CARBON CORROSION OF ALLOYS AT HIGH TEMPERATURE

Transcription

1 D.J. Young, J. Zhang CARBON CORROSION OF ALLOYS AT HIGH TEMPERATURE D.J. Young University of New South Wales, Sydney J. Zhang University of New South Wales, Sydney Abstract Alloys used at high temerature must resist both cree and corrosion. Design for corrosion resistance is based on formation of a slow growing, rotective oxide scale by selective oxidation of an aroriate alloy comonent, usually chromium or aluminium. A successful scale will exclude other corrodents, notably carbon, which can otherwise cause extremely raid corrosion at high temeratures. Selective oxidation of an alloy comonent necessarily lowers the concentration of that metal in the alloy subsurface region. Under thermal cycling conditions, mechanical damage to the scale leads to renewed oxide growth and accelerated alloy deletion. Eventually, a oint is reached where diffusion of a corrodent into the alloy becomes cometitive with the outward diffusion of alloy metal to reair the rotective scale. Two examles of alloy failure by carbon attack are considered. In the steam cracking (yrolysis) rocess, centrifugally cast tubes of heat resisting alloy are exosed to a gas stream of hydrocarbon and steam, at a carbon activity of unity. Formation and reair of the surface chromia scale causes alloy deletion, Kirkendall void formation and subsequent internal reciitation of chromium-rich carbides. Their formation makes chromia scale formation much more difficult, and generates internal stress. Eventually, the tubes fail by cree ruture. In other rocesses (e.g. steam reforming, heat treatment), synthesis gases are suersaturated with carbon at intermediate temeratures. Once the alloy s rotective scale is breached, carbon attacks the deleted substrate. In the case of ferritic alloys, it forms a surface scale of Fe 3 C. As this scale thickens, the suersaturated carbon reciitates as grahite within its outer regions. The resulting volume exansion causes disintegration of the cementite in a rocess known as metal dusting. In the case of austenitic alloys, no metal carbide is formed. Instead, carbon dissolves in the deleted metal to diffuse inward and reciitate as grahite within the metal matrix. Again, volume exansion causes disintegration of the alloy, and dusting results. Dusting occurs at an extraordinarily raid rate, and leads to failure by section loss or even enetration. 1. Introduction Carbon is unique among the common oxidants in being stable as a solid if the environment is sufficiently reducing. For this reason, the thermodynamic reference state is chosen as ure, solid grahite, for which a C = 1. If the carbon activity is less than one, but large enough to stabilise a metal carbide, the reaction is described as carburisation. In such rocesses, internally reciitated carbides develo raidly and alloy destruction can result. Because carbides are much less stable than the corresonding oxides, carburisation is a roblem only under reducing conditions. The necessary conditions arise in articular gases encountered in e.g. etrochemical rocesses such as steam cracking (yrolysis). They can also develo at an alloy-scale interface, where the oxygen otential is maintained at a low value by the local equilibrium This situation is a roblem only if carbon can enetrate the oxide scale. M + ½O (g) = MO (1) If the gas is suersaturated with resect to carbon ( a 1), an even greater threat emerges. If the gas can be c equilibrated, carbon is released from the gas hase and deosits in a rocess described as coking. Often, 1

2 however, gases remain suersaturated. In this event, catalysis of carbon deosition by the metal can lead to its disrution and fragmentation in an extremely raid corrosion rocess known as metal dusting.. Carbon Activities Gas hase rocesses roducing carbon include the synthesis gas reaction the Boudouard reaction and hydrocarbon cracking, e.g. CO CO CH H H O C () CO C 4 H C Their standard free energies are listed in Table 1. All three reactions are very slow as homogenous gas hase rocesses, and they will not reach equilibrium in a tyical laboratory reactor, unless catalysed. Although many materials of ractical interest, such as iron, nickel, cobalt and their alloys, are catalytically active to these reactions, their oxide scales are inert. As seen in Table 1, temerature effects are very different for these carbon roducing reactions. Thus methane and hydrocarbons in general, can roduce significant carbon activities only at high temeratures. On the other hand, the synthesis gas and Boudouard reactions roduce increasing carbon activities as temeratures are lowered. Table 1-Standard free energies of reactions Reactions G o A BT (Jmol 1 ) A B CO + H = H O + C CO = CO + C CH 4 = H + C (3) (4) 3. Carburisation Reactions in Steam Cracking Conditions Pyrolysis or steam cracking tubular reactors are used for making ethylene or roylene, e.g. ( 6 4 O H O) C H C H H ( H ) (5) Hydrocarbon cracking also roduces carbon, and a C 1. Steam acts as a diluent in order to reduce the amount of solid carbon roduced by gas hase yrolysis, and also rovides an oxidant to assivate tube metal surfaces. Rising rices of feedstock hydrocarbons and the need for imroved rocess efficiencies have led to higher oerating temeratures. Simultaneously, tube wall thicknesses have been reduced to imrove heat transfer efficiency. Thus rocess engineering changes have led to higher tube metal temeratures and reduced loadbearing sections. These increased demands on material roerties have been met by a series of advances in alloy design. The tubes are centrifugally cast, austenitic chromia-formers. Their comositions (Table ) have evolved from the old HK 40 grade (Fe-5Cr-0Ni), through the HP grades (Fe-5Cr-35Ni) to high nickel alloys containing 45 or even 60% Ni. The increased alloy levels have rovided significant imrovements in cree roerties, but the alloys are still subject to corrosion by carbon.

3 Table -Some nominal heat-resisting alloy comositions (wt%) Alloys C Si Mn Ni Cr Nb Al Other HK HP Zr,0.Ti HP Mod Nb H Cr Micro Ce 35/ Pa Mo, Hf 45HT Hf The general aearance of cast heat resistant alloys after exosure to simulated reaction conditions is shown in Figure 1. Both alloys shown have develoed chromium-rich oxide scales, causing deletion from the underlying alloy and dissolution of chromium-bearing carbides from the subsurface zone. Major differences are aarent deeer within the alloys: in Figure 1(a) no damage has occurred, aart from coarsening of the original interdendritic carbides, whereas in Figure 1(b) massive internal carburisation is obvious. The latter effect leads to swelling of the metal. If allowed to continue, this rocess eventually leads to cree ruture. a b 100µm 100µm Figure 1-General aearance of cast heat resisting alloys after 500 h exosure to steam-hydrocarbon mixture at 1100 o C: (a) high silicon, aluminium bearing version, and (b) low silicon version of HP grade. The different behaviour of the two alloys is due to their different chemistries and the resulting subsurface oxidation. Minority comonents silicon and aluminium lead to develoment of a continuous oxide layer beneath the chromia scale in one case (Fig. 1(a)). This slows the chromia scale growth rate and simultaneously imroves its ability to block carbon entry. The alloy shown in Fig. 1(b) has a low silicon content (0.6 wt %), insufficient to form a continuous layer, and internal reciitation of SiO results. As a result, the chromia scale grows more raidly, leading to Kirkendall voids develoing in the subsurface alloy region. In the absence of a continuous silica (or alumina) sublayer, the chromia scale is unable to block carbon entry, and internal carbide reciitation results. Many laboratory studies of carburisation have led to a good understanding of the rocess 1. It is commonly observed -4 that exosure of heat resisting alloys to gas comositions such that no chromia scale can form leads to internal chromium carbide reciitation rather than external scale formation. Carbon dissolves in the alloy and diffuses inwards to react with chromium and reciitate its carbides. Carbon diffusion controls the rate at which the carbide reciitation zone deth, X i, increases with time, t, and arabolic kinetics result: X i ( i) k t (6) According to Wagner s theory 5 when the carbon ermeability in the alloy is high 3

4 k ( i) N v D ( s) c c ( o) N M (7) (s) where NC is the concentration of dissolved carbon at the alloy surface, D C the carbon diffusion coefficient in (o) the chromium-deleted alloy matrix, N M the original alloy concentration of metal M which forms carbide MC and ε a factor accounting for diffusional blocking by reciitates. Thus carburisation rates are redicted v to vary inversely with concentration of reactive solute metal. Carburisation of Fe-Cr alloys 6,7 roduces chromium-rich (Cr,Fe) 7 C 3 reciitates, and follows arabolic kinetics. ( o) Plots of k against 1/ N were shown to be linear excet at high N values. The sloes of these lines (i) Cr were used together with v (for ( Cr 0.6Fe0.4 ) 7C3 formed by low chromium alloys) and the assumtion 1 to calculate carbon ermeabilities. Comarison in Table 3 with values found from indeendently measured N 8 and D 9 c values shows good agreement, demonstrating the utility of (7) in (s) c describing carburisation rates. (o) Cr Table 3-Carbon ermeabilities N ( s ) D C C (cm s -1 ) in Fe-Cr 900 C 1,000 C 1,100 C From Equation (7) From N and D C (s) C Equation (7) is based on the assumtion of very stable carbide reciitates and the comlete removal of chromium from the alloy by carburisation. In fact, this is a oor aroximation, and significant levels of chromium remain in the alloy matrix 1. This has the effect of decreasing the effective value of N (o) Cr aearing in (7), leading to the rediction of an enhanced value of k. However, the effect is not large, and uncertainties in measured values of N (s) C, D C and k (i) (i) total at least the same amount of error. It is concluded that Wagner s simle result (7) allows good order of magnitude rediction for model alloy carburisation. More imortantly, it also rovides good redictions for commercial alloy carburisation rates, as seen in Table 4. Table 4-Carburisation rate constants 10 7 k P(cm s -1 ) 900 C 1,000 C 1,100 C Measured Calculated Measured Calculated Measured Calculated G G H Fe-35Cr-45Ni Pa HT Alloy comosition affects carburisation rates in a number of ways. Carbon ermeability, N (s) C D C, varies strongly with alloy Fe/Ni ratio, as shown in Fig.. A minimum is found for nickel base alloys of Inconel tye comositions, this value being about 1/5-1/3 times that of a standard HP grade. 4

5 1 10 N C (s) DC, 10-9 cm s N Ni Figure -Carbon ermeability as a function of alloy comosition for Fe-Ni alloys at 1000 C with unit carbon activity. Minority alloy comonents are also imortant. Carbide-formers such as molybdenum articiate in reciitation reactions. The effect on kinetics can be redicted 10 (o) by adjusting the effective value of NM in (7) to account for the additional reciitating metal. Other carbide-formers such as niobium and reactive elements are often resent, although at low levels. They have unexectedly large effects in reducing carburisation rates, for reasons which are unclear 1. Silicon slows carburisation 11, even when the gas is not oxidising to the solute metal. The effect results from the deression of carbon solubility by silicon, and its negative effect on D C. Whilst alloy comositional effects on carburisation are reasonably well understood, it remains the case that accetable carburisation resistance cannot be achieved in the absence of an external oxide scale. 4. Metal Dusting in Synthesis Gas Synthesis gas can be roduced in the steam reforming reaction CH 4 HO CO 3H (8) at the oerating temeratures of o C, where a C < 0.5. However, as the roduct gas cools below about 700 o C, the equilibrium constant K increases, a C becomes suersaturated, and the ossibility of metal dusting arises. Metal dusting is a catastrohic form of corrosion in which metals exosed to carbon-suersaturated gas disintegrate, forming metal-rich articles (the dust ) disersed in a voluminous carbon deosit. Rates of alloy consumtion can be remarkably fast, leading to otentially dangerous conditions. Although heat resisting alloys are initially able to withstand this attack, the subsequent onset of the dusting rocess is difficult to redict, and its develoment can go undetected. Early reorts of industrial failures 1-15 were followed by the research of Hochman on dusting of iron, nickel, cobalt, and chromia-forming ferritic and austenitic alloys. Subsequent work by Grabke 19-3 quantified 5

6 and extended Hochman s observations. More recent work by the authors 4-35 and others has clarified some of the more mysterious asects of dusting rocess. Iron, nickel, cobalt and their alloys are catalytic to reactions () (4), and carbon deosition results from contact of these gases with common structural metals. Understanding how catalysis of carbon release from a gas disruts and fragments the catalytic metal requires a detailed consideration of reaction morhologies and mechanisms. Practical alloys form rotective oxide scales, usually chromia, which can act as a barrier to carbon ingress. However, at the relatively low temeratures involved, alloy diffusion is rather slow, and the ability of these alloys to reheal damaged scales is limited. Once this caacity is exhausted, carbon attacks the chromiumdeleted substrate. The mechanism and morhology of this reaction is in most resects characteristic of the dusting of alloy basis metals: iron, nickel or Fe-Ni. Different mechanisms aly to ferritic and austenitic alloys. 4.1 Dusting of Iron Iron and low alloy steels exosed to carbon-suersaturated gas grow external scales of cementite, Fe 3 C. In the early stages of reaction, the scale thickness, X, grows according to arabolic kinetics: X = k t (9) and the rate is controlled by inward diffusion of C through the scale 4. At the same time, the mass of coke deosited on to of the scale and the quantity of iron consumed both increase. The coke deosit is highly orous, consisting largely of grahite nanotubes and filaments which are decorated with cementite nanoarticles 4-6,31,33. Thus iron is consumed in roducing the cementite scale and the large numbers of Fe 3 C articles (the dust ). The cementite dust forms by disintegration of the scale outer surface. At longer reaction times, the overall change in scale thickness with reaction time reresents the net outcome of the growth and disintegration rocesses: dx dt k kd (8) X where k d is the linear rate constant for cementite scale loss. The disintegration is attributed to nucleation of grahite at favourable sites within the cementite, and the resulting volume exansion. Grahite reciitation is ossible because carbon activities are necessarily greater than unity within the Fe 3 C, in order for that hase to exist. Growth of the grahite is suorted by continuing diffusion of carbon through the Fe 3 C lattice. The source of the carbon is catalysis at the scale surface of reactions such as () and (3). Continued growth of coke filaments occurs via the same mechanism. Exosed facets of Fe 3 C nanoarticles catalyse release of carbon from the gas hase. This carbon diffuses through the article toward other facets which are energetically favourable for grahite deosition. At these ositions, the carbon attaches to the grahite filament, lengthening it and dislacing the cementite article outward. The overall rocess is shown schematically in Figure 3. Because the resulting coke structure is highly orous, the gas is able to enetrate it, reaching the cementite scale surface where further catalysis of carbon release is therefore ossible. 4. Dusting of Nickel Metal dusting of nickel and austenitic alloys differs from the reaction of ferritic materials in that cementite is not formed, and the corresonding nickel carbide is unstable. An examination of the dusting behaviour of ure nickel rovides a good basis for understanding the reaction of austenitic, heat resisting alloys. Dusting of nickel is much slower than the corresonding iron reaction. It roduces the reaction morhology shown in Fig. 4, where the carbon deosit is in direct contact with the metal, and is growing into it. 8,30,39,40 The metal-grahite interface is seen in Fig. 4 to be faceted, and reflects the referred orientation relationshi (0001) Gr //(111) Ni. It is the develoment of this energetically favoured interface at an angle to the metal 6

7 surface which allows grahite nucleation within the carbon-suersaturated metal. Growth of the grahite nuclei causes volume exansion and disrution of the metal. Its disintegration roduces the nickel nanoarticles seen in Fig. 4. When exosed to gas, these articles catalyse more carbon release, and its deosition at crystallograhically favoured sites to grow the nanofilaments, as shown in Fig. 4. The continued growth of grahite into the metal requires a suly of carbon, which is available only by diffusion through the substrate nickel. This is ossible because the nickel surface remains in contact with a high carbon activity, which is available from the gas. The dusting reaction roduct is a two-hase mixture of grahite lus nickel which is gas ermeable, and at uniform carbon and nickel activities throughout its thickness. Carbon is transorted through the coke via the gas hase. Nickel transort is not via diffusion; instead, nickel articles are dislaced outwards by accumulating grahite. The requirement for dusting that grahite basal lanes be oriented at an angle to the surface is not related to nickel surface orientation, and cannot thereby be controlled. Prevention of dusting requires either a surface barrier to carbon entry, or disrution of the {0001}gr //{111} Ni relationshi by alloying. 7,9,34,35 Grahite nucleation is made more difficult by alloying with Cu, which forms no carbide and is non-catalytic to coke. Adding coer to Ni 7, austenitic Fe-Ni 35 and austenitic stainless steels 9 markedly decreases carbon utake. X i (1) X () -Fe Fe 3 C Porous Coke: Grahite+Fe 3 C J C v 1 J C v J gas (a) Grahite Fe 3 C J C J C a C (gas) a C =1 (b) 7 Figure 3-Schematical diagram showing (a) the rocess of iron dusting and (b) carbon diffusion through Fe 3C articles for continuous grahite nano-tube formation.

8 a Coke Ni b Ni articles c Ni articles Carbon filaments Grahite Ni Figure 4-(a) and (b): Metallograhic and TEM cross-sections, resectively, and (c) TEM surface coke morhologies of ure nickel after metal dusting. 5. Protection against Carbon Corrosion Surface oxide scales lay a critical role in the long term behaviour of ractical alloys exosed to carbon-rich gases. Isotoe exeriments have shown 43 that the solubility of carbon in Cr O 3 and Al O 3 at 1000 C is extremely low. Scales of these oxides are therefore exected to rovide effective diffusion barriers to carbon. Furthermore, most observations suggest that these oxides rovide no significant catalysis of carbon deosition from the gas 44. To remain effective, such a scale must retain its mechanical integrity and chemical stability. In the atmosheres of interest, Cr O 3, Al O 3, SiO, and FeCr O 4 are all stable (and stable with resect to the corresonding carbides), and the usual heat resisting alloy designs are on this basis exected to succeed. However, many alloys nonetheless fail by carburisation or dusting. In the case of dusting, an imortant factor 8

9 is the rather modest temerature involved, and the consequently slow alloy diffusion. For examle, a standard material such as Alloy 800 (about 0 wt. % Cr) dusts raidly at temeratures of o C 45,46. The onset of heat resisting alloy dusting has been observed by Grabke and co-workers 0,,47 under isothermal conditions, and by Toh et al. 4,45,48 under temerature cycling conditions. Selective oxidation of chromium roduces a Cr O 3 scale and a chromium-deleted subsurface alloy region, until local scale damage allows gas access to the metal. If sufficient chromium remains, the Cr O 3 scale reheals; if not, other reactions follow. Commonly, is too low for nickel or iron oxides to form and, instead, carbon enters the alloy, reciitating O chromium carbides. Immobilisation of chromium in this way renders future oxide healing of the surface imossible, and gas access to the chromium-deleted surface continues. The surface is now essentially an Fe- Ni alloy. At high nickel levels it undergoes grahitisation and disintegration in the same way as ure nickel; ferritic alloys form cementite and dust in the same manner as iron. Imroved rotection requires scales which resist sallation and cracking, and which remain imermeable to carbon. Suerior rotection against high temerature carburisation is achieved with alloys containing relatively high silicon levels (about %) lus small amounts of reactive elements. These alloys form a silica sublayer beneath their chromia scales, and maintain good scale adherence as a result of the reactive element effect. Even better resistance to carbon enetration is rovided by alumina scales. Comositions which lead to alumina scale formation can be better suited to coatings than to structural alloys. An alternative aroach to resisting dusting attack (at temeratures where oxide scale formation can be difficult to maintain) is the use of coer alloying to inhibit grahite nucleation. New coer-bearing alloys have recently been develoed for this urose 44,49,50. The dusting erformance under cycling conditions of two such alloys is comared with that of Ni, Ni-0Cr and the high erformance Haynes alloy 14 in Figure 5. Figure 5-Carbon utake during reeated 1 h cycles of exosure to a dusting gas at T = 650 o C. Alloy comositions in wt. %. Reference 50. 9

10 6. Summary and Conclusions Carburisation reactions at a C 1 are successfully described by classical theory: solid-state diffusion of dissolved carbon controls the rate and arabolic kinetics result. Wagner s diffusion theory rovides good quantitative rediction of rates, even for comlex commercial alloys, roviding that no oxide scale forms. Metal dusting reactions at a C 1 involve comlex mechanisms. Iron and ferritic alloys grow a cementite scale by inward carbon diffusion, a rocess which also leads to grahite nucleation within the cementite. Disintegration of the carbide roduces articles which further catalyse the grahite deosition rocess, leading to the accumulation of large amounts of orous coke. Nickel and austenitic alloys do not form cementite, but catalyse the direct nucleation and growth of grahite. This occurs within the metal, disruting its structure and roducing a dust of austenite. In all cases, the major mass transfer rocess is inward transort of carbon, either as a gas secies through orous coke, or as a solute in catalytic material, either metal or cementite. The key to rotection against both carburisation and dusting is therefore blockage of inward carbon transort. In the case of lower temerature dusting reactions, the additional ossibility of reventing the catalytic rocess of carbon release is available. Catalysis can be revented by altering the surface chemistry. Feasible methods of doing this are alloying austenitic materials with coer, coating with coer or tin rich materials, or oisoning with sulhur. Carbon transort can be blocked by a scale of chromia or alumina. Achieving a chemically stable and mechanically resilient scale requires successful alloy design and sometimes high temerature re-oxidation. The alloy design necessary to achieve the necessary oxide erformance can require imlementation as a coating rather than a structural alloy. 7. Acknowledgements Suort from the Australian Research Council and Schmidt & Clemens GmbH (Germany) is gratefully acknowledged. References 1. D.J. Young, High Temerature Oxidation and Corrosion of Metals, First edition 008, Elsevier.. H.J. Grabke, U. Gravenhorst, W. Steinkusch, Werkst. Korros., 7, 91 (1976). 3. A. Schnaas, H.J. Grabke, Oxid. Met., 1, 387 (1978). 4. G.M. Smith, D.J. Young, D.L. Trimm, Oxid. Met., 18, 9 (198). 5. C. Wagner, Z. Elektrochem., 63, 77 (1959). 6. O. Ahmed, D.J. Young, in High Temerature Corrosion and Materials Chemistry II, eds. M.J. McNallan, E.J. Oila, T. Maruyama, T. Narita, The Electrochemical Society, Inc., Pennington, NJ (000), D.J. Young, O. Ahmed, Mater. Sci. Forum, , 93 (001). 8. T. Wada, H. Wada, J.F. Elliott, J. Chiman, Met. Trans., 3, 865 (197). 9. R.P. Smith, Acta Met., 1, 578 (1953). 10. D.R.G. Mitchell, D.J. Young, W. Kleeman, Mater. Corros., 49, 31 (1998). 11. R.H. Kane, Corrosion, 37, 187 (1981). 1. E. Cam, C. Phillis and L. Cross, Corrosion, 10, 149 (1954). 13. W.G. Hubbell, The Iron Age, 157, 56 (1946). 14. O. L. Burns, Corrosion, 6, 169 (1950). 15. P.A. Lefrancois and W.B. Hoyt, Corrosion, 19, 360t (1963). 16. R.F. Hochman, in Proc. 3 rd Int. Cong. Met. Corrosion, University of Moscow Press, Moscow, (1969). 17. R.F. Hochman and M.G. Klett, in Proc. 5 th Int. Cong. Met. Corrosion, NACE, Houston, TX (1974). 18. R.F. Hochman, in Proceedings of Symosium Proerties of High Temerature Alloys with Emhasis on Environmental Effects, eds. Z.A. Foroulis and E.S. Pettit, Electrochemical Society, Pennington, NJ (1977), H.J. Grabke, J. Hemtenmacher and A. Munker, Werkst. Korros., 35, 543 (1984). 10

11 0. J.C. Nava Paz and H.J. Grabke, Oxid. Met., 39, 437 (1993). 1. H.J. Grabke, R. Krajak and J.C. Nava Paz, Corros. Sci., 35, 1141 (1993).. H.J. Grabke, R. Krajak and E.M. Muller-Lorenz, Werkst. Korros., 44, 89 (1993). 3. H.J. Grabke, C.B. Brancho-Trochonis and E.M. Muller-Lorenz, Werkst. Korros., 56, 81 (007). 4. C.H. Toh, P.R. Munroe, D.J. Young, Oxid. Met., 58, 1 (00). 5. J. Zhang, A. Schneider, G. Inden, Corros. Sci., 45, 81 (003). 6. J. Zhang, A. Schneider, G. Inden, Corros. Sci., 45, 139 (003). 7. J. Zhang, D.M.I. Cole, D.J. Young, Mater. Corros., 56, 756 (005). 8. J. Zhang, D.J. Young, Corros. Sci., 49, 1496 (007). 9. J. Zhang, D.J. Young, Corros. Sci., 49, 1450 (007). 30. J. Zhang, P. Munroe, D.J. Young, Acta Mater., 56, 68 (008). 31. J. Zhang, D.J. Young, Oxid. Met., 70, 189 (008). 3. J. Zhang and D. J. Young, Corros. Sci., 51, 983 (009). 33. M.A.A. Motin, J. Zhang, P.R. Munroe, D.J. Young, Corros. Sci., 5, 380 (010). 34. D.J. Young, J. Zhang, C. Geers, M. Schutz, Mater. Corros., 6, 7 (011). 35. J. Zhang, D.J. Young, Corrosion Science, 56, 184 (01). 36. E. Piel, J. Woltersdorf, R. Schneider, Mater. Corros., 49, 309 (1998). 37. Z. Zeng, K. Natesan, V.A. Maroni, Oxid. Met., 58, 147 (00). 38. C.M. Chun, T.A. Ramanarayanan, J.D. Mumford, Mater. Corros., 50, 634 (1999). 39. C. M. Chun, J. D. Mumford and T. A. Ramanarayanan, J. Electrochem. Soc, 147, 3680 (000). 40. P. Szakalos, M. Lundberg and R. Petterson, Corros. Sci., 48, 1679 (006). 41. C. Rosado, M. Schütze, Mater. Corros., 54, 831 (003). 4. D.J. Young, M.A.A. Motin, J. Zhang, Defects and Diffusion, 89-9, 51 (009). 43. B. A. Baker, G. D. Smith, Corrosion 000, NACE, Houston TX, Paer 57 (000). 44. Y. Nishiyama, K. Moriguchi, N. Otsuka, T. Kudo, Mater. Corros., 56, 806 (005). 45. C.H. Toh, P.R. Munroe, D.J. Young, K. Foger, Mater. High Tem., 0, 19 (003). 46. H.J. Grabke, Mat. High Tem., 17, 4, 483 (000). 47. H. J. Grabke, R. Krajak, E. M. Müller-Lorenz and S. Strauss, Mater. Corros., 47, 495 (1996). 48. C. Toh, P.R. Munroe, D.J. Young, Mater. High Tem., 0, 57 (003). 49. T.A. Ramanarayanan, C.M. Chun, J.D. Mumford: US Patent 6,737, P. Seck, D.J. Young, US Patent 13/55,93. The Author David John Young, Professor, University of New South Wales David Young is Emeritus Professor in the School of Materials Science and Engineering, UNSW. He has worked in the field of high temerature corrosion for more than 40 years, ublishing extensively in the research literature. His early career was in Canada (University of Toronto, McMaster University, National Research Council of Canada). Returning to Australia, he worked for BHP Steel Research, then joined UNSW, first in Chemical Engineering, and later in the School of Materials Science & Engineering, where he was Head for 15 years. His resaearch on high temerature corrosion has focused on comlex gas mixtures, articularly the effects of carbon, sulhur and water vaour. He has ublished two books, High temerature oxidation and corrosion of metals, Elsevier (008), and (with J.S. Kirkaldy), Diffusion in the condensed state, Institute of Metals (1988). 11