Material Science

Improving Modeling of Corrosion Products in the Primary Circuit of PWRs through the Experimental Determination of CRUD Characteristics

Abstract 

High-temperature water flowing through the coolant circuits in pressurized water reactors (PWRs) creates a harsh environment, deteriorating these materials. This corrosion results in the formation of oxide layer(s) primarily composed of nickel ferrites, iron oxides, nickel oxides, and other nickel-iron-chrome spinel oxides. Being exposed to fluid with high flow rates and pressure, these oxide layers can be eroded, resulting in particulate fouling due to the release of corrosion products (Chalk River Unidentified Deposits, CRUD) into the coolant circuit. When CRUD subsequently deposits on fuel surfaces, it negatively affects the fuel performance (heat transfer, and fuel failure); in addition, the particles can undergo neutron activation, which is problematic when the particles detach and travel to out-of-core regions, contributing to worker’s radiological exposure.

Several factors affect CRUD deposition in these environments. As deposition of these particles depends in part on the surface charge of the particles and the nearby surfaces, tuning coolant chemistry and/or the composition of the primary circuit materials has been one of the empirical levers for CRUD mitigation. While the benefits of modified water chemistries such as Zn addition and using alloys with low Ni composition are already seen in some operating PWRs, the underlying mechanisms leading to these benefits are not fully understood. This work aims to build a physics-based CRUD deposition model. The model will predict deposition in environments and geometries representative of coolant circuits of PWRs, and the main goal is to understand the role of water chemistry parameters, the effect of Ni composition, Zn addition, and reducing agents on the CRUD deposition.

To ensure the relevance of this model, the surface properties of CRUD particles have been measured and incorporated into simulations to study the impact of the parameters that are affected by water chemistry. To this end, a library of particles, specifically the trevorite-franklinite-magnetite ternary system (nickel-zinc-iron spinel oxides), has been synthesized, covering a range of compositions and water chemistry (Zn addition, and reducing environment). The reaction products were screened for phase purity using X-ray diffraction, and their surface properties as a function of size and composition were evaluated by electrophoretic light scattering. Similarly, stainless steel coupons representing the interior surfaces of coolant circuit materials were exposed to hydrothermal conditions at 200°C with varying concentrations of Zn, Ni and reducing agents in the water to understand the growth and possible detachment of nanoparticles from these surfaces.

Surface characterizations were performed on the single-phase ferrite nanoparticles to obtain Isoelectric Point (IEP) and zeta potential. Increasing Nickel content in the particles shifted their surfaces toward a more basic character, resulting in a positive shift in the IEP. Higher concentrations of Zinc shifted the surface towards more acidic character, effectively lowering the IEP. These IEP values of the primary circuit materials and CRUD particles, along with the operational pH of the coolant, gave a qualitative understanding of the corrosion product deposition and the effect of Zn and Ni compositions.

To simulate the transport, deposition and re-entrainment process of CRUD particles, a  COMSOLTM simulation was designed for a simple geometry representing piping with fluid properties consistent with the PWR environment. In bulk, the transport of these particles was predominately influenced by turbulent diffusion, implemented using a k-ω turbulence model. Near the pipe wall, the particle trajectory was largely governed by electric double layer forces and Van der Waals interactions. To account for the differing length and time scales of the physical processes involved, sequential multiscale modeling was implemented. A macro-scale model handled the fluid flow, traced the trajectory of the particle inside the pipe, and obtained parameters for the constituent fine-scale models as needed. 

Consistent with expectations from DLVO (Derjaguin, Landau, Verwey, and Overbeek) theory, these results showed a decrease in the deposition with increasing Stern potential. The overall probability of deposition was obtained by combining the results from macro-scale simulations and the fine scale simulations. The smaller particles (i.e., below 500 nm-600 nm) followed a nearly digital response to the value of the Stern potential (i.e., they stuck if there was no Stern potential, but the probability dropped rapidly to zero when potential was above 25 mV), while the larger particles had relatively slow decay with increasing Stern potential. 

The formulated CRUD deposition model serves as a mechanistic model to understand the effect of surface charge on deposition. While the current work dealt with the sensitivity of Ni and Zn content on deposition, the methodological framework established in this work can serve as a template to study the effect of any future modifications to the coolant chemistry on CRUD deposition.