I I I ] -20 -1.5 -I.@ -0.5 0 0.5 1.0 1.5 velocity (cmlsec) a I I 1 I I 1 I I X W t I 2 I I I I I I 4 35 7 6 8 -1 I _ I _ r - " - - - 1 I I I I I I I -1.5 -1.2 -0.9 -06 -03 0 0.3 0.6 09 1.2 velocity (cmlsec) b FIG. 4.-The spectrum of Te12902 (a) and Te129(N03)4 (6) sources against a absorber at 80°K. Note the effect of q # 0 on lines 4 and 8 in (a). ROLE OF THE ANISOTROPIC RECOILLESS FRACTION In the same way that the nuclear quadrupole moment and radius serve as probes for the measurement of the e.f.g.and the s density at the nucleus the angular correlation of the nuclear dipole transition can be used as a probe for the anisotropy of the mean square displacement ( x 2 ) in a polycrystalline sample.A necessary and sufficient 124 APPLICATION OF THE M ~ S S B A U E R EFFECT IN P29 condition for the deviation of the line intensities in a quadrupole splitting spectrum from the (C.G.)2 is that thefis anisotropic.15 The deviation of the line intensities from the (C.G.)2 values will be reflected by the various Am transitions (see fig. 2). In fact one may express the relative intensities as wheref(8,#) is the anisotropic recoilless fraction with respect to the electric field axis and FArn(O,$,q) is the dipole angular correlation function corresponding to Am.The values of F are given in table 4. It is possible to predict the direction of the anisotropy off with respect to cos 8 from the behaviour of I. for Am = 0 and Am = 1. From table 4 it is seen that the Am = 0 function has a maximum at 8 = n/2 therefore iff(@ also has a maximum at n/2 the Am = 0 transition intensity will be enhanced. When thef(8) has a maximum at O" the Am = 1 transitions will be en- hanced. The TeO source is typical of an enhanced transition Am = 0 i.e. the recoilless fraction is larger in the direction perpendicular to the electric field axis (see fig. 4). For the group IV tetraiodides,16 the Am = 1 transitions are amplified indicating thatfhas its maximum in the direction of the electric field axis which is the direction of the X-I bond (X = C Si Ge and Sn).Due to the high spin transitions in 1129 we encounter three types of Am transitions viz. Am = f 1 0 and the mixed transition & 3+ & 4. The latter has a similar 8 distribution as the Am = 0 transition but less pronounced. Therefore in comparison to Fe57 S d i 9 or Te125 the probe for the detection of an anisotropicfis more sensitive in INTERPRETATION OF THE QUADRUPOLE SPLITTING CONSTANTS Most of the iodine compounds reveal a quadrupole splitting spectrum. The reason for the large e.f.g. at the iodine nucleus is the deficiency of p electrons or " p holes " in the spherically symmetric 5s25p6 configuration. The electric field gradient in iodine due to one p hole (hJ is 11 1.6 x ~ m - ~ as compared to the e.f.g.of 0.1 x from the nucleus. The e.f.g. due to the p holes may vary according to the ionicity of the bond the amount of the s character and the distribution of the p electron defect around the x y and z axes thus the spherical symmetry around the iodine atom increases and the e.f.g. decreases. The s electrons do not contribute directly to the e.f.g. however due to the renormalization of the orbital wave functions the fraction of p electrons decreases and the e.f.g. is attenuated. Based on the above arguments Townes and DaiIey l7 developed a theory which relates the molecular quadrupole coupling (e2qmOlQ) to the charge distribution around the atom. According to their formulation produced by a unit charge at a distance of 2 (3) 2 e ~ o i Q = - U p e 2 4 p t Q , M.PASTERNAK 125 where e2qat Q is the quadrupole coupling due to onep electron (+ 2293 Mc/sec for 112') and Up is defined as thep electron " excess " or " defect " and is directly related to the charge distribution l8 Where the iodine has a single bond e.g. in the z direction (a bond) one may express the electron population in terms of chemical bond parameters as where s2 is the amount of s character and i is the bond ionicity positive for negative ions and negative for positive ions and ;TI is the amount of the x character in the bond. up= -Uz+(Uy+U,)/2. (4) U = I+s2+i; Ux = 2 - x x ; u, = 2-ny ( 5 ) The asymmetry parameter y is related to Uy and U as follows :18 11 = ( 4 x x - - y y ) / 4 z z = 3 W x - UY)l2UP. (61 Using the values of e2qQ and y one may calculate Up and U,- Uy.In order to deduce the amount of sp hybridization the x character and the ionicity solely from the quadrupole interaction as is done by the n.q.r. spectroscopist it is necessary to state apriori two of these parameter values and then from the experimental results calculate the third one. In fact most of the ionicity sp hybridization and n character values ' ' 9 2o as deduced from the n.q.r. data of the halogen compounds were cal- culated on this basis. However from the i s . data which is unique to the ME spectroscopy valuable information regarding the s character and ionicity can be extracted. The isomer shift is primarily related to the s density at the nucleus by the expres- sion :21 In IlZ9 AR/R is p~sitive,~ therefore an increase in the s density at the absorbing nuclei a will result in an increase in the isomer shift.Assuming that the chemical bond involves only thep valence electrons of iodine (a situation which is energetically favourable) the effect of increasing the p density- which is zero at the nucleus-is to screen the nucleus from the s electrons and therefore decrease the effective s density; i.e. the p density variation will have the opposite effect on the i s . to that of the s density. However the effect of thep electrons on the i.s. will be secondary compared with that of the s electrons. Where onlyp electrons are involved in the bonding a relation between the isomer shift and the number h ofp holes in the 5s25p6 outer configuration of I- can be derived. The s density I $s I is related to the effective 2 by the Fermi-SegrC formula,21 I $s I = a z (8) where a is a constant.The effective charge can be found from the Slater shielding coefficients. Hafemeister et aL3 calculated the Zeff for the I- configuration and found that Ze.f = 7.25 + @35hP. (9) Substituting expressions (8) and (9) in eqn. (7) and neglecting terms in hi one gets a linear dependence of the i s . on hp. This relationship has been experimentally con- firmed by Pasternak et aL4* 22 in the range 0 < hp < 1.3. The explicit linear relation found was 6 = 0.136hP - 0.054 (cmlsec) (10) 126 where the isomer shift is referred to the standard ZnTe source. The validity of expres- sion (10) for higher values of hp has not yet been experimentally confirmed. However from eqn. (9) for h = 3 the contribution of hg terms to I $s I is only 10 %.Fortun- ately most of the iodine compounds especially the organic compounds have h values close to 1. The p holes are related to the p electron population by APPLICATION OF THE MOSSBAUER EFFECT IN By combining eqn. (4) (6) and (1 I) one may in principle find thep electron distribution around the iodine atom. The calculations of the n bonding character and the ionicity are straightforward. Where sp hybridization is involved e.g. in 102 the isomer shift will be chiefly influenced by the s electrons. An indication of the removal of s electrons from iodine is given by a large and negative value of the isomer shift. One may empirically express the isomer shift in terms of s and p holes as The value of Q was calculated 23 from the results of the isomer shift in 10; (pure p bonding) and in 10; and I0zv (sp and spd hybridization respectively) assuming that the amount of charge removed for the 1-0 bond is the same in each case.We found that 6 = -ah,+bh,+c. (12) 6 = - 0.82hS + 0.I36hp - 0.054 (cmlsec). (13) The dependence of S on h needs further confirmation. An important contribution could be is. data of IF7 since a detailed theoretical calculation has already been made 24 on the number of the s and p electrons involved in the I-F bonds. EXAMPLES OF STRUCTURE DETERMINATION FROM ME DATA H6TeO6 A monoclinic polycrystalline orthotelluric acid source was used in this experiment. In H6Te06 the Te atom in situated in a distorted octahedron25 formed by six oxygens. The spectrum obtained with this source (Te natural abundance) against the non-split CuI12’ absorber at 80°K is shown in fig.5. The parameters found from the quadrupole splitting spectrum were where the quadrupole coupling is in terms of the 1127 ground-state quadrupole moment (Q(129)/Q(127) = 0.701 26) and the isomer shift is referred to the standard ZnTe source. The isomer shift is the same as that obtained for Na3H2106 27 (= -0*335_+0*002) which indicates that the s density at the nucleus is not affected either by the distortion of the octahedron or by the substitution of Na for H in H6Te06. The existence of an e.f.g. acting at the nucleus is consistent with the distorted structure. From the positive sign of e2qQ we conclude that U > (U + Uy)/2 (see eqn. (3) and (4)). Despite the nuclear transformation of Te the distortion of the octahedron is preserved. From ME studies of the I 0 z v structure 27 a non-split line was detected suggesting a non-distorted octahedron The possibility of using the ME in 1129 to obtain structural properties of tellurium compounds was studied by Pasternak and Bukshpan l4 who found that it is possible with compounds where the tellurium is in a positive ionic state.Experiments with Te02 (orthorombic) Te(N03)4 and Te metal both as Te129 sources as Te12’ absorbers e2qQ(127) = 268+ 10 Mc/sec ; 6 = -0-3343.0.005 cm/sec, M. PASTERNAK 127 showed that Up which is a probe for the electronic distribution is conserved for the high ionic states of Te. With Te129 metal it was found that the value of Up for the source is reduced by a factor of 4 as compared to that of the Te125 absorber. Another interesting result regarding the oxy-tellurium compounds was obtained with Te205 which is produced by heating H6Te06 to 500°C.The spectrum of this source against a absorber is shown in fig. 6. It is concluded that this substance 1 I I 1 I 1 FIG. 60 I I I I # I 1 *'** :. . .. . f " lot- .* # '. .- 0 50 too 150 20 0 250 channel FIG. 6.-Spectrum of TexOp The composition of this compound was found to be TezOs (see ref. (29)). is composed of TerV and Tevl species. The existence of a mixture of these two tellurium states in Te,O has been confirmed by Dutton and Cooper 28 by conventional chemical analyses in the liquid state. Measurements carried out with a Te';L'O absorber could not resolve these two Te species (see fig. 7). One notes the potentiality of the ME in 112' for detecting in the solid state the actual ionic states of Te ions as well as their concentrations.MOLECULAR CRYSTALS Measurements with various simple inorganic molecular crystals containing iodine were carried out by our group. The molecules studied were 12 IBr IC1 ICN 12Cls 23 and the group 4 tetraiodides.16 Results were deduced regarding the electron density (ionicity) hybridization n-bonding and lattice dynamics. It was found that the 128 APPLICATION OF THE M~SSBAUER EFFECT I N r129 intramolecular frequencies play a significant role in the recoilless fraction and its anisotropy ; e.g. the recoilless fraction of 12C16 is significantly larger than that of 12. This was explained on the basis of the high frequencies of the I-Cl bond which are not excited by the recoilless process at 100°K temperature. Indeed the ( x 2 ) for the I-Cl bond as calcuIated by Nagarajan 2 9 from Raman and i.-r.data was less than - I2O* ) I l l I I 1 l I I l I I I I I I I I I.1 to 0.9 09 0.7 0s 05 0-4 0.3 0.2 ci 0 -01 -0.2 -0.3 -w -05 -06 -0.7 -0B velocity (cmlsec) a I I l l / l l l ! / l l l l l l l / l I.1 K) 09 08 Q7 CH3 05 04 O q 0 2 01 0 -01 -0.2 -0.3 -0rl -05 -0.6 -0.7 -08 velocity (cmlsec) b FIG. 8.-The spectrum of Cr13 above (a) and below (b) the Curie temperature. The line positions marked A.BC correspond to the following values A Hi = 27 KOe cos p = 0.4 ; B Hi = 23-5 KOe cos = 0.5 ; C Hj = 16 KOe cos = 0.6. M. PASTERNAK 129 that deduced from the absolute f. It was concluded that in the recoilless effect mechanism the whole (ICI,) mass participates and the frequencies excited are those of the C1-C1 bonding and the intermolecular modes.CrI3-A FERROMAGNETIC COMPOUND Investigations of the bulk properties and n.m,r. of Cr5 have been performed on the insulating magnetic crystals of the chromium halides.,O. N.m.r. studies of Br79 and Brs1 in CrBrs (T = 3205°K) revealed 31 that the Hi acting at the Br nucleus is 37 KOe and cos p = 0.67 where p is the angle between qzz and Hi. Cr13 has the same TABLE TH THE ISOMER SHIFT (WITH RESPECT TO A ZnTe SOURCE) QUADRUPOLE COUPLING CONSTANTS AND q VALUES OF THE 1129 comoums INVESTIGATED TO DATE compound LiI NaI KI RbI CSI CUI AgT PbTe CH31 Cr13 CI4 SiT4 Ge14 Sd4 Te2 I2 Ba(I0312 (NH4)103 m03 Te02 Te(NOd4 IBr JC1 ICN 12Cl 6 12C14Br2 KI04 Na3H210 H6TeO6 is. (cmlsec) -0.038 f0.0025 -0046 f04025 -0051 f0.0025 - 0.043 3_04025 - 0.037 f0.0025 -000446 f0.0005 - 0.026 f0402 + 0.022 f0.005 + 0.020 f0.003 + 0.023 ~t0405 + 0.065 f0.003 + 0.026 &0*003 4-0.048 f0.003 + 0.043 f0.003 +0*071 f0.002 + 0.083 f0-001 + 0.1 11 350.02 +0.131 f0-02 + 0.1 56 f0.02 +0*152 f0.001 + 0.252 f0.001 + 0.123 f0.002 +0*173 f0.005 + 0.1 19 rtO.OO5 + 0.350 f0.002 + 0.282 f0-002 - 0.205 f0.005 - 0.234 f0.006 -0.335 f0.002 - 0.334 &0*005 - 1739 f10 + 662 f8 -2102flO - 1335 f 10 - 15OOf10 - 1364 510 - 532 f10 - 2156 f 10 3- 1030 f20 I + 1121 A10 3.1139 f 10 - 2892 f 10 -3113f20 -2640f10 + 3060 f 10 4-3040f10 + 291 6 f 10 0 0 0 + 268A10 rl 0 0 0 0 0 0 0 0.35 f0.05 0 0 0 0 0.80 f0.05 0.1 6 f0.03 - - 0.55 f0.05 0 006 f0.02 0.06 f0.03 0 0-06 f0.02 - - 0 0 0 I crystal structure as CrBr, a hexagonal closed packed arrangement of" sandwiches " of Cr and I.ME experiments in CrIi27 were carried out by Kalvius et aZ.32 Above the Curie temperature (T' = 70°K) a broad line was detected from which the e2qQ has been deduced.From the further broadening of the line below T, a value of ITl = 140rfr 30 KOe was calculated. The ME pattern of CrIi29 above and below T is shown in fig. 8. Above the Curie temperature the following parameters were calculated e2qQ(127) = + 662 & 8 Mc/sec ; q = 0-352 0.05 ; 6 = + 0.23 +0*05 mm/sec 130 APPLICATION OF THE MOSSBAUER EFFECT IN If29 The sign of the e.f.g. is consistent with the Cr-1-Cr bond arrangement i.e. the principal axis of the e.f.g. is perpendicular to the bonding plane. Below the Curie temperature by using the eigenvalues 33 of the spin Hamiltonian governed by both magnetic and electrostatic interaction we found lo that Hi = 20f4 KOe and cos /3 = 0.55+0*05.FUTURE DIRECTION OF ME SPECTROSCOPIC STUDIES WITH 1129 The number of iodine compounds that have been investigated to date is negligible in comparison with those of iron and tin. The ME spectroscopic data determined for the three If2’ compounds studied are summarized in table 5. Important fields which may be exploited with the ME in 1129 are (i) studies of trapped radicals and molecules in inert matrices. This field has been extensively studied by other spectroscopic techniques. ME results may supply information regarding the influence of the intermolecular bonds on the molecular charge distribu- tion and density. Similar investigations can be performed on frozen solutions. (ii) Investigation of simple organic species followed by extension to more complex compounds.The potentiality of application to compounds of medical interest is evident. Of special interest are the charge transfer complexes where one may obtain and calibrate the amount of the charge transferred via the i.s. Also the inductive effect can be studied as was done with the ME in Sn119. The field or organic semi- conductors is of interest since iodine often plays a significant role in the conduction mechanism. (iii) Studies of the iodine-fluorine compounds for IF, e.g. a structure where the iodine has a five-fold symmetry (q = O?) one expects to measure the largest negative i.s. These data are essential for the calibration of the is. dependence on the number s holes. In IF5 one expects to obtain a very large e2qQ and possibly an upper limit for the isomer shift dependence on hp.S. Jha R. Segnan and G. Lang Physic. Reu. 1962,128 1160. H. de Waard G. de Pasquali and D. Hafemeister Physics Letters 1963 5 217. D. W. Hafemeister G. de Pasquali and H. de Waard Physic. Rev. B 1964,135 1089. M. Pasternak A. Simopoulos and Y. Hazoni Physic. Rev. A 1965,140 1892. T. P. Das and E. L. Hahn Nuclear Quadrupole Resonance Spectroscopy suppl. 1 of Solid State Physics (Academic Press Inc. 1958). C. I. Perlow and M. R. Perlow J. Chem. Physics 1966,45 2193. C. E. Bemis and K. Fransson Physics. Letters 1965 19 567. R. Sanders and H. de Waard Physic. Reu. By 1966,146,907. M. Pasternak unpublished thesis. lo C. Goldstein and M. Pasternak to be published. li H. de Waard and J. Heberle Physic. Rev. B 1964,136 1615. l2 R. Bersohn J.Chem. Physics 1962,20,1505. l3 M. H. Cohen Physic. Reu. 1954,96 1265. l4 M. Pasternak and S. Bukshpan Physic. Rev. 1967 163 297. l5 S. V. Karyagin Dokl. Akad. Nauk. S.S.S.R. 1963 148 1102. l6 S. Bukshpan unpublished thesis. l7 C. H. Tomes and B. P. Dailey J. Chem. Physics 1949,17,782. l9 W. Gordy Disc. Faraday SOC. 1955,19 14. 2o D. L. Schawlow J. Chem. Physics 1954,22,1211. 21 For a general review on the i s . see D. A. Shirley Reu. Mod. Physics 1964 36 339. 22 M. Pasternak and T. Sonnino paper presented to the h i . Con$ Hyperfine Nuclear Spectroscopy 23 M. Pasternak and T. Sonnino to be published in J. Chem. Physics. 24 R. L. Oakland and G. H. Duffey J. Chem. Physics 1967,46 19. 25 L. C. Pauling 2. Krist. 1935 91 367. see ref. (9 p. 140. (Wellington New &aland 17-20th October 1966). M. PASTERNAK 131 26 R. Livingston and H. Zeldes Physic. Rev. 1953,90,609. 27 K. Rama Reddy F. de Sousa Barros and S. Debenedetti Physics Letters 1966 20,297. 28 W. A. Dutton and W. C. Cooper Chem. Rev. 1966,66,657. 29 G. Nagarajan J. Mol. Spectr. 1964 12 198. 30 A. Narath Physic. Rev. A 1965 140 854. 31 S. D. Senturia and G. B. Benedek Physic. Reu. Letters 1966 17,475. 32 G. M. Kalvius L. D. Oppliger and S. L. Ruby Physics Letters 1965 18 241. 33 R. M. Steffen E. Mathias and W. Schneider US. Atomic Energy Commission Report TID-15749 34 P. Zory Physic. Rev. A 1965,140,1401. 35 M . Pastern& unpublished results. 1962.

ISSN:0430-0696    

DOI:10.1039/SF9670100119

出版商:RSC    

年代:1967

数据来源: RSC