Atomistic study of grain boundary sink strength under prolonged electron irradiation
Highlights
► Prolonged irradiation until steady state is simulated in molecular dynamics simulations. ► Rate theory is revisited with all parameters derived from atomistic simulations. ► The defect accumulations in nanograin Mo from molecular dynamics simulations agree well with the rate theory predictions. ► The grain boundary structures remain unchanged after prolonged irradiation.
Introduction
Grain boundaries can act as sinks and sources of mobile defects [1], [2], [3]. Under irradiation, such as neutron irradiation in nuclear reactors or ion irradiation during implantation, mobile defects – vacancies and self-interstitial-atoms (SIAs) and their clusters – form and interact with the GBs [4], [5], [6], [7]. The GBs in a typical polycrystal can be of various natures, ranging from high-energy boundaries with disordered structures to high-symmetry boundaries with highly ordered crystalline structures and low energies [8]. High-energy GBs are quite prevalent in structural materials, and their role as sinks and sources of mobile defects is of both scientific and technological interest.
The GB sink strength, and changes thereof, is less an issue when the grain size is large. For large grains, many mobile defects cluster in the grain interior, annihilate at interior sinks such as dislocations, or undergo mutual recombination of vacancy–SIA pairs [9]. As the grain size decreases to the nanoscale, it approaches the mean free path for mutual recombination. Moreover, interior sinks such as dislocations become scarcer. Consequently, mobile defects have a much higher likelihood of reaching GBs before recombination or annihilation in the grain interior [6]. Due to the high density of GBs that are sinks of irradiation-produced defects, nanograins are expected to be more resistant to irradiation damage than their coarse grain counterparts, and experimental evidence does support this expectation [10], [11], [12], [13], [14]. For example, under the same irradiation dose, the defect accumulation at 300 K in nanocrystalline gold is much lower than that in coarse grain gold [11]; and the tolerance against irradiation-induced amorphization of nanocrystalline Gd2(Ti0.65Zr0.35)2O7 decreases with increasing grain size [14]. However, it is also possible that the GBs could saturate as sinks of mobile defects for nanograins, after prolonged periods of defect absorption.
In this paper, we use a combination of rate theory and MD simulations to examine the interaction of mobile defects with GBs, and in particular whether the ability of a GB to act as a sink saturates. As a start, we focus on point defects, i.e., single vacancies and SIAs. We measure the defect accumulation in BCC Mo near high-energy GBs using MD simulations, and use rate theory to reproduce the MD results by assuming constant sink strength of the GBs. The agreement between MD and rate theory is indirect evidence that the GB does not saturate, at least at the MD scale of time and defect production rate. The conclusion is then extended to realistic production rates using rate theory analysis. Further, both the atomic density and the width of the GBs remain unchanged during irradiation simulations, directly showing that the GB structure does not evolve during prolonged absorption of point defects. In the following, we will describe the MD simulations in Section 2, the rate theory in Section 3, and the results in Section 4. Finally we conclude in Section 5.
Section snippets
Molecular dynamics simulations
Due to the relevance of BCC metals in the irradiation environments of nuclear reactors, we take BCC Mo as a prototype material. The interatomic interaction is described by the Finnis–Sinclair type [15] potential developed by Ackland et al. (the Ackland potential) [16]. This potential well captures the correlation between the GB energy and GB structure in BCC Mo [8]. Further, the calculated formation energies of vacancy (2.54 eV [17]) and SIA (6.96 eV [16]) are in good agreement with the ab initio
Rate theory formulation
To allow further analysis of our MD simulations, in this section we present the rate theory that will be used to describe the defect evolution. With point defects created randomly, it is meaningful to consider the average concentration of vacancies (Cv) and SIAs (Ci); here the concentrations are those in excess of thermal equilibrium. The rate of change in defect concentrations with time t is described by Eq. (1) [28], [29]:Here P0 is the rate
Results and discussion
In this section we show that GBs do not saturate, using a combination of MD simulations and rate theory. First, by assuming constant sink strength k2 the rate theory reproduces MD simulation results, for various grain sizes, GB types, and production rates. This reproduction provides indirect evidence that high-energy GBs do not saturate as sinks of irradiation produced mobile defects. Further, the pair distribution function and planar structure factors show that the atomic density and the width
Conclusions
Using a combination of MD simulations and rate theory, we have shown that high-energy GBs in BCC Mo do not saturate under irradiation damage, at least for nanograins. In particular, the rate theory with constant sink assumption at GBs predicts the same defect behaviors as MD simulation, indicating that GBs indeed do not saturate as sinks to irradiation produced defects. Further, the GB structure analyses show that both the atomic density and the width of the GBs remain unchanged after prolonged
Acknowledgement
The authors gratefully acknowledge the DOE/BES support through the Computational Materials Science Network (CMSN) project on “Multi-scale simulation of thermo-mechanical processes in irradiated fission-reactor materials”. Huang acknowledges supports from the National Science Foundation support (DMR-0906349) and the Defense Threat Reduction Agency (HDTRA1-09-1-0027). Millett and Zhang thank the support of the LDRD project at INL on “Irradiation-induced evolution of defect and microstructure in
References (37)
- et al.
J. Nucl. Mater.
(1994) - et al.
Nucl. Instr. Meth. Phys. Res. B
(1997) - et al.
J. Nucl. Mater.
(2001) - et al.
J. Nucl. Mater.
(2004) - et al.
Phys. Lett. A
(1979) - et al.
J. Nucl. Mater.
(1993) - et al.
J. Nucl. Mater.
(2010) - et al.
Nucl. Instr. Meth. Phys. Res.
(1983) - et al.
J. Nucl. Mater.
(1978) - et al.
Science
(2010)