ESR and Mössbauer studies of Fe3+ ion in calcium boro-aluminate glasses

ESR and Mössbauer studies of Fe3+ ion in calcium boro-aluminate glasses

0038-1098/85 $3.00 + .00 Pergamon Press Ltd. ,Solid State Communications, Vol. 53, No. 11, pp. 985-988, 1985. Printed in Great Britain. ESR AND MOSS...

341KB Sizes 0 Downloads 10 Views

0038-1098/85 $3.00 + .00 Pergamon Press Ltd.

,Solid State Communications, Vol. 53, No. 11, pp. 985-988, 1985. Printed in Great Britain.

ESR AND MOSSBAUER STUDIES OF Fe 3 ÷ ION IN CALCIUM BORO-ALUMINATE GLASSES C.S. Sunandana* and R. Jagannathan t *School of Physics ~'School of Chemistry, University of Hyderabad, Hyderabad - 500134, India

(Received 21 August 1984 by E.F. Bertaut) A correlatedESR and MOssbauer spectroscopic study of the glass system 2CaO-3B203 .A1203 .xFe203 (0.1 -~< x ~ 1) is reported. At lowFe contents (x ~< 0.1) Fe 3 ÷ ions are found only in tetrahedral environments, giving a single, sharp ESR peak at g = 4.13. For higher x, occupation by Fe 3÷ of other-symmetry sites become possible, as seen by the evolution of g = 2.0 and g = 6.0 ESR features. A careful examination of M6ssbauer isomer shift data has yielded an average value of 1.96 A for the Fe a +-O 2 distance, besides confirming the tetrahedral site symmetry. Using this value and quadrupole splitting data the extent of angular distortion (at) from perfect tetrahedral symmetry has been deduced as 30 v from a calculation based on Newman's superposition model. 1. INTRODUCTION DANCE et al. recently reported an ESR study of transition metal ions V 4 +, Fe 3 ÷ and Cu ~* doped in 2CaO.3B203 .Al~O3 glass [1]. In the case of Fe 3 ÷ dopant (0.01%) a single ESR line atg = 4.3 is reported although earlier works on iron-doped glasses do reveal lines corresponding to g = 2.0 and g = 6.0 as well [2-6]. Attempts to explain their origins have also been made [3.7]. This prompted us to undertake the present, concentration- and temperature-dependence study of Fe dopant in the above base-glass system, by means of the ESR and M6ssbauer spectroscopic probes, in the hope of deducing supplementary and corroborative information on the glassy environment. The work of Griffiths is highlighted in the interpretation of ESR results in such systems [7]. Interpretations of M6ssbauer isomer shifts of iron in borate-, borosilicate., and silicate glasses are somewhat ambiguous with respect to the mierosymmetry of the transition ion. The discrepancies are pointed out and an attempt is made to measure the M6ssbauer spectra in these glasses and to assign the site symmetries of the dopant. It is shown that isomer shift data can provide a method of obtaining an estimate of the average distance of Fe 3 +-O 2 -. Using this value the extent of deviation from cubic symmetry, characterised by the twist angle, is deduced from the recent model of Brodbeck and Bukrey [8]. Accounts of IR and photoacoustic spectral studies on this system have already been published [9]. 2. EXPERIMENTAL Calcium boroaluminate glasses containing 10 to

90% Fe20s were prepared as reported in our earlier

work [9]. The colour of the products varied from amber through light and dark brown to brownish-black. They were characterized by X-ray diffraction and i.r. spectra to be glassy as described earlier [9]. M6ssbauer spectra were recorded in a constant acceleration 'Elscint' spectrometer with a 'Promeda' multiehannel analyser using 25 m.C~ 7 Co/Rh source. The calibration of velocity is carried out using sodium nitroprusside as standard for paramagnetic spectra. Magnetic susceptibility measurements were carried out using a Guoy magnetic balance. ESR spectra were recorded on a JEOL PE-3X X-band spectrometer equipped with a 100 KHZ field modulation unit. High temperature spectra (27 < T < 200°C) were recorded with a variable temperature accessory. 3. RESULTS AND DISCUSSION 3.1. E.S.R. Spectra E.S.R. spectra of 2CaO.3B20 3 .AlzO3 .xFe203 (x = 0.1, 0.3, 0.5, 0.7 and 0.90) were recorded in the temperature range 300-473 K. Representative room temperature spectra are reproduced in Fig. 1 along with the spectrum of Dance etal. f o r x = 0.01 in the inset [1]. It is observed that for low iron concentrations the resonances at g ~ 2 and g ~ 6 are quite weak. However for higher dopant concentrations two derivative type resonances, a relatively sharp peak at g = 4.13 (A) and a broad one at g = 2.00(B) appear. In addition, an absorption peak at g = 6.44(C) occurs as a shoulder to A. The observation of B and C peaks is a feature in common with the observations of Castner et al. and other workers [2, 6]. The E.S.R. spectra also show the following additional features. (i) With increasing iron concentration the intensity (/) of the A-peak decreases


Vol. 53, I~o. 1 t



resonances A and B with increasing Fe 20 3 concentration. The numbers NA and Nn are estimated from the product I ( A H ) 2 of respective resonances. An inverse correlation between NA and Ne is evident. As AH is expected to increase for both the A as well as B sites due to dipolar and exchange mechanisms as concentration of Fe 3 ÷ ions increase, the inverse relation suggests preference for A sites for low concentrations upto x = 0.1 and preference for B sites for subsequent addition of iron. This is also in accordance with the observation by Dance et al. of only the A peak at very low concentrations [1 ]. The dependence of line-width of A-peak on the dopant concentration is shown in Fig. 3 (a). Although the line-width increases initially, it tends to saturate showing increasing exchange interaction. Magnetic susceptibility (×) supports this inference which shows a parallel trend with increasing x (Fig. 3(b)).



x =0.1





















2 50







/// // o,'



(:3 /

Fig. 1. Room temperature ESR spectra of 2CaO.3B 20 3. A1203xFe203 glass f o r x = 0.1,0.3, 0.5, 0.7 and 0.9. Inset shows the spectrum o f x = 0.01 [1].

-o o




q !


r I /


while its peak-to-peak width (AH) increases, (ii) both A H and I of B peak increase and (iii) the I of C increases but tends to saturate. Fig. 2 shows the relative changes in the numbers of Fe a + ions in the two conspicuous 0.2 6




14 1.0

400 Fe-conc.(X)

300 I'





;oo 7. m

z ~

+X 0









Fe conc. (x) Fig. 2. Relative concentrations of A, B and C spectrum as a function of Fe-concentration (x) see text.

Fig. 3. (a) Line width (AH) vs. x plot for A-peak (b) Magnetic susceptibility (dc) vs. x at 300 K. The spin Hamiltonian H = gBH + DS2z + E [Sx2 --S~ ] adequately accounts for the three distinct g-values obtained for the present case as well as in other iron doped glasses [3, 6 - 8 ] . The g = 4.13 peak arises from a tetrahedral or octahedral rhombic distortions implying E/D = 1[3, the g = 2.0 and 6.0 peaks resulting from sites with weak crystal field terms and axial distortions thereof and would be regarded as the parallel and perpendicular components of the Kramer's doublets where E/D = 0, i.e. E = 0. Griffiths has given a thorough discussion of the origin o f g = 4.13 in the e.s.r, spectrum based on group theoretical arguments [7]. It is shown

yol. 5'3, No. 11


that three distinct ligand environments with Ta, On and C2v symmetries which could result in a situation which is more symmetric than the overall symmetry group of the Fe 3 ÷ site can account for the experimental observation. But in none of the three cases the assumption of Castner et al. viz D = O, E ~ 0 is valid. On the other hand D is small but non-zero with D ,¢ E. In the light of MOssbauer data discussed later, supporting distorted tetrahedral symmetry for Fe 3 + ions, the environment of the type MA2B 2 suggested by Griffiths is applicable for the microsymmetry in the present case. The g = 2.00 resonance could arise from (i) isolated Fe a + ions in an axially symmetric field and (li) spin-spin interaction between Fe 3 + ions. The minimum Fe203 content being 10 mol% in the present case, spin-spin interactions should dominate, since it was shown by Moon et al. that 3 mol% is just about the limit at which isolated Fe 3 + ions could cease to exist [ 10]. However, the absence o f g = 2.00 resonance in 1 mol% Fe203 containing 2CaO.3B203A1203 glass would tend to push this limit to 10% in this system so that in 10 tool% Fe 20 3 containing glass one has only isolated Fe 3 +, at room temperature. It is worth mentioning that there is a monotonic decrease of intensities of all the ESR signals with increasing temperature, for all the compositions. This feature which has also been observed in sulphate and oxychloride glasses arises due to the approach of the system towards glass transition when the entropy of the system tends to rapidly reach a maximum value by way of an efficient coupling between the magnetic fields caused by ionic motion and the spin magnetic field [ 11 ]. 3.2 MOssbauer Spectra MOssbauer spectra of all compositions consist of broad, paramagnetic doublets with no magnetic splitting detectable down to 80 K. A representative spectrum is given in Fig. 4. The values of isomershifts, quadrupole splittings and line-widths for various compositions are listed in Table 1. The isomershifts for all cases lie in the range 0.31 to 0.34 mm sec-1, relative to natural ion. These values are typical of highspin trivalent iron. As for the coordination of the trivalent irons, a survey of Fe ~+


to I0

"; 2.24 u)


?-.20 2,16 2.12 2.08 -2.0








VELOCIT Y(mm/sec)

Fig. 4. STFe MOssbauer spectrum of 2CaO, 3B20 aA12Oa-0.3Fe203 at 300K. ions in octahedral and tetrahedral sites in crystalline materials and in glasses reveals the following features. In well established tetrahedral oxide lattices such as CaBaFe40s and KFeO2 for which X-ray data are available the values lie close to 0.15 mm see-t.(Fe) [ 12]. On the other hand Kurkjian's tabulated data show that in glasses the value extends upto 0.32 mm see -1. [13] ; The lowest value for octahedral high spin Fe 3+ in crystalline materials is exhibited by CaFe 204 which is 0.38 mm see -1 . [12]. The tabulated data of Kurkjian for Fe 3 + in glassy materials show values greater than 0.38 mm/sec, with an upper limit of 0.48 mm see- 1. with respect to iron. Thus both in four and six coordinations isomershifts of Fe 3 + in glasses tend to be higher than in similar crystalline environments, apparently due to averaging effect in the former. In view of such an analysis the values of 0.37 to 0.48 mm sec -1.(Fe) observed for iron in alkali borate glasses and 0.41 and 0.52 mm sec-l.(Fe) for the inequivalent sites in lead oxide borate glasses fall well within the range for octahedral sites indicating the suggested change in coordination in the former or difference incoordination in the latter is not evident [14, 15]. The values observed for the present system also can be unambiguously assigned to tetrahedral

Table 1. MOssbauer parameters o f Fe a + in 2CaO. 3B20a .Al2OaxFe20a at room temperature. T1 and T2 are the line-widths o f the individual lines.

0.1 0.3 0.5 0.7 0.9

I.S. (Fe) 3+ (+ 0.01 m m sec -1 )

Q.S. (_+ mm sec -1 )

T1 (_+ 0.02 mm sec -1 )

% (_+ 0.02 mmsec -1 )

0.34 0.33 0.29 0.28 0.31

1.08 1.11 1.14 1.14 1.08

0.82 0.82 0.82 0.85 0.82

0.71 0.68 0.73 0.73 0.68

+ relative to natural iron




coordination. In addition to assignment of site symmetry, isomershift data can be used to obtain an estimate of average Fe3+-O 2- bond distance from the work of Jagannathan et al. [12]. These authors have shown in different coordinations that in ionic systems a close correlation exists between isomershift and interatornic distance due to dominant contribution coming from the overlap terms. Using the relationship between isomershifts and bond distances for tetrahedral coordination in the present isomershift values lead to an average F e - O distance of 1.96 A in good agreement with the values deduced by other methods in similar systems

[81. The quadrupole splittings for Fe 3 + observed lie around ~ 1.00 mm see- 1. and is typical of glassy matrices. As in Fe 3 +(3d s) the q.s. arises only due to lattice contribution, it is indicative of significantly larger asymmetry is glasses compared to crystalline systems where the value is usually ~ 0.5 mm sec- 1. A measure of distortion from cubic symmetry may be obtained either from crystalline strain data or more easily from quadrupole splitting data provided a knowledge of the average bond distance [Ro] is available [8]. Based on the Newman's superposition model, the angular twist (or) produced by the rotation of the vertical plane containing the upper pair of oxygens with respect to Fe 3 + relative to the plane containing the other is given by Ib°l = Ib21 = 21bi(Ro)lSin2,~


larger x values sites giving rise to g ~, 2.0 and g ~ 6.44 also evolve. The intensities of these e.s.r, lines exhibit concentration and temperature dependence. MOssbauer spectra show evidence only for tetrahedral site occupance for Fe 3 + indicating that all the three e.s.r. signals observed arise due to differences in the tetrahedral environments. In the light of Griffiths' work [7] the sharp signal at g = 4.13 is attributable to environment of the type FeA2B2. It is shown that isomershift data can given an estimate of the average distance of Fe37-O 2distance. In the present case a value of 1.96 A is deduced in good agreement reported for glassy systems. The distortion from cubic symmetry is significantly larger in glass systems compared to similar environments in crystalline systems as shown by large quadrupole splitting. Combining the average value of Fe 3 +-0 2 distance deduced from isomershift data, quadrupole splitting. Combining the average value of Fe3+-O 2Bukrey [8] on the basis of Newman's superposition model an upper limit of a = 30.5 ° is obtained for the twist angle giving a measure of distortion from cubic symmetry. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.


EQs =Ze2Q(l--r~)[l+3"-~-211'2[R-~o]lbg.(Ro)[,


9. 10.

where bT are the ligand field splitting constants. Ib g (Ro)[ in (3) is related to the lattice sum contributing to electric field gradient. Using the Ro value obtained from isomer shift data, Q = 0.21 x 10 - ~ 7** = - 9 . 1 4 1 1 6 , 17] and 77= 1 corresponding to maximum asymmetry a value of 30.5 ° is obtained for the average twist angle (~). 4. CONCLUSION E.S.R. Spectral studies of Fe 3 + doped in the glass 2CaO.3B 20 3.A120 3 .x Fe 20 3 show that for x ~< 0.1 only sites corresponding to g = 4.3 are occupied. For

Vol. 53, No. 1 |

11. 12. 13. 14. 15. 16. 17.

J.M. Dance, J.P. Darnaudery, H. Baudry & M. Monneraye, Solid State Commun., 39, 199 (1981). R.H. Sands, Phys. Rev., 99, 1222 (1955). T. Castner, G.S. Newell, W.C. Holton & C.P. Slichter, J. Chem. Phys., 32, 668 (1960). C. Hirayama, J.G. Castla & M. Kuriyama, Physics Chem. Glasses 9, 109 (1968) D. Loveridge & S. Parke, Ibid., 12, 19 (1971). C.R. Kurkjian & E.A. Sigety, Physics Chem. Glasses 9, 73 (1968). J.S. Griffiths, Mol. Phys., 8,213 (1964). C.M. Broadbeck & R.R. Bukrey, Phys. Rev. B, 24, 2334 (1981). C.S. Sunandana & R. Singh, Phys. Stat Sol. (a), 75, K91 (1983). D.W. Moon, J.M. Aitken, R.K. McCrone & G.S. Cieloszyk, Physics Chem. Glasses, 16, 91 (1973). R. Parthasarathy, K.J. Rao & C.N.R. Rao, Chem. Phys., 68,393 (1982). R. Jagannathan & K.N. Shrivastava, Hyperfine Interactions 7,377 (1979). C.R. Kurkjian, J. Non Cryst Solids, 3, 157 (1970). T. Raman, G.S. Rao & D. Chakravorthy, J. Non-crystalline Solids, 29, 85 (1978). E. Burzo & I. Ardelean, Phys, Chem. Glasses, 20, 15 (1979). N.N. Greenwood & T.C. Gibb, MOssbauer spectroscopy, P. 97, Chapman and Hall (1971). J.O. Artman, A.H. Muir, Jr., & H. Wiedersich, Phys. Re~., 173, 337 (1968).