Abstract
The optimal contents of C, S, Cu and Sb in a new low alloy steel for the purpose of improving the sulphuric acid dewpoint corrosion resistance were investigated by weight loss and potentiodynamic polarisation in 50 wt-%H2SO4 at 70°C. The optimal elemental contents of C, S, Cu and Sb in the low carbon sulphuric acid dewpoint corrosion resistant steel (SADPS) are 0·09, 0·016, 0·3 and 0·09 respectively. The corrosive products of Fe2(SO4)3 and FeSO4 were found on the surface of plain low carbon steel while corrosive products of (Cu,Fe)SO4, FeSbO4, FeSb2O4, CuSbO4 and CuO were found on the surface of SADPS. Potentiodynamic polarisation test also revealed that the higher corrosion resistance of SADPS was due to inhibition of the cathodic reaction due to surface enrichment by the alloying elements.
Keywords
Introduction
In the metallurgical, electric power generation, petrochemical and other industrial fields, sulphur (S) containing fuels are burnt, producing sulphur and nitric oxides: SO2, SO3 and NOx.1 – 4 When the exhaust gas temperature drops below temperatures corresponding to a dewpoint, or when the gas is in contact with a low temperature wall such as flues, smokestacks, boiler air preheaters and other facilities exposed to exhaust gas,5,6 these acid gases combine with moisture to form highly concentrated acids. This ‘dewpoint’ corrosion causes serious corrosion failures with both general corrosion and pitting corrosion occurring.
Low alloy steels present improved mechanical properties and are used for a wide range of structural uses; they frequently have modestly improved corrosion resistance when compared with plain carbon steels.7 The addition of small quantities of single alloying elements, such as Cu, Sb, Co, Ni and Cr, has been shown to improve the corrosion resistance of low alloy steels.1,2,5,8–12 For example, it is well known that addition of copper to steel can improve the performance in the sulphuric acid dewpoint corrosion environment.3,13 – 16 Furthermore, additions of antimony favour the development of a Cu containing intermetallic compound that may inhibit the anodic and cathodic reactions.3,5 Cementite (Fe3C) aids hydrogen evolution and therefore a low carbon steel will generally corrode more slowly than high carbon grades.4,17 Using these principles, S-TEN 1 steel,3 containing Cu and Sb was developed by Nippon Steel, and is the most effective for use in the sulphuric acid dewpoint corrosion environment prevalent in Japanese power generation. However, for higher sulphur content fuels, such as are used in China, the usefulness of S-TEN 1 steel in China is limited.
Within the literature, there is a lack of information on the synergy between the elements C, S, Cu and Sb on the corrosion resistance of low carbon alloy in sulphuric acid. In this study, an orthogonal experimental design method was adopted to optimise the steel composition for sulphuric acid dewpoint corrosion with respect to these four elements.
Experimental
Chemical compositional design and preparation of specimens
Nine candidate sulphuric acid dewpoint steels (SADPSs), with varying compositions of carbon, sulphur, copper and antimony, were prepared by reference to an orthogonal experimental design using four parameters and three levels (Table 1). Other elemental contents were controlled in the ranges: 0·1-0·2Si, 0·9-1·0Mn, 0·01-0·02P, 0·45-0·55Cr, 0·1-0·2Ni, 0·01-0·02Ti and balance Fe (wt-%). The steels, numbered through 1 to 9, were produced by a vacuum induction furnace in Wuhan Iron and Steel (Group) Corporation, and the corresponding compositions were measured by a high frequency infrared ray carbon sulphur analyser (EMIA 820V; Horiba, Japan) and an inductively coupled plasma atomic emission spectrometer (Ultima 2C; Horiba, Japan). A plain low carbon steel (PLCS) containing 0·1-0·22C, 0·12-0·2Si, 0·4-0·6Mn, 0·03S, 0·03P and balance Fe (wt-%) was selected for comparison. The PLCS and the SADPS were air cooled from the melt and their microstructures are similar, composed of mainly ferrite with some pearlite, as shown in Fig. 1.
Microstructures of a PLCS and b SADPS
Chemical composition of four elements/wt-%
Rectangular specimens with a dimension of 30×30×4 mm were prepared for immersion testing. Additionally, specimens with a dimension of 10×10×4 mm were embedded in epoxy resin leaving a working area of 1 cm2 for anodic polarisation curve measurement. Both specimens were ground to a 1200 grit finish using SiC paper and subsequently cleaned with distilled water before testing. The test conditions were chosen to simulate service conditions. Hence, as the temperature of many chimneys and flue ducts is around 70°C, at which the corresponding H2SO4 concentration at the sulphuric acid dewpoint is ∼50 wt-% in accordance with Taylor vapour–liquid equilibrium diagram, 18 18,19 all measurements were carried out under these conditions.
Immersion tests
Immersion tests were carried out in accordance with ASTM G31-71.20 The area of each tested specimen exposed to 50% sulphuric acid was measured to a precision of 0·02 mm, and specimen mass to a precision of 0·1 mg before and after exposure. Before corrosion tests, specimens were cleaned ultrasonically with acetone and absolute ethyl alcohol and dried in cold air. Specimens were immersed in 50 wt-%H2SO4 water solutions at 70°C for 24 h.
The corroded mass losses per unit area and per unit time were calculated by the following equation18
Electrochemical measurements
Anodic polarisation tests were carried out on an electrochemical workstation (IM6ex; ZAHNER, Germany) with a standard three electrode system. All potentials were expressed with respect to the saturated calomel electrode (SCE). Before any data were recorded, an initial time delay of 1800 s was set to allow the specimens to achieve equilibrium in the solution at open potential. The scanning started at a potential of ∼250 mV below the open circuit potential of each sample with a scanning rate of 0·5 mV s−1.
Surface analyses
Scanning electron microscopy (SEM; Tecnai G2 20; FEI, The Netherlands) was performed after the immersion and anodic polarisation tests. The corresponding component phases on corroded surfaces were determined using X-ray diffraction (XRD; X'Pert PRO; PANalytical B.V., The Netherlands) with Cu Kα radiation. Electron probe microanalysis (EPMA; EPMA-1720; SHIMADZU, Japan) was performed to examine the relationship between cross-section and surface morphology of the specimens after immersion for 24 h.
Results
Orthogonal array test and analysis
Immersion tests in 50%H2SO4 water solution at 70°C were carried out according to the L9 (34) orthogonal array (Table 1), and corrosion rates are listed in Table 2. These results revealed that the influence of antimony on the corrosion rate is the most significant among C, S and Cu, with an optimal chemical composition of 0·09C–0·016S–0·3Cu–0·09Sb (wt-%). The corrosion rate of the optimised SADPS in 50% sulphuric acid at 70°C for 24 h is 0·7821 mg cm−2 h−1, in comparison, the corrosion rate of the PLCS is ∼7 times greater (5·2513 mg cm−2 h−1).
L9 (34) results of specimens immersed in 50 wt-%H2SO4 water solution
*The corrosion rate of the PLCS was 5·2513 mg cm−2 h−1at 70°C.
Analysis of specimens after immersion corrosion
The surface morphologies of the PLCS and the optimal SADPS after 24 h exposure were observed by SEM, as shown in Fig. 2. The corroded PLCS surface is covered by a loose, cracking and discontinuous layer. By contrast, the corroded SADPS surface is characterised by a dense and continuous layer, which supports the observation that the SADPS has superior corrosion than the PLCS.
Morphologies of corroded a PLCS and b SADPS after immersing in 50 wt-%H2SO4 water solution at 70°C
The chemical compositions of corrosion products on the PLCS and the SADPS, analysed by SEM/EDX (see Table 3) indicate that the contents of five elements (Si, S, Sb, Cr and Cu) on the corroded SADPS are much higher than the SADPS substrate while only two elements (Si and S) are higher than the PLCS substrate. X-ray diffraction patterns of corrosion products on the PLCS and the SADPS further reveal that corrosion products are predominantly ferric and ferrous sulphate on the plain carbon steel but a complex mixture of copper–iron sulphate, ferrous and ferric antimonite, copper antimonite and copper oxide on the optimised steel (Fig. 3). The EPMA mapping was also conducted to examine the elemental distributions along the cross-sections of the corroded surfaces of the optimised steel, as shown in Fig. 4. Enrichment and uniform distribution of elements (Cu, Sb and S) on the corroded surface indicates that the corrosion products contain elements Cu, Sb and S, which is in accordance with the results of EDX analysis and X-ray examination. All results mentioned above confirm that the addition of small quantities of S, Sb and Cu into the sulphuric acid resistant steel changed the morphology and composition of corrosion products and contributed to the formation of much denser corrosion products in comparison with the plain carbon steel.
Electron probe microanalysis mapping of certain elements on cross-section of SADPS obtained for corrosion products after 24 h immersion in 50 wt-%H2SO4 water solution at 70°C
X-ray diffraction patterns of corrosion product of a SADPS and b PLCS
Energy dispersive X-ray analysis of corroded surfaces after 24 h immersion
Potentiodynamic polarisation
The potentiodynamic polarisation curves of the PLCS and the SADPS at 70°C in 50% sulphuric acid are given in Fig. 5. The curves for both steels show a reduction in current typical of either an active–passive transition or to salt film formation (e.g. due to the corrosion products). A second anodic current maximum I a2 was observed at 620±10 mV(SCE) in addition to that for the ordinary anodic dissolution I a1. Seo et al.14,21 attributed this to the dissolution of the corrosion products layer/salt film deposited on the surface, but Davies14,22 claimed that it related to the oxidation of adsorbed hydrogen atoms formed during cathodic polarisation. It is likely to be due to both reasons. The corrosion potentials E corr and the corrosion currents I corr of both steels calculated from Fig. 5 are listed in Table 4 and show that the corrosion current of the SADPS is about 5-6 times greater than that of the PLCS, consistent with the immersion experiments. The polarisation data indicate that the added alloying elements predominantly reduce the cathodic reaction kinetics with the anodic reaction largely unaffected. This is consistent with the observed surface enrichment of copper, antimony, etc., and the reduced exchange current density for hydrogen evolution on these elements compared with on iron. Interestingly, in this experiment, the salt film appeared to be less stable, requiring a greater peak anodic current density and resulting in a higher current in the potential independent region. This would appear to be contrary to the microscopy evidence of a compact surface film on the optimised steel and require further study (Fig. 6).
Potentiodynamic polarisation curves of specimens in 50 wt-%H2SO4 water solutions at 70°C
Observation (SEM) of corroded a PLCS and b SADPS after potentiodynamic polarisation tests in 50 wt-%H2SO4 water solution at 70°C
Corrosion potential E corr and corrosion current I corr calculated from Fig. 5
Discussion
The chemical compositions of corrosion products obtained by EDS and EPMA confirm that the enrichment of elements (Cu, Sb, Cr, S, etc.) on the SADPS, and new types of corrosion products [(Cu,Fe)SO4+FeSbO4+FeSb2O4+CuSbO4+CuO] were detected by XRD analysis, by contrast, simple corrosion products [(Fe2(SO4)3+FeSO4] were detected on the PLCS due to the lack of important elements such as Cu, Sb and S. Meanwhile, some other corrosion products, such as Cu2S and CuS probably existed. Figure 7 displays Gibbs free energy changes of several possible sulphides changed during corrosion process at various temperatures.23 It indicates that Cu2S is the easiest formation phase and the stablest phase among these sulphides for its lowest Gibbs free energy. As is known, H2S will be released when MnS24 and FeS appear in H2SO4 water solution while Cu2S and CuS can be stable. Results suggest that the synergic effect of Cu and Sb elements changed the types of corrosion products and improved the corrosion resistance. The EPMA mapping showed that Cu and Sb distributed uniformly over the steel surface, indicating Cu and Sb containing products with high density. The SEM observations of the corroded specimens revealed that the corrosion layer with multiphases of the SADPS is characterised by a dense, continuous and well bonded structure while that of the PLCS is characterised by numerous pits and microcracks. Therefore, the PLCS presented much higher corrosion rate than that of the SADPS in 50%H2SO4 water solution.
Gibbs free energy changes of different sulphides
Potentiodynamic polarisation curves of the PLCS and the SADPS showed that the corrosion potential E corr of the SADPS is slightly nobler than that of the PLCS and the corrosion current I corr of the SADPS is much lower than that of the PLCS, as shown in Table 4. The changes of both E corr and I corr profit from the synergic effect of the addition of elements (such as Cu, Sb and S). The preferential dissolution and redeposition process occurred during the potentiodynamic polarisation test. The process of Cu and Sb redeposition is thought to be accelerated by the cathodic polarisation.5,14 As a consequence of the Cu and Sb accumulation on the surface of the SADPS, the corrosion current density decreased through the suppressed cathodic reactions. Eventually, the cathodic overpotential in the hydrogen evolution is increased through Cu and Sb enrichments.14,25 In Ref. 5, Le et al. concluded that sufficient addition of Sb not only created a protective Sb2O5 oxide layer, but also stimulated the development of high corrosion inhibiting, Cu containing compounds that further inhibited the anodic and cathodic reactions. Nippon Steel noted that sufficient amount of Sb in Cu containing steel considerably promoted the formation of Cu2Sb film, inhibiting not only anodic reaction but also cathodic reaction.3
Schematic diagrams of protective layer forming in H2SO4 water solution are shown in Fig. 8. The process is split into three steps:
Schematic diagrams of protective layer forming mechanism in H2SO4 solution
under ambient conditions, dissolved oxygen and hydrogen cations coexist in the sulphuric acid solution. Near to the solution/metal interface, the composition of SADPS mainly consists of Fe with a small amount of added Cu and Sb noble elements. When temperature reached 70°C, part of dissolved oxygen escapes
the active dissolution process of SADPS in sulphuric acid occurs. After Fe dissolves into H2SO4 water solution, Cu and Sb elements subsequently become detached from the metal surface and preferential dissolution occurs along the continuous paths of less noble atoms
dissolved Cu and Sb are deposited on the surface, causing repassivation of the material. The selective dissolution and redeposition process will repeat until the Cu and Sb enriched layer is stabilised.
Conclusions
Orthogonal chemical of the low carbon SADPS compositions (in mass-%) are 0·09C–0·016S–0·3Cu–0·09Sb, which was validated performing good corrosion resistance.
The Cu and Sb containing steel presents much lower corrosion rates than Cu and Sb free steel in H2SO4 water solution. Enrichment of Cu and Sb is confirmed through the use of EPMA and surface EDX analyses. The accumulated Cu, Sb, Cr and S compounds on the surface improve the corrosion resistance of the low alloy steel by reducing the active dissolution of the substrate. A dense compact corrosion product layer formed on the surface of the SADPS which is proved to be the best protection to the substrate.
Footnotes
Acknowledgements
The authors acknowledge the financial support from Major State Basic Research Development Program of China (973 Program) under contract No. 2012CB619600.