- Email: [email protected]
EPR of Mn2+ in single crystals of calcium hexachlorostannate hexahydrate
Volume I
CI1EMICAL
I I, number 6
EPR OF M$+
IN SiN6LE
CRYSTALS
PHYSICS
LETTERS
OF CALCIUM HE~~HLOROST~ATE
16 November
I9S4
HEXAHYDRATE
R. HRABANSKI
Received 6 July 1984; in finz11form 3 September 1984 hesahydratc An EPR study has been carried out on kIn2* -doped single crystals of cslcium hcsachforostannate [Cn(Hz0)6 1 SKI6 in tbc temperature range 77-378 K at X-band frequencies (9.3 GHz). hln 2+ ions substitutin_e the divalenr metal exhibit a unique - manetic compfes with the z-axis directed along the c-axis of the crystal. The temperature of the zero-field splitting parameter is discussed. 1
~exac~llorostannate hexahydrates of several divalent metals have been considered to be isamorphous with a large number of salts of general fommla [M&0)6] XY6, where IM is a divalent metal, X is a quadrivalent element such as Si, Sn, Ti, Zr and Ge, and Y may be F or Cf. Many crystals of fluorosilicates, fhrorotitanates and ~uoroge~~l~ates conta~ing MnZ+ as ionic impurities have been studied by EPR [ 1-S]_ However, very little work has been reported on the EPR of chlorostannates [9, IO] _ In recent publications. we have presented EPR spectra ofMn~* in [Mg(H20)6] SnCI, over a wide range of temperature [ lo]_ In this paper, the results of EPR measurements on M$‘-doped single crystakr of range 77-378 K fca(H2 016 I SnC16 in the temperature are reported. Kitazume et al. [ 1 l] detennined the crystal structure of[Ca(H20)6] SnC& at room temperature by Xray diffraction. The crystal is trigonal with space group R3. The structure consists of lle~aaqu~~ciunl cations and hexachlorostannate anions arranged alternately in cohunns along the threefold axes. The ligands in the caions form
slightly
distorted-octahedra,
while
the ligands
in the anions
form almost regular octahedra about the central metal atoms. It was suggested that a weak OH ___Cl hydrogen bond exists. both along the columns of ions and between these columns. The single crystals of [Ca(H20)6} SnCI, employed in this work were grown by slow evaporation at room temperature of an equimolar_solution of reagent-grade CaCI, - 6Hz 0 aird SnC14;5H,O in distilled water. The M& was introduced into-the host lattice by adding a 0 Op9-2614/84/S 03.00 Q Flsevier Science Publishers (North-Holland Physics Publishing Division)
B-V,
dependence
small amount (O-5 wt%) of manganese chloride to the solution. Neariy colourless crystals in the foml of rhombohedra with well developed faces were obtained. The crystals were hygroscopic and were protected from exposure to air by coating with paraffin oil. EPR experiments were performed on the standard SE/X-28 Polish-made spectrometer operating at X-band (9.3 GHz) frequencies, provided with 100 kHz modulation. The magnetic flux density was measured by proton resonance_ 1: I-diphenyl-2-picryl-hydrazyl (DPPH) has been taken as the reference forg-factor detemlination (g = 2.0036 2 0.0003). EPR spectra recorded with the external magnetic field applied along directions pamIle to and perpendicular to the trigonal axis (c-axis) of the crystal are shown in figs. 1 and 2. The EPR spectrum of Mn’+ consists of five fiie-structure (fs) components, which are split into six lines of hyperfme structure (hfs). At intermediate directions, the spectrum was more complicated due to forbidden transitions. Studies of the anguIar variation of the hfs sextets showed that the spectrum is isotropic for rotation about the c-axis. Thus, the spectrum clearly corresponds to the unique axially symmetric magnetic complex [Mn(H20)6] 2t with the z-axis directed along the c-axis of the crystals_ The EPR spectrum was analysed using the foIlowing spin Hamiltonian appropriate for the trigonal symmetry of the crystal field:
607
volume
i Il. number
6
CIIEhliCAL
PHYSICS
112m S -3i2.m
-112.m i -519,m’ ’
*
I
;
1
I
’
1984
-112 m
-3-m
’
16 November
LEITERS
*
’
112.m
I2.m
I
s 3/2.m I I 1 1
’
2OmT
D PPH I+_
1. Room-temperature
EPR spectrum
of hln*+- m [GI(H~O)~
where t!le symbols have their usual meanings 1121. The spin-Hamiltonian parameters obtained at room temperature and liquid-nitrogen temperature are listed in table 1. The parameters were obtained by using the resonance field positions up to the second-order perturbation for the above spin Hamiltonian. The value of the parameter b$ was found to be zero within experimental error. Tbeg-value was found to be slightly less than the free-electron value, implying an absence
I
312.m
I
1/2,m =e -1 I2 .m I I I I I I zkl I2 m I , I 1 -1I2.m e
I
I
I
-: 312 m
I?
~
ZOmT DPPH Fig. 2. Room-temperature EPR spectrum IC~(tl20)~ 1 SnClg wit11 B perpcndiculsr
of htn”in to the c-asis
(D =90”).
] SnC16 with5
prlrallel
to the c-ads
(0 = 0).
Table 1 Spin-Hamiltonian parameters for hln’+ in [Ca(Hz0)6 1SnC16 single crystals at 293 and 77 K. AU the crystal-field end I~ypcrtine pararnctcrs are in units of IOm4 cm-’ Parameter
293 Ii 1.0007 -176.5 -2.4
77 I( F 0.0005 f. 1.0 t 0.3
2.0005 -188.8 -2.9
+_0.0005 f I.0 f 0.3
-93.6
*OS
-91.3
f 0.5
-89.6
f 0.5
-905
f 0.5
of covalent bonding between the Mr?+ ion and the surrounding ligands [I;]_ The relative signs of the spin-Hamiltonian parameters were determined by considering the variation of the spacings of the byperfine lines. The sign ofA,, was found to be negative from spacings of adjacent hyperfine lines. The spacing of successive (21+ 1) hyperfiie lines corresponding to the fine transition M c-, M - 1, for B IIz. is the absolute value ofA,, + (At/B,) (M - m), where BO = h/go [l] _Thus, if A,, is negative, the spacing should decrease as M increases for a given value ofm. The measurements show that this effect occurs; for example, for n2 = l/2 the spacing varies from 10_30mTforM=--312 to9.15mT forM=5/2. A difference of 1_15 mT is in good agreement with the value of 1.06 mT determined from the second-order perturbation expression. From this expression, it follows also that the line corresponding td ~4 = 5/i should lie at a field greater than that fdr m = --5/2 by &%I,,
Vohm~e 11 1. number 6
C!lEMlCAL
PHYSICS
- (SA~/2Bo) (Ul1). The absolute value of this overall separation for a hyperfiie set is greater for negative M if A ,, is negative_ From standard theory [ 12]1 it follows that forf3 ]Iz theM @M - 1 transitions for negative by appear at lower magnetic field values for negativeM relative to those for which M is positive. This means that, if the overall splitting of-hyperfme lines decreases with increasing field, b! and Al, have the same sign. This effect occurs because the overall splitting decreases from 50.92 mT forM= -3/Z to 45.67 mT forM = 5/I. A difference of 5.25 mT is also in good agreement with the value of 5.36 mT determined from the second-order perturbation expression. Thus, the parameters bg and A ,, (also A,) have the same sign. For an absolute detemunation of the sign, data at liquid-helium temperature are required_ which is not possible in our laboratory. The spin-Hamiltonian parameters were then used for the calculation of the positions of the resonance lines corresponding to the allowed AM = -+ 1. LMB= 0 transitions_ In fig. 3. a comparison of the observed and calculated angular variation of fs transitions M ++ M - 1 in the (1 i0) plane is presented_ The temperature dependence of the EPR spectra in the temperature range 77-378 K was studied. As the temperature was decreased, only an increase of the absolute value of the zero-field splitting parameter b$ and small changes of the widths of the resonance lines were observed_ The other spin-Hamiltonian parameters were nearly the same, within experimental error. PreviousIy, Graybeal et al. [ 141 reported a doublet splitting of the NQR frequency of 3”Cl in [Ca($0)6] SnCl6 below 113 K, which they attributed to the occurrence of a phase transition or two different crystallographic environments of the chlorine atoms. However, the non-equivalence of the positions of the chlorine atoms may also be associated with the disorder of SnClzgroups, and it would be expected for compounds having the rhombohedral structure [ 151. Recently, Kitazume et al. [ 1 l] found several successive phase transitions in [Ca(&0)6] SnCI, using the differential thermal analysis (DT!A) method. We propose that the peaks of the DTA curves observed by Kitazume et al. [ 1 l] may instead arise from the gradual dehydration at temperatures of 360 and 401 K and decomposition of the crystals at 413K. Our measurements have not confirmed the occurrence of any phase transition in these crystals.
LE-l-l-ERS
16 November
1984
r
1
LOO
E al
35c
3oc k-312--5
cl’
30‘
60’ o---
90’
Fis. 3. Arqlar
variation of the allowed fs transitions in the (170) plane of hln’+in [Ca(HzO)e) SnCl6. The circles arc the esperimental values while the solid curves are calculated from the parameters listed in table I.
The temperature variation of the zero-field splitting parameter bt was studied. Following the work of Walsh et al. [ 161, the temperature dependence of the zerofield splitting parameter can be discussed in terms of implicit contributions caused by static crystal distortions from themlal expansion or contraction of the lattice and explidit contributions resulting from lattice vibrations_ Assuming a harmonic motion of the nearest-neigbbour ions, the explicit part is usually described by 117-191 &T)
= b:(O)
•t i? coth @lo/XT),
(2)
where b$ (0) is the splitting in the absence of lattice vibrations (at 0 K), 6 involves information about the coupling of the ion to the lattice vibrations and w is 609
Volume 11 1. ntrmbrr 6
CIiEhlICAL
I’IiYSlCS LIXTERS
16 November
1984
So far we have considered the exjkit effect, but the implicit effect may also cont&bute to the tempera:
ture dependence of $_ Unfortunately, the separation of the two effects requires a study of the pressure dependence of b$ and knowledge of the equation of state of the material. Such data are not available up to this moment. However. only a small anisotropy of thermal expansion of the lattice is observed [ 1 I] _This means, in turn, that the contributions of the implicit effect to the temperature dependence ofb! are not significant_
L___.
~;OO
_..
_._._I___.
200
J
~._.__l___.__.
300
T(K)
This work was supported in part by the Polish Academy of Sciences under the Project MR.I.9.
-
Fig. 4. Tcmpcrature
dcpendcncc of the zero-field splitting paThe piottcd circtcs correspond to the cspcrintcnr;ll values. Tbc solid curve is a tit to eq. (2) with parameters +cn in the test. =
ramctcrbq.
the vibrational frequency of the particular mode. The experimental values ofby were fitted to eq. (2). Fig.4 shows that the measured temperature dependence of-b; can be SatiSfaCtOdy described by eq. (2) over a wide range of temperature. The poor description in
the high-temperature region is due to the failing of the harmonic approximation. The following values for the parameters in eq. (2) were obtained: b;(O) = -196
X IO-’ cm-* _
6 =6.2X
lo-“cm-‘.
no = 135
cm-~.
This would mean that coupling of the impurity corn-plcses ~MI~(H~O)~]~+ with a vibrational mode with frequency abc%t 135 cm- 1 is responsible for the observed temperature dependence orb:. No IR or Raman ~neasuren~ents of the vibrational spectrwn of [Ca(H20)6] SnC16 doped with Mn3-+ have, to our knowledge. been performed. However, Raman studies ofvibrational spectra in nickel chlorostannate [20] show that one of the frequencies assigned to the N-0 bending mode is around 204 CI~-~ _On the other hand, a value of 260 con-* , assigned to the Mn-0 bending mode. was found in manganese fluorotitanate [2 l] _ The lower values of the vibrational frequencies in these systems were assigned to the lattice modesThus, the value of 135 CIN-~ obtained by us seems to be too low. 610
References [I] B. Blcancy and D.J.E. Ingram. Proc. Roy. Sot. A205 (19.51) 336. 121 E. Friedman and \V. Low, Pbys. Rev. 120 (1960) 404. [3] R. Ilrabariski, P.R. Sczaniccki and J. Stankowslii, Pbys. Star. Sol. 5 I ;I (1979) 143. [41 Yu.V. Yablokov. hl.hl. Zaripov. A.hI. Ziatdinov and R.L. Dzrvidovicb. Cbcm. Phys. Letters 48 (1977) 443. [S j G. Jayaram and G.S. !&try. C&m. Pbps. Letters 77 (1981) 314. [6] G. Jayaram and G.S. Sastry. CIICIK Pbys. Lcttcrs 97 (1983) 431. 171 R.S. Rubins. J. Chem. Pbys. 70 (1979) 4363. IS] SK;. hlisra, Solid State Commun. 45 (1983) 967. [9] Y. Ajiro. S.A. Friedberg and N.S. van der Ven, Phys. Rev. 12 (1975) 39. [IO] R. IIrabaAski, Pbys. Stat. Sol. 77a (1983) Kl43. Ill ] T. Kitaznmc, hf. Sckizaki and hi. Subara. J. hlol. Strucr. 58 (1980) 161. [ 121 A. Abragam and B. B&any, EIectron paramagetic resonance of transition ions (Chrendon, Osl‘ord, 1970). [I31 E. Simanck and #.A. hluller, J. Pbys. Cbem. Solids 31 (1970) 1027. [ 141 JDD. Graybcal. RJ. hIcKown and S.D. Ing, J. Pbys. Cbcm. 74 (1970) 1814. [ 15 1 S. Ray, A. Zalkin and D.11. Templeton. Acta Crystl B29 (1973) 2741. [ 161 N’.hl- Walsh Jr., J. Jeener and N. Rloembergen. Pbys. Rev. 139A (1965) 133s. [ 171 W.hI. Walsh Jr.. Phys. Rev. 1 I4 (1959) 1473. [IS] G. Ptistcr, W. Dreybrodt -and W. Assmus, P11ys. Stat. Sol. 36 (1969) 35 1. [IV I R.A. Serway, Phys. Rev. B3 (1971) 606. [20 J D.M. Adams and W.R. TrumbIc. Inog. Cbim. Acta 10 (19743 235. [?I] P. Cboudbury, B. Ghosh. R.S. Ra~huvansbi and 1I.D. Bist. J. Raman Spectry. 6 (1983) 99.
CI1EMICAL
I I, number 6
EPR OF M$+
IN SiN6LE
CRYSTALS
PHYSICS
LETTERS
OF CALCIUM HE~~HLOROST~ATE
16 November
I9S4
HEXAHYDRATE
R. HRABANSKI
Received 6 July 1984; in finz11form 3 September 1984 hesahydratc An EPR study has been carried out on kIn2* -doped single crystals of cslcium hcsachforostannate [Cn(Hz0)6 1 SKI6 in tbc temperature range 77-378 K at X-band frequencies (9.3 GHz). hln 2+ ions substitutin_e the divalenr metal exhibit a unique - manetic compfes with the z-axis directed along the c-axis of the crystal. The temperature of the zero-field splitting parameter is discussed. 1
~exac~llorostannate hexahydrates of several divalent metals have been considered to be isamorphous with a large number of salts of general fommla [M&0)6] XY6, where IM is a divalent metal, X is a quadrivalent element such as Si, Sn, Ti, Zr and Ge, and Y may be F or Cf. Many crystals of fluorosilicates, fhrorotitanates and ~uoroge~~l~ates conta~ing MnZ+ as ionic impurities have been studied by EPR [ 1-S]_ However, very little work has been reported on the EPR of chlorostannates [9, IO] _ In recent publications. we have presented EPR spectra ofMn~* in [Mg(H20)6] SnCI, over a wide range of temperature [ lo]_ In this paper, the results of EPR measurements on M$‘-doped single crystakr of range 77-378 K fca(H2 016 I SnC16 in the temperature are reported. Kitazume et al. [ 1 l] detennined the crystal structure of[Ca(H20)6] SnC& at room temperature by Xray diffraction. The crystal is trigonal with space group R3. The structure consists of lle~aaqu~~ciunl cations and hexachlorostannate anions arranged alternately in cohunns along the threefold axes. The ligands in the caions form
slightly
distorted-octahedra,
while
the ligands
in the anions
form almost regular octahedra about the central metal atoms. It was suggested that a weak OH ___Cl hydrogen bond exists. both along the columns of ions and between these columns. The single crystals of [Ca(H20)6} SnCI, employed in this work were grown by slow evaporation at room temperature of an equimolar_solution of reagent-grade CaCI, - 6Hz 0 aird SnC14;5H,O in distilled water. The M& was introduced into-the host lattice by adding a 0 Op9-2614/84/S 03.00 Q Flsevier Science Publishers (North-Holland Physics Publishing Division)
B-V,
dependence
small amount (O-5 wt%) of manganese chloride to the solution. Neariy colourless crystals in the foml of rhombohedra with well developed faces were obtained. The crystals were hygroscopic and were protected from exposure to air by coating with paraffin oil. EPR experiments were performed on the standard SE/X-28 Polish-made spectrometer operating at X-band (9.3 GHz) frequencies, provided with 100 kHz modulation. The magnetic flux density was measured by proton resonance_ 1: I-diphenyl-2-picryl-hydrazyl (DPPH) has been taken as the reference forg-factor detemlination (g = 2.0036 2 0.0003). EPR spectra recorded with the external magnetic field applied along directions pamIle to and perpendicular to the trigonal axis (c-axis) of the crystal are shown in figs. 1 and 2. The EPR spectrum of Mn’+ consists of five fiie-structure (fs) components, which are split into six lines of hyperfme structure (hfs). At intermediate directions, the spectrum was more complicated due to forbidden transitions. Studies of the anguIar variation of the hfs sextets showed that the spectrum is isotropic for rotation about the c-axis. Thus, the spectrum clearly corresponds to the unique axially symmetric magnetic complex [Mn(H20)6] 2t with the z-axis directed along the c-axis of the crystals_ The EPR spectrum was analysed using the foIlowing spin Hamiltonian appropriate for the trigonal symmetry of the crystal field:
607
volume
i Il. number
6
CIIEhliCAL
PHYSICS
112m S -3i2.m
-112.m i -519,m’ ’
*
I
;
1
I
’
1984
-112 m
-3-m
’
16 November
LEITERS
*
’
112.m
I2.m
I
s 3/2.m I I 1 1
’
2OmT
D PPH I+_
1. Room-temperature
EPR spectrum
of hln*+- m [GI(H~O)~
where t!le symbols have their usual meanings 1121. The spin-Hamiltonian parameters obtained at room temperature and liquid-nitrogen temperature are listed in table 1. The parameters were obtained by using the resonance field positions up to the second-order perturbation for the above spin Hamiltonian. The value of the parameter b$ was found to be zero within experimental error. Tbeg-value was found to be slightly less than the free-electron value, implying an absence
I
312.m
I
1/2,m =e -1 I2 .m I I I I I I zkl I2 m I , I 1 -1I2.m e
I
I
I
-: 312 m
I?
~
ZOmT DPPH Fig. 2. Room-temperature EPR spectrum IC~(tl20)~ 1 SnClg wit11 B perpcndiculsr
of htn”in to the c-asis
(D =90”).
] SnC16 with5
prlrallel
to the c-ads
(0 = 0).
Table 1 Spin-Hamiltonian parameters for hln’+ in [Ca(Hz0)6 1SnC16 single crystals at 293 and 77 K. AU the crystal-field end I~ypcrtine pararnctcrs are in units of IOm4 cm-’ Parameter
293 Ii 1.0007 -176.5 -2.4
77 I( F 0.0005 f. 1.0 t 0.3
2.0005 -188.8 -2.9
+_0.0005 f I.0 f 0.3
-93.6
*OS
-91.3
f 0.5
-89.6
f 0.5
-905
f 0.5
of covalent bonding between the Mr?+ ion and the surrounding ligands [I;]_ The relative signs of the spin-Hamiltonian parameters were determined by considering the variation of the spacings of the byperfine lines. The sign ofA,, was found to be negative from spacings of adjacent hyperfine lines. The spacing of successive (21+ 1) hyperfiie lines corresponding to the fine transition M c-, M - 1, for B IIz. is the absolute value ofA,, + (At/B,) (M - m), where BO = h/go [l] _Thus, if A,, is negative, the spacing should decrease as M increases for a given value ofm. The measurements show that this effect occurs; for example, for n2 = l/2 the spacing varies from 10_30mTforM=--312 to9.15mT forM=5/2. A difference of 1_15 mT is in good agreement with the value of 1.06 mT determined from the second-order perturbation expression. From this expression, it follows also that the line corresponding td ~4 = 5/i should lie at a field greater than that fdr m = --5/2 by &%I,,
Vohm~e 11 1. number 6
C!lEMlCAL
PHYSICS
- (SA~/2Bo) (Ul1). The absolute value of this overall separation for a hyperfiie set is greater for negative M if A ,, is negative_ From standard theory [ 12]1 it follows that forf3 ]Iz theM @M - 1 transitions for negative by appear at lower magnetic field values for negativeM relative to those for which M is positive. This means that, if the overall splitting of-hyperfme lines decreases with increasing field, b! and Al, have the same sign. This effect occurs because the overall splitting decreases from 50.92 mT forM= -3/Z to 45.67 mT forM = 5/I. A difference of 5.25 mT is also in good agreement with the value of 5.36 mT determined from the second-order perturbation expression. Thus, the parameters bg and A ,, (also A,) have the same sign. For an absolute detemunation of the sign, data at liquid-helium temperature are required_ which is not possible in our laboratory. The spin-Hamiltonian parameters were then used for the calculation of the positions of the resonance lines corresponding to the allowed AM = -+ 1. LMB= 0 transitions_ In fig. 3. a comparison of the observed and calculated angular variation of fs transitions M ++ M - 1 in the (1 i0) plane is presented_ The temperature dependence of the EPR spectra in the temperature range 77-378 K was studied. As the temperature was decreased, only an increase of the absolute value of the zero-field splitting parameter b$ and small changes of the widths of the resonance lines were observed_ The other spin-Hamiltonian parameters were nearly the same, within experimental error. PreviousIy, Graybeal et al. [ 141 reported a doublet splitting of the NQR frequency of 3”Cl in [Ca($0)6] SnCl6 below 113 K, which they attributed to the occurrence of a phase transition or two different crystallographic environments of the chlorine atoms. However, the non-equivalence of the positions of the chlorine atoms may also be associated with the disorder of SnClzgroups, and it would be expected for compounds having the rhombohedral structure [ 151. Recently, Kitazume et al. [ 1 l] found several successive phase transitions in [Ca(&0)6] SnCI, using the differential thermal analysis (DT!A) method. We propose that the peaks of the DTA curves observed by Kitazume et al. [ 1 l] may instead arise from the gradual dehydration at temperatures of 360 and 401 K and decomposition of the crystals at 413K. Our measurements have not confirmed the occurrence of any phase transition in these crystals.
LE-l-l-ERS
16 November
1984
r
1
LOO
E al
35c
3oc k-312--5
cl’
30‘
60’ o---
90’
Fis. 3. Arqlar
variation of the allowed fs transitions in the (170) plane of hln’+in [Ca(HzO)e) SnCl6. The circles arc the esperimental values while the solid curves are calculated from the parameters listed in table I.
The temperature variation of the zero-field splitting parameter bt was studied. Following the work of Walsh et al. [ 161, the temperature dependence of the zerofield splitting parameter can be discussed in terms of implicit contributions caused by static crystal distortions from themlal expansion or contraction of the lattice and explidit contributions resulting from lattice vibrations_ Assuming a harmonic motion of the nearest-neigbbour ions, the explicit part is usually described by 117-191 &T)
= b:(O)
•t i? coth @lo/XT),
(2)
where b$ (0) is the splitting in the absence of lattice vibrations (at 0 K), 6 involves information about the coupling of the ion to the lattice vibrations and w is 609
Volume 11 1. ntrmbrr 6
CIiEhlICAL
I’IiYSlCS LIXTERS
16 November
1984
So far we have considered the exjkit effect, but the implicit effect may also cont&bute to the tempera:
ture dependence of $_ Unfortunately, the separation of the two effects requires a study of the pressure dependence of b$ and knowledge of the equation of state of the material. Such data are not available up to this moment. However. only a small anisotropy of thermal expansion of the lattice is observed [ 1 I] _This means, in turn, that the contributions of the implicit effect to the temperature dependence ofb! are not significant_
L___.
~;OO
_..
_._._I___.
200
J
~._.__l___.__.
300
T(K)
This work was supported in part by the Polish Academy of Sciences under the Project MR.I.9.
-
Fig. 4. Tcmpcrature
dcpendcncc of the zero-field splitting paThe piottcd circtcs correspond to the cspcrintcnr;ll values. Tbc solid curve is a tit to eq. (2) with parameters +cn in the test. =
ramctcrbq.
the vibrational frequency of the particular mode. The experimental values ofby were fitted to eq. (2). Fig.4 shows that the measured temperature dependence of-b; can be SatiSfaCtOdy described by eq. (2) over a wide range of temperature. The poor description in
the high-temperature region is due to the failing of the harmonic approximation. The following values for the parameters in eq. (2) were obtained: b;(O) = -196
X IO-’ cm-* _
6 =6.2X
lo-“cm-‘.
no = 135
cm-~.
This would mean that coupling of the impurity corn-plcses ~MI~(H~O)~]~+ with a vibrational mode with frequency abc%t 135 cm- 1 is responsible for the observed temperature dependence orb:. No IR or Raman ~neasuren~ents of the vibrational spectrwn of [Ca(H20)6] SnC16 doped with Mn3-+ have, to our knowledge. been performed. However, Raman studies ofvibrational spectra in nickel chlorostannate [20] show that one of the frequencies assigned to the N-0 bending mode is around 204 CI~-~ _On the other hand, a value of 260 con-* , assigned to the Mn-0 bending mode. was found in manganese fluorotitanate [2 l] _ The lower values of the vibrational frequencies in these systems were assigned to the lattice modesThus, the value of 135 CIN-~ obtained by us seems to be too low. 610
References [I] B. Blcancy and D.J.E. Ingram. Proc. Roy. Sot. A205 (19.51) 336. 121 E. Friedman and \V. Low, Pbys. Rev. 120 (1960) 404. [3] R. Ilrabariski, P.R. Sczaniccki and J. Stankowslii, Pbys. Star. Sol. 5 I ;I (1979) 143. [41 Yu.V. Yablokov. hl.hl. Zaripov. A.hI. Ziatdinov and R.L. Dzrvidovicb. Cbcm. Phys. Letters 48 (1977) 443. [S j G. Jayaram and G.S. !&try. C&m. Pbps. Letters 77 (1981) 314. [6] G. Jayaram and G.S. Sastry. CIICIK Pbys. Lcttcrs 97 (1983) 431. 171 R.S. Rubins. J. Chem. Pbys. 70 (1979) 4363. IS] SK;. hlisra, Solid State Commun. 45 (1983) 967. [9] Y. Ajiro. S.A. Friedberg and N.S. van der Ven, Phys. Rev. 12 (1975) 39. [IO] R. IIrabaAski, Pbys. Stat. Sol. 77a (1983) Kl43. Ill ] T. Kitaznmc, hf. Sckizaki and hi. Subara. J. hlol. Strucr. 58 (1980) 161. [ 121 A. Abragam and B. B&any, EIectron paramagetic resonance of transition ions (Chrendon, Osl‘ord, 1970). [I31 E. Simanck and #.A. hluller, J. Pbys. Cbem. Solids 31 (1970) 1027. [ 141 JDD. Graybcal. RJ. hIcKown and S.D. Ing, J. Pbys. Cbcm. 74 (1970) 1814. [ 15 1 S. Ray, A. Zalkin and D.11. Templeton. Acta Crystl B29 (1973) 2741. [ 161 N’.hl- Walsh Jr., J. Jeener and N. Rloembergen. Pbys. Rev. 139A (1965) 133s. [ 171 W.hI. Walsh Jr.. Phys. Rev. 1 I4 (1959) 1473. [IS] G. Ptistcr, W. Dreybrodt -and W. Assmus, P11ys. Stat. Sol. 36 (1969) 35 1. [IV I R.A. Serway, Phys. Rev. B3 (1971) 606. [20 J D.M. Adams and W.R. TrumbIc. Inog. Cbim. Acta 10 (19743 235. [?I] P. Cboudbury, B. Ghosh. R.S. Ra~huvansbi and 1I.D. Bist. J. Raman Spectry. 6 (1983) 99.