First-principles investigation on dissolution and diffusion of oxygen in tungsten Abdullah Alkhamees, Yue-Lin Liu, Hong-Bo Zhou, Shuo Jin, Ying Zhang, Guang-Hong Lu * Department of Physics, Beijing University of Aeronautics and Astronautics, Beijing 100191, China article info Article history: Received 16 April 2009 Accepted 25 July 2009 PACS: 21.10.Dr 21.10.Ft 71.15.Mb 81.05.Bx 61.72.y abstract Using a first-principles method, we have investigated dissolution and diffusion properties of oxygen (O) in tungsten (W). Single O atom prefers to occupy the tetrahedral interstitial site (TIS). Two interstitial O atoms are attractive and tend to be paired up at two neighboring TIS with a distance of 0.228 nm and a large binding energy of 1.60 eV, which indicates a strong tendency of O clustering in W. O is preferred to diffuse between the most nearest neighboring TIS with a diffusion barrier of 0.17 eV. By the estimation of pre-exponential factor according to an empirical theory, the diffusion coefficient as a function of temper- ature has been determined, which is 1.50  10 9 m 2 /s at a typical temperature of 500 K. The results pro- vide a good reference to understand the behavior of O in intrinsic W. Ó 2009 Elsevier B.V. All rights reserved. 1. Introduction It is an unquestionable fact that the shortage and the increase of demand of energy in the 21st century have evoked international cooperation to consider other ways to meet the mankind needs. For example, fusion energy, a kind of clean, infinite energy for fu- ture generation, has been developed for its construction via ITER Project [1]. Tungsten (W) and W alloys are considered among the most promising plasma facing materials (PFMs) because of their low sputtering erosion and good thermal properties such as high thermal conductivity and high melting temperature. The light ele- ments such as oxygen (O) are regarded as the common impurities in metals. The presence of even very little amount of O impurity, e.g., as low as 30 atomic parts per million (appm) [2–4] in W, can change the microstructure significantly. Further, O can be eas- ily trapped by some extended defects such as vacancies, disloca- tions and grain boundaries [5], leading to the degradation of the mechanical properties. Moreover, recent experimental studies re- ported that O can be regarded as the trapping sites, which makes large amount of hydrogen (H) isotope ions aggregate and further form H blisters [6–8]. However, fewer studies have been devoted to understanding the interaction of between O and W, which has a direct impact on the design and operation of the PFM. The interaction of O with metal as well as metal–alloys is of great scientific and technological interest. In the previous experi- ments, Hatakeyama et al. studied the diffusion behavior of O in vanadium-alloy [9]. They found that diffusion of O can enhance the precipitations of vanadium-alloy. Filius and van Veen reported the effect of impurity O on the helium (He) retention in W and molybdenum [10]. They found that O can decrease the binding en- ergy between He and vacancy-type cluster. As to theoretical pre- dictions, the electronic structure and bonding properties of O with metal–alloys such as NiAl [11–14] have been investigated using first-principles method since O is considered as one possible reason for the room temperature brittleness of the NiAl. Despite many years of research of O-metal interaction, many fundamental aspects underlying the O-metal interaction are still not fully under- stood, especially for the behaviors of O in W. Moreover, Under- standing O–W interaction will be helpful to understand the effect of O on the structure and diffusion properties of other metals. In order to understand the physical mechanism underlying the interaction of O with intrinsic W, in this paper, we investigate the structure, energetics, and diffusion properties of O in W using first- principles calculations. This work serves as the first step to inves- tigate the effect of O on the blistering behaviors of H in W. 2. Computational methods We employ a total-energy method based on density functional theory [15,16] with generalized gradient approximation developed by Perdew and Wang [17] and the projector-augmented-wave po- tential. We use a kinetic energy cutoff of 350 eV for all calculations. The uniform grids of k-points are sampled by 5  5  5 for 54-atom supercell and 3  3  3 for 128-atom supercell. The calculations have been carried out by VASP [18,19]. The calculated equilibrium lattice constant is 0.317 nm for a bcc W, in good agreement with the corresponding experimental value of 0.316 nm. The energy 0022-3115/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jnucmat.2009.07.012 * Corresponding author. Tel./fax: +86 10 82339917. E-mail address: lgh@buaa.edu.cn (G.-H. Lu). Journal of Nuclear Materials 393 (2009) 508–512 Contents lists available at ScienceDirect Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat