Stainless steel corrosion by molten nitrates : analysis and lessons learned.

Abstract

A secondary containment vessel, made of stainless 316, failed due to severe nitrate salt corrosion. Corrosion was in the form of pitting was observed during high temperature, chemical stability experiments. Optical microscopy, scanning electron microscopy and energy dispersive spectroscopy were all used to diagnose the cause of the failure. Failure was caused by potassium oxide that crept into the gap between the primary vessel (alumina) and the stainless steel vessel. Molten nitrate solar salt (89% KNO{sub 3}, 11% NaNO{sub 3} by weight) was used during chemical stability experiments, with an oxygen cover gas, at a salt temperature of 350-700 C. Nitrate salt was primarily contained in an alumina vessel; however salt crept into the gap between the alumina and 316 stainless steel. Corrosion occurred over a period of approximately 2000 hours, with the end result of full wall penetration through the stainless steel vessel; see Figures 1 and 2 for images of the corrosion damage to the vessel. Wall thickness was 0.0625 inches, which, based on previous data, should have been adequate to avoid corrosion-induced failure while in direct contact with salt temperature at 677 C (0.081-inch/year). Salt temperatures exceeding 650 C lasted for approximately 14 days. However, previous corrosion data was performed with air as the cover gas. High temperature combined with an oxygen cover gas obviously drove corrosion rates to a much higher value. Corrosion resulted in the form of uniform pitting. Based on SEM and EDS data, pits contained primarily potassium oxide and potassium chromate, reinforcing the link between oxides and severe corrosion. In addition to the pitting corrosion, a large blister formed on the side wall, which was mainly composed of potassium, chromium and oxygen. All data indicated that corrosion initiated internally and moved outward. There was no evidence of intergranular corrosion nor were there any indication of fast pathways along grain boundaries. Much of the pitting occurred near welds; however this was the hottest region in the chamber. Pitting was observed up to two inches above the weld, indicating independence from weld effects.