revisting caustic cracking of steel vessels and pipes for alumina
Transcription
revisting caustic cracking of steel vessels and pipes for alumina
REVISTING CAUSTIC CRACKING OF STEEL VESSELS AND PIPES FOR ALUMINA PROCESSING R.K. Singh Raman a,b* , Sarvesh Pal a,1 a Department of Mechanical and Aerospace Engineering b Department of Chemical Engineering Monash University, Vic 3800, Australia 1 Current Affiliation: Department of Mechanical and Mining Engineering The University of Queensland, Brisbane Australia 4072 * Presenter and corresponding author: Email Raman.singh@monash.edu Ph + 61 3 990 53671 Fax + 61 3 990 51825 ABSTRACT Caustic cracking tests were conducted using Bayer solutions of different chemistry at different temperatures (that are used in Alumina industry for extraction of Alumina from Bauxite ores). The validity of the commonly used caustic cracking susceptibility diagram for steels exposed to plain caustic solutions has been assessed using notched and precracked specimens. The study presents first results towards the development of a model susceptibility diagram for actual Bayer solutions, improving upon applicability of such diagrams over the traditional plain caustic diagram. INTRODUCTION In the Bayer process, steel is the commonly used material for construction of reaction vessels and pipes for different processing components for extraction of alumina from bauxite ores, such as digesters, decomposer and precipitator (Burstein et al., 1984, Flis et al., 2008, Le et al., 1990, Le et al., 1992, Le et al., 1993, Raman et al., 2003, Raman et al., 2007, Shin et al., 2004, Singbeil et al., 1982, Sriram et al., 1985, Sriram et al., 1985). Stress corrosion cracking (SCC) is the premature cracking of materials under the synergistic action of a tensile stress and corrosive medium, neither of which would cause cracking when acting alone. Caustic embrittlement is a form of SCC that results from embrittlement of material exposed to caustic environment. The most fundamental and detrimental feature of SCC is that a ductile material that would have undergone considerable elongation before fracture may suffer embrittlement in the presence of the corrosive environment, leading to a premature brittle fracture (i.e., without much elongation). Shown in Figure 1 is an example of caustic embrittlement and intergranular crack propagation over the considerable area-fraction of the fracture surface of a steel specimen, and pure mechanical failure (as evidenced by ductile dimples over rest of the area) due to overloading of the cross-section reduced by intergranular stress corrosion cracking. 1 Fig.1 SEM fractograph of a caustic embrittled mild steel: showing a transition from corrosion-assisted intergranular cracking (lower part of the micrograph) to the pure mechanical failure (ductile dimples). Magnification: X850 (Raman, 2005). Caustic embrittlement continues to be a concern for steels in hot caustic service, however, low carbon steels is still the most frequently used material (Metals Handbook, 1987). Berk et al., 1950 and Mazille et al., 1972 investigated the limits of caustic concentrations and temperatures for caustic embrittlement. For example, steels were found to show immunity to caustic embrittlement at 5wt% NaOH when temperatures were below 95 ºC, whereas this temperature limit was 40 ˚C at 50wt% NaOH. Caustic cracking susceptibility (CS) diagram (Figure 2) (Metals Handbook, 1987) is a plot of caustic concentration against temperature, based on industrial as well as laboratory tests for a maximum of 62 days. This diagram identifies different regions of caustic cracking susceptibility. The upper hatched area is the severe cracking zone, whereas in the regime immediately below this area the cracking may or may not take place. The upper hatched area was developed in the laboratories whereas the lower curve is based on the field experience (Schmidt et al., 1951). Understanding of caustic cracking susceptibility is less clear for the regions between the upper hatched area and the lower curve. By developing an improved understanding of the effect of caustic concentration and temperature on caustic cracking, it may be possible to expand the limits for safe application of carbon steel in various steps of Bayer process (Gontijo et al., 2009). Caustic cracking of in-service components is highly likely to be influenced by the sharp notches and other stress-raisers that are often present in the fabricated components. Therefore, to investigate the validity of the CS diagram it may be necessary to test appropriately the notched and pre-cracked specimens. Impurities/additions in caustic solutions can also changes caustic cracking susceptibility. For example, aluminate ions (AlO2-) increase caustic cracking susceptibility of steel (Le et al., 1989) because the incorporation of AlO2- decreases the stability of the passive film that develop on steels in the plain NaOH solution. On the other hand, some oxidizing agents, such as KMnO4, and NaNO3 assist the formation of γ-Fe2O3 films, which favourably shifts the electrochemical potential and retards caustic cracking ( Humphries et al., 1967). 2 Fig.2: Caustic cracking susceptibility diagram and caustic cracking susceptibility data generated using CNT specimens (data indicated as stars) superimposed caustic cracking susceptibility diagram (Raman et al., 2010). The stars represent definite cracking region. The squares are potential cracking and circles are the definite absence of cracking This paper establishes the need for developing a model caustic cracking susceptibility diagram for typical Bayer solutions, using pre-cracked specimens and circumferential notch tensile (CNT) technique. This diagram will serve as a new guideline for addressing the ongoing concerns of caustic cracking in alumina processing industry. EXPERIMENTAL PROCEDURES Caustic Cracking Susceptibility Tests Caustic cracking susceptibility tests were conducted in plain NaOH and Bayer solution at different concentrations and temperatures that were selected on the basis of the data in Figure 2. The test material was AS/NZ 3678-grade 250 steel having the chemical composition (wt.%), C: 0.17, Si: 0.27, Mn: 11.19, P: 0.19, S: 0.09, Cr: 0.02, Ni: 0.01, Mo: 0.02, and the mechanical properties, yield strength: 338 MPa, ultimate tensile strength: 492 MPa and elongation: 35%. Tests were conducted using circumferential notch tensile (CNT) specimens. The details of this technique can be referred in literature (Rihan et al., 2006, Pal, et al. 2009). The CNT specimens were fatigue pre-cracked using rotating bending machine before being installed in CNT testing rig. Tests were carried out using specimens with different pre-crack depths and applied loads that resulted in different stress intensities (KI) at the pre-cracks. A tests was either continued until the specimen failed or was terminated after a considerable length of time for given test condition. Fractography The fracture surface was ultrasonically cleaned with a cleaning solution that contained 6 ml conc. Cleaned surface was observed under JEOL 840 scanning electron microscopy (SEM) in order to investigate the presence of the well established fractographic evidence of caustic cracking of steel. 3 RESULTS Caustic Cracking Behaviour of Grade 250 Steel To investigate the applicability of caustic cracking susceptibility (CS) diagram, four concentrations at various temperatures were chosen. 10-40 wt% caustic solution will have a good chance to caustic cracking. However, it is not possible to precisely control the precrack depth and hence stress intensity (KI). Therefore, the applied loads were in the range of the fracture toughness and the threshold stress intensity factor for SCC (KISCC). Investigation of Caustic Cracking in Plain Caustic Solutions Temperatures and caustic concentrations for these tests were selected such that their combinations would fall in the regimes of ‘definitely no cracking’, ‘definitely cracking’ and ‘may be cracking’ in the susceptibility diagram (Figure 2). The distinctive fractographic features for caustic cracking of the pre-cracked specimens was the presence of intergranular cracking immediately adjacent to the fatigue pre-crack. Typical fractographs showing intergranular cracking described in Figure 1. The overall fracture surface in Figure 3a has four regions that are located an order (in moving from edge to centre): the machined notch at the specimen edge, fatigue pre-crack, caustic cracking and exclusively mechanical fracture. The distinctive features of these regions are established at higher magnifications. In the light of the features for the presence and absence of inter-granular cracking (clearly identified in Figure 3a and b), the fractographic features of the CNT specimens tested using the chosen combinations of caustic concentration and temperature were examined. The results are summarized in Table 1. Machine Notch Fatigue Crack CC Zone Mechanical Failure (a) (b) 4 (c) Fig. 3: SEM fracture surface of CNT specimens tested in caustic solutions showing: (a) overall surface showing different regions, (b) at higher magnification, clear inter-granular features and (c) dimples (in the lower half) adjacent to the beach marks (in the upper half) and the absence of inter-granular features in between. Table 1 Caustic cracking tests in plain NaOH solutions Caustic concentration (wt%) 10 20 30 40 Caustic solution temperature (˚C) Applied stress intensity (MPa m1/2) 100 35.8 70 32.9 Time to failure (h) Evidence of intergranular fractographic features Caustic cracking susceptibility 100 39.4 Did not fail (in 3192 h) Did not fail (in 6960 h) 134 h Yes Yes 80 35.7 408 h Yes Yes 80 29.5 49.7 Yes No Yes 65 950 h Did not fail (in 2064 h) 100 27 815 h Yes Yes 70 39.9 Did not fail (in 2000 h) No No 55 32.6 Did not fail (in 5520 h ) No No 100 27.8 33.7 Yes No Yes 70 45 32.5 184 h Did not fail (in 6219 h) Did not fail (in 3192 h) No No No No No No No No 5 Investigation of Caustic Cracking in Bayer Solutions Similar to those in plain NaOH, caustic cracking tests were conducted in different Bayer solutions with different caustic concentrations. The results are summarised in Table 2. Based on the test results presented in Table 1, the caustic cracking susceptibility of steel in plain NaOH is predicted in various test conditions. This is largely consistent with the common caustic cracking susceptibility diagram (Figure 2). However, the data in Table 2 indicate a greater susceptibility of steel in Bayer solution. For example, the steel was found to be highly susceptible to caustic cracking even at a considerably low temperature (55 ˚C) in 30wt% FC Bayer solution. But the cracking susceptibility diagram for steel in plain caustic solution (Figure 1) would predict immunity under this condition. However, as the FC concentration of the Bayer solution increased to 40wt%, the caustic cracking susceptibility decreased drastically. The caustic cracking at this concentration was observed only at a high temperature (100 ˚C). Table 2 Caustic Cracking tests in Bayer solutions Free caustic concentration (wt%) 10 20 30 40 Caustic solution temperature (˚C) Applied stress intensity (MPa m1/2) 100 37.3 70 33.1 100 80 Time to failure (h) Evidence of intergranular fractographic features Caustic cracking susceptibility Yes Yes 31.7 2088 h Did not fail (in 2323 h) 241 h Yes Yes 27.3 2280 h Yes Yes 80 32.3 Yes Yes 65 62.9 950 h Did not fail (in 4000 h) 100 27.3 1801 h Yes 70 42.2 87 h Yes 55 53.2 210.3 h Yes 100 75.1 Yes 70 41.2 45 30 192 h Did not fail (in 2112 h) Did not fail (in 1632 h) No No No No No No Yes Yes Yes Yes No No Figure 4a shows the fracture surface of specimen loaded in 30wt% Bayer solution at 100 º C. It took a relatively longer time (1801 h) to fail, presumably because of the relatively low KI employed (27.3 MPa m1/2). However, as predicated by the susceptibility diagram, fracture surface did possess the feature for caustic cracking (i.e., inter-granular cracking). The specimen tested at 70 ºC failed in 80 h at an applied stress intensity of 42.2 MPa m1/2, and possessed an intergranular feature, confirming caustic cracking (Figure 4b). The CNT 6 specimens tested at 55 °C failed in 210 h at an applied stress intensity of 53.2 MPa m½. As per caustic susceptibility diagram 30wt% NaOH at 70 °C lies in the ‘may be cracking’ zone and 55 °C is considered to be safe. However, the present results would suggest a possibility of caustic cracking both at 70 and 55 °C. DISCUSSION The greater caustic cracking susceptibility in Bayer solutions than in plain NaOH can be attributed to the role of impurities and additives in the Bayer solution. Consistent with the very recent work on pure NaOH by the same authors ( Raman et al., 2010), caustic cracking susceptibility in 30wt% NaOH in this study was found to be limited to 100 ºC (Table 1), whereas in Bayer solutions the caustic cracking susceptibility was observed at much lower temperatures (55 ºC). Sriram et al ( Sriram et al., 1985) conducted caustic cracking tests in a Bayer solution (overall concentration 14.4% , Free caustic concentration 9% ) at 92 ºC using notched specimens and observed that the presence of AlO2- species moves SCC susceptibility towards a lower regime of stress intensities. Le et al ( Le et al., 1993) and Raman ( Raman, 2005) have conducted the experiment in Bayer solutions and different caustic aluminate solutions in conjunction with tests in plain NaOH at 100 ºC and reported the steel to be more susceptible to caustic cracking in the Bayer solutions. Among the impurities dissolved in the Bayer solutions, aluminate ions have the greatest damaging influence on the susceptibility of steel ( Le et al., 1993). The detrimental effect of aluminate ions is attributed to the formation of an amorphous film of Fe 3-xAlxO4, where X ≤ 2, which is less stable. (a) (b) 7 (c) Fig. 4: SEM fracture surface of CNT specimens tested in 30wt% Bayer solution at: (a) 100 ˚C, (b) 80 ˚C, and (c) 55 ˚C, each showing inter-granular features immediately after fatigue pre-cracked area, suggesting caustic cracking. CONCLUSIONS 1. A caustic cracking susceptibility diagram has been developed for representative Bayer solutions. 2. Steel is more susceptible to caustic cracking in Bayer solution than in plain caustic solutions of similar caustic contents. ACKNOWLEDGEMENTS Authors record a special word of appreciation to Alcoa World Alumina for providing a Bayer solution for testing. REFERENCES 1987 1987 Metals Handbook Book Metals Handbook 13 (1178) (1987) pp.p.1178 Berk et al., 1950 Berk, A. and Waldeck, W., Caustic danger zone Chemical Engineering 57 (6) (1950),pp. 235 Burstein et al., 1984 Burstein, G. and Woodward, J., Examination of the Stress Corrosion Cracking of a Low Alloy Steel by Auger Electron Spectroscopy, Corros. Rev. 6 (1) (1984),pp. 81-96 Flis et al., 2008 Flis, J. and Ziomek-Moroz, M., Effect of carbon on stress corrosion cracking and anodic oxidation of iron in NaOH solutions, Corros. Sci. 50 (6) (2008),pp. 1726-1733 8 Gontijo et al., 2009 Gontijo, G., Araújo, A., Prasad, S., Vasconcelos, L., Alves, J. and Brito, R., Improving the Bayer Process productivity - An industrial case study, Miner. Eng. 22 (13) (2009),pp. 1130-1136 Humphries et al., 1967 Humphries, M. and Parkins, R., Stress-corrosion cracking of mild steels in sodium hydroxide solutions containing various additional substances, Corros. Sci. 7 (11) (1967),pp. 747-761 Le et al., 1989 Le, H. and Ghali, E., Active-passive behaviour and stress corrosion cracking of A516 steel in Bayer solution, J. Appl. Electrochem. 19 (3) (1989),pp. 368-376 Le et al., 1990 Le, H. and Ghali, E., The electrochemical behaviour of pressure vessel steel in hot bayer solutions as related to the scc phenomenon, Corros. Sci. 30 (2-3) (1990),pp. 117-134 Le et al., 1992 Le, H. and Ghali, E., Slow strain rate and constant load tests of A 285 and A 516 steels in Bayer solutions, J. Appl. Electrochem. 22 (1992),pp. 396-403 Le et al., 1993 Le, H. and Ghali, E., Stress corrosion cracking of carbon steel in caustic aluminate solutions of the Bayer process Corros. Sci. 35 (1-4) (1993),pp. 435-442 Mazille et al., 1972 Mazille, H. and Uhlig, H., Effect of temperature and some inhibitors on stress corrosion cracking of carbon steel in nitrate and alkaline solutions, Corrosion-NACE 28 (11) (1972),pp. 427-433 Raman, 2005 Raman, R. K. S., Evaluation of caustic embrittlement susceptibility of steels by slow strain rate testing, Metall. Mater. Trans. A 36 (7) (2005),pp. 1817-1823 Raman et al., 2003 Raman, R. K. S. and Muddle, B., Caustic stress corrosion cracking of a spheroidal graphite cast iron: characterisation of ex-service component, Mater. Sci. Technol. 19 (2003),pp. 1751-1754 Raman et al., 2010 Raman, R. K. S. and Pal, S., Investigations Using Smooth and Notched Specimens into Validity of Caustic Cracking Susceptibility Diagram, Metall. Mater. Trans. A 41 (9) (2010),pp. 2328-2336 Raman et al., 2007 Raman, R. K. S., Rihan, R. and Ibrahim, R., Role of imposed potentials in threshold for caustic cracking susceptibility (KISCC): Investigations using circumferential notch tensile (CNT) testing, Corros. Sci. 49 (12) (2007),pp. 4386-4395 Rihan et al., 2006 Rihan, R., Raman, R. K. S. and Ibrahim, R., Determination of crack growth rate and threshold for caustic cracking (KISCC) of a cast iron using small circumferential notched tensile (CNT) specimens, Materials Science and Engineering: A 425 (1-2) (2006),pp. 272-277 Schmidt et al., 1951 Schmidt, H., Gegner, P., Heinemann, G., Pogacar, C. and Wyche, E., Stress Corrosion Cracking in Alkaline Solutions, Corrosion 7 (1951),pp. 295-302 Shin et al., 2004 Shin, H., Lee, S., Kim, S., Tran, T. and Kim, M., Study on the effect of humate and its removal on the precipitation of aluminium trihydroxide from the Bayer process, Miner. Eng. 17 (3) (2004),pp. 387-391 Singbeil et al., 1982 Singbeil, D. and Tromans, D., Caustic Stress Corrosion Cracking of Mild Steel, Metall. Mater. Trans. A 13A (6) (1982),pp. 1091-1098 9 Sriram et al., 1985 Sriram, R. and Tromans, D., The anodic polarization behaviour of carbon steel in hot caustic aluminate solutions, Corros. Sci. 25 (2) (1985),pp. 79-91 Sriram et al., 1985 Sriram, R. and Tromans, D., Stress corrosion cracking of carbon steel in caustic aluminate solutions—crack propagation studies, Metall. Mater. Trans. A 16 (5) (1985),pp. 979-986 Sriram et al., 1985 Sriram, R. and Tromans, D., Stress corrosion cracking of carbon steel in caustic aluminate solutions - slow strain rate studies, National Association of Corrosion Engineers 41 (7) (1985),pp. 381-385 BRIEF BIOGRAPHY OF PRESENTER Professor Raman Singh currently works at Monash University in the Department of Chemical Engineering and the Department of Mechanical & Aerospace Engineering. His interdisciplinary research expertise comprises: • • • • • • • • • Role of Nano-/Microstructure in Corrosion, Stress Corrosion Cracking, Corrosion and Corrosion-assisted Cracking of Weldments, Failure Analysis of Metallic Industrial Components, High Temperature Corrosion in the Industrial Environments, viz., Steamgenerators/boilers of Thermal and Nuclear Power Plants, Petroleum/Petrochemical, Advanced Coatings and Surface Modifications for Corrosion Resistance, Microbiologically-induced Corrosion and Cracking, Corrosion of Magnesium Alloys, Surface and Sub-surface Characterisation of Corrosion. Professor Raman Singh’s primary research interest is microstructure-corrosion relationship. He has also worked extensively on stress corrosion cracking, and corrosion and corrosionmitigation of magnesium alloys, including for the use of magnesium alloys for bio-implant applications. Over 70% journal publications of a total of a total of his over 90 peerreviewed journal publications of the last 10 years appeared in the top 10% of the journals in the category, Metallurgical Engineering and Metallurgy. His professional distinctions include: editorship of a book on cracking of welds, membership of the Review Board of the prestigious, Metallurgical and Materials Transactions A, leadership/co-chairmanships of several international conferences and guest co-editorship of international journals, regular keynote/invited lectures at several international conferences, over 110 peer-reviewed international journal publications, 14 book chapters and over 90 reviewed conference publications, and several competitive research grants. 10