Magnetic and Electric Hyperfine Interactions of FeS7” Nuclei in Compounds Y3- Ca Fe,- Sn O, (O.O<x<2-0) BY I. S. LYUBUTIN* E. F. MAKAROV~’ and V. A. POVITSKII~’ Inst. of Chem. Physics and Crystallography Institute Academy of Science Leninskii prospekt Moscow U.S.S.R. Received 27th July 1967 The effective magnetic fields at the Fe5’ nuclei in the octahedral and tetrahedral sub-lattices of the tin-substituted YIG YJ~xCaxFe5~xSn,012 (0.0 < x< 2.0) were investigated by the Mossbauer method. The dependence of Heff on x obtained from the experimental results is compared with the correspondent dependence of the magnetic moments of the Fe3+ ions derived from the Gilleo theo- retical model. The reasons for the difference in the concentration dependence of Heff and the magnetic moment for the a-sub-lattice are discussed.The contribution to the Heff from the dipolar interaction of neighbouring magnetic ions was negligible in the iron garnets. The main reason for the difference in the values of Heff in the a- and d-sub-lattices of the iron garnets is connected with the fact that the chemical bonding of Fe3+ ions in tetrahedral sub-lattice is partly covalent. The quadrupole effects in YIG could not be observed on the polycrystalline samples below the Curie temperature. The quadrupole splitting and the isomer shifts for the both sites of Fe3+ at T>T were observed. Froin the line intensity the distribu- tion of Sn4+ and Fe3+ ions in the sub-lattices was obtained. It is known that Fe3+ ions are located on the two crystallographically non- equivalent sites in different quantities in the yttrium-iron garnet (YIG) structure.Three Fe3+ ions are at the tetrahedral sites (d-sub-lattice) and two Fe3+ ions are located on octahedral sites (a-sub-lattice) in the formula unit (Y3][Fe,] (Fe,) O12. In the YIG substituted by tin it appeared possible to measure the Mossbauer effect on the nuclei of magnetic (Fe57) and non-magnetic (Sn119) ions and to study the effective magnetic fields He, acting on those nuclei in the different sub-lattices. The special interest of this system is due to the presence of the large effective magnetic fields at the Sn119 nuclei.1-4 This paper deals with the investigation of the Mossbauer effect on the FeS7 nuclei in polycrystalline samples of the Sn4+-substituted yttrium-iron garnets Y,-xCa,Fe,-,Sn,O, (x == 0 ; 0,l; 0,3; 0,5; 0,7; 0,9; 1,O; 1,l ; 1,2; 1,5 and 2,O).The source CoS7 in a Cr matrix was at the room temperature. MAGNETIC HYPERFINE INTERACTIONS In fig. 1 the Mossbauer absorption spectra of Y3-,Ca,Fe,-,Sn,01 for different values of x at 77°K are shown. The spectra up to x = 1.1 can be interpreted as two Zeeman effects arising from the Fe57 nuclei in the a- and d-sub-lattices where they are subjected to the different effective magnetic fields. We consider that the more intensive lines are produced by the d-~ub-lattice,~ where there are more Fe3+ ions than in the a-sub-lattice. When x = 1.1 on the *Crystallography Institute Academy of Sciences U.S.S.R. Moscow. ?Institute of Chemical Physics Academy of Sciences U.S.S.R. Moscow. 31 32 MAGNETIC AND ELECTRIC HYPERFINE INTERACTIONS Zeeman background an intensive doublet appears; this is due to the transition of some part of the Fe3+ ions to a paramagnetic state.When the magnetic splitting is close to zero the spectra are transformed to give the asymmetric doublet. Non-magnetic ions Sn4+ are located in the octahedral a-sub-lattice in the Y3-,Ca,Fe,-,Sn,01 system.6 The part of Fe3+ ions in the tetrahedral d-sub-lattice have as first neighbours (in the a-sub-lattice) only nonmagnetic ions Sn4+ when x= 1.2 v mm/sec FIG. 1 .-Mossbauer absorption spectra of the iron garnets Y3-xCaxFe5-,SnxO12 for different values of x at 77°K. Sn4+ ions are randomly distributed at the a-sites. These Fe3+ ions are in the paramagnetic state because d-a-interaction is the main effect in the iron garnets. We consider that the appearance of the doublet in the central part of the Moss- bauer spectrum is connected with the Fe3+ ions in the d-sub-lattice ; these are excluded from exchange interaction since their first neighbours are non-magnetic Sn4+ ions.The dependence of the effective magnetic field at the nuclei Fe3+ ions in a- and d-sub-lattices on x at 77°K is shown in fig. 2. In the interval O<x< 1.1 the field behaviour for both sub-lattices is similar. Then values of the field sharply decreases over the small interval of x 1.1 <x< 1.2 and only the broadening of the doublet lines shows that in the compounds Y1.,Cal.2Fe3.8Snl.2012 (T = 290°K) and Y1.5Ca1.5Fe3.5Sn1.5012 (T = 140°K) ' the effective magnetic field is still acting at the Fe nuclei. This field disappears only when x = 2.0. The existence of this field up to x = 2-0 may be explained by the appearance of the superparamagnetic regions for 1-2 < x < 2.0 in the ferrite.The superparamagnetic transition at 130°K for the compound Y1.5Cal.5Fe3.5Sn1.5012 was predicted recently by Ishikawa.' 1. S. LYUBUTIN E . F . MAKAROV AND V. A . POVITSKII 33 It is interesting to compare the values of the effective magnetic fields with the values of the magnetic moments of the sub-lattices. To estimate these magnetic moments we used the theoretical model of Gi1le.0~ Fig. 3 shows the theoretical curves (lines) of the magnetic moments corresponding to one Fe3+ ion in the a- 0 0 . 5 1.0 I *5 2.0 X FIG. 2.-The dependence of the effective magnetic field Hcff at the nuclei of the Fe3f ions in a- and d-sub-lattices of Y3-xCaxFe,-,SnxOlz against x at 77°K.Also shown is the concentration dependence of the Htff at the Sn1I9 nuclei in these garnets3 X FIG. 3.-The theoretical dependence of the magnetic moments of the Fe3+ ion in the a- and d-sub- lattices (lines) and values of magnetic moments obtained from the experimental values of the Heff and d-sub-lattices calculated by us using Gilleo model for the (Y,-xCax)[Fe,-,Snx] (Fe,) OI2 garnets. The value of Heff depends on the effective magnetic moment of the ion.'' Assuming that the effective magnetic field H& in the d-sub-lattice is proportional to the magnetic moment of Fe3+ ion in this sub-lattice we have calculated the values of the magnetic moments of d-sub-lattice Md(x) for different x using the measured (crosses) against x for the Y3-xCaxFe5-xSnxOl system. 2 34 MAGNETIC AND ELECTRIC HYPERFINE INTERACTIONS values of H&(x).The values of the magnetic moments of Fe3+ ions in the d-sub- lattice obtained from the relation are shown in fig. 3 (crosses); where Md(0) = 5pB-is the magnetic moment of Fe3+ ion and H&(O) is the field at the nucleus of this ion in the d-sub-lattice of the pure YIG (x = 0). The values of the calculated magnetic moments are in a good agree- ment with the theoretical values in the interval O<x< 1.0 (fig. 3). The theoretical values of the magnetic moment of Fe3+ ion in the a-sub-lattice do not depend on the substitution and are equal to 5,uB for all x (fig. 3). However from experimental data He, (at 77°K) in the a-sub-lattice decreases when x rises in the same way as in the d-sub-lattice (fig. 2). Analysis shows that this result cannot be explained using only the temperature effects i.e.by the approach of the Curie point to the temperature of the measurement. The decrease of He, in the a-sub-lattice in comparison with He, expected theo- retically from values of the magnetic moments of Fe3+ can be probably explained by the " reverse " influence of the d-sub-lattice on the a-sub-lattice when the substitution takes place in the a-sub-lattice. The Gilleo theory does not take into consideration this influence. Actually according to the Gilleo model when the magnetic ions are substituted by non-magnetic ions in the a-sub-lattice the effective magnetic moment of Fe3+ ion in the d-sub-lattice decreases with decreasing number of the exchange linkages with the magnetic ions in the a-sub-lattice.However the variation in the value of the effective magnetic moment of the magnetic ion may probably take place not only due to the decrease of the number of the exchange linkages but also due to the decrease of the effective magnetic moment of the neighbouring ion with which the exchange interaction takes place. We now discuss the differences between He, at the nuclei of Fe3+ ions in octahedral site and Heff in the tetrahedral site of YIG. Freeman and Watson lo suggested that possible reasons for the decrease of He, at the nuclei of the Fe3+ ions in the d-sub-lattice in comparison with the field in the a-sub-lattice were (i) distortions of cubic symmetry (ii) covalent bonding effects. For the pure YIG this difference is aboutc90 kOe and when we change the concentration of the Sn4+ ions this difference remains approximately constant up to x = 1.0 (fig.2). However at the same time when x rises the number of the magnetic ions in the a-sites decreases (because they are substituted by Sn ions) which are the nearest neighbours for the d-sites. This should decrease the value of the dipolar field in the d-sub-lattice and consequently decrease the difference between He, in the a- and d-sites. As the measured difference in the He, in the a- and d-sites remains constant up to x = 1.0 (see fig. 2) in spite of the fact that the number of the neighbouring magnetic ions which cause the dipolar field decrease two-fold for x = 1.0 we may conclude that the dipolar fields slightly influence the He, value. The values of the isomer shifts for Fe3 ions in the tetrahedral and the octahedral sub-lattices of the garnets Y,-,Ca,Fe,-,Sn,O, and unusual large value of the quad- lupole constant (see table 1) confirm that the chemical bonding with oxygen ions has partly the covalent character l1 of Fe3+ in the tetrahedral sites.This covalent bonding is probably the main reason of the difference between He, in the octahedral and tetrahedral sub-lattices of rare-earth iron garnets. When the x rises the values of the isomer shifts and the quadrupole constants for the a- and d-sub-lattices of the Sn4+-substituted YIG remain approximately constant (see later). This result explains the constancy of the difference between He, in the a- and d-sub-lattices up to x = 1.0 1. S . LYUBUTIN E . F . MAKAROV A N D V . A . POVITSKII 35 ELECTRIC HYPERFINE INTERACTIONS Alf and Wertheim l2 were the first to measure the quadrupole interaction of Fe5' nuclei in a single crystal of the YIG using the Mossbauer method in an external magnetic field of 0.5 kOe applied along the [ 11 11 and [ 1001 axes.The measurements were made at 300'K i.e. lower than the Curie point (Tc = 550°K). No quadrupole interaction was found in the polycrystalline samples of the YIG at 300°K nor at 80'K,5 and thus the authors concluded that e2qQ = 0. Our measurements at the temperatures T<Tc also do not reveal any quadrupole interaction. We explain this contradiction in the following way when electric quadrupole and mag- netic dipole interactions exist simultaneously and e2qQQp.H and I = 3/2 the (e'4Q) megwed e2qQ (3 cos*O- 1)/2 where 8 is the angle between He, direction (which coincides with the direction of the easy magnetization axis in the absence of the external magnetic field) and the direction of the symmetry axis of the electric field gradient (EFG).13 When 8 = 54'44' the term 3 cos20- 1 = 0 and consequently (e2qQ)measured = 0- This case is realized in polycrystalline samples of YIG at temperatures below 7'.In fact in the YIG the easy magnetization axis is in the [ l l 11 direction l4 while the axis of symmetry of the EFG for the d-sub-lattice coincides with the [loo] direction and one for the a-sub-lattice coincides with the [ I l l ] direction. Thus in the absence of the external magnetic field the angle 8 is 54'44' for all Fe3+ ions in the d-sub-lattice and in the a-sub-lattice 6 = 70'32' for 75 % of the Fe3+ ions i.e. (e2qQ)measured = 4 e2qQ and 6 = O only for 25 % of Fe3+ ions i.e.Thus observation of the quadrupole interaction at the Fe57 nuclei in poly- crystalline YIG below Tc is practically impossible. The precise values of e2qQ for both sites of Fe3+ -ions in this case can be obtained in the paramagnetic state. Such measurements were made on some rare-earth iron garnet.15 Mossbauer spectra of the different tin-substituted YIG at temperatures T > T are shown in fig. 4. All of them except the last consist of 3 lines of different intensity. The width of outside lines of the spectra are always the same and in the pure YIG they are equal to 0.3 rninlsec and in the substituted garnets 0.4+-0-45 mm/sec. These spectra may be interpreted as consisting of the two quadrupole doublets where the lines corresponding to the larger velocities coincide and so form the most intensive third line of the spectrum.The more intensive doublet with the greater splitting and the lower isomer shift belongs to the Fe3+-ions in the d-sub-lattice. The results of the present work on pure YIG are shown in the table 1 with the data from ref. (ll) (12) (15). When the tin content rises T decreases and for x = 1.0 T = 2 6 0 K 6 Thus for this sample we can measure the temperature shift in the 300-573°K range. The value of the shift is 0.08+0.03 mm/sec for both Fe3+ sites. On the other hand the isomer shifts for both sites of Fe3+ are almost indepen- dent of the tin content. This allowed one to correct the data of Nicholson and Burns (see table 1) in which the isomer shifts of YIG have been measured at 610"K but corrected for 300°K using the value 0.22 mnilsec obtained for pure iron.I6 For the spectra of the tin-substituted YIG the decrease of the intensity of the second and the third lines of the spectrum is observed with x increasing (see fig.4). When x = 2 the second line almost disappears. We therefore suggest that the remaining two lines correspond to Fe3+ ions in the d-sub-lattice only. Their areas are equal. The doublet of the a-sub-lattice appears to be asymmetrical with a ratio of the areas of 1 0.75. This asymmetry is probably connected with the anisotropy of the Debye-Waller factor for Fe3+ ions in the a-sub-1atti~e.l~ The comparison of (e2qQ>measured e'qQ. 36 MAGNETIC AND ELECTRIC HYPERFINE INTERACTIONS the total areas of the components of the both doublets in the pure YIG (when the relative number of the Fe3+ ions in the both sub-lattices is taken into account) allows an estimate of the ratio of the Mossbauer probability in the sub-lattices fa/h=l.l.This ratio probably does not change when substitution takes place. The total number of Fe3+ ions corresponding to the formula unit Y,-,Ca,Fe,-,Sn,012 3 I ~ s " ~ ~ k ~ .~ .r 7 '"-"**-*-. :.,. ~". ~ ~ ~ ',,. ,.'.*.."*..'. I 1 3 5 .-*.':;.. . p #. * . f 33 # . . . *...-. . . . 5 . '.-. XE2.O 29 X=0.9>-. * *.' I 1 -1.0-0.5 0 0.5 1.0 -1.0-0-5 0 0.5 1.0 velocity (mrnlsec) FIG. 4.-Mossbauer spectra for two Fe3f ions sites of Y3-xCaxFes-xSnxO12 in the paramagnetic temperature region. TABLE 1. quadrupole A = 1.e2qQ splitting (mmlsec) isomer shifts d* at 300°K (mmlsec) a-site d-site a-site d-site data l2 0.94 f0.19 0.78 f0-16 0.57 f0-0511 0*26f0.0511 data l5 0.52 f0-04 0.92 f0-04 0.49 f0.04 0.32 f0-04 present work 0.47 f0.02 0-93 f0-02 0.46 f0.05 0.23 f0.05 T = 300°K - - T = 610°K 0-36 f0.07t 0.19 f0.071- T = 573°K *relative to stainless steel ; t the values corrected by us (see text).changes from 5 when x = 0 to 3 when x = 2. Hence as above we can obtain the number of Fe3+ ions in each sub-lattice from the experimental spectra (see fig. 5). These results prove that Sn4+ ions substitute for Fe3+ ions in the octahedral sub-lattice only. This fact confirms the data of Geller et aL6 obtained earlier by X-ray investigation for x = 1 and x == 2. The Mossbauer method is however simpler and more convenient than X-ray analysis for studying substitution in rare- earth iron garnets. I .S. LYUBUTIN E . F . MAKAROV AND V. A . POVITSKII 37 The isomer shifts and the quadrupole interactions for both sites of Fe3+ ions were almost independent on the tin contents in garnets. The quadrupole splitting for both a- and d-sub-lattices is independent on temperature in the range from the Curie point to 600°K. X FIG. 5.-Number of Fe3+ ions in the a- and d-sub-lattices of Y3-,CaxFe,-xSnxO12 against the concentration of the tin ions x. We thank Prof. K. P. Belov Prof. L. M. Belyaev and Prof. V. I. Goldanskii for their interest in this work and for useful discussions. We are grateful to Y. V. Baldochin for help in performing the experiment. K. P. Belov and I. S. Lyubutin JETP Letters 1965 1,26. V. I. Goldanskii M. N. Devisheva V. A. Trukhtanov and V. F. Belov JETP Letters 1965,1 31 ; Physics Letters 1965 15 317.K. P. Belov and I. S. Lyubutin JETP 1965,49 747. V. I. Goldanskii M. N. Devisheva E. F. Makarov G. V. Novikov and V. A. Trukhtanov JETP Letters 1965 4 63. R. Bauminger S. G. Cohen and A. Marinov S. Ofer Physic. Rev. 1961 122 743. S. Geller R. M. Bozorth M. A. Gilleo and C. E. Miller J. Physics Chem. Solids 1960,12,111. K. P. Belov and I. S. Lyubutin Kristullographiya 1965 10 351. M. Gilleo J. Physics Chem. Solids 1960 13 33. * Y. Ishikawa J. Appl. Physics 1964 35 1054. l o R. E. Watson and A. J. Freeman Physic. Rev. 1961 123 2027. I1 L. R. Wakler G. K. Wertheim and V. Jaccarino Physic. Rev. Letters 1961 6,98. l2 C. Alf and G. K. Wertheim Physic. Rev. 1961 122 1414. l3 G. K. Wertheim M&buuer Eflect (Acad. Press. N.Y. London 1964). l4 J. F. Dillon Physic. Rev. 1958 111 1476. l6 R. S. Preston S. S. Hanna and J. Heberle Physic. Rev. 1962 128 2207. I7S. V. Karaygin Doklady 1963,148 1102. W. J. Nicholson and G. Burns Physic. Rev. A 1964,133.