Deposit Induced Corrosion in Biofuel Combustion
Abstract: Deposit induced corrosion is one of the main mechanisms for attack on heat exchangers in biofuel boilers. This severe corrosion decreases the obtainable steam temperature in boilers, thereby the efficiency in production of electricity is reduced. Deposit formation from the gas on superheater tubes are is generally referred to as fouling. Fouling decreases the heat transfer from the gas to the steam. Alkali and chlorine rich biofuels, and especially waste, will form deposit rich in alkali chlorides and sulphates that are corrosive. High sulphur contents in the fuel and gas reduces the corrosion attack. Waste can also be rich in heavy elements such as lead and zinc that can form low-melting-point eutectic chloride melts. Alloys rich in nickel, chromium and molybdenum are used to improve the corrosion resistance in biofuel boilers.
In order to investigate candidate alloys for waste boilers a field test was performed including two test panels in the waterwall at 300 and 338 °C as well as a superheater test coil at 460-540 °C. The alloys ranged from carbon steel to nickel based Sanicro 65. The deposits and corrosion products were characterised by X-ray microanalysis (EDX), Scanning Electron Microscopy (SEM), Auger Electron Spectroscopy (AES) and X-ray diffraction (XRD). Secondary Ion Mass Spectrometry (SIMS) and X-ray Photoelectron Spectroscopy (XPS) were done on selected samples. Both in the waterwall and on the superheater thicker deposits formed on the ferritic steels. In the waterwall the addition of chromium and nickel only decreased the corrosion rate at the higher temperature. Protective Cr-Ni-Fe oxide separated the chlorides from the metal on Sanicro 28 and Sandvik 625. In the superheater the corrosion rates were high, up to 34 mm/year for T22. Sanicro 28 and the more Cr and Ni rich alloys performed the best. The austenitic alloys were preferentially attacked on the side of the deposit crest and Sanicro 65 directly under the deposit crest. Alloy 310 suffered severe pitting corrosion in a line at the side of the deposit crest. The scales in the superheater were porous and cracked. Chlorides were observed in the metal/scale interface. Sanicro 28 formed chromium oxide that separated the chlorides from the metal in the interface. Grain boundary attack and local attack under a thick porous chromium oxide was observed on Sanicro 65.
Laboratory tests were done with coupons of the austenitic steels 310 and Sanicro 28 imbedded in KCl-tablets at 500 °C in 5%O2-10%H2O-N2 gas. The aim was to mimic the deposit and its morphology on superheater tubes and to investigate how this affects the corrosion. Consistent with field behaviour, it was shown that transport through the KCl-deposit controlled the corrosion rate. The corrosion rate decreased with increasing thickness after reaching a maximum at 1 mm. The rim of the coupons was preferentially attacked due to microcracks through the tablet. Scale growth was investigated after 24, 168 and 672h. The corrosion products after 24h consisted of FeCr oxide and potassium chromate. The chromates grew laterally, consuming the chromium in the surrounding oxide. After 168h an oxide was formed under the chromate and chlorine was enriched in the scale/metal interface. Finally after 672h the oxides were 5 μm thick and few chromates were seen. Similar to fouling, crystals of KCl formed near thicker iron oxide after 168h in these tests.
This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.