oxydation zircaloy 4

download oxydation  zircaloy 4

of 6

Transcript of oxydation zircaloy 4

  • 8/9/2019 oxydation zircaloy 4

    1/6

    Journal of Nuclear Materials 202 (1993) 239-244

    North-Holland

    Improvement in oxidation resistance of Zircaloy-4

    by surface alloying with a thin layer of palladium

    G.A. Eloff a, C.J. Greyling a and P.E.Viljoen b

    aAt omic Energy Corporation of South Afr ica Lt d., P.O. Box 582, Pretori a 0001, South Afri ca

    b Uni versit y of t he Or ange Free State, P.O. Box 339, Bloemfont ein 9300, South Af ri ca

    Received 13 October 1992; accepted 18 December 1992

    Economic considerations and environmental pressures have forced reactor operators world-wide to consider higher

    burnup, to the extent that standard Zircaloy-4 fuel cladding is now approaching its design limitations. One of the main

    factors limiting the safe in-reactor lifetime of PWR fuel is the waterside corrosion of the fuel cladding. In this investigation

    the oxidation resistance of Zircaloy-4, surface alloyed with a thin layer of palladium by annealing in vacuum at 95O”C,was

    studied. Short term oxidation in air shows a retardation of the transition to linear oxidation kinetics. It is argued that this

    retardation is caused by the incorporation of intermetallic particles into the oxide layer which increases the ductility of the

    oxide, thereby retarding the onset of cracking from the outside surface of the oxide layer.

    1 Introduction

    The performance of Zircaloy4 fuel cladding tubes

    in pressurized water reactors (PWRs) all over the world

    has been extremely satisfactory over the past two

    decades. In the past decade, however, economic con-

    siderations have been forcing reactor operators to con-

    sider higher bumup levels of their fuel in order to

    contain reload cost and to minimize reactor downtime.

    An added attraction of higher burnup is that the

    amount of spent fuel is accordingly reduced. The latest

    PWRs are designed to operate at higher coolant tem-

    peratures than the previous generations of PWRs with

    longer fuel cycles than before, which means that the

    fuel is now required to withstand a much more hostile

    environment for considerably longer periods [l].

    It is generally accepted that waterside corrosion

    (oxidation and hydriding) of fuel cladding represents

    the single mosti important limitation on the safe in-re-

    actor lifetime of fuel [2-41. The developments referred

    to in the previous paragraph have spurred various

    programs to develop alternative zirconium base alloys

    capable of meeting the demands of reactor operators

    through the next decade. Considerable improvements

    in corrosion resistance have already been achieved

    with new alloys such as ZirloTM (Zr-l%Nb-l%Sn-

    O.l%Fe) [5,6]. The search for better materials, mean-

    while, continues unabated as economics and environ-

    mental concerns about nuclear safety keep the pres-

    sure up. This investigation was carried out with the

    purpose of determining whether the oxidation resis-

    tance of Zircaloy-4 can be improved by surface alloying

    with palladium.

    2. Background

    The effect of a small addition of one of the PGMs

    to some alloys can be quite dramatic, as evidenced by

    the significant improvement in corrosion resistance ob-

    tained by alloying pure titanium with a small amount of

    palladium [7]. This effect, generally referred to as

    cathodic modification or cathodic alloying, is normally

    brought about by either adding a small amount of the

    PGM to the alloy, or by ion implantation of the desired

    PGM into the surface where the cathodic protection is

    required [8,9].

    The comparable chemical and metallurgical be-

    haviour of titanium and zirconium suggests that addi-

    tion of palladium to Zircaloy-4 may have a similarly

    beneficial effect. Very little reference to any such work

    can, however, be found in the literature. This is proba-

    0022-3115/93/ 06.00 0 1993 - Elsevier Science Publishers B.V. All rights reserved

  • 8/9/2019 oxydation zircaloy 4

    2/6

    240

    G.A. Eloff et al. / improv ement in oxidat ion resistance of Zircaloy- 4

    bly due to the results published by Dixon [lo] who

    found no improvement in the corrosion resistance of

    Zircaloy-2 enriched with small amounts of palladium.

    Alloying pure zirconium with varying small amounts of

    palladium, according to Schleicher [ll], brought about

    an improvement in corrosion resistance through anodic

    protection by galvanic corrosion of Zr,Pd intermetallic

    precipitates.

    3. Experimental procedure

    3 I Sample preparation

    A number of 10 mm pieces were cut from a stan-

    dard Zircaloy-4 fuel cladding tube in the cold-worked,

    stress-relieved condition. A 2-p,rn thick palladium layer

    was electroplated onto the surfaces of these samples.

    They were divided into five groups of four each, and

    these groups, designated Al-A4 through El-E4, were

    then separately vacuum-annealed at 950°C for 30 min,

    1, 2, 4 and 8 h, respectively, inside a quartz tube which

    was evacuated to a vacuum of better than 10m4 Pa.

    After annealing the samples were quenched, still in

    vacuum, to room temperature by withdrawing the evac-

    uated tube from the furnace. The average cooling rate

    through the (a + p&phase was approximately 18 K/s.

    It was essential to perform the annealing in vacua to

    contain the high affinity of zirconium for oxygen at

    elevated temperatures.

    Fig. 1. Cross section of a Zircaloy-4 tube, surface alloyed with

    a 2 km thick palladium layer, after vacuum annealing at

    950°C for an hour (sample B2) (200

    x 1.

    4. Results

    4.1. M icrostructur e

    3.2. Sample characteri zati on

    One sample of each group (Al-El) was subjected

    to Auger line scan analysis over a cross section of the

    tube, to determine the depth distribution of palladium

    in the samples. Another sample of each group (A2-E2)

    was mounted end-on in resin, abraded down to 4000

    grit finish, and given a 15-s etch using a solution of 5%

    HF/45% HN0,/50% H,O. The microstructures of

    these samples were then studied under a metallograph-

    ical microscope.

    The microstructures of samples B2, D2 and E2 are

    shown in figs. 1 to 3, respectively. In fig. 1, the zirco-

    nium/palladium interface is difficult to distinguish.

    The outer bands of the zirconium matrix consists of a

    fine a-Widmannstltten structure with a high concen-

    tration of intermetallic precipitates on the sub-grain

    boundaries of the WidmannstHtten structure, while the

    inner band consists mostly of equiaxed a-zirconium

    grains. There is a clear distinction between the two

    types of microstructure. As may be seen in fig. 2, there

    is no more evidence of a separate surface layer after 4

    h of annealing, and the band of the Widmannstgtten

    structure is considerably wider than is the case of fig. 1.

    Some grains in the inner band, closer to the boundary

    between the two types of microstructure, have also

    transformed to the WidmannstPtten structure. In fig. 3

    it is shown that the bands of the WidmannstHtten

    structure are even wider and that several of the

    equiaxed grains throughout the inner band have also

    transformed.

    3.3. Oxidat ion

    4.2. Auger el ectron spectro scopy l i ne prof i l es

    The other two samples of each group (A3,A4-

    E3,E4) were oxidized at 450°C in a quartz tube furnace

    open to the atmosphere. During oxidation the samples

    were periodically removed from the furnace and

    weighed to determine the oxide weight gain as a func-

    tion of time.

    Two AES

    line profiles, obtained on samples Al and

    Dl, are shown in figs. 4 and 5, respectively. In both

    cases, both the zirconium and the palladium profiles

    are shown for scans from outside of the outer surfaces

    of the samples, up to several microns into the bulk of

    the sample. From fig. 4 it appears that there is still a

  • 8/9/2019 oxydation zircaloy 4

    3/6

  • 8/9/2019 oxydation zircaloy 4

    4/6

    242 G.A. Eloff et al. / Improv ement in oxidat ion resistance of Zircaloy- 4

    50

    100

    150 200 250 300 350

    POSITION pm)

    Fig. 5. Auger line scan of a cross section of sample Dl,

    showing complete dissolving of the initial palladium surface

    layer into the zirconium matrix after 4 h of annealing at 950°C

    in vacuum.

    h of oxidation at a weight gain of approximately 40

    mg/dm*. This is still considerably better than the

    performance of the as-received samples, which started

    1

    I

    I DI

    OXIDATION TlME (h)

    Fig. 6. Average weight gain curves as indicated for groups A

    to E respectively. Curve F represents the average weight gain

    of two “as-received” samples. Error bars have not been in-

    cluded as they are smaller than the symbols in the graph.

    I I I I I I I I I

    0 10 20 30 40 50 60 70 60 90 100

    OXIDATION TIME (h)

    Fig. 7. Average weight gain curves showing the initial oxida-

    tion kinetics for two sets of palladium-treated samples (curves

    A and B) and the as-received samples (curve F).

    breaking away at the expected weight gain of 30

    mg/dm’ which was reached after about 80 h.

    5 Discussion

    The solid solubility of palladium in a-zirconium is

    virtually zero at room temperature and is still lower

    than 0.2 at% at 800°C. In the P-phase, however, the

    solid solubility of palladium is almost two orders of

    magnitude higher and reaches a maximum of 11.5 at%

    at 1030°C [12]. The (a + P)/p-phase boundary temper-

    ature is lowered with increasing palladium content to a

    minimum of 755°C at 7 at% palladium. It is therefore

    expected that the palladium which had gone into solu-

    tion in p-zirconium during annealing, would virtually

    completely precipitate in the form of one or more of

    the known Zr-Pd intermetallic compounds. This is

    confirmed by the dense agglomeration of intermetallic

    precipitates observed on the sub-grain boundaries of

    the a-Widmannstltten structure.

    The distinct boundary between the two observed

    types of microstructure observed in figs. 1, 2 and 3 is a

  • 8/9/2019 oxydation zircaloy 4

    5/6

    G.A. Eloff et al. / Improv ement in oxidati on resistance of Zi rcaloy-4 243

    manifestation of the concentration gradient created by

    the inward diffusion of palladium from the sample

    surfaces. This boundary represents the distance from

    the surface at which the palladium concentration falls

    below the value needed to cause complete transforma-

    tion to the P-phase.

    From fig. 5 one may assume a significant diffusion

    distance of approximately 200 pm after 4 h at 950°C.

    Using the well-known solution for Fick’s second law

    x=J2Dt,

    (I)

    the diffusion coefficient of palladium in B-zirconium

    was calculated to be 1.4 x 10m8 cm’/s, which is an

    order of magnitude higher than the self-diffusion coef-

    ficient of 95Zr in pure p-zirconium at this temperature

    [12], and which is comparable to the self-diffusion

    coefficients of many metals close to their melting points.

    This is not surprising when considering that there is a

    eutectic point at 24.5 at% palladium on the

    zirconium-palladium phase diagram, and that the

    melting point of zirconium decreases almost linearly

    from 1852°C for pure zirconium to 1030°C at the

    eutectic point.

    The oxidation kinetics of zirconium-based alloys are

    usually described by a rate law equation of the general

    form

    (AW)” = kt,

    2)

    where n is the exponent which is characteristic of the

    rate law being followed. If n = 2 then eq. (1) is the

    well-known parabolic rate law of Tammann, Pilling and

    Bedworth [13] which indicates that oxide growth takes

    place by a simple diffusion mechanism. More often

    than not, however, published results fit a cubic (n = 3)

    rate law which, though many possible explanations

    have been offered, still remains without a solid theoret-

    ical base [14]. The experimental data represented in

    fig. 6 was fitted to eq. (2) by linearizing the model and

    applying a “least squares” fitting technique. For the

    purpose of comparison, only the data points up to

    weight gains of 30 mg/dm* were used. The results

    obtained in this way correspond more closely to a

    fourth power rate law (n = 4) as can be seen in table 1.

    The improvement in the oxidation resistance

    brought about by surface alloying Zircaloy4 with palla-

    dium is quite substantial, and can be associated with an

    extension of the protective nature of the oxide layer to

    a substantially thicker oxide. The results suggest that

    the most favourable palladium surface concentration

    was obtained after 2 h of annealing, as further anneal-

    ing resulted in an earlier transition to linear oxidation

    kinetics.

    Table 1

    Kinetic parameters n, k obtained by fitting data to eq. (1)

    Sample Annealing

    n

    group

    time (h)

    A 0.5 3.92 10.2

    B

    1

    3.94 9.6

    C

    2

    3.86 9.8

    D

    4 3.70

    9.6

    E

    8 3.51 9.2

    As received

    _

    2.85 6.8

    A possible explanation for the delay of the transi-

    tion to linear oxidation may be found by considering

    the role of intermetallic precipitates in preventing

    cracks in the oxide layer. The ratio between the molar

    volumes of ZrO, and o-zirconium (the so-called

    Pilling-Bedworth ratio) is 1.56, which means that the

    growing oxide is under a compressive stress at the

    metal-oxide interface. The formation of new oxide at

    this interface creates a radially outward pressure on

    the oxide layer, which eventually causes a tensile stress

    on the oxide-atmosphere interface. Although the tran-

    sition to linear oxide kinetics is still not fully under-

    stood, it is generally accepted that this is caused, at

    least partially, by cracks or pores in the oxide which

    develop as a result of the tensile stress in the outer

    part of the oxide layer [15].

    Douglass [16] has shown that the presence of impu-

    rity atoms increases oxide plasticity, presumably by

    enhancing dislocation mobility in the oxide matrix.

    This enhancement in dislocation mobility could be due

    to a lowering of the Peierls-Nabarro stress, which

    depends on the atomic structure and the nature of the

    atomic bonding forces. An increase in oxide plasticity

    or ductility implies that the oxide should accommodate

    higher stress conditions by plastic deformation before

    cracking commences.

    Bangaru et al. [17] suggested that a homogenous

    distribution of fine precipitates, such as is obtained by

    P-quenching, gives rise to the formation and mainte-

    nance of a fine-grained structure for the oxide layer,

    which results in a tougher oxide with improved crack

    resistance under tensile stress conditions; a view shared

    by Glazkov et al. [18]. In addition, a fine oxide grain

    structure implies a high diffusion rate for oxygen ions

    through the oxide layer, as oxygen transport occurs

    mainly through grain boundary diffusion. According to

    Nowok [19], this would benefit the formation of new

    oxide at the metal-oxide interface at the expense of

    internal oxidation of intermetallic precipitates within

  • 8/9/2019 oxydation zircaloy 4

    6/6

    244

    G.A. Eloff et al. / Improvement in oxidation resistance of Zircaloy 4

    the oxide layer, thereby leading to a reduction in the

    internal stresses in the oxide layer.

    All the above-mentioned arguments support the

    idea that the improvement in oxidation resistance

    achieved by surface alloying with a thin layer of palla-

    dium is brought about by postponing the transition to

    linear oxidation kinetics through the formation, during

    B-quenching, of a high concentration of intermetallic

    particles near the surface, which act to prevent crack-

    ing of the oxide layer by increasing oxide plasticity and

    reducing stresses in the oxide layer.

    6. Conclusion

    This investigation has demonstrated that the oxida-

    tion rate of Zircaloy-4 in air can be significantly re-

    duced by surface alloying with a thin layer of palla-

    dium. The microstructural effect of this treatment is

    the formation of a region of a-Widmannstatten struc-

    ture near the surface with an abundance of intermetal-

    lit precipitates on the subgrain boundaries which, when

    incorporated in the oxide layer during oxidation, act to

    retard the transition to linear oxidation kinetics. This

    retardation is ascribed to increased ductility of the

    oxide which allows a higher deformation tolerance

    before cracking.

    Acknowledgements

    The authors are indebted to Dr. W.J. de Wet, Dr.

    E.T. van der Kouwe, Mr. J.G.M. Bresser and Miss M.

    van Reenen for their contributions to this study.

    References

    [l] 0. Ozer, EPRI J. (1990) 47.

    [2] R. Traccucci, A. Berthet and J.F. Marin, Nucl. Eur.

    Worldscan 3-4 (1990) 14.

    [31

    [41

    [51

    [6

    171

    RI

    [91

    [lOI

    illI

    [l-a

    [I31

    [I41

    1151

    [I61

    [171

    [I8

    [I91

    J. Thomazet, J.P. Mardon, D. Charquet, J. Senevat and

    P. Billet, IAEA Report IWGFPT/34 (IAEA, Vienna,

    1990) p. 255.

    F. Garzarolli, F. Schlemmer, E. Steinberg and H.G.

    Weidinger, Proc. 9th Int. Symp. on Zirconium in the

    Nuclear Industry, Kobe, Japan, November 1990, to be

    published.

    R.S. Miller, Nucl. Energy 31 (1992) 47.

    G.R. Kilp, D.R. Thornburg and R.J. Comstock, IAEA

    Report IWGFPT/34 (IAEA, Vienna, 1990) p. 145.

    M. Stern and H. Wissenberg, J. Electrochem. Sot. 106

    (1959) 755.

    J.H. Potgieter and M.U. Kincer, S. Afr. J. Chem. 44

    (1991) 47.

    C.R. Clayton, in Surface Alloying by Ion, Electron, and

    Laserbeams, eds. L.E. Rehn, S.T. Picraux and H.

    Wiedersich (American Society for Metals, Metals Park,

    Ohio, 1985) p. 325.

    C.F. Dixon, Report PM-I-62-5, Mines Branch, Depart-

    ment of Mines and Technical Surveys, Ottawa, Canada

    (April 1962).

    H.W. Schleicher, Proc. AEC-EURATOM Conf. on

    Aqueous Corrosion of Reactor Materials, October 1959,

    Report TID-7587, p. 108.

    0. Kubachewski-von Goldbeck, in Zirconium: Physico-

    chemical Properties of its Compounds and Alloys, ed. 0

    Kubachewski, Atomic Energy Review Special Issue no. 6

    (International Atomic Energy Agency, Vienna, 1976).

    C. Wagner, Corrosion Sci. 13 (1973) 3.

    B. Cox, in: Advances in Corrosion Science and Technol-

    ogy - vol. 5, eds. M.G. Fontana and R.W. Staehle

    (Plenum Press, New York, 1976).

    B. Cox, Report AECL 5548 (Atomic Energy of Canada

    Ltd., August 1976).

    D.L. Douglass, J. Less-Common Metals 5 (1965) 1.

    N.V. Bangaru, R.A. Busch and J.H. Schemel, Seventh

    Int. Symp. on Zirconium in the Nuclear Industry, ASTM

    STP 939 (American Society for Testing and Materials,

    Philadelphia, 1987) p. 341.

    A.G. Glazkov, V.M. Grigor’ev, V.F. Kon’kov, A.S.

    Moinov, A.V. Nikulina and V.I. Sidorenko, IAEA Re-

    port IWGFPT/34 (IAEA, Vienna, 1990) p. 158

    J. Nowok, Oxidation of Metals 18 (1982) 1.