Experimental Multiscale Methodology for Predicting Material Fouling Resistance

Fouling, or the unwanted buildup of corrosion products on a surface, complements corrosion as a surface-based material degradation mechanism. Fouling is particularly detrimental in energy systems, as the surfaces which transfer the most enthalpy are typically the most susceptible. In essence, the presence of fouling deposits, often porous and/or thermally insulating, impedes the efficient flow of heat through a system. The high flow rates and large temperature differences in the hot and cold sections of a heat transfer system or coolant loop provide ideal environments to induce flow-assisted corrosion, and subsequent transport of corrosion products from areas producing soluble & particulate species, to the areas in which deposition is most favorable. In most energy systems, corrosion is most severe in the hottest sections of a coolant loop, since corrosion rates often accelerate exponentially with temperature. In the case of systems undergoing boiling, however, the formation and cavitation of bubbles on a material surface may induce species precipitation via microlayer dryout underneath bubbles. Here the growth of fouling deposits can be particularly rapid, and the resultant effects unusually devastating. Economically, heat exchanger fouling costs many industrial countries roughly 0.25% of its gross domestic product in annual lost productivity due to fluid treatment, cleaning, component replacement, and loss of production ¹, equivalent to approximately $42 billion in the United States. Figures for worldwide industries such as boilers, marine shipping, and oil/gas production are far higher.

In nuclear light water reactors (LWRs), foulants composed of soluble and particulate nickel- and iron-bearing species deposit on fuel cladding rods to form CRUD, or Chalk River Unidentified Deposits. CRUD takes the form of highly porous deposits roughly tens to one hundred microns thick, with porosities on the order of 50%. In pressurized water reactors (PWRs), CRUD is largely formed from deposits of nickel ferrites (Ni_xFe_3−xO₄) in addition to small amounts of NiO, Ni metal, and some iron and nickel chromites (FeCr₂O₄, NiCr₂O₄, and Ni_xFe_1−xCr₂O₄) ^1,19. Images of CRUD at different length scales are shown in Figure 1. ² It takes the form of brownish-grey deposits on fuel rods, which when viewed in cross section exhibits multiple, self-similar length scales of porosity. The larger pores, known as "boiling chimneys," provide a path for vaporized water to escape the CRUD, while the smaller pores throughout serve to draw water within. This porous network of CRUD normally draws water into its pores via capillary action, but dryout may occur for very high values of heat flux. The smallest constituents of CRUD are particles, some crystalline, of the iron-nickel-chromium oxides which form the nanoporous framework. Zinc oxides and zirconium oxide are also sometimes present in CRUD.

CRUD can cause multiple negative side effects in PWRs, including CRUD-induced localized corrosion (CILC), which can lead to sudden fuel failure, elevated cladding temperatures due to CRUD's additional thermal resistance, and downward axial power shifts in the reactor ⁹. Previous studies ³ have shown that Zircaloys (of which PWR fuel cladding is made) corrode far more quickly at elevated temperatures. Recent multiphysics modeling efforts ⁴ have shown that the presence of CRUD tens of microns thick can elevate cladding temperatures by 15°C, which corresponds to roughly a doubling of the corrosion rate. The last problem of axial power shifting is due to the accumulation of boron, present to control the neutron population in the reactor, in the pores of the CRUD. Recent massively parallel, multiphysics simulations of CRUD's effects ⁵ demonstrated this effect by simulating CRUD formation and boron accumulation along 3,000 fuel rods in a reactor, demonstrating a downward shift in reactor power. This behavior was already known from nuclear plant measurements ⁶.

Previous attempts have been made to develop materials resistant to CRUD formation. Solutions have included manipulating the growth of CRUD to improve heat transfer ¹⁰, manipulating water chemistry to minimize CRUD growth and corrosion ¹¹, pretreating zircaloy ¹², and using magnetic separation as a filtration system ²¹. One recent patent ⁷ proposed to electropolish fuel cladding surfaces, to remove nucleation sites for bubbles which can induce CRUD. However, so far no method has resulted in completely preventing the growth of CRUD. We believe that uncovering the mechanism of the initiation event of CRUD formation, the adhesion of the first particles to a substrate, is the key to preventing its formation altogether. Therefore, precise measurement of surface adhesion forces, along with proof-of-principle experiments at higher length scales, is required to predict and confirm the fouling resistance of candidate materials.

This paper presents an experimental, multiscale methodology to discover materials that can be applied to nuclear fuel cladding and prevent the growth of CRUD. In order to achieve this goal, a variety of materials are being tested, shown in Table 1, representing different classes of potential anti-fouling and model materials. First, sputtered or grown specimens of each material are exposed an environment similar to that of a nuclear reactor, specifically a PWR (Pressurized Water Reactor), and "cooked," or indirectly heated to induce pool boiling with excess amount of CRUD forming particulates (NiO and Fe₃O₄) to simulate the boiling process on the surfaces of nuclear fuel rods. The cooked samples containing simulated CRUD are then imaged top-down using a high-resolution scanning electron microscope (SEM) and in cross section using a focused ion beam (FIB), allowing for visualization of any CRUD-substrate bond that may be present.

The experimental pool boiling setup that has been developed possesses certain unique benefits that make experiments repeatable, effective, and more representative of in-reactor conditions. A vertically oriented sample holder, as shown in Figure 2 prevents nickel- and iron-based particulates that fall out of solution from depositing on the surface of the sample by gravity. Strongly sonicating the CRUD-bearing solution prior to the experiment allows for longer test times and less precipitation of particulates. In addition, for each of the four samples being tested, the heat flux and temperature is recorded throughout the experiment. The pH of the simulated PWR water is also recorded throughout the experiment.

In order to quantify the bond between the CRUD deposits and the surface, we also perform AFM force spectroscopy (AFM-FS) measurements in both air and simulated PWR water on clean surfaces of the same materials. First, AFM cantilevers are functionalized with particles made of one micron spheres of compounds commonly found in CRUD (NiO or Fe₃O₄). In the MFP3D-BIO, the AFM used in this procedure, the cantilever is modeled so that the force applied to the cantilever tip is proportional to its deflection when in contact with the surface. A laser reflects off the cantilever to a photodetector to measure its deflection. In the procedure below, two constants, the spring constant and the deflection InvOLS (inverse optical lever sensitivity) must be calibrated in order to convert the deflection read by the photodetector to nanoNewtons.

In order to record adhesion of the tip to the surface, the cantilever tip is pressed against the surface of the sample with a force of 25nN. In the experiments recorded below, the tip will then remain at the surface for a period of either 0, 5, 10, 30, or 60 seconds. As the cantilever is being retracted, the tip will often adhere to the sample surface, causing the cantilever to deflect until the force of the cantilever deflection is greater than the adhesion between the CRUD particle and the surface under investigation. Then the cantilever snaps back to zero deflection. It is this maximum deflection, and its corresponding adhesion force, that the AFM can quickly and repetitively record ¹⁸. The area under the force-distance curve produced by AFM-FS is a direct measure of the work of separation between CRUD and a substrate, and therefore quantitatively estimates the adhesion of that CRUD particle to each surface. Using the same AFM cantilever on multiple samples eliminates variability in particulate size and composition.

This unique combination of simulated CRUD growth and AFM measurements provides multiple, coupled datasets that allow us to understand how and why CRUD may develop on certain material surfaces. Rather than undergo time-consuming pool boiling experiments, which can take up to a day to complete, a large number of materials can be pre-screened using AFM-FS to estimate their fouling resistance. We also show initial results confirming the efficacy of this procedure, whereby materials which exhibit little to no CRUD adhesion in pool boiling also show very low values of CRUD-substrate adhesion in AFM-FS. Conversely, materials which grow large amounts of adherent CRUD in pool boiling also adhere strongly to CRUD particles in the AFM.