Characterisation of interfacial segregation to Cu-enriched precipitates in two thermally aged reactor pressure vessel steel welds

Highlights

• Characterisation of interfacial segregation of Ni, Mn and Si to Cu-enriched clusters.

• Analysis method gives information on interface composition and widths of large numbers of clusters.

• Reduction in interface energy due to segregation of Ni, Mn and Si is calculated.

Abstract

To understand the contribution of long term thermal ageing to Reactor Pressure Vessel (RPV) embrittlement two high Cu steel welds with different Ni contents were thermally aged for times up to 100,000 h at 330 °C and 365 °C. Microstructural characterisation using Atom Probe Tomography was performed. Thermal ageing produced a high number density of nano-scale Cu-enriched precipitates. The precipitate–matrix interfaces were enriched in Ni, Mn and Si. The characterisation of these interfaces using a double cluster search approach is the subject of this work.

The interface region around thermally-induced precipitates was found to be wider in steels with higher bulk Ni contents and where precipitates had larger core radii. The effect of ageing temperature on interface width was small when comparing precipitates of equal core radius. The narrower interface width in the lower Ni steels is reflected in the composition of the interface, which has a lower Ni content than in the higher Ni material. The reduction in interfacial energy due to the segregation of Ni, Mn and Si has been calculated and shows enhanced reductions in interfacial energy with increasing precipitate size, but no obvious effect of temperature.

Introduction

Cu precipitation in Reactor Pressure Vessel (RPV) and High Strength Low Alloy (HSLA) steels has been studied for many years [1], [2], [3], [4], [5]. It is well known that after thermal ageing or irradiation a high number of Cu-enriched clusters (CECs) form in steels containing >0.1 at% Cu. In addition, alloying elements such as Ni, Mn and Si can also influence their formation and evolution.

Segregation of Ni, Mn and Si to the interfaces of CECs with the surrounding matrix has been observed [4], [6], [7], [8]. In the early stages of CEC nucleation interfacial segregation of Ni, Mn and potentially Si could play an important role [9], [10]. The consensus of opinion from the literature from both modelling and experimental work is that segregation of Ni, Mn and Si reduces the interfacial energy of the clusters with the surrounding matrix [9], [10], [11], [12]. Sn segregation to CEC interfaces has also been observed by Sha et al. [13].

Whilst there have been many studies which report interfacial segregation at precipitate–matrix interfaces there are relatively few which report quantitative data on the degree of segregation. Typically composition profiles or proxigrams [14] are used to identify segregation. Whilst these provide evidence of segregation, and of the maximum concentrations of segregated solutes it can be difficult to obtain absolute values of the composition and width of the interface over large numbers of particles.

In early work, Worrall et al. [15] and Buswell et al. [6] both reported segregation of Ni to the interfaces of CECs in model alloys (Fe–Cu–Ni) using Atom Probe Field Ion Microscopy (APFIM). Later, Osamura et al. [16] used a combination of Small Angle Neutron Scattering (SANS) and APFIM to study a thermally aged Fe–Cu–Ni–Mn alloy. The latter authors found that, for precipitate sizes above 2.1 nm SANS measurements of the Guinier radii obtained from the nuclear and magnetic scattering ratio differed between a binary Fe–Cu and an Fe–Cu–Mn–Ni alloy. They associated this with interfacial segregation of Ni and Mn to precipitates which have transformed from bcc to fcc. Their atom probe results also suggest some interfacial segregation.

More recently, Kolli et al. [12], [17], [18], [19] have studied interfacial segregation in HSLA steels using the proxigram methodology. With this method they report interface compositions in [12], [19], but it is not clear how these values were obtained from the proxigrams.

In later papers by the same group on Ni–Al alloys (Plotnikov et al. [20]) and Al–Li–Sc–Yb (Monachon et al. [21]) used composition profiles to calculate interface widths. In Plotnikov et al., they fit a spline curve to the profile and a take the width as the distance between the 10th and 90th percentile concentrations. In Monachon et al. they use composition profiles and it appears that they take the interface as the distance between the points of inflection in the Sc and Yb profiles. These papers show it is possible to obtain quantitative information from concentration/proxigram profiles but it is reliant on producing composition profiles for each precipitate which can be time consuming.

Hyde et al. [22] showed that clusters in irradiated and thermally aged materials exhibited a core shell structure. By comparing the radius of gyration of different elements (Cu, Ni and Mn) with that of the whole cluster the authors could distinguish between clusters with a core–shell structure, those where Ni, Mn and Cu were homogenously distributed and those which had a Cu-rich region and Ni and Mn rich regions. Whilst this method can demonstrate there is/is not a core shell structure, it cannot provide information on the interface width, nor the composition.

In order to study the formation and evolution of Cu clusters in RPV steels (and welds) in the absence of irradiation two high Cu steel welds, one with high and the other with low Ni content were thermally aged for 90,000–100,000 h at temperatures of 330 °C and 365 °C. Atom Probe Tomography (APT) was used to characterise the CECs in the resulting microstructures. Furthermore, an approach to characterising the interface width and composition is presented which can be easily applied to all clusters in a dataset, irrespective of their size, without the requirement to generate individual composition profiles.