AECL-9382

A T O M I C ENERGY [ < Y W ) L'ENERGIEATOMIQUE

OF CANADA LIMITED V * ^ / DU CANADA LIMITEE

ARE ZIRCONIA CORROSION FILMS A FORM OFPARTIALLY STABILISED ZIRCONIA (PSZ) ?

Les pellicules d'oxyde de zircone sont-elles une forme dezircone en partie stabilisee (PSZ) ?

B. COX

Chalk River Nuclear Laboratories Laboratoires nucleaires de Chalk River

Chalk River, Ontario

March 1987 mars

ATOMIC ENERGY OF CANADA LIMITED

Are Zirconla Corrosion Films a Form of

Partially Stabilised Zlroonla (PSZ)?

by

B. Cox

Reactor Materials DivisionChalk River Nuclear LaboratoriesChalk River, Ontario KOJ 1J0

1987 March

AECL-9382

L'ENERGIB ATOMIQUE DU CANADA, LIMITEE

Leg pellioules d'oxyde de zircone sont-ellea une forme*

zlrcone en partie stabilisee (PSZ)?

par

B. Cox

RESUME

On examine la question de la comprehension de la formation de

porosites dans une pellicule d'oxyde de zirconium encore sous compression

biaxe. On compare la pellicule d'oxyde avec la zircone en partie stabilisee

(PSZ) dans laquelle on a observe que la transformation par les contraintes

de la zircone quadratique conduit a la microfissuration de la structure. On

enumere les similitudes entre PSZ et la pellicules d'oxyde thermique formee

sur les alliages de zirconium et on propose une hypothese pouvant expliquer

la penetration des pores ou des micr~fMssures de la pellicule d'oxyde sur

zircaloy-2 jusqu'a un point tres voisin de la surface de separation

oxyde/metal et expliquer I1observation que ce phe'ncaene ne se produit pas

dans une pellicule d'oxyde sur alliage de Zr-2,5?Nb. On peut verifier oette

hypothese par spectroscopie de Raman au laser de la pellicule d'oxyde au

cours de la orolssance a des temperatures elevees.

Division des Materiaux des reacteursLaboratoires Nucleaires de Chalk River

Chalk River, Ontario KOJ 1J01987 raars

AECL-9382

ATOMIC ENERGY OF CANADA LIMITED

Are Ziroonla Corrosion Films a Form of

Partially Stabilised Zlrconla (PSZ)?

by

B. Cox

ABSTRACT

The problem of understanding the development of porosity in a zirconiumoxide film still under biaxial compression is discussed. The oxide film iscompared with partially stabilised zirconia (PSZ) where stress inducedtransformation of tetragonal zirconia has been observed to lead tomicrocracking of the structure. The similarities between PSZ and thethermal oxide films formed on zirconium alloys are enumerated, and anhypothesis is proposed that can both explain the penetration of pores ornicrocracks in oxides on Zircaloy-2 to a point very close to the oxide/metalinterface, and explain the observation that such a phenomenon does not occurin oxide films on Zr-2.5?Nb. This hypothesis could be tested by laser Ramanjpectroscopy on oxide films during growth at elevated temperatures.

Reactor Materials DivisionChalk River Nuclear LaboratoriesChalk River, Ontario KOJ 1J0

1987 March

AECL-9382

1. INTRODUCTION

It has been known for many years (1-15) that the oxide films formed onzirconium and its alloys during thermal oxidation were largely crystalline,and were mainly monoclinic ZrO;,, with small quantities of cubic (ortetragonal) ZrOa present. Because of the small crystallite size in theoxide (very much smaller than the grain size in the metal), and overlappingdiffraction peaks, in few instances has it been established clearly whetherthe minor phase was cubic or tetragonal. The careful analysis ofdiffraction patterns by Ploc (11) showed the minor phase to be cubic in hisspecimens. Also, the fractional phase composition of oxide films variesaccording to the manner in which they are formed (16,17). Thus, oxide filmsformed anodically tend to have much higher proportions of cubic ZrO2 thanthose formed thermally, and those formed thermally show increasing fractionsof the cubic (or tetragonal) form as the temperature increases. Limitedevidence suggests that there is little or no change in the phase compositionduring cooling from oxidation temperatures (18,19), but that some of thecubic (tetragonal) crystallites transform during oxidation (16) attemperatures of S0O-72O°C.

Oxide films on zirconium alloys grow under high compressive stresses,resulting from the high Pilling-Bedworth ratio (- 1.56), and containsignificant quantities of impurities, especially in commercially usedalloys. It has thus proved difficult to establish whether the cubic (ortetragonal) phase present in the oxide was stabilized by the impuritycontent, the small crystallite size, the stresses present, or a combinationof these. Observations of increasing crystallite size with oxide thickness,and the preferential formation of a fibre texture perpendicular to theoxide/metal interface (16,18,22), have suggested that the growth stressesare the primary driving force for this "ripening" of the oxide structure.

A further change in oxide structure, that has a profound influence on thelong-term corrosion and hydrogen uptake properties of zirconium alloys, isthe development of pores and/or cracks in the protective oxide film once acritical thickness is surpassed. The proportion of pores to cracks in theox?de (differentiated by virtue of their different aspect ratios), is also afunction of the conditions of oxidation (23,24), and leads to distinctivefeatures in the oxidation kinetics. However, from the standpoint of thehydrogen absorption during corrosion in water, the precise size and shape ofthe flaw may be less important than the closeness of approach to theoxide/metal interface of a continuous path for the migration of molecularspecies. This determines the thickness of the residual, imperviousbarrier-layer of oxide through which hydrogen atoms must diffuse to enterthe metal (25,26). Evidence (27) suggests that different zirconium alloyscan show quite different characteristics in the fraction of their oxidethicknesses that become permeable to molecular species after the transitionin the oxidation kinetics (equated with the development of porosity in theoxide film).

Earlier work has usually correlated the development of porosity in oxidefilms with the recrystallisation of the oxide and the development of

the fibre texture that are both driven by the growth stresses (18,24). Theappearance of cracking in the oxide has been equated with an effect of thetensile stresses that can develop at the oxide/environment interface as aresult of the relaxation of the growth stresses initially induced at theoxide/metal interface, because of the mode of growth of successive oxidelayers by inward migration of oxygen ions. However, such an hypothesis, inits simplest form, should lead to the cracks in the oxide penetrating onlyto (or perhaps a little beyond) the neutral stress axis in the oxide (28).Since the latest layers of oxide formed (i.e. those closest to theoxide/metal interface) are those with the highest compressive stresses,cracking should not be able to propagate through these. Similarly, theselayers, as the last to form, should have undergone the leastrecrystallisation (under the residual growth stresses) and, therefore,should not have had a chance to develop porosity. The above hypothesis foroxide breakdown, therefore, does not predict the development of cracks orpores in the layers of oxide closest to the oxide/metal interface, andconversely, therefore, there should always be a significant thickness ofbarrier-layer oxide remaining at this interface.

If we leave aside, for present purposes, the arguments about whether cracksseen in oxide films at or near to the metal/oxide interface aftermetallographic preparation of the specimen resulted from damage induced bythose preparation techniques (29), we are still left with unequivocalevidence (obtained at the oxidation temperature and during the oxidationprocess) that the pore structure in oxides formed on some zirconium alloys,under some conditions of oxidation, penetrates virtually right up to theoxide/metal interface (27,30). We, therefore, need a modified oxidebreakdown hypothesis that can explain this.

In recent years the development of advanced ceramics based on partiallystabilised zirconia (PSZ), has shown that much tougher materials ("ceramicsteel") can be produced in this manner (31-36). The presence of adistribution of small metastable tetragonal ZrO2 crystallites in thesematerials is a critical aspect of the blunting of incipient cracks in theceramic (37-11). Should we be considering the corrosion films formed onzirconium alloys as a type of PSZ, and, if we do, will this lead us to anydifferent conclusions about the possible modes of breakdown of these filmsfrom those already espoused above? It is the purpose of this paper toconsider these questions in the light of the burgeoning literature on PSZ.

2. What is PSZ?

The phase diagrams of Zr02/Mx0y binary systems were established manyyears ago (42). A significant number of group II and III oxides (e.g. CaO,MgO, ThO2, Y203 and rare earth oxides) when added in sufficient quantitiesstabilize the high temperature cubic (fluorite) modification of ZrO2 down toroom temperature. This eliminates the phase transformation and concomitant- 7% volume change that is responsible for the disintegration ofunstabilised zirconia ceramics following thermal cycling. Thus was born the

stabilised zirconia refractory industry. If insufficient stabilizer isadded for full stabilisation at all temperatures, then a partialdecomposition of the cubic solid solution can occur following long hightemperature anneals.

It was discovered more recently that, if an insufficiency of certainstabilising additives were made (especially Y20,) and if a low temperatureanneal was given after sintering (- 13QQ°C), then a product could beobtained consisting predominantly of cubic ZrOj,, but with a dispersion ofsmall precipitates of the high temperature tetragonal form of unstabilisedzirconia. These tetragonal ZrO2 precipitates were stabilised not by theirsolute content, but by their small size (surface energy term) and thecompressive stress field applied by the matrix (37-41). Partiallystabilised zirconias (PSZ) fabricated in this fashion were found to havesignificantly higher thermal shook resistance and fracture toughnesses thanthe fully stabilised ceramic (4;:,44), and were christened "ceramic steel"(31). The improved fracture toughness was shown to be associated with aphase transformation from tetragonal to monoolinic in a zone around anypropagating crack. It was believed that the increase in volume associatedwith this phase-transformation relieved the tensile stresses at the cracktip and thereby arrested it. A full description of the development ofhypotheses explaining the improved fracture toughness is given in the nextsection.

It was found that by adding a sufficiently finely dispersed zirconia phaseto mullite (45) or alumina (46,47) in which it was essentially insoluble,similar stabilised tetragonal ZrQ2 particles could be produced. Theseplayed an apparently similar role in toughening the glass or alumina, by anapparently similar mechanism. The products are usually calledtransformation (or zirconia) toughened aluminas (TTA or ZTA). The smallvolume fraction of tetragonal ZrO* in either PSZ or TTA ceramics, coupledwith the 1% volume change that could be achieved following transformation tomonoclinic ZrO2, limited the extent of the toughening that appearedtheoretically possible (48). A further step then was the development of atechnique (using very fine particle size powders) for producing fullytetragonal zirconia polycrystals (TZP), that were not completely stabilisedchemically, and hence would undergo stress induced transformation to themonoclinic form (35,48-52). In order to explain the extent of fracturetoughness improvement that could be achieved in these various forms ofzirconia and alumina the initial hypothesis, that the volume changefollowing transformation relieved the stresses at the crack tip, has beenquestioned.

3. Mechanism of transformation toughening

The volume change resulting from the tetragonal-monoolinlc phase trans-formation in zirconia is large (- 7 V0I5G), and the mechanism involved hasbeen shown to be basically a diffusionless martensltic transformation(39-41, 53i54). Although the hysteresis in a heating/cooling loop can bevery large, it is basically impossible to stabilise the high temperature

phase in any large body with a large grain size by rapidly cooling. Thevolume change involved does mean that the tetragonal phase can be stabilisedby high hydrostatic stresses (55-57) that are insufficient to causeformation of the high pressure orthorhombic structure (58-60). Early workon the recrystallisation of precipitated hydrous zirconias also showed thata combination of small particle size and adsorbed ionic species at thecrystallite interfaces could stabilise the metastable tetragonal phase(formed on precipitation from aqueous solution and drying) to temperaturesclose to the normal phase transformation temperature (61-66).

Initial observations of the behaviour of small tetragonal ZrO2 particles inPSZ and TTA bodies (where a zone of particles transformed to monoclinic wasobserved out to a critical distance from the propagating crack) gave rise tothe transformation toughening hypothesis already mentioned. Theoreticalcalculations of the extent of toughening that should be possible for a givenvolume fraction of tetragonal ZrO2, a reasonable maximum size for thetransformed region around the crack, and the expected volume change ontransformation, suggested that the observed improvements in fracturetoughness were larger than could be predicted on this hypothesis (67). Morerecent observations have shown that the martensitic twinning in individualtetragonal zirconia particles, that accompanies the transformation tomonoclinic, can give rise to microcracking at the boundary of the originalparticle (Figure 1) in PSZ (68), and a similar microcracking of the actualzirconia particle (that results in the development of small intergranularoracks in the adjacent alumina matrix) has been seen in TTA (67). Theseobservations have resulted in an hypothesis by which relaxation of stressesat the primary crack tip (and hence increased fracture toughness) resultsfrom the development of a microcrack network in the critical zone around thecrack, where transformation of tetragonal-monoclinic zirconia is observed.These microcracks are individually too small to propagate under the stressintensity that can be generated at each one, but serve to lower the stressintensity at the tip of the main crack (67,69). The "microcracking"hypothesis seems to be supported, at present, by the electron microscopeobservations, and seems to be capable of explaining the large increases infracture toughness that have been observed. Peak fracture toughnesses of~20MPa/m have been reported for processed materials having toughnesses ofonly 3-5 MPa/m in the un3tabilised condition (70). These figures are trulygetting close to those typical of metals.

H. Similarities of zirconia films on Zr to PSZ

The oxide film formed thermally on a zirconium sample shows some similarityto PSZ because it consists mainly of small crystallites of monocliniczirconia with a small percentage of cubic (or tetragonal) zirconiacrystallites (10-12). In few instances has it been clearly demonstratedwhether the second phase is tetragonal or cubic. Ploc was able to showcubic ZrO2 in his films, and the minor phase is usually referred to as

cubic ZrO2. By reason of thbir small size, chemical analysis of individualcrystallites has not been possible so as to demonstrate whether or not thecubic crystallites are stabilised chemically (by the impurities or alloyingelements) or mechanically (by the high compressive stresses (28) resultingfrom the formation of a ZrQ2 layer on Zr). Despite the apparent evidence tothe contrary (16,18,19,22), it is possible that there could be moretetragonal ZrO2 in thick oxide films than is measured at room temperature(because it transforms on cooling) or than is seen by x-ray analysis duringoxidation (because the x-rays come predominantly from near the oxide surfacewhere the stress has already relaxed).

However, observation of stripped thin-oxide layers in the TEM (10-15) hasshown very few instances of heavily twinned grains (Figure 2) with aresemblance to the transformed tetragonal zirconia particles in PSZ or TTA.Most of the features seen in this figure are probably Moire fringes.Perhaps this lack of heavily twinned crystallites arises because their sizein oxide films is usually much smaller than in PSZ or TTA. Mark-site t''insin transformed ZrO, are typically 30-50 nm wide; a size similar to thecrystallite size in oxide films. Thus the absence of twinned crystallites inoxide films on zirconium alloys would not imply that those crystallites hadnot undergone a transformation. Hence, the observations of Ploc, that theepitaxiil relationships (between the oxide and Zr matrix) were consistentwith the monoclinic crystallites having formed by the transformation ofpreexisting cubic crystallites (11), and X-ray measurements during hightemperature oxidation (16) that suggest a tetragonal-monoclinic trans-formation of some crystallites occurs as the oxide thickens, are consistentwith this hypothesis. X-ray diffraction studies of the development of atexture in oxide-films of Increasing thickness, either after cooling fromthe oxidation temperature (18) or at the oxidation temperature (16,19,20),showed that the Traction of cubic (or tetragonal) crystallites in athickening oxide did not remain constant. . There was a tendency for thequantity of cubic (or tetragonal) ZrO2 to increase during the formation ofpre-trans it ion oxide films and then for the diffraction-peak intensities toroach a maximum near the transition thickness. During this time the peakintensities of preferred orientations of monoclinic ZrO2 continued toincrease almost linearly. Tetragonal zirconia appeared first on thestrongest alloy (Zr-2.5tNb).

Since fresh oxide is formed at the Zr/ZrO2 interface by the inward migrationof oxygen anions, the outer layers of the oxide are those first formed. Theobserved saturation of the cubic (tetragonal) peak intensities withincreasing oxide thickness is consistent, therefore, with a situation wherea significant fraction of cubic or tetragonal zirconia is formed initiallyat the oxide/metal interface under the influence of the high compressivestressos induced there. As successive layers of fresh ZrOz are formed atthis interface, any individual ZrO2 layer moves away from the interface, andthe biaxial compressivo stress on it diminishes. Once this stress hasdecreased below the critical value needed to stabilise the cubic (ortetragonal) crystallites, then these should transform to monoclinic ZrO2,

and may generate micro-cracks by a mechanism similar to that observed in PS7.or TTA. Even if microtwinning following transformation does not occur themartenaitic shear could still cause tracking at crystallite boundaries.That the cubic (or tetragonal) crystallites are stabilised by stress ratherthan composition is suggested by the observation that in both Zircaloy-2(18,19) and Zr-2.5?Nb (71) a vacuum anneal after oxidation, that reduces theresidual stresses in the oxides, causes the 111f X-ray peak to disappear.An alternative explanation of this observation, that the tetragonal ZrO2crystallites are restricted to a very thin layer of oxide close to theoxide/metal interface that dissolves in the metal during the vacuum anneal,is also consistent with the observations. After oxidation at 500°C, longvacuum anneals at the same temperature eliminate the (iii)j peak onZircaloy-2, but not on Zr-2.5$Nb. An anneal at 700°C is needed to eliminatethis peak from the oxide on the niobium alloy (18,71). Since littledifference in the amount of oxide dissolution after a given vacuum anneal isexpected between Zircaloy and Zr-2.5?Nb, the implication is that thestabilised tetragonal crystallites are present at much greater distancesfrom the interface in the latter than in the former alloy.

An hypothesis based or. the transformation of E-ZrO2 crystallites near theoxide/metal interface would provide a mechanism by which a microcrack orpore network might develop throughout the oxide film on an alloy. It wouldbe necessary only for the transformation stress needed to maintain thecrystallite in the metastable cubic (or tetragonal/ form to be close to themaximum stress generated at the oxide/metal interface for such stabilisedcrystallites to transform and crack immediately they move away from theinterface (where the stress was highest). The factors controlling thisbehaviour would be, firstly, the critical stress needed to maintain thecrystallite in its cubic (or tetragonal) form. This stress would beaffected by the size of the crystallite, and the extent to which it waschemically stabilised by impurities or alloying additions; and, secondly,the maximum biaxial compressive stress that could be generated at theoxide/metal interface. The latter would be a function of the strength ofthe underlying metal, and should be much higher (for instance) forZr-2.5!fNb than for Zircaloy-2.

It is appropriate at this point to ask whether or not the evidence on theeffect of pressure on the phase transformation temperature and the stressesactually measured in oxide films make such an hypothesis reasonable.Figure 3 plots the monoclinic/tetragonal and the monoclinic/orthorhombicZrO2 phase boundaries as a function of temperature and pressure, using thepublished data (55-59) together with the maximum average stress in the oxidefilm measured by Roy et al. (18,19,28) on zirconium, Zircaloy-2 andZr-2.5?Nb alloy specimens. It can be seen that the maximum average stressmeasured increases with the strength of the metal substrate, and in thethickness range 3-5 ym (where the kinetic transition occurs) is close tothat required to stablise the high pressure phase only for Zr-2.5?Nb.Several factors, however, make it plausible that the maximum stress at theoxide/metal interface could be well above this value.

(i) The stress measured by either the bending of a thin beam, or thedisplacement of an X-ray peak will be the average over the oxidethickness (or the mean range of the x-rays if less than this). Sincethe local stress decreases from the oxide/metal interface to theoxide/environment interface, and may be close to zero at the latterfor oxides reaching the rate transition, the maximum stress at theoxide/metal interface could be close to twice the average value.

(ii) The phase boundary lines were measured for bulk oxide samples withcrystallite sizes much larger than the crystallites in a typical oxide 'film. Since small crystallite size is known to stabilize the highpressure phase the position of the phase boundary could be shiftedto lower pressures for the oxide films compared with the bulk oxide,

(iii) Measurement of the stress generated in the oxide by the bending of athin beam of the alloy will tend to underestimate the average stressbecause of the relaxation resulting from the bending of the beam,especially for thicker oxide films where stresses in the beam werehigh enough to induce plastic deformation (28), Support for thisargument can be found in the generally lower average oxide stressesmeasured by Roy et al. (18,71) on their thinner specimens, whereplastic deformation ensued at an earlier stage in the oxidation. Intheir X-ray measurements the signal was coming predominantly from theouter - 2 vim of oxide, so that the method had reduced sensitivity forthe structure of crystallites near the oxide/metal interface inpost-transition oxides (18,19).

It thus appears quite reasonable that tetragonal ZrO2 might be formed underthe high stresses generated at the oxide-metal interface and transform tomonoclinic as that stress decreases during subsequent oxidation, or duringcooling to room temperature.

The literature on zirconia phase diagrams shows (72) that niobium would tendto stabilize an orthorhombic mixed oxide (6Zr0z.Nb205), that appears to beidentical with the high pressure form of ZrOz reported by Bendeliani (58)and Bocquillon (54) (Table 1). None of the alloying additions in Zircaloy-2(Sn.Fe.Cr,Ni) provide any stabilisation of the more symmetrical ZrO2structures (42,73). Thus the critical stress to maintain a high-densityform of ZrO2 (of a given crystallite size) may be higher in Zircaloy-2 thanin Zr-2.55tNb. Both the differences in strengths of the alloys and in theexpected m * t ZrO2 transformation pressures (at any given temperature),therefore, would tend to make cubic (or tetragonal) crystallites in theoxide on Zircaloy-2 transform at a point nearer to the oxide/metal interfacethan similar sized crystallites in the oxide on Zr-2.5$Nb. There is littleevidence on the respective sizes of crystallites in thin oxides onZr-2.5fcNb and Zircaloy-2, but if any difference is evident it appears to bein the direction of smaller oxide crystallites on Zr-2.5?Nb (74). Thisfactor would further enhance the difference in stability of cubic (ortetragonal) crystallites in oxides on the two alloys.

Observations of the development of porosity in oxides on the two alloys hasshown (25,27,30) that on Zircaloy-2, as the oxide thickens, the pore (ormicrocrack) network appears to extend right through the oxide film to (orvery near to) the oxide/metal interface. In contrast, on Zr-2.5$Nb undersimilar conditions of oxidation a relatively thick barrier layer of oxidecontinues to exist at the oxide/metal interface (27). Could it he that theoxides are behaving like PSZ and that transformation cf cubic (ortetragonal) crystallites to monoclinic ZrO2 stimulates miorocracking veryclose to the interface in post-transition oxides on Zircaloy-2, whereaj thecompre.iP've stresses continue to be too high to permit this on Zr~2.5?Nbwithin - 1 \tm of the oxide/metal interface? If this difference is at leastpartly the result of differences in the strength of the underlying metal,what will be the effect of irradiation induced strengthening competing withirradiation induced creep and growth of the matrix on this behaviour - or onthe deformation of the oxide crystallites?

5. Is this hypothesis testable?

In order to test this hypothesis we would require a method for measuring thephase-structure of a ZrO2 film, in-situ on the metal, since, if the cubic(tetragonal) is stabilised by the stress in the oxides then stripping theoxide should cause it to transform. Measurements at the oxidationtemperature and preferably during the oxidation process itself would also bedesirable since cooling may cause some transformation. Most diffractiontechniques have difficulties distinguishing the various zirconia phases inreflection because of the number of overlapping diffraction lines and thefew unique lines that can oe used to characterise cubic or tetragonalzirconia in the presence of monoclinic and possibly also orthorhombicphases. This becomes even more difficult if solid-solution effects andstresses are causing shifts in the relative positions of individual peaks(18,19,71). It would appear, therefore, that we should be looking for atechnique that does not depend upon the diffraction of some form ofelectromagnetic radiation.

Laser Raman spectroscopy may satisfy this requirement. The Raman spectra ofthe various forms of zirconia (Figure H) are very distinct (75-79), thetechnique can be performed in reflection, on a microscopic area of surface,and the exciting radiation (up to visible red wavelengths) will probablypenetrate far enough through the relatively opaque ZrO2 films to see theoxide-metal interface through post-transition oxides. Furthermore, becausei*. can be carried out through an optical microscope, laser Raman spectra canbe obtained preferentially at different depths in the oxide by focussing theexciting beam below the surface. A final advantage is that laser Ramanapectroscopy is unaffected by thermal emissions from a hot sample attemperatures below 1000°C. The technique has been used to resolve thearguments (57-60) about the existence of a high pressure orthcrhombic phasein favour of it (80) and to study the phase of ZrO2 precipitates in glassesand ceramics (81,82).

6, Consequences and Predictions

If such an hypothesis explaining the generation of porosity throughout oxidefilms on the Zircalcys, and predicting the persistence of a relatively thickbarrier layer of oxide at the metal/oxide interface on 2r-?,5^Nb weredemonstrated, the following consequences and predictions would ensue forZtrcaloys:-

(i) Any effect of cold-work and a-annealing on \e time to transitionshould be In the direction of shorter times to transition for theweaker matrix, unless Its strength were reduced to the point wheretetragonal ZrO2 could not be stabilized. The observed effects ofoold^working are in the opposite direction (83), suggesting thatprecipitate size and distribution are more important than matrixstrength in affecting the corrosion kinetics of the Zircaloys.

(il) A. small effect of wall thickness on the time to transition andcorrosion race, in the direction of shorter times to transition andhigher post-transition corrosion rates for thinner material.

(iii) A significant effect of thermal cycling (i.e. frequency of specimenweighing) on time to transition and post-transition rate should beobserved if cooling to room temperature causes t •» m ZrO2 transitionthat results in microcracks that do not heal on returning totemperature.

For Zr-Nb alloy specimens:-

(i) An increase in barrier layer oxide thickness with increasing matrixstrength, possibly leading to reduced hydrogen uptake rates. Thusincreased cold-work or niobium concentration (in the a-phase) shouldbe beneficial. The advantages of increased niobium content shouldpersist only up to the point where the ct-Zr matrix is saturated (orsupersaturated) in Nb; beyond that point (in Zr-2.5?Nb) changingvolume fractions of 6-Zr may have a detrimental effect on thecorrosion behaviour. Nevertheless, small variations in niobiumcontent in the vicinity of 2.5? may affect the behaviour.

(ii) Because of the factors leading to Its thick, continuously presentbarrier oxide, Zr-2.5?Nb should be less affected by thermal cycling(i.e. weighing frequency) than the Zircaloys.

For other zirconium alloy systems:-

(i) Alloyiig additions that improve the stability of tetragonal ZrO2 atlow te .peratures and pressures should improve the time to transition.Yttrium alloys, that might be expected to fall into this category,however, give very poor corrosion resistance and hydrogen uptake rates.These have been ascribed to an effect of yttrium containing

intermetallics (84). An alloying addition (such as a rare earth) thatcould be added in quantities below the solubility in the matrix, andthat would segregate to the oxide side of the oxide/metal interface,might affect the m <- t ZrO2 transformation without producing detri-mental Intermetallics, and should be looked for and tested.

In the corrosion of zirconium alloys in general the observation of theenormous detrimental effect of high concentrations of LiOH in the water isnoteworthy (85). Li+ is much more effective than any other alkali metal ionin the solution, seems to operate by concentrating in the oxide (86)(possibly at the oxide/metal interface), and results in oxide films that areporous from almost the beginning of oxidation (87). Such a specific effectof Li+ might be understood in terms of the effect of the incorporation ofsmall concentrations of Li into ZrO2 on the temperature and pressure of them * t phase transformation. The generation of Lij,ZrO3 during corrosion hasbeen observed, but generally only at LiOH concentrations well above thethreshold for the effect (85).

Acknowledgements

The author is indebted to V.F. Urbanic and R.A. Ploc for their helpfulsuggestions on the subject of this report.

REFERENCES

1. F.H. Krenz, Proc. 1st Int. Cong, on Met. Corr., London, 1961, p.462.

2. J.E. Bailey, J. Nucl. Mat., 1963, 53, 259, and Proc. 3rd Eur. Reg. Cong,on E.M, Prague, 1961, A, 395.

3. M. Denoux and J.J. Trillat, C.R. Acad. Sc. Paris, 1961, 25J3, 4683,

1965, 260, 5003, and IXe Colloque de Metallurgie, Saclay, 1965,p.173.

4. F. Barbesino, R. Di Pietro, G. Perona and R. Sesini, En. Nucleare,1964, n_, 567.

5. D.L. Douglass and J. van Landuyt, Acta. Met., 1965, j_3, 1069, and IXe

Colloque de Metallurgie, Saclay, 1965, p.155.

6. G. Beranger, F. Duffaut, B. de Gelas and P. Lacombe, J. Nucl. Mat,1966, 21f 290, and 1968, 28, 185.

7. P. Gondi and G.F. Missiroli, II N. Cim., 1967, 48, 223.

8 A'' F.W. Vahldiek, J. Less Comm. Met., 1967, 1_2> 19-

9. M.G. Cowgill and W.W. Smeltzer, J. Electrochem. Soc, 1968, 1J_5. ̂ 71 -

10. R.A. Ploc, "Interpretation of a Transmission Electron DiffractionPattern from Thin (S2000A) ZrO2 films", Atomic Energy of Canada Ltd.,Report AECL-2794 (Nov. 1967).

11

11. R.A. Ploo, J. Nucl. Mat., 1968, 28, 48; 1982, TJ_O, 59; 1983, n_3, 75;and 1983, V]5, 110.

12. R.A. Ploo, Proc. 27th Annual Conf. of EMSA, p. 140, (1969).

13. R.A. Ploc, Proc, 7 m e Cong. Int. de Micr. Elec., Grenoble, 1970,p.373.

1 it. A.W. Urquhart and D.A. Vermilyea, ASTM-STP-i>51, 1974, p.463,

15. G.P. Sabol, S.G. McDonald and G.P. Airey, ASTM-STP-551, 197*», P.435 andProc. of IMS Meeting, Detroit 1974, "Stress Effects and the Oxidationof Metals", (Ea, J.V. Cathcart) p.352.

16. J.B. Lightatone and J.P. Pemsler,. Proc. of Conf. on "Kinetics ofReactions in Ionic Solids" Alfred, N.Y., 1967, Plenlum, Chapter 27,p.461.

17. R.A. Ploc and M.A. Miller, J. Nucl. Mat., 1977, 6_4, 71, and Proc.7 m e Cong. Int. de Micr. Elec. Grenoble, 1970, p.371.

18. C. Roy and G. David, J. Nucl. Mat., 1970, 37, 71.

19. G. David, R. Geschier and C. Roy, J. Nucl. Mat., 1971, 38. 329.

20. P. Boisot, G. Béranger and R. Penelle, C.R. Acad. Sc. Paris, Ser. C,1970, 271. 25?-

21. R. Penelle, P. Boisot, G. Béranger and P. Lacombe, J. Nucl. Mat., 1971,38. 340.

22. J.B. Lightstone and J.P. Perasler, Proc. 6th Int. Conf. on Reactivity ofSolids, Schenectady, 1968, p.615.

23. B. Cox, Prog. In Nuclear Energy, Ser. IV, vol.4, Ed. C M . Nicholls,Chapter 3-3, "Recent Developments in Zirconium Alloy CorrosionTechnology", p. 166-190, Pergamon, Oxford, 1961.

24. B. Cox, "Oxidation of Zirconium and its Alloys", Adv. in Corr. Sei. andTech., Eds Fontana and Staeble, Plenum, N.Ï., Vol. 5, (1976) p.173-

25. B. Cox., J. Nucl. Mat., 1968, 27, and 1969, 29, 50.

25. B. Cox "Mechanisms of Hydrogen Absorption by Zirconium Alloys",Presented at MRS Fall Meeting, 1984, Boston, Atomic Energy of CanadaLtd., Report AECL-8702 (1985).

27. B. Cox., J. Nucl. Mat., 1987, in the press.

12

28. C, Roy and R. Burgess, Oxid. Met., 1970, 2_, 235.

29. B. Cox and A. Donner, J. Nucl. Mat., 1971, 41, 96 and 1973, 47, 72.

30. B. Cox, J. Less. Comm. Met., 1963, 5_, 325.

31. R.C. Garvie, R.H. Hannink and R.T. Pascoe, "Ceramic Steel?", Nature,1975, 58, 235.

32. G.K. Bansal and A.H. Heuer, J. Amer. Ceram. Soc, 1975, 5_8, 235.

33. N. Claucsen, ibid., 1976, 59, 179, and 1978, 61_, 85.

34. D.L. Porter and A.H. Heuer, ibid., 1977, 6_0, 183, and 1979, 62, 298.

35. T.K. Gupta et al., J. Mat. Sei., 1977, V2, 2421, and 1978, ]_3, 1474.

36. R.H.J. Hannink, ibid., 1978, 1_3, 2487.

37. D.L. Porter, A.G. Evans and A.H. Heuer, Acta. Met., 1979, 2]_, 1619.

38. A.G. Evans and A.H. Heuer, J. Amer. Ceram. Soc., 1980, 6_3, 241.

39. F.F. Lange, J. Mat. Sei., 1982, 1_7, 225, 235, and 240.

40. A.G. Evans, N. Burlingame, M. Drory and W.M. Kriven, Acta. Met., 1981,29, 447.

41. A.H. Heuer and M. RUhle, Acta Met., 1985, 33, 2101.

42. R.C. Garvie, "Zirconium Dioxide and some of its Binary Systems",Refractory Metals Ed. A.M. Alper, Vol. 5, High Temperature Oxides,Pt. II, Acad. Press, N.Y., 1970, Ch.4, p.117.

43. R.G. Garvie, U.S. Patent 3,620,781, Nov. 16, 1971.

44. R.C. Garvie and P.S. Nicholson, J. Amer. Cerarn. Soo., 1972, 55, 152.

45. E. Bischoff and M. RUhle, ibid., 1983, 66, 123.

46. A.H. Heuer, N. Claussen, W.M. Kriven and M. RUhle, ibid., 1982, 65,642.

47. E.P. Butler and A.H. Heuer, ibid., 1982, 65, C206.

48. T.K. Gupta, Sei. Sinter., 1978, 1_0, 205.

4 9 . T.K. Gupta , R .B . G r e k i l a and E .C. Subbarao , J . E l e c t r o c h e m . S o c , 1 9 8 1 ,128, 929.

50. F.F. Lange, in "Fracture Mechanics of Ceramics", Vol.4, Krta. ür-i :*.,Hasselman, and Lange, Plenum N.Y., 1978, p.877.

51. K. Tsukuma and M. Shimada, J. Mat. Sei., 1985, 20, 1178.

52. T. Sato, S. Ohtaki and M. Shimada, ibid., 1985, 20, 1406.

53. G. Wolten, J. Am. Ceram. S o c , 1963, U6, 419, and Aeta Jry:<i., ' *> ••,17, 763-

5M. J.E. Bailey, Proa. Roy. S o c , A, 1964, 27JK 395.

55. E. Dow Whitney, J. Am. Ceram. Soc., 1962, 45, 612.

56. G.L. Kulcinski, ibid., 1968, 51.. 582.

57. L.M. Lityagina, S.S. Kabalkina, T.A. Pashkina and A.I. Kh r.'.y..:i: v, /.-,Phys., Sol. State, 1978, 20, 2009.

58. N.A. Bendeliani, S.V. Popova and L.F. Vereshchagin, Geikhir-iiya, '•».'',§_, 677.

59. G. Bocquillon and C. Susse, Rev. Int. Hautes Temp, et Refract., T%>> ,5, 247 and 1969, 6, 263.

60. L.G. Liu, J. Phys. Chem. Sol., 1980, in., 331.

61. C.B. Amphlett, L.A. McDonald and M.J. Redman, J. Inorg. Nucl. Chen.,1958, 6, 236.

62. R.C. Garvie, J. Phys. Chem., 1965, 6J_, 1991.

63. K.S. Mazdiyasni, C.T. Lynch and J.S. Smith, Am. Ceram. Soc., 1965, ^372.

64. E. Dow Whitney, Trans. Far. Soc., 1965, p_t_, 1991.

65. J.E. Bailey, D. Lewis, Z.M. Librant and L.J. Porter, Trans. Brit.Ceram. Soc., 1972, jH_, 25.

66. R.C. Garvie, J. Phys. Chem., 1978, 82, 218.

67. M. RUhle, "Microcrack and Transformation Toughening in ZrO2 -Containing A1SO3", Proc. of Fall 1986 MRS Meeting, Boston, 1-6 Dec.1986 (Abstract E2.H).

68. B.C. Muddle, "Displacive Phase Transformation in Zirconia-basedCeramics", Proc. of Fall 1986 MRS Meeting, Boston, 1-6 Dec. 1986(Abstract E1.1).

• •. :.;:. ;•'.••.•;,,•.•;, "A!>.i'Jt th<> Influence of Rlastic Interaction of

'•': •!•:•!•!>::; on v.i>: TJURIUIOSS of Ceramics", Proc. of Fall 1986 MRS

'•'.•.•! :rlt;, ••< :;t ,>ri V , Dec, 1986 (Abstract E2.9).

' . K. Tij'ik.irn.i, '!'. T.'ikahata and H. Yamamui'a, "Mechanical Properties and

"i •'•• :it:M]i'Uiri; "f TZP and TZP/A12O3 Composites", Proc. of Fall 1986 MRS

••'.•.'i:.w, iv .;'.,:ii, i-i, nee. 1986 (Abstract E3.1).

''. ', •'. y .i:. ! •'.!). Wu, "3trG.MS Generation in Zr-2.5 wt? Nb During

>.: : t' i m at hi)')0"., Research Report, University of W. Ontario, Faculty

! ••::.,.•. .;."., I, ii'lnn, Canada (flay 1970).

'. . '•...•. '••: ' !• a n d L .w. C o u g h a n o u r , J . R e s . NBS, 1 9 5 5 , 5 5 , 2 0 9 .

A. '>'.-4 .'."'. a n d H. T o b e r , B e r . D e u t . Ke ram. G e s . 1 9 5 3 , 3_0, ^ 7 .

' - . ' • ' . . ; . 'w . i .M!! m a W.W. S m e l t z e r , J . E l e c t r o c h e m . S o c , 1 9 6 8 , U_5, ^ 7 1 .

'. '•!. i'-ii i 1 i pp i and K..S. M a ; : d i y a s n i , J . Am. C e r a m . S o c . , 1 9 7 1 , 5 £ , 2 5 4 .

•• . V. ".. Y.'>r amid IM a n d W.B. W h i t e , i b i d . , 197-M, 57_, 2 2 .

'•:. ••.-: ri". l.-iSiXi a , B. P a p a n i c o l a o u a n d I . M . A s h e r , J . P h y s . Chem. s o l . ,

' • . '•'.. : : ; ; ; i y.'ir.o a n d T. S a k u r a i , J . Am. C e r a m . S o c , 1 9 7 7 , 6_0, 3 6 7 .

• ' ' , "'.\-.. "\.irk>? a n d F . A d a r , A d v . i n M a t . C h a r a c t e r i s a t i o n , E d s .••.(•.:}:;ir.Etor., ^ o n d r a t e and S n y d e r , P l e n i u m , N . Y . 1 9 8 3 , p . 1 9 9 .

'• . ii. Ar-.-i ;'-ii ind M. I s h i g a m e , P h y s . s t a t . s o l . ( a ) , 1 9 8 2 , 7j_, 3 1 3 .

• • . • . . ' ' . . ::ar'ki-" a n d F . A d a r , , ] . Am. C e r a m . S o c . , 1 9 8 2 , 65_, VM.

* '. !.. o o t o a n d F . A d a r , i n " M i c r o b e a m A n a l y s i s - 1 9 8 4 " E d s . Romig J r . a n di j o M s t e i : i , San F r a n c i s c o P r e s s , San F r a n c i s c o , 1 9 8 4 , p . 1 2 1 .

H.J,. G . P . M a r i n o a n d R . L . F i s c h e r , B e t t i s A t . Power L a b . , R e p o r t

WAPD-TM-1322 ( 1 9 7 8 ) .

y i . H. C o x , U . K . A . E . A . , A . E . R . E . H a r w e l l , u n p u b l i s h e d d a t a ( 1 9 6 3 ) .

8 ' j . H. C o r i o u , L . G r a i l , J . M e u n i e r , M. P e l r a z a n d H. W i l l e r m o z , J . N u c l .

M a t . , 1 9 6 2 , 7 , 3 2 0 .

fJ>. S . G . M c D o n a l d , G . P . S a b o l a n d K . D . S h e p p a r d , ASTM-STP-824 ( 1 9 8 4 )

p . 5 1 9 .

8 7 . N. R a m a s u b r a m a n i a n and N. P r e o c a n i n , A t o m i c E n e r g y o f C a n a d a L t d . ,

R e p o r t AECL-8108 ( 1 9 8 3 ) P . 1 1 3 -

15

TABLE 1

ORTHORHOMBIC Z r O ,

Author

Bendeliani

Booquillon

Liu

Roth & Coughanour

5.

(5

3.

cell

a

110

.16

328

964

constants

b

5.073

5.12

5.565

5.120

c

5.26H

5.3)

6.503

5.289

Remarks

estimated from d111

"cotunnite" structure

Zr\(Nb2Zr2)0l60

16

(a)

(b)

MICROCRACKS FORMED DURING TRANSFORMATION

(o)

MICROCRACKS

F i g u r e 1: P a r t i a l l y s t a b i l i s e d z i r c o n i a c o n t a i n i n g t - Z r O 2 p a r t i c l e s ( a ) ,t r a n s f o r m e d ZrO2 p a r t i c l e in A12O3 (b ) and mechanism ofm i c r o c r a c k fo rma t ion d u r i n g t w i n n i n g on t r a n s f o r m a t i o n ( c ) .

17

Figure 2: Transmission electron micrograph of oxide film near theoxide/metal interface of a post-transition oxide on Zircaloy-2(Ref. 14).

1000 —TETRAGONAL

© REF. 55CD REF. 58CD REF. 59© REF. 57® REF. 56

O ZrA ZIRC. 2D Zr-Nb• Zr/Nb 15 pm'

cc.

<UJQ_

AVERAGE STRESSIN 3-5 jjm OXIDE(REFS. 18, 28)

X--

ORTHORHOMBICOR

ORTHO. + MONO.

10 20 30PRESSURE (kbar)

Figure 3: Effect of pressure on monoclinic-tetragonal ZrO2 transformation.

19

CUBIC(ZRO2-157«CA0)

TETRAGONALZROO

MONOCLINICZRO2

Figure 4: Raman spectra of the common polymorphic forms of

ISSN 0067-0367

To identify individual documents in the serieswe have assigned an AECL- number to each.

Please refer to the AECL- numberwhen re-questing additional copies of this document

from

ISSN 0067-0367

Pour identifier les rapports indi viduels faisantpartie de cette serie nous avons assignsunnumeroAECL- achacun.

Veuillezfaire mention dunumero AECL- sivous demandez d'autres exemplaires de cerapport

au

Scientific Document Distribution OfficeAtomic Energy of Canada Limited

Chalk River, Ontario, CanadaKOJ1J0

Service de Distribution des Documents OfficieisL'Energie Atomique du Canada Limitee

Chalk River, Ontario, CanadaK0J1J0

Price: A Prix: A

©ATOMIC ENERGY OF CANADA LIMITED, 1987

857-87