Intergranular penetration and embrittlement of solid nickel through bismuth vapour condensation at 700°C

Abstract

Grain boundary penetration of liquid bismuth in polycrystalline nickel is investigated at 700°C. The contact between the two metals is ensured by bismuth transport through vapour phase. The formation of a nanometre-thick Bi-rich layer on external surfaces of solid nickel is revealed by glow discharge optical spectroscopy. This layer is the consequence of Bi vapour condensation on nickel substrate at 700°C. The liquid Bi/solid Ni contact leads to grain boundary penetration of liquid bismuth in the form of a film of nanometric thickness as revealed by Auger electron spectroscopy. The presence of these intergranular films causes strong room temperature brittleness of nickel, as shown by tensile and bending tests. The comparison with results due to the direct contact between solid nickel and bulk liquid bismuth-rich alloy indicates that in both cases intergranular penetration rate and embrittlement are of the same magnitude. Based on these results, a concept of a new device for liquid metal embrittlement (LME) tests is outlined.

Introduction

Liquid metal spallation target is a complex device where neutrons are produced by highly accelerated protons [1], [2]. Neutrons are needed either in fundamental research (European spallation source (ESS) [3], spallation neutron source (SNS) [4]) or in accelerator driven system (ADS) for power generation [5] or nuclear waste incineration [6]. Spallation targets constitute an alternative, as opposed to the nuclear reactors, for neutron production with an advantage of a pulsed source. Neutrons are produced by spallation reaction i.e., collisions between pulsed high-energy protons (of the order of GeV) and heavy metal nucleus. Liquid metal spallation target (Hg, Pb or Pb–Bi) are preferred because of the higher neutron production rate, easier cooling system and increased lifetime with respect to the solid target. However, one of the critical points of such a system is the compatibility between liquid metal target and solid metal container, usually called window (i.e., the entrance for high-energy protons). This compatibility must be ensured under complex loading conditions including residual stresses, irradiation and thermal fatigue. It also requires: (i) limited liquid metal corrosion i.e., relatively homogeneous attack and dissolution of solid metal by stagnant or flowing liquid metal and (ii) resistance to the liquid metal embrittlement (LME) [7] i.e., preferential transgranular (rare) or intergranular (frequent) rapid penetration of liquid metal under stress and consecutive failure. In some model systems (Al–Ga [8], Cu–Bi [9]), this penetration can even occur without external stress and result in low temperature brittleness.

Two types of liquid target/solid container systems are considered for the design of the spallation target:

(i) liquid mercury in contact with 316L austenitic stainless steel, in SNS project [4], [10]. Liquid mercury is a good candidate as spallation target, because of relatively low-working temperatures (typically between 60°C and 150°C) and materials lifetime and neutronic performance are increased compared to water-cooled solid target [4]. 316L stainless steel is chosen as mercury container, because it exhibits good performance while simultaneously exposed to the intense flux of high-energy protons and neutrons [4] and to the flowing liquid mercury [10].

(ii) liquid lead [11], [12], [13] or lead–bismuth eutectic alloy (Pb–55.5 wt% Bi) [12], [14] in contact with ferritic steel, in the MEGAPIE project [2] and GEDEON program [15], [16]. Liquid Pb–Bi spallation target is also considered as possible neutron source in the concept of undercritical hybrid reactor [6]. The range of working temperatures (350–550°C) is higher than for mercury target, but an important advantage of Pb–Bi target is the increased neutron production rate as compared to pure lead or mercury. Nevertheless, the compatibility between these liquid metals and ferritic steels as container is the critical point and some papers [12], [17], that compared the damaging effects of liquid Pb and liquid Pb–Bi on different steels, underlined that the addition of Bi was at the origin of LME.

Our effort is aimed at the understanding of LME phenomenon on a model system: liquid bismuth/solid nickel [18], [19], [20]. The advantage of this couple is that it is possible to separate grain boundary liquid penetration, taking place at elevated temperature, from the resulting intergranular embrittlement at room temperature, which allows grain boundary analysis by Auger electron spectroscopy. Grain boundary liquid penetration was typically obtained by direct contact between solid Ni and liquid Bi-rich alloy [18], [20]. Direct contact results in the formation of relatively short intergranular films of micrometric thickness and long intergranular films of nanometric thickness [20]. Although only micrometric films are visible by scanning electron microscopy (SEM), both films result in strong intergranular brittleness [20]. These results underline the key role of nanometric films in the assessment of LME phenomena, but it requires quite complex experimental device due to the direct contact procedure.

The alternative way of ensuring the contact between solid and liquid metals is vapour transport in temperature gradient and subsequent condensation on solid substrate. This way can apply for elements with high equilibrium pressure and will be called indirect contact. It has been successfully used in several studies. In particular, bismuth vapours from Cu–1 wt% Bi were used for the controlled doping of copper bicrystals [21], mercury vapours condensation was shown to induce intergranular penetration in alpha-brass [22] and Zn, Cd and Hg were theoretically shown to have the equilibrium vapour pressure high enough to ensure their transport between the specimen surface and the crack tip in the LME phenomena [23].

The aim of this paper is to point out that the indirect contact between solid nickel and liquid bismuth through bismuth vapour condensation results in grain boundary penetration and consecutive room temperature embrittlement of the same magnitude that in the case of the direct contact. It will be also suggested that mechanical testing, as opposed to standard SEM observations, is necessary to reveal grain boundary penetration in the form of nanometre-thick films and that bismuth vapour condensation can be used in a simple device to perform LME experiments.