Article pubs.acs.org/EF
Mitigation of Fireside Corrosion of Stainless Steel in Power Plants: A Laboratory Study of the Influences of SO2 and KCl on Initial Stages of Corrosion Sofia Karlsson,* Torbjörn Jonsson, Josefin Hall, Jan-Erik Svensson, and Jesper Liske Energy and Materials, Department of Chemical and Biological Engineering, Chalmers University of Technology S-412 96, Göteborg, Sweden ABSTRACT: The effect of SO2(g) on the initial oxidation of the stainless-steel 304L, sprayed with 0.1 mg/cm2 KCl and exposed in 5% O2 and 40% H2O at 600 °C, was investigated. In the absence of SO2(g), KCl accelerates the corrosion attack by the formation of K2CrO4. The reaction with KCl depletes the oxide in chromium and converts it into an iron-rich, poorly protective oxide. When SO2(g) was introduced to the gas flow, KCl rapidly transformed into K2SO4. In contrast to KCl, K2SO4 does not form K2CrO4. Hence, it does not accelerate the corrosion rate. Although the conversion of KCl to K2SO4 is fast, the corrosion rate of KCl samples exposed in the presence of SO2(g) is higher than samples exposed in the presence of K2SO4. It is therefore suggested that small amounts of unreacted KCl react initially with the protective oxide, forming K2CrO4, which depletes the oxide in chromium. However, because of the presence of SO2(g), K2CrO4 immediately reacts with SO2(g) to form K2SO4. This study shows that the initial stages of the corrosion attack are of great importance. The initial complex interactions between the flue gas, deposit, and oxide scale affect the future corrosion resistance of the steel.
1. INTRODUCTION Combustion of waste and biofuels usually results in a flue gas with significant amounts of alkali chlorides and hydrogen chloride (HCl), while the sulfur dioxide (SO2) content is relatively low.1,2 This flue gas composition favors the formation of alkali-chloride-rich deposits on the heat-exchanger parts of the boiler, e.g., the superheaters. It is well-known that alkali chlorides accelerate the corrosion of the high-temperature alloys used for superheaters. The alkali part of the alkali-rich deposits has been reported to play an important role in alkali-chloride-induced corrosion of stainless steels.3−8 The corrosiveness of alkali chloride was explained by the formation of alkali chromates in a reaction with the protective scale. The reaction depletes the oxide in chromium and destroys its protective properties. As a result, a fast growing, iron-rich, and poorly protective oxide forms. Other authors have designated chlorine (Cl2) as the source of the aggressive behavior of the salt in the “active oxidation” mechanism,9,10 where Cl2(g) acts as a catalyst for corrosion. The mechanism supposes that Cl2 forms in a reaction between oxygen (O2) and alkali chloride at the scale surfaces. Gaseous Cl2 then penetrates the scale and generates volatile transition metal chlorides at the scale−metal interface. The transition metal chlorides diffuse outward toward the scale−gas interface, where they are oxidized by O2, precipitating metal oxide. The reaction releases Cl2(g), which enters the process again. In contrast to this reaction scheme, it has recently been suggested that the transport of chlorine through the scale is supported by ions (Cl−).11,12 The issue of alkali-chloride-induced corrosion has been previously studied, and different ways to mitigate its effect have been addressed in recent years. A successful way to minimize the corrosiveness of alkali chlorides is to sulfate them to corresponding alkali sulfates. This can be performed using © 2014 American Chemical Society
elemental sulfur, sulfur-rich additives, or co-combustion with a suitable fuel, e.g., sludge and coal.2,13−20 The presence of sulfur in the fuel changes the flue gas chemistry; alkali chlorides react with SO2/SO3 forming alkali sulfates, and chlorine is released as HCl. There are several papers that investigate the gas-phase reactions between alkali chlorides and sulfur-containing species.21,22 In addition to these publications, an investigation of the sulfation of solid alkali chlorides (i.e., deposits) and how they affect the corrosion has recently been published.23 In that paper, the effect of adding SO2(g) to the exposure with potassium chloride (KCl)-treated 304L samples (Fe18Cr10Ni), was studied. The results show that the alkali-chloride-induced corrosion was mitigated by the conversion of KCl to potassium sulfate (K2SO4). The reaction kinetics were very fast, and in contrast to KCl, K2SO4 does not deplete the protective oxide in chromium by forming potassium chromate (K2CrO4).5 However, despite the fast conversion of KCl to K2SO4, the corrosion attack of KCl-treated 304L steel samples exposed in the presence of SO2 was much more severe than that in corresponding samples exposed directly toward K2SO4. However, there was no evidence of chromate formation. The present paper investigates the initial stages (from 15 min to 24 h) of corrosion of 304L exposed in the presence of KCl and SO2(g) at 600 °C. The aim is to understand the complex interplay and kinetics of the reactions between the flue gas, deposit, oxide scale, and steel during the initial stages of exposure. The accelerated corrosion of 304L in this complex environment is compared to corresponding salt-free exposures and samples exposed directly to K2SO4. Hence, another aim is to present a mechanism of the initial corrosion attack and show Received: October 28, 2013 Revised: April 8, 2014 Published: April 8, 2014 3102
dx.doi.org/10.1021/ef402127h | Energy Fuels 2014, 28, 3102−3109
Energy & Fuels
Article
added directly into the hot zone of the furnace, through a separate gas inlet. The flow rate of SO2 was controlled by a digital mass flow regulator. The atmosphere was equilibrated for at least 2 h before the exposure in the hot furnace. During the insertion of the samples to the equilibrated furnace, the water vapor content was temporarily decreased to avoid hydration of the salt. Samples were mounted three at a time using an alumina sample holder and positioned parallel to the gas flow direction in the middle of the furnace. The samples were exposed isothermally, and the temperature was kept at 600 °C. Exposure times ranged from 15 min to 168 h, and the samples were introduced into a hot furnace. The mass change of the samples was measured prior to and after exposure using a six decimal Satorius balance. After exposure, the samples were stored in desiccators with P2O5. Reference exposures without salt were also carried out, in both dry and wet O2.
how similar scenarios can affect the corrosion rate of the heatexchanger parts in a boiler.
2. EXPERIMENTAL PROCEDURES 2.1. Sample Preparation. Chemical composition of the investigated material, austenitic stainless-steel 304L, is listed in Table 1. Samples were cut into coupons (15 × 15 × 2 mm3) and ground with
Table 1. Chemical Composition of 304L element
Fe
Cr
Ni
Mn
Si
Mo
(wt %) (atomic %)
68.0 67.0
18.5 19.5
10.2 9.5
1.4 1.4
0.55 1.08
0.49 0.28
320 grit SiC and deionized water. The samples were then polished with 9, 3, and 1 μm diamond solution, including lubricant liquid, on a polishing cloth until a mirror-like surface was obtained. Prior to exposure, the samples were cleaned in acetone and ethanol using ultrasonic agitation. After the polishing procedure, the samples were sprayed with a saturated solution of the desired salt (KCl or K2SO4) in a water−ethanol mixture. The amount of salt applied was controlled, after drying with air, using a six decimal Satorius balance. The amount of salt corresponded to a potassium equivalent of 1.35 μmol of K+/ cm2, which for KCl equals 0.1 mg/cm2. Desiccators with phosphorus pentoxide (P2O5) were used for storage. 2.2. Exposures. Exposures were performed in horizontal furnaces equipped with quartz tubes (see Figure 1). The atmosphere consisted
3. CORROSION PRODUCT CHARACTERIZATION AND SURFACE MORPHOLOGY X-ray Diffraction (XRD). Crystalline corrosion products formed on the sample surfaces after exposure were analyzed using grazingincidence X-ray diffraction (GI-XRD) using a Siemens D5000 powder diffractometer, equipped with grazing-incidence beam attachment and a Göbel mirror. The radiation used was Cu Kα, and the angle of incidence was 2°. The estimated depth of the XRD analysis is about 3−5 μm. The measurement range was 20° < 2θ < 70°. The XRD results given in Table 2 are divided into three levels with respect to their intensity: weak, medium, and strong. These levels are derived from the strongest peak in the diffractogram. Hence, the intensity of the strongest (i.e., highest) peak for each phase was divided by the intensity of the strongest peak in the diffractogram. All values between 1 and 0.3 were denoted strong; all values between 0.3 and 0.1 were denoted medium; and all values of