RELAXATION PEAKS PRODUCED BY DEFECT COMPLEXES IN CERIUM DIOXIDE DOPED WITH TRIVALENT CATIONS

(1)

HAL Id: jpa-00220998

https://hal.archives-ouvertes.fr/jpa-00220998

Submitted on 1 Jan 1981

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

RELAXATION PEAKS PRODUCED BY DEFECT

COMPLEXES IN CERIUM DIOXIDE DOPED WITH

TRIVALENT CATIONS

M. Anderson, A. Nowick

To cite this version:

(2)

JOURNAL DE PHYSIQUE

Colloque C5, suppZ6ment au nOZO, Il'ome 4 2 , octobre 1901 page C5-823

RELAXATION PEAKS PRODUCED BY DEFECT COMPLEXES I N CERIUM D I O X I D E DOPED WITH TRIVALENT CATIONS

M.P. ~ n d e r s o n * a n d A . S . N o w i c k

Henry Kmunb School o f Mines, Colwnbia University, New York, Iieu York 10027,

U.S.A.

Abstract.- Ceriuii dioxide doped with y 3 + ions, which is of interest for its ionic conductivity, is here studied by anelastic

~~~ethods. The internal friction peaks observed are analyzed intc

peals A , B, C (with Hr = 0.64, 0.71 and 0.78 eV, respectively). Peak A, present at low y 3 + concentrations is established to be

due to a W pair (V = oxygen-ion vacancy), which is a charged

defect conpensated by an isolate2 Y. Peal: B is believed to be

due to a Y2V complex with the V-Y pair in the (1,0,0) configuxa-

tion. For this defect, only relaxations involving the V notion are observable, giving an example of the "frozen-free split" phenomenon.

1. Introduction.- Cerium dioxide (CeOz), which has the fluorite struc-

ture, is interesting because it can be doped quite heavily $:ith both divalent and trivalent cations. Such doping introduces oxygen vacancies (V) for charge compensation. These vacancies migrate xith relative ease at elevated tenperatures to ~ r o d u c e good oxyqen-ion conducting

solid electrolytes; these are of interest for high--temperature fuel

cells and oxygen sensors.' Figure 1 shows one-half the unit cell of the fluorite structure containing a dopant cation and an oxysen vacancy. ca2+-doped Ce02 has already been studied and found to produce an anelastic peak due to Ca-V pairs in nearest neighbor position (as in Piq.

The case of

x 3 +

doping is of special interest because it yields one V for every two 1.1" ions. If the latter were roobile, one coula then expect to obtain fully compensated M2V triplets. But the lack of ri~obility on the cation sublattice suggests more like a randon distribution of bI3+ ions. There have been many studies of the electrical prop- erties of ?13+-doped CeO2 (as well as Zr02 and ThC2).' A recent detailed

study3 of conductivity in y3f-doped CeO2 as a function of temperature

and Y203 concentration gave a quantitative analysis for the range lm/o Y2O3. But the sharp naximun in conductivity near 4% Y2O3 is only describable in qualitative terms.

The tools of anelastic relaxation have never previously been ap- plied to this problem. V:e felt that they could be helpful to reveal some of the various defect clusters that form.

*

_Now_{at 3xxon Corporate Research Laboratories, Linden,}_N.J.

(3)

JOURNAL DE PHYSIQUE

0

0'- I O N HOST CATION ( 4 + )

VACANCY

@

DOPANT CATION

( 2 + or 3 + )

Fig. 1: Diagram of one-half unit

cell of Ce02 which contains one dopant cation and an oxygen vacancy in nn position.

2. b:ethods. --

--

Sar~ples were prepared as sintered polycrystalline con-

pacts after chertiical precipitation of the oxides fron solution, as

previously d e s ~ r i b e d . ~ Earaples were in the form of bars - 2 0 ~ 3 ~ 0 . 6 rm3,

coated with gold by sputterins. Internal friction, $ , was aeasured in

flexural vibration at - 6 kHz by electrostatic excitation and detection,

in an apparatus described previously.'

3. Results and Spectral Analysis.- Internal friction was measured as

a function of tenperature over the range 20°

-

3 5 0 ~ ~ for samples con-

taining 0.5, 1.0, 2.5, 4, 6, and Em/o Y2O3. Peaks were observed in all

cases, the height increasing rapidly with increasing Y203 con.Lent. The

two l o v ~ s t compositions Gave essentially single Debye peaks located at

77Oc. Using the expression

- '

=

-ti1

exp (-fi,/kT)

'1 r (1)

for the relaxation rate T;', and substituting the expected value TO' =

10" sec-', then gives Hr = 0.64 eV for the activation enthalpy. Fs

the dopant level increases the peak gross but also broadens toward the high-temperature side. Peaks for the lowest three conpositions are shown in Fig. 2. By the case of 8% Y2O3 the peak is broadened much

r~lore and has reached the relatively huge heigh'c of 210 x lo-'.

We expect that the initial; nearly Debye, peak A represents a relatively simple defect, while the broadening at higher concentrations reflects the development of inore conplex clusters. Accordinqly, we have analyzed the peaks as a collection of Debye peaks appearing at

successively hiqher temperatures (i. e. , higher IIr1s)

.

I:e assume peak

A is always present, and allow it to be slightly broader than a Debye

peak. Then, by subtraction and analysis of the residual internal fric-

(4)

Fig. 2: Internal friction versus 1/T, for Ce02 containing 0.5, 1.0 and 2.5m/o Y2O3.

samples. These have Hr = 0.71 and 0.78 eV, respectively. (IJon-linear

least squares optimization is used to obtain accurate values for the

peak height, Q m r and Ifr for each conponent peak.) Such a spectral ana-

lysis is probably no longer valid when the dopant concentration becomes too high; therefore we confine ourselves to the first three compositions.

The results of this spectral analysis appear in the first 4 colui.~ns of

Table 1. It shows that peak A is prominent for all 3 cases, that peak

B appears ninimally for 1% Y2O3 and prominently for 2 . 5 % : and that peak

C appears in this latter case. In the next section, we will seek models

to explain peaks A and B.

4. Eodeling of Feaks A and B.- As already mentioned, one oxygen-ion vacancy V is introduced for every two Y 3 + ions. Lie postulate that the y 3 + ions are randonly distributed on the cation sites. The reason is

the very large activation energy (-4-5 eV) found for cation diffusion

in such oxide systems.' It is also observed that cationic ordering

takes place only after very long anneals at -1000 - 1 2 0 0 ~ ~ ; the present

preparation method is then much more consistent with a completely dis- ordered state on the cationic sublattice. Oxygen vacancies, on the

other hand, are very nobile with a low activation energy (-0.6 e ~ ) 3 ;

therefore, they nay be expected to settle in on minimum energy sites

defined by the y 3 + distribution.

In considering pairs of Y ions, it is reasonable to look for a

Y-Y separation beyond which a V attached to one Y is no longer affected

by the second Y. This distance is computed by displacing the vacancy

(5)

C5-826 JOURNAL DE PHYSIQUE

Table 1. Feak heights, $I,, of peaks A, B and C for three Ce02;Y203 compositions. Calculated concentrations of singlet Y (cs) and Y2 pairs

(cp) and their ratios to the peak heights. (All $I, are in units of

found to be 8 A . (Note that the lattice parameter of Ce02 is

a = 5.41 A.) Based on the concept of this interaction volume, we have

calculated the concentration of single Y and Y-Y

airs

(cs and cp,

respectively) in the cation sublattice. (For concentrations up to 2.5% Y203 triplets turn out to be unimportant.) The second half of Table 1 gives cs and cp and the ratio of the peak height $I: to cs and of $:I

to cp. It is seen that these ratios give nearly constants in each

case. It is therefore reasonable to assign peak A to a YV pair and

peak B to a Y2V triplet. No firm conclusion can be d r a m about peak C. Feak A : The defect involving a single dopant cation and a vacancy in nn position is shown in Fig. 1. This is a trigonal defect, and is well known for the case of the divalent dopant as the Wachtnan 8-posi-

tion nodel.' For Y dopincj, however, this defect is inconpletely con-

pensated: it possesses a net positive charge which is compensated by another isolated Y ion elsewhere in the crystal. The latter has cubic synnetry and is not an anelastically active defect. The trigonal W pair gives rise to a single relaxatiorrof the Sq4 modulus, for which

T;' = 4 ~ 1 2 (w12 = the nn jump frequency of the vacancy). The value Hr = 0.64 eV is consistent with the fact that the migration energy of a free V in Ce02 is 0.61 eV.3

Additional support for the model comes from dielectric relaxation experiments7 which show that T, (diel) = 2 T, (anel), as expected for this model.5 It therefore appears to be well established that the

dominant defects at low concentrations are the YV pair and a compen-

sating isolated Y, the former giving rise to the observed anelastic peak.

B

Peak B: From the ratios om/cp in Table 1, it is reasonable to

-

interpret peak B as due to Y2V complexes. We have considered various

Y-Y pair configurations that fall within the interaction distance

(6)

position. The following Y-Y pairs fall into this category. For the

first Y at (0,0,0), the second can be at (+,+,O), (1,0,0), (1,+,+,) or

(1,1,0) (in units of 5 ) . All four form low symmetry Y2V defects which are capable of giving rise to anelasticity. However, for the low temperatures involved, we can only consider reorientations involving the

jump of the V and must take the Y's as immobile. We, therefore, have

the possibilities for a "frozen-free splitIn6 if all of the possible

relaxations cannot occur through the agency of the V motions. In fact,

examination of the (+,%,0) and (1,+,4) Y-Y configurations shows that no contribution to relaxation via V migration can occur for these two cases. This leaves only the remaining two Y-Y configurations as the possible origins of peak B.

The (1,0,0) case is pictured in Fig. 3. The vacancy has 8 equivalent sites for each of three equivalent Y-Y orientations. The defect symmetry is 110 monoclinic, which gives rise to two relaxations of modulus S44 and one of S11-S12.6 However, the S11-Sl2 relaxation cannot occur by V motion alone, though both S44 relaxations can. Their rates are readily shown to be

T - I r = 4 ~ 1 2 and 2w12

+

2w17

where the V sites 1, 2 and 7 are shown in Fig. 3. Though wl2 and w17

are inequivalent jumps, their activation enthalpies may be expected to be close enough as to contribute a single, perhaps broadened, Debye peak.

Finally, the (1,1,0) Y-Y pair must be considered. This involves 4 equivalent sites for the vacancy,giving a 110 monoclinic defect. Only a single S44 relaxation can occur via V motion, involving T;' =

4w12, i.e., characterized by the same type of jump as the nn pair Fig. 1 . Because the influence of the second Y ion is rather weak in this case due to its large distance, the presence of this defect may only contribute to the broadening of peak

A.

We therefore believe that the (1,0,0) Y-Y pair with V located as shown in Fig. 3 is the principal contributor to peak B.

Fig. 3: Model of the Y2V defect

with the Y-Y pair in the (1,0,0) configuration. Striated atoms are y3+. Number 1 is the vacancy while 2 and 7 mark two others of

(7)

JOURNAL DE PHYSIQUE

Acknow1edgr;lents.- T h i s work was s u p p o r t e d by t h e U.S. D e p a r t n e n t

o f E n e r g y . One of t h e a u t h o r s ( H . P . A . ) i s now a t t h e Exxon C o r p o r a t e

R e s e a r c h S c i e n c e L a b o r a t o r y , L i n d e n , N . J .

R e f e r e n c e s

1. E t s e l l , T.13. and F l e n ~ a s , S . N . , Chen. Revs. 70 ( 1 9 7 0 ) 339.

2 . L a y , K.W. and Whitmore, D . H . , Fhys. S t a t . S o l i d i 43 ( 1 9 7 1 ) 175.

3 . F7ang, D . Y . , P a r k , D.P., G r i f f i t h , J . and Nowick, A . S . , S o l i d S t a t e I o n i c s , i n p r e s s .

4 . K i m , K . K . a n d Nowiclr, A . S . , J. Phys. C . 1 0 ( 1 9 7 7 ) 509.

5 . Machtnan, J . B . , Fhys. Rev. 1 3 1 ( 1 9 6 3 ) 517.

6 . Nowick, A . S . , Adv. Phys. 1 6 ( 1 9 6 7 ) 1.