RADIATION DAMAGE IN OXIDESElectron irradiation damage in MgO

(1)

HAL Id: jpa-00220096

https://hal.archives-ouvertes.fr/jpa-00220096

Submitted on 1 Jan 1980

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are published 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.

RADIATION DAMAGE IN OXIDESElectron irradiation damage in MgO

R. Youngman, L. Hobbs, T. Mitchell

To cite this version:

R. Youngman, L. Hobbs, T. Mitchell. RADIATION DAMAGE IN OXIDESElectron irradiation damage in MgO. Journal de Physique Colloques, 1980, 41 (C6), pp.C6-227-C6-231.

�10.1051/jphyscol:1980658�. �jpa-00220096�

(2)

JOURNAL DE PHYSIQUE Colloque C6, supplkme~tt au no 7 , Tome 41, Juillet 1980, page C6-227

R/ZD/A TION DAMAGE /N OXIDES.

Electron irradiation damage in MgO

R. A. Youngman, L. W. Hobbs and T. E. Mitchell

Department of Metallurgy and Materials Science, Case Western Reserve University, Cleveland, OH 44106, U.S.A

RBsumB. - Des monocristaux de MgO ont Ctt irradies et examines par microscopie klectronique haute tension entre 300 et I100 K et jusqu'a 650 kV. On interprkte la diminution observee de I'hnergie seuil de deplacement B travers la destruction thermiquement activee de paires de Frenkel Mg sur des sites voisins. Aux fortes doses et aux temperatures Clevtes, des boucles interstitielles parfaites de vecteur de Burgers 112 ( 110

>

sont nucltees et s'allongent selon les directions ( 001 ). La croissance des boucIes obeit & une loi d'Arrhenius et conduit

a

une valeur de l'energie d'activation de la mobilite des lacunes d'oxygkne d'environ 3,O eV. La mobilitk d'interstitiels dans des conditions non stationnaires explique probablement la decroissance apparente de l'knergie d'activation aux basses tempiratures. Le caracttre anisotrope de la croissance des boucles est attribue a la difference de carac- tere entre crans sur les dislocations coin selon que leur ligne est orientee dans les directions ( 001

>

^ou⁽¹¹⁰^).

Abstract. - MgO crystals have been irradiated and examined in a high voltage electron microscope at tempe- ratures from 300 to 1100 K and voltages up to 650 kV. An observed decrease in threshold displacement energy with increasing voltage is interpreted in terms of thermally-activated break-up of close Mg Frenkel pairs. k t high doses and temperatures, perfect 112 ( 110 ) ( 110 ) interstitial edge dislocation loops are observed to nucleate and grow in an elongated mode along ( 001 ). Arrhenius plots of the growth rate yield an activation energy

-

3.0 eV for the mobility of oxygen vacancies. An apparent decrease in activation energy at lower temperatures is probably due to interstitial mobility under non-steady-state conditions. The anisotr_opic growth of loops is ascribed to differences in the character of jogs on edge dislocations in ( 001 ) and ( 110 ) directions.

1 . Introduction. - Although a great deal is known from spectroscopic evidence about point defect species produced by irradiation of ceramic solids [I], relatively little is known about the aggregation of displaced ions in ceramics compared with metals.

However, because of the many actual and potential applications of ceramics in nuclear environments, there has been increased activity in recent years to study the defect clusters produced by irradiation.

The use of transmission electron microscopy to reveal the types of damage produced in ceramics by neutron, ion and electron irradiation has recently been review- ed by Hobbs [2]. In some cases, notably the alkali halides, quartz and other silicates, the damage is caused by ionization (radiolysis). In most ceramic oxides, the damage is caused by direct knock-on displacement with a well-defined threshold energy.

The point defects so produced (cation and anion Frenkel pairs) may then diffuse, if the temperature and mobility are high enough, leading to either mutual recombination, annihilation at sinks or aggregation into new extended defects. Both interstitial dislocation loops and voids are commonly observed to result from the aggregation process.

In the present paper, we describe observations on interstitial dislocation loop growth in MgO irradiated with high energy electrons in a high voltage electron microscope (HVEM). This is part of a program design- ed to investigate radiation damage in a variety of

ceramic oxides, some results of which have been published previously [3, 41.

2. Materials and procedures. - Single crystal boules of MgO containing

-

100 ppm impurities were obtained from Oak Ridge National Laboratory.

Thin sections, 3 mm in diameter, were cut, polished and ion-thinned to electron transparency for irradiation and observation in a Hitachi HU-65033 HVEM operating at voltages up to 650 kV. The specimen holder was a double-tilting heating device from Gatan Inc., capable of temperatures up to 950 OC.

Irradiations were generally performed at a flux of 5 x loz3 elm2 and micrographs were taken periodi- cally at a lower flux in order to measure loop growth rates.

3. Results and discussion. - At low temperatures (around room temperature) a high density of unresol- vable strain centers are produced by electron irradiation. As the irradiation temperature is increased, the density of these defect clusters decreases and they are able to grow into well-defined dislocation loops at temperatures above 800 K. Threshold voltages for displacement were determined by measuring loop densities and areas as a function of accelerating voltage at constant dose and by extrapolating to zero damage rate. The threshold voltage, E,, along

<

¹⁰⁰⁾

was found to decrease with increasing temperature from 460 kV at 300 K to 290 kV at 1000 K. The

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1980658

(3)

C6-228 R. A. YOUNGMAN, L. W. HOBBS AND T. E. MITCHELL

Fig.

5 x

1. - Loop growth in MgO irradiated along

<

100 ) at 600 OC and 650 kV for (a) I, (b) 20, (c) 45 and (d) 60 min. (Electron flux = loz3 elm2.) A typical loop growing along ( 001 ) is arrowed. The perpendicular set of loops is out of contrast.

A B C D

Fig. 2. - Loop growth In MgO irradiated along ( 100 ) at 700 OC and 650 kV for (a) 1, (b) 20, (c) 45 and (d) 60 min. (Electron flux =

5 x loz3 e/m2,)

(4)

ELECTRON IRRADIATION DAMAGE IN MgO C6-229

corresponding displacement energy Ed can be calculat- ed from the relationship

where M is the mass of the displaced ion, m o is the rest mass of an electron and c is the velocity of light.

Depending on whether the observed E, corresponds to 0 or Mg displacement, we derive the following values of Ed :

Ed for 0 has been determined by optical absorption to be 60 eV [5]. Hence it is likely that the observed E, = 460 kV corresponds to Mg displacement (Ed = 61 eV) since both ions would then be displaced and the resulting defect clusters could be stoichiometric. The same argument applies to the lower value of E; at higher temperatures, i.e. Ed(Mg) = 34 eV and Ed(0) = 5 1 eV at 1 000 K. Pells and Phillips [6]

have also observed that Et decreases with increasing temperature in A1203 ; they ascribe the low temperature value to 0 with Ed = 75 eV and the high temperature value to A1 with Ed = 18 eV, arguing that it is unnecessary to displace both ions to form aggregates at higher temperatures. However, in view of the fact that the decrease in E, occurs over the temperature range where rapid loop growth occurs in both A1,O3 and MgO, it is more likely that thermally-activated break-up of close Frenkel pairs is occurring at high temperatures.

Irradiation to high dose at high temperatures causes the rapid nucleation and growth of large dislocation loops, as shown in the typical sequences of figures 1 and 2. Standard contrast analysis shows that the loops have the following properties :

a ) They are perfect edge disloction loops with 112 ( 110 ) Burgers vectors lying on ( 110 } planes.

b) They are stoichiometric and interstitial in character.

c ) They have an anisotropic growth mode, elongat- ing along ( 001 ) and having a major to minor axis of about ten to one.

The nature of the loops is predictable since 112 ( 110 ) is the shortest lattice vector in the MgO structure and partial Burgers vectors would require a high energy stacking fault. The fact that the loops are interstitial type implies that interstitials are more mobile than vacancies, as in metals. Typical loop growth kinetics are shown in figure 3 where the major axis is

lotted

against irradiation time t for several temperatures. It is seen that there is an initial parabolic growth region where steady state conditions are being established, followed by a linear growth region

3aQ

'2

-

V) 2-

4 LL

a S Imo

600 1200 I800 2400 3000 500 QOO 4800 5400 IRRADIATION TIME (sl

Fig. 3. - Typical plots of loop radius (major axis) versus irradiation time a t 500, 600 and 700 OC. Note the linear growth which follows the initial non-steady-state parabolic growth.

corresponding to steady state conditions. This is in accord with standard models of loop growth control- led by vacancy motion [7], where the linear growth rate is given by

where K is the displacement rate, D, is the diffusivity of vacancies and AH, is the activatioq energy for vacancy motion. An Arrhenius plot of L versus T-' is shown in figure 4 and indicates two regions with a transition at about 600 OC. Above 600 OC the straight line yields an activation energy AH,

-

3.0 eV. The tailing off below 600 OC may be due to not achieving steady-state conditions or limited loop growth con- trolled by interstitial motion with an activation energy Loop growth is due to the biased attraction of AHi

-

^{1.0 eV.}

interstitials to the dislocation strain-field. Under steady-state conditions an equal number of vacancies must go to alternative'sinks, namely the foil surfaces or voids. Voids were either not formed or too small

REIPRCEAL TEMPERATURE ( X l f l l / " ) O

Fig. 4. - Arrhenius plot of loop growth rate versus reciprocal temperature.

(5)

C6-230 R. A. YOUNGMAN, L. W. HOBBS AND T. E. MITCHELL

the foil (Fig. 5). These pits are apparently caused by radiation-induced sublimation or sputtering in accor- dance with the known high vapor pressure of MgO.

A final discussion point is the relevance of kinetic models developed for metals to crystals such as MgO where anions and cations have to be taken into account Firstly the value of K in equation (2) should correspond to the species with the lower damage rate in order for the loops to remain stoichiometric. Assuming Ed to be approximately the same for Mg and 0," the heavier Mg ion would have the lower K value. (Under steady-state conditions oxygen would have a higher density of Frenkel defects with a higher recombination, rate.) Secondly the value of AH, 1:3.0 eV should correspond to the motion of the rate-controll- ing slower species, namely oxygen vacancies [8].

Again .this is necessitated by stoichiometry and

Fig. 5 . - ~ tpits with sides ~ h to ( 001 ) formed by radia- electrostatic considerations. The elongated growth

tion-induced sputtering of MgO irradiated along ( 100 ) for mode has been observed previously in NaCl 191 but

30 min. at 800 OC. is not understood. We believe it is due to the diffe-

rence in character between the ioas formed on the to be detected. Unfortunately, in the high tempera- ( 001 ) and ( 170 ) sides of the -edge dislocation

loops but the precise mechanism is not obvious.

ture, high dose region where void growth might be

expected, observations are obscurred by the very Acknowledgment. - This research is supported high density of dislocations (Fig. 2) and by the for- by the U.S. Department of Energy, Contract No.

mation of square etch-pits on the bottom surface of EY-S-02-2119. A004.

DISCUSSION Question. - A. E. HUGHES.

Have you tried any electron irradiations outside 'the TEM in bulk samples, to see if voids are actually

produced at high temperatures ? Reply. - T . E . MITCHELL.

No, it is difficult to attain a high enough dose by this method. However, we plan to irradiate thicker regions of specimen in the HVEM and then to re- thin for examination. We have successfully detected voids in A120, by this technique.

Question. - M . PULS.

1. Would you not underestimate the Ed-values at the high temperatures because of annealing effects and,

2. Are you saying that you are displacing only one of the ions, what happens to the other one ?

Reply. - T. E . MITCHELL.

The underestimation of Ed is small because the electron flux in the HVEM is so high and because, even just above the threshold, the supersaturation of interstitials is high enough to favor clustering of interstitials. I believe that it is necessary to displace both types of ions, since stoichiometric clusters are observed. Thus the measured threshold corresponds to the more difficult displacement - in this case the heavier Mg ions.

Question. - W . C . MACKRODT.

Your value for the energy of vacancy migration, namely 3.0 eV, seems to be appreciably greater than the most recent measurements. Could your value correspond to the motion of an Mg2+ vacancy in a locally strained field around a dislocation ?

Reply. - T. E . MITCHELL.

That is unlikely. Although both vacancies and interstitials are attracted to the stress-field of interstitial dislocation loops, the rate-controlling process is the migration of excess vacancies through the bulk to sinks such as the foil surface or voids.

Question. - J. H . CRAWFORD.

G. P. Summers, G. S. White, K. H. Lee and I have investigated the displacement energy in MgA1,0, and we also observe a temperature dependence of Ed, though in an opposite sense than you report at 77 K.

The value of 0= displacement, monitored by the F-band, is 56 eV while at 300 K the value is 130 eV.

We attribute this difference to easier trapping and stabilization of oxygen interstitials at 77 K rather than at 300 K.

Reply. - T . E. MITCHELL.

The lower value at low temperatures may also be due to a replacement collision sequence mechanism.

Anyway, I would expect Ed to decrease at higher temperatures (5000C perhaps) where vacancies become mobile.

(6)

ELECTRON IRRADIATION DAMAGE IN MgO

References

[l] SONDER, E. and SIBLEY, W. A., Point Defects in Solids, Vol. 1, Plenum Press, New York, p. 201 (1972).

[2] HOBBS, L. W., J. Amer. Ceram. Soc. 62 (1979) 267.

[3] MITCHELL, T. E., BARNARD, R. S., HOWITT, D. G . and HOBBS, L. W., in High Voltage Electron Microscopy 1977, T . Imura and H. Hashimoto, eds., Japan Soc. of Electron Micro- scopy, p. 503 (1977).

141 HOWITT, D.

P.

and MITCHELL, T. E., in Electron Microscopy 1978, Vol. 1, Microscopical Society of Canada, p. 276 (1978).

[5] CHEN, Y. and ABRAHAM, M. M., J. Amer. Ceram. Soc. 59 (1976) 101.

[6] PELLS, G. P. and PHILLIPS, D. C., J. Nucl. Muter. 80 (1979) 207.

[7] KIRITANI, M., YOSHIDA, N., TAKATA, H. and MAEHARA, Y., J. Phys. Soc. Japan 38 (1975) 1677.

[8] WUENSCH, B. J., in Mass Transport Phenomena in Ceramics, A. R. Cooper and A. H. Heuer, eds., Plenum Press, p. 221 (1975).

191 HOBBS, L. W., in Developments in Electron Microscopy and Analysis, J . A. Venables, ed., Academic Press, p. 287 (1976).