HIGH TEMPERATURE DAMAGE ACCUMULATION IN HOT PRESSED ALUMINA

HAL Id: jpa-00225496

https://hal.archives-ouvertes.fr/jpa-00225496

Submitted on 1 Jan 1986

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.

HIGH TEMPERATURE DAMAGE ACCUMULATION IN HOT PRESSED ALUMINA

A. Robertson, D. Wilkinson

To cite this version:

A. Robertson, D. Wilkinson. HIGH TEMPERATURE DAMAGE ACCUMULATION IN HOT PRESSED ALUMINA. Journal de Physique Colloques, 1986, 47 (C1), pp.C1-661-C1-666.

�10.1051/jphyscol:19861101�. �jpa-00225496�

HIGH TEMPERATURE DAMAGE ACCUMULATION IN HOT PRESSED ALUMINA

A.G. ROBERTSON and D.S. WILKINSON

Department of Metallurgy and Materials Science, McMaster University, Hamilton, Ontario L8S 4L7, Canada

Resume - On a 6tudi6 le d6veloppement de la cavitation aux joints de grains, dans l'alumine pressee ii chaud, ii la fois durant le fluage en flexion et pen- dant les recuits sans contrainte dans l'air. L'endomagement par fluage est provoqud ii la fois par la contrainte appliquee et par la pression du gaz dans les cavites. La pression interne est engendr6e par le gaz piegg durant le pas- sage 2 chaud et par le gaz produit par reaction chimique durant le test. Sur les surfaces usinges, une peau resistant l'endomagement d'environ 15-20 p d16paisseur recouvre une region 2 forte cavitation.

Abstract - The development of grain-boundary cavitation in hot-pressed A1203 both during flexural creep and tress-free annealing in air has been studied.

Creep damage accumulation in driven both by the applied stress, and by inter- nal gas pressure in the cavities. The internal pressure results from gas trapped during hot pressing, and from gas-generating chemical reactions acti- ve during testing. At machined surfaces, a damage-resistant skin about 15-20 pm thick overlies a heavily cavitated region.

INTRODUCTION

Hot pressed alumina (HPA) is an attractive model material for basic research into creep fracture. The work reported here represents the first stages in applying a damage accumulation approach to this topic. The long term objective is a fundamental understanding of crack nucleation and propagation in the creep fracture regime, applicable to predicting reliability in structural ceramic components in high temperature service. The immediate aims are to evaluate HPA as a model material, and to establish geometries and conditions for future tests. In this work, the levels of internal grain boundary cavitation (area fraction and number density) were determined for HPA both after creep (at 1350'C) and after stress-free annealing in air (1250, 1350 and 1470°C). The profiles were generated by image analysis of SEM micrographs taken in step scans across internal sections of test specimens.

EXPERIMENTAL

All specimens were taken from a single billet of hot-pressed A1203 obtained from AVCO, Systems Division. Bend specimens 5 ~ 1 2 . 5 ~ 6 5 mm were machined from the billet. The procedure used is outlined elsewhere /l/. Two parallel straight lines, 10 mm apart, were diamond-scribed across the centre of one 12.5 X 60 mm side of each creep bar. From measurements of the separation of these lines,

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19861101

JOURNAL DE PHYSIQUE

before and a f t e r a c r e e p t e s t , f i b e r s t r a i n s were c a l c u l a t e d /2/. Bars were c r e p t i n a i r , a t 1350°C, i n a four-point bend r i g / l / . Curvature a t high f i b e r s t r a i n s was minimized by creeping b a r s "on edge". The l o a d i n g d i r e c t i o n was p a r a l l e l t o t h e o r i g i n a l d i r e c t i o n o f hot-pressing. Control specimens were placed on t h e lower block o f t h e bend r i g . Creep s t r a i n was monitored by a 3-probe extensometer. S t r e s s e s and s t r a i n s were c a l c u l a t e d a f t e r Hollenberg e t a l . / 3 / ; applied s t r e s s e s r e p o r t e d h e r e a r e maximum, s t e a d y s t a t e values. After creep, samples were s e c t i o n e d p a r a l l e l t o t h e loading d i r e c t i o n . The c u t f a c e s were diamond p o l i s h e d , thermally etched (1600 " C , a i r , 34 min.), and coated w i t h gold.

S t e p s c a n s o f 12 SEX micrographs each were taken a c r o s s t h e p o l i s h e d , etched f a c e s . On each micrograph, t h e average a r e a f r a c t i o n and number d e n s i t y o f g r a i n boundary c a v i t i e s was measured.

FIBER STRAIN (%l

to.

z S

U g

LL U W U

*

m 5

-10.

-8. 0 8.

FIBER STRAIN (%l

-8. 0 8.

FIBER STRAIN I%) -

- ^{40 MPa} -

0,.

10. I I I , ,

12 MPa

-

-

. ^{I I}

-8.

-8. 0 8.

FIBER STRAIN I%) a -

I _

l 8 .

Fig. 1: Creep damage p r o f i l e s , with s w e l l i n g e f f e c t s removed, f o r two a p p l i e d s t r e s s e s : ( a ) and ( c ) , a r e a f r a c t i o n ; ( b ) and ( d ) number d e n s i t y .

Creep Damaae

Damage measurements on c r e p t samples shows t h a t t h e r e is a d e f i n i t e c r e e p component t o g r a i n boundary c a v i t a t i o n : t h e damage on any f i b r e i n c r e a s e s with s t r a i n . However, t h e average damage a t t h e n e u t r a l a x i s is a l s o non-zero, and i n c r e a s e s w i t h time: t h e m a t e r i a l is s w e l l i n g . Near t h e specimen f a c e s , damage

Fig. 2: Swelling cavitation in HPA during air anneals: a ( area fraction vs.

time, b) number density (voids.cm-2) vs time, c) average grain size (m) vs time, d ) number density (voids-gain-l) vs time.

1350 C 1250 C

1

l o g 1 0 P e f f (MPa)

Fig. 3: The swelling rates of average individual pores vs effective pressure, for the p O f ~ O in Table 2, and n=l.

JOURNAL DE PHYSIQUE

profiles are perturbed. A damage-resistant 'skin' about 15-20 p thick overlies a region in which cavitation is heavier than would be expected by extrapolating from the interior of the bar.

The creep component of damage, with swelling effects removed, is shown in Fig. 1. These figures were generated from raw creep damage data by converting position into strain /l / and by subtracting the average damage level at the neutral axis. The resulting net damage plots for a single stress were superimposed. This approach assumes that the damage level at the neutral axis in a crept bar can be taken to be representative of the swelling component of damage across the bar.

The figures show that near-surface damage consists of a cavitation-resistant 'skin' overlying a heavily-cavitated layer. The average damage level for the combined 'skin and spike' is close to that extrapolated from internal damage. On the area fraction plots (Figure la,c) the rate of increase of the 'spike' damage with strain is similar to that for internal damage.

Swelling Damage

Data and typical microstructures for swelling in the absence of applied stress at 1250 "C, 1350 "C and 1470 'C are shown in Fig. 2. The area fraction of cavities ( f ~ , Fig. 2a) increases with time at all three temperatures. However, this is not the case for the number densities (Fig. 2b). In this figure, NA initially increases rapidly at all temperatures. This may reflect growth into the detectable size range of pre-existing voids. At 1470°C however, NA peaks near 30 hours, after which it decreases rapidly. Fig. 2c shows the variation of grain size, LQ, with time. The dependence of this parameter on time and temperature is similar to that for f ~ . At temperatures of 1250°C and below, grain size is effectively constant. Fig. 2d shows the variation in number density of cavities, expressed as voids per grain. This plot was generated from NA and L3 data by converting the average 2-dimensional grain size into the area of a regular hexagon. The data suggest that NA decreases at 1470°C, at times longer than 30 h, because rapid grain growth causes cavities to coalesce.

DISCUSSION Creep Damage

The area fraction of grain boundary cavities in the interior of crept bars is strain-controlled at both stress levels. This suggests that the growth of larger cavities is controlled by grain boundary sliding. However, the number density (a parameter dominated by the smaller cavities) is strain-controlled only in the 40 MPa tests. This is not the case in the 12 MPa tests. This indicates that the mechanism controlling the growth of small cavities changes with stress (possibly to diffusion control at low stress). The limited amount of data currently available precludes any firm conclusions.

Near-surface damage profiles suggest that the surface damage gradients result from transient load shedding which occurs early in a creep test, but which is not active subsequently. Work is proceeding to determine whether chemical or mechanical processes are dominant.

Two types of chemical effects may be active. Boundary segregation of solutes, modulated by near-surface concentration gradients over distances of tens of microns, could affect grain boundary cohesion, and so, cavity nucleation.

Alternatively, oxygen attack down grain boundaries could generate high-pressure CO or C02 gas within cavities /4-6/. Near the bar surface, however, the gas may be able to escape, and the carbon would become depleted relatively rapidly.

seen in regions near tensile faces. However, the machining schedule used would be expected to create a 15-20 p thick surface layer, in which residual compressive stresses on the order of 100-200 MPa were present /7,8/. Such a layer would overlie a region subject to a compensating tensile stress field. Also, the surface layer would be charged with a high density of potentially mobile dislocations and active sources / g / . Until residual fields and mobile dislocation densities were annealed out during either a heat treatment or a creep test, or were swept away early in a test by creep deformation, the overall fiber stress gradients and matrix 'softness' (that is, its ability to deform without cavitating) would change sharply near machined surfaces.

Swelling

The swelling damage data has been tested against the hypothesis that the driving force for swelling is generated solely by gas trapped within the compact by pore closure /10/, during hot pressing in air. Following pore closure, the mass of gas in pores is assumed to be constant. The hypothesis implies that the residual cavities in the as-received billet are charged with high-pressure gas.

The cavities are assumed to be spherical and closed, and the trapped gas to behave ideally. An analysis of the model using the current swelling data /l/ leads to several conclusions. First, we conclude that the amount of gas trapped in the pores is much less than one would expect. It therefore appears that much of the trapped gas is dissolved into the A1203 during hot pressing. When this effect is included in the model, the swelling at 1250°C shows a linear dependence of cavity growth rate v and the effective pore pressure, Peff (Fig. 3). However, at 1350 and 1470 "C, the swelling rates do not decrease linearly with the calculated Peff.

At these temperatures, gas-generating reactions capable of sustaining pore pressures somewhat smaller than those initially present due to trapping, apparently occur within the HPA. As a cavity grows, and the pressure component due to trapped gas falls, and chemically-generated gas pressures progressively become a larger fraction of the total driving force.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the contributions of G. Sutherland, who performed the swelling tests, and the assistance of Norton Research (Canada), Niagara Falls, Ontario, in the use of their Joyce-Loebl image analyzer. This work was supported by funding from the Natural Science and Engineering Research Council, Ottawa, Canada.

CONCLUSIONS

Hot-pressed alumina, containing dissolved carbon and machined with 35 p diamond, accumulates low levels of cavitation damage internally when crept in air at 1350 "C. The damage has 3 superimposed components: one related to creep, one due to swelling, and one occurring near machined faces. Damage accumulation cannot be characterized from surface measurements alone. Up to total damage levels of about f~ = 0.004, the creep component of flexural damage is strain- controlled at applied steady state stresses of 12 and 40 MPa. However, while the number density of cavities is strain-controlled at 40 MPa, this is not the case at 12 MPa. Swelling in air is driven both by the pressure of gas trapped within the compact during hot-pressing, and by internal gas-generating reactions. The reactions are significant at 1350 and 1470 "C, but are negligible at 1250 °C.

Swelling can be used to predamage HPA uniformly before creep testing.

JOURNAL DE PHYSIQUE

REFERENCES

A.G. Robertson and D.S. Wilkinson, Intl. Conf. on Fracture Mechanics of Ceramics, Virginia Poly. Inst., June 1985.

J.R. Porter, W. Blumenthal, and A.G. Evans, Acta. metall. 29, 1899-1906 (1981).

G.W. Hollenberg, G.R. Terwilliger and R.S. Gordon, J. Am. Ceram. Soc. 54[4], 196-199 (1971 1.

P.G. Shewmon, H. Lopez, and T.A. Parthasarathy, Scripta Metall. 17, 39-43

( 1983).

S.J. Bennison and M.P. Harmer, Grain growth and cavity formation MgO-doped Al2O3, in "Advances in Ceramics, Vol. 6, Character of Grain Boundaries", Am.

Ceram. Soc. , Columbus, OH ( 1983).

R. Raj, Acta Metall. 30, 1259-1268 (1982).

D.B. Marshall, A.G. Evans, B.T. Khuri Yakub, J.W. Tien, and G.S. Kino, Proc.

R. Soc. Lond. A385, 461-75 (1983).

F.F. Lange, M.R. James and D.J. Green, J.Am. Ceram. Soc. 66 [2], 6-16-17 (1983).

B.J. Hockey, Proc. Br. Ceram. Soc. 20, 95-115 (1972).

A.G. Robertson and D.S. Wilkinson, to be published.