Solid State Ionics 9 & 10 (1983)73-76 73 North-Holland Publishing Company

MOBILITY OF T1 + IONS IN A DEFECT PYROCHLORE STRUCTURE TiNb205 F

s P.P. Man, H. Theveneau, P. Papon

Laboratoire de Dispositifs Infra-Rouge et Physique Thermique, ERA CNRS 365 ESPCI, i0 rue Vauquelin 75231 Paris cedex 05, France

J.L. Fourquet

Laboratoire des Fluorures eL Oxyfluorures Ioniques, ERA CNRS 609 Facult@ des Sciences, Universit@ du Maine 72017 Le Mans cedex, France

NMR measurements on the two naturally abundant isotopes of thallium and on fluor have been made in the defect pyrochlore TINb2OsF. An Arrhenius activation energy of 0.08 eV was obtained from the temperature dependence of the 205T1 absorption linewidth. CW data indicated the presence of the electron-nuclear magnetic interactions. The spin-lattice relaxation times showed two motions of T1 + ions with the activation energies of 0.25 eV and 0.14 eV. These relaxation times were interpreted with the exchange and chemical shift interactions.

i. INTRODUCTION

The compound TINb2OsF studied here possesses the defect pyrochlore structure A+M2X6 where A+=T1 + N=Nb and X=O or F. It consists of corner shared MX6 octahedra which form a rigid network [~. It has been shown that the T1 +ions are found in two sets of high multiplicity positions 32e I and32~ inside the same cavity of this rigid network (Fig.l). The radius of the cavity and the bottle- neck between two csvities are I.RO ~ and 1.30 respectively whereas the ionic radius of T1 + ion is 1.50 ~. The mobility of T1 + ions was in- vestigated by means of dielectric relaxation and conductivity measurements [2]. The conductivity obeys the relation oT=~oexp(-Ea/kT)with Ea=O.23eV and ~o =0.375~-lcm-lK. This mobility is attribu- ted to ~he high polarizability of T1 + ions.

In order to obtain more microscopic information about thallium mobility, we have performed NNR measurements on both isotopes of thallium (203TI and 205T1) and 19F as well. These nuclei have a nuclear spin I=I/2, therefore quadrupolar ef~cts are absent. But the thallium isotopes are heavy nuclei which, usually, exhibite the electron- nuclear magnetic interactions [3].

2. EXPERIMENTAL

Powder of TINb205F was compressed under a hydro- static pressure of 5kbar. Most of the peak to peak linewidth (~Hpp) measurements were obtained with a Varian Nodel V-4200 broad line spectrome- ter at 7.6MHz between 100K and 50OK. Relaxation time measurements of 203T1 and 205T1 were made at 27MHz and 14.6MHz with a Bruker SXP 4-60MHz pulse spectrometer between 200K and 60OK. Spin

spin relaxation times (T2) were determined from the free induction deca~ (F.I.D.). Spin-lattice relaxation times (TI,T 1 ) were measured by the inversion recovery technique.

• 16a A32el •Sb t32e2

Fig. 1 A cavity of the defect pyrochlore structure, the anions 0 and F are not represented

0 167-2738/83/0000-0000/$ 03.00 © 1983 North-Holland

74 P.P. Man et al. /Mobility o f T1 + ions

3. RESULTS AND DISCUSSION

3.1 C.W. data

We have investigated the temperature dependence of the 19F absorption line at 7.6MHz in order to verify the rigidity of the sublattiee Nb205 F on one hand and to know if this nuclear spin could be a NMR probe for detecting thallium motion on the other hand. The result of peak to peak line- width AHpp measurements from 100K to 500K is re- presented in Fig.2. The data show that the line- width remains essentially constant. The experi- mental value of the second moment of 19F(<AH2> F) has been estimated using the relation <&H2> = (AHpp/2) 2. We obtained <&H2>F= II.22G 2 which is comparable to the theoretical value (12.39G2) calculated from crystallographic data with the Van Vleck relation [4] valid for magnetic dipo- lar interaction only. These results clearly in- dicate that we cannot investigate the thallium mobility by studying 19F resonance.

It is the reason why we have performed direct measurements on both isotopes of thallium nu- cleus. It is well known that for a heavy nucleus, we must consider the chemical shift interaction as well as the nucleus-electron-nucleus indirect interaction. The former interaction exhibits a square magnetic field dependence of the second moment and a shift of the resonance magnetic field for a fixed frequency. The latter creates a difference of AHpp of the two isotopes.

A saturated solution of CH3COOT1 was used as a reference for the magnetic field shift measure- menL which is defined as

H-Hre f

~is°= Hre f

The measured values of ~iso are temperature in- dependent and remain equal Lo +0.06%.

This value represents a diamagnetic shift corresponding to an ionic environment which is consistent with the structure [5]. The experimen- tal second-moments of the 205Tf "absorption lines at room temperature were plotted versus the square of the resonance frequency (~o) in Fig.3. This frequency dependence suggests that the elec- tron cloud about the thallium ions is far from spherical, which agrees with X-ray data [6].

A plot of &Hpp of both isotopes of thallium nu- clei versus temperature is shown in Fig.2. The linewidth of 203T1 spin is larger than that of 205T1 spin. This indicates the presence of nu- cleus-electron-nucleus indirect interaction bet- ween thallium nuclei. The temperature dependence of the 205T1 linewidLh shows that Tl+ ion motion is thermally activated. The activation energy E a and jump frequency vs can be determined with the relation

AHpp = 1.154y<&H2> / vs

The experimental points are filled with the Ar- rhenius expression

~s = vo exp(-Ea/kT)

and with <AH2>T1 = 1.8G 2 which is the experimental second moment of 205T1. Ea is found to be 0.08 eV and ~o=106see-l.

3.2 Pulsed NMR data

Pulsed NMR measurements were made over a wide range of temperature (200K to 60OK). The spin lattice and spin-spin relaxation times plotted on a logarithmic scale versus reciprocal tempe= rature are shown in Fig.4 for 205T1 at 27HHz and at 14.6HHz. The 203TI relaxation times at 26.7HHz has the same temperature dependence than the 205T1 relaxation times at 27HHz. The FID has

= . .4 ff 'q2

i i i i

6 "F

OI -200

2O3Ti

r I I I , I

-100 0 100 200

T ( ° c )

6O

~= 40

I N

Z '~ 20

0 0

~O,Ti

50 100

Qo ~ (M H ~') 100

150

Fig. 2 Temperature dependence of the peak to peak derivative linewidth for 19F, 203T1 and 205T1

Fig. 3 Frequency dependence of the estimated second-moments for 205TI

always been observed to be exponential. The spin lattice relaxation is found to be non-exponential over the range 223K-473K. The non-exponentiality is analyzed for two relaxation times denoted as T1 and TI*. Above 223K, the relaxation times are of activated type. The activation energies Ea estimated from T1, Tl* and T 2 are summarized in Table 1. We are unable to measure a well defined minimum in the spin-lattice relaxation times but about 523K, the three relaxation times TI, TI* and T 2 become equal. Fig. 5 shows the Vo2 depen- dence of T 1 and the frequency independence of Tl* at room temperature of 205T1 for three fre- quencies (14.6MHz, 27MHz and ll5MHz).

3o

10

5 ?

~,, = 0 .5

Spin-lattice relaxation times measurements clear- ly indicate two kinds of T1 + motion characterized by two activation energies of 0.25±0.02eV and O.14±O.02eV. The larger energy is comparable to that obtained from conductivity measurements (0.23eV) and may correspond to a diffusion pro- cess involving a jump of the ion from a cavity to a neighbor cavity. The lower energy would corrrespond to a motion inside the same cavity which can not be detected by conductivity expe- riments.Such a motion could be a jump from one 32e site Lo the bottleneck (16d site) where the electronic cloud of T1 + would be distorted be- cause of the difference of radius. This situa- tion is very favorable for the spin-lattice re- laxation by the chemical shift anisotropy. A more detailed analyze of the relaxation times will be given elsewhere [7] . The results were fitted with the following relations [8,9] @

(TI)-I = (2~) 2 2Tc 1 + (~o ~ )2

C

* 1 2(A0)2 2T (T 1

)- : ~ c

+ (~o Tc )2 i

where ~/h is the exchange coupling constant, ~oo the difference in the Larmor frequencies of the two isotopes, Te the correlation time, and A~ the chemical shift anisotropy constant. Usually, ~c is identify to i/~s. Ue obtained #/h = 16 kHz, &~ = 0.02 and the prefactors vo are gathered in Table i.

Spin-spin relaxation times of 205Tl show a motional narrowing behaviour. The data were fit- ted with the following relation

1 A (T2)205 : 0.3 (2~T-~-)2T c

With Ea = O.08eV and ~/h = 16kHz, the prefactor vo is reported in Table i. These voare comparable to that obtained from C.W. measurements (lO 6 see-l). The small value of the activation energy (O.08eV) shows that it may be due to a local motion, probably a swing from a 32e site to a 32e site in the same cavity.

0.1 0 .05

3 0 0

T Cc) 3 0 0 2 0 0 100 0 - 5 0

I I I L I

0 0 O

I I r I I I I I

2 3 4 5

100(~/1r (K -1 )

Fig. a Temperature dependence of the relaxation times of 205Tl at :

27 MHz ( •: TI, ,: fl: , .: T 2 ) la.~ MHz ( x TI, a T 1 , o : T 2 ).

~" 2 0 0

g

~ , 100

0

0 I I

I I

,OSTi

P.P. Man et al. /Mobility o f T1 + ions 75

5 0 100 150 ,$we)

Fig. 5 Frequency dependence of the spin-lattice times of 205TI ( n : TI, o: TI* ).

76 P.P. Man et al. /Mobility o f T1 + ions

Vo(T l) Ea(T 1) vo(TI*) Ea(TI*) vo(T 2) Ea(T 2) (sect 1) (eV) (see-l) (eV) (sec-~ (eV)

27MHz 205TI 0.6xlO 9 0.26 0.6xlO II 0.14 0.3xlO 7 0.08

26.7MHz 203Ti 0.7xlO 9 0.27 0.7xlO II 0.12 107 0.08

14.6MHz 205Ti 0.2xlO 9 0.23 0.2xlO II 0.15 O.2xlO 7 0.08

Table i : Activation energies Ea and prefactors Vo obtained from relaxation times data.

The fact that Ea(T1) = 3Ea(T2) in general sug- gests that the ionic motion is one dimensional when the ionic motion is long range [lO] . But in the defect pyrochlore, we do not believe that the T1 + ions motion could be one dimensional because the rigid network Nb205F has a three dimensional structure.

4. CONCLUSION

The present work shows that the thallium ions are good NMR probes for detecting long range mo- tion like those determined by conductivity expe- riments as well as local motions which do not lead to diffusion. However, due to the complexi- ty of the interactions experienced by the Tl+ ion, and the large number of sites in a cavity, a definitive interpretation of the NMR results is difficult. Other measurements are required at higher frequencies to dlscriminate between the different types of interactions. It would be in- teresting to study the NMR signal of another ca- tion : for example, 87Rb+ (I=3/2) which has the same ionic radius but a lower polarizability and which is sensitive to quadrupolar interaction, to obtain a better identification of the motions. Nevertheless we suggest that the Tl+ ions in the defect pyrochlore TINb205F exhibit two kinds of motion : a jump from cavity to cavity of the ri- gid network Nb205F characterized by an acti- vation energy of 0.25eV which is comparable to that obtained from conductivity measurements (0.23eV), and local motions inside the same ca- vity characterized by the activation energies of O.14eV and O.08eV.

5. ACKNOWLEDGEMENTS

We thank Mrs Lacour for compressing the sample. Thonks are due to Dr H. Estrade and Dr D. Muller for the high frequency NMR measurements on Tl + ions.

6. REFERENCES

1. J.L. Fourquet, C. Jacoboni and R. De Pope, Acts Cryst. B35, 1570, (1979).

2. J.L. Fourquet, M. Rousseau and R. De Pape, Mat. Res. Bull. 14, 937, (1979).

3. N. Bloembergen end T.J. Rowland, Phys. Rev. 97, 1679, (1955).

4. J.H. Van Vleck, Phys. Ray. 7_~4, i184, (3948).

5. J.L. Bougher and P.J. Bray, Phys. end Chem. of Glosses i0, 77, (1969).

6. G.E. Gellison Jr. and S.G. Bishop, Phys. Rev. BI9, 6418, (]979).

7. P.P. Man, H. Th@veneau, P. Papon and J.L. Fourquet, (to be published).

8. M. Villa and A. Avogadro, Phys. Stat. Sol(b) 75, 179, (1976).

9. T.C. Farror and E.D. Beaker, Pulsed and Fourier transform NNR, Academic Press (1971)

i0. P.M. Richards, Physics of superionic conduc- tors, Topics in Current Physics, (Springer Verlag, 1979).