Some thoughts on the mechanisms of in-reactor corrosion of zirconium alloys
Introduction
Zirconium alloys were developed for use as nuclear fuel cladding (amongst other nuclear reactor components) in the early 1950s [1], [2]. A number of reviews of their corrosion mechanisms, both out- and in-reactor, have been published since then [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]. The most comprehensive of the recent reviews [9], [11] are unfortunately out of print, and so a further review at this time seems appropriate, especially as recently published research has resulted in significant changes to ideas of what is affecting corrosion in-reactor. It is not the intent here to match the comprehensiveness of IAEA-TECDOC-996 [11] which comprises 308 pages and 538 references. The aim will be to follow the development of an understanding of the corrosion mechanisms that operate in nuclear reactors, by citing the key pieces of research that have moulded this understanding. Wherever possible references that give a good listing of previous work in a specific area are used, so that readers will have access to the primary references – although the availability of such early references can no longer be guaranteed.
Before an assessment of the in-reactor corrosion mechanisms can be attempted, a summary of the early zirconium alloy development, and the mechanisms that operate out-reactor is necessary. If the oxidation mechanism out-reactor is not understood, useful interpretations of the in-reactor behaviour cannot be offered.
Section snippets
Development of zirconium alloys
Zirconium was chosen for use in the cores of water-cooled nuclear reactors because of its low thermal neutron capture cross-section, reasonable mechanical properties and adequate corrosion resistance in high temperature water [1], [3]. It was initially thought that the poor corrosion resistance of some batches of unalloyed zirconium produced by the van Arkel process was a result of stray impurities. However, it was found that improving the purity did not eliminate these problems [1]. The
Corrosion out of reactors
Because the pre-transition corrosion kinetics are independent of pH between about 1 and 13 (if no specifically aggressive species such as LiOH is present) and of the source of the oxygen (O2, H2O, O in Na/K), if other factors such as temperature and pressure are constant, it has been argued that the oxidation processes are controlled entirely by diffusion of species through the oxide film [5], [6]. The recent suggestion by Bossis et al. [156] that the rates of surface reactions can be rate
Mobile charge carriers in zirconium oxide films
In the absence of external applied potentials on the specimen there can be no net current through the oxide. Thus, the negative and positive oxidation currents must be equal and opposite. If this is not the case initially, then a potential will develop across the oxide that will equalise these currents. In ceramic zirconia specimens oxygen was established to have much higher diffusion rates than zirconium ions. Measurements using radioactive oxygen isotopes [23] showed that oxide grain boundary
Rate determining processes
It is not easy to deduce, from separate measurements of the oxygen diffusion coefficients and the electrical conductivity of oxide films formed on zirconium alloys, which process is rate determining. However, since charge balance must be maintained between the two processes (at least on a general basis, if not on the local ionic scale as required by the Wagner/Hauffe oxidation mechanism [37]), then, if the two processes do not initially proceed at identical rates, a potential will develop
Corrosion kinetics
During the pre-transition period of corrosion (i.e. prior to the accelerated, roughly linear, post-transition period indicating the development of gross porosity) the kinetics at reactor temperature (300–360 °C) have repeatedly shown an approximately cubic rate law, rather than the parabolic law predicted for a process controlled by bulk diffusion of oxygen through the oxide [37]. Analysis of oxygen isotope diffusion profiles [23], [43] showed that the crystallite boundary diffusion was more
Oxide breakdown processes
In general it is not possible to grow oxide films on zirconium alloys that are thicker than 2 μm without a change in the kinetics to either an approximately linear or a cyclic kinetic stage (post-transition) that is considered to indicate the development of some sort of porosity in the previously protective oxide. Although the pre-transition oxide is regarded as a diffusion barrier this does not mean that it is necessarily perfect. A few small flaws in the pre-transition oxide will not
Effects of hydrogen
Hydrogen absorption during corrosion in aqueous environments was one of the earliest observations made during the study of zirconium alloy corrosion [3]. The mechanism of hydrogen absorption is more appropriate for a separate review than for this one, but the question of whether or not, and under what circumstances hydrides in the metal phase can affect the corrosion reaction is appropriate for consideration here. The corrosion of solid zirconium hydride in high temperature steam has been
In-reactor corrosion morphology
A number of different corrosion morphologies have been observed under irradiation either in test reactors or power reactors.
(1) Uniform oxide growth is usually observed in Pressurised Water (or Heavy Water) Reactors provided enough dissolved hydrogen is present to suppress water radiolysis. When these films become very thick (⩾100 μm) oxide delamination and spalling occur [79]. Some cladding has been observed to spall at lower oxide thicknesses.
(2) Nodular corrosion has been a perennial problem
Enhanced diffusion
Because the growth of oxide films on zirconium alloys requires the diffusion of oxygen ions through the films it was expected from the start of the exposure of zirconium alloys in-reactors that displacement of ions from their lattice sites by fast neutron or heavy particle (primary knock-on) damage would lead to enhanced point defect concentrations, enhanced diffusion and hence enhanced corrosion in-reactors [88]. Mechanisms similar to those leading to enhanced rate processes in the metal were
Galvanic effects in-reactor
In order to see galvanic effects between dissimilar metals a potential difference and an electrically conducting path between the two metals are necessary. In the case of zirconium alloys out-reactor even the ‘air-formed’ oxide film present on all zirconium surfaces in aqueous environments has sufficient resistivity to prevent significant galvanic currents passing to dissimilar metals. In-reactor it was learned very early [116], [117] that this was not the case when Zircaloy-2 flow-tubes in the
Corrosion rates and oxide morphologies in commercial reactors
With the advent of eddy current oxide thickness measurements [11] an explosion of available data on oxide thickness measurements in both PWRs and BWRs has become available. Only in BWRs are major corrections to the ‘lift-off’ measurements necessary if magnetic crud deposits are present [131].
Water chemistry effects
If nodular corrosion is a micro-galvanic corrosion effect between the SPPs and the surrounding α-Zr matrix then the water chemistry becomes important for several reasons:
(i) Changing the redox potential of the coolant (e.g. by hydrogen additions) should eventually eliminate the galvanic potentials driving the effects. The evidence (Fig. 32) suggests [9], [11] that this condition will not be reached until the hydrogen content of the water exceeds 10 cc/kg. Prior to this point it is not clear
Conclusions
Opinions about the mechanisms that are important in determining the corrosion rates of zirconium alloys have changed over the years. Some factors that were thought to be important during early studies have not revealed supporting evidence. Thus:
- •
There seems to be little or no supporting evidence for the idea that displacement damage from fast neutrons and the correspondingly increased diffusion rates in the protective oxide result in direct enhancement of the corrosion rates [9], [11]. Evidence
References (156)
J. Nucl. Mater.
(1968)J. Nucl. Mater.
(1968)J. Electrochem. Soc.
(1968)- et al.
Solid State Ion.
(1995) - et al.
J. Electrochem. Soc.
(2001) - et al.
J. Nucl. Mater.
(2001) - et al.
J. Nucl. Mater.
(2001) - et al.
J. Nucl. Mater.
(1995) J. Nucl. Mater.
(1969)J. Nucl. Mater.
(1970)
Oxidation of Metals
J. Nucl. Mater.
J. Nucl. Mater.
J. Nucl. Mater.
J. Less. Comm. Metals
J. Mater. Sci.
J. Nucl. Mater.
J. Nucl. Mater.
Corrosion
J. Amer. Ceram. Soc.
J. Nucl. Mater.
J. Nucl. Mater.
Zirconium Alloys for Nuclear Energy
J. Aust. Inst. Metals
J. Electrochem. Soc.
J. Electrochem. Soc.
J. Nucl. Mater.
Cited by (512)
Oxidation behavior of Zr-2.5Nb alloy exposed to steam in the temperature range of 600–1200 °C
2024, Journal of Nuclear MaterialsFormation of large loop-hydrogen complexes and related effects on mechanical properties of Zirconium investigated with molecular dynamics method
2024, Materials Today CommunicationsPeculiarities of interaction between oxygen and solute/vacancy in commercial-type zirconium-based alloys: A first-principles study
2024, Materials Today CommunicationsHydrogen diffusion in zirconium hydrides from on-the-fly machine learning molecular dynamics
2024, International Journal of Hydrogen Energy