Corrosion Protection of Metallic Waste Packages Using Thermal Sprayed Ceramic Coatings

Ceramic coated carbon steel coupons were corrosion tested in water with dissolved salts to simulate exposure to evaporation concentrated groundwater in an underground nuclear repository. Metallography revealed no corrosion at the ceramic metal interface of dense coatings, even though electrical measurements demonstrated that the coatings were slightly porous. Experimental results and a model to predict corrosion rates influenced by a porous ceramic coating and coating lifetimes are presented. INTRODUCTION Certain refractory ceramic oxides have desirable properties for the construction of containers for long-term use in nuclear waste disposal applications such as the proposed Yucca Mountain repository. Ceramics are thermodynamically stable against further oxidation and far less prone to environmental corrosion than metals under realistic repository conditions. The aqueous corrosion rates of oxides such as spine1 (MgAl,O,), alumina (Al,O,) and titania (TiO,), fall in the range of a few millimeters per million years. Oxide ceramics are unlikely to be subject to microbiologically influenced corrosion (MIC), which may attack most, if not all, of the available engineering metals over time. Ceramics have a reputation for poor mechanical performance, and sufficiently large, impermeable vessels are not easily fabricated in large numbers. Current waste package designs are based on multiple metallic layers to provide handling capability and “defense in depth” because of varying corrosion mechanisms between layers. The most promising approach for incorporating ceramics in large waste packages is as a low porosity protective coating to a supporting metallic structure, such as the steel “corrosion allowance material” (CAM) that has been part of the primary design focus of the Yucca Mountain Project (YMP). Coatings applied by thermal-spray can be effectively seamless and offer a method for final closure of the package while maintaining low average temperatures within the waste package. Without liquid or vapor phase water, electrochemical corrosion and MIC processes are considered impossible, so an impervious ceramic coating should protect the metal vessels indefinitely. Even an imperfect coating should extend the life of the package, delaying the onset and reducing the severity of corrosion by limiting the transport of water and oxygen to the ceramic-metal interface. Logically, if the oxygen transport is impeded, the corrosion rate will decrease. There is a presumption that all thermal sprayed coatings will be porous at some level, so the model which follows is an attempt to account for increased impedance to oxygen transport due to a porous coating and predict a resulting corrosion rate. EXPERIMENT Porous (19% porous) alumina coatings were produced on cylindrical steel substrates using conventional plasma spray. Higher density alumina, titania and spine1 coatings were manufactured via high velocity oxy-fuel (IIVOF) spraying (2% porous) and detonationspraying (6% porous). Coatings up to 1.5 millimeters thick were tested. Some coatings were applied over a bond coat of a nickel-based alloy. Optical and SEM metallography of sample cross sections demonstrate the morphology of the various coating types. Image analysis was used to estimate the total fraction of porosity. Apparently circular pores might be elongated pores revealed in cross section. Radial separations (microcracks or pores) appear to be submicron in thickness and a few microns long. Rounded inclusions are particles which apparently melted but solidified before impact and sharp-edged inclusions are unmelted particles trapped in the coating. I Approximate spray directiol Figure 1: Optical and SEM micrographs of a low-density (-19% porous) plasma sprayed coating after 6 month exposure to simulated concentrated J-13 well water at 90°C. Coating thickness 1.1 mm Figure 2 : Optical and SEM micrographs showing the structure of a high density HVOF coating after 6 months exposure in simulated concentrated J-13 water at 90°C. Corrosion Testing Corrosion testing was carried out in stirred baths containing simulated 10x concentrated, J-13 well water at 90°C (total dissolved solids -1500 ppm, pH between 10.0 and 10.2). Samples straddled the water line, exposing them to water, oxygen, and deposited salts. Some were put in whole and some were slotted in two places (above and below the water line) to induce local corrosion. Samples were withdrawn from the baths at 3 month intervals for examination. Slotted regions were filled with epoxy to trap any corrosion products which might be present prior to sectioning. A B Figure 3: (A) Plasma sprayed coating (-19% porosity) and (B) Dense HVOF (-2% porosity) coatings exposed to simulated concentrated J-13 well water at 90°C for 6 months. AC Impedance Spectroscopv Simple contact conductivity (DC resistance) measurements conducted in salt water demonstrated that a conductive pathway to the substrate could be established quickly for all coatings, verifying the presence of interconnected porosity. AC impedance spectroscopy using a potentiostat was performed on several variations of thermal-sprayed samples immersed in 1000x concentrated, simulated J-13 water (-130,000 ppm dissolved solids). Since conductivity through liquid-filled channels is directly related to the ease with which ionic species (including dissolved oxygen) can pass through the channels, the electrical impedance measured at very low frequencies should correspond directly to the impedance of oxygen transport from the outside to the substrate, whereas the impedance at very high frequencies is primarily due to ionic conductivity of the electrolytes in the pores. At high frequencies, there is electrical conductivity due to charge transfer, but one can picture the ions themselves merely oscillating in place, so that there would still be reduced net transport of oxygen. At low frequencies (around 10 Hz), the electrochemical impedance was found to be increased by eight orders of magnitude for HVOF and detonation coatings, corresponding to a reduction of the corrosion rate of the substrate by eight orders of magnitude (~10~). The reduction in corrosion rate can be explained by impeded oxygen transport through the electrolyte in the pores as well as by an increase in the interfacial impedance due to blockage by the coating. Results are in general agreement with the lack of corrosion observed beneath dense coatings. The plasma sprayed coating had little effect on the measured impedance or apparent corrosion rate because its 19% porosity was highly interconnected. q q !i! BXXX Detonation (6% porous) (2% pot&) Llcl“ HVOF u n n -4 6