April 13, 2007

High Temperature Corrosion of Superheater Materials Below Deposited Biomass Ashes in Biomass Combusting Atmospheres

By Cha, S C

The aim of this work is to investigate the corrosion behaviour of superheater materials below deposited ashes in biomass combusting atmospheres. The materials were exposed in three different settings (without deposits, with filter ash and with cyclone ash). Exposure tests were carried out with variation of water and HCl content for 360 h at 535C. The HCl addition led to higher mass losses of materials than those without HCl addition. The rate of mass loss is proportional to the deposition aggressiveness and water content in the HCl containing atmospheres. NiCr rich superalloys are more resistant than austenitic and ferritic steels.

Keywords: Biomass, Ferritic steels, Austenitic steels, NiCr based superalloy, Water, HCl effect


The use of biomass fuels for power generation is increasing as a result of the attempt to reduce the emissions of greenhouse gases to the atmosphere. Biomass is a renewable resource with almost zero net CO2 emission since carbon and energy are sequestered during the biomass growth.1,2 The amount of biomass used in production is huge and geographically widespread, compared with other renewable resources.3 Furthermore, biomass is an easy to use energy source considering the present technical level and economics. Therefore biomass has been the focus of many countries for sustainable energy production. The goals for 2010 for energy production through combustion in the EU are not only to increase the efficiency of power plants but also to increase the total amount of energy originating from renewable sources from 6 to 12%. In the long term, biomass is expected to have a 20% share of the current primary energy supply.4,5 One of the problems in achieving these aims is corrosion of superheaters, caused by the deposition of alkali chloride containing ashes.6,7 Biomass combustion is known to produce a rather corrosive flue gas. This is because of the large amounts of fly ash and volatile alkali species, which are produced and form deposits on cooled boiler surfaces. Data on typical fuel compositions and on ash formation were reported in a number of works.8-12 The problem of high temperature corrosion is a crucial impedinent for the commercialisation of biomass fired power generation. In order to reduce corrosion, the maximum steam temperatures are kept considerably lower than in fossil fuel fired power plants. Using highly alloyed steels is another way to prevent superheater corrosion in biomass combustion. From this it follows that there is a need to investigate the resistance of different superheater materials to high temperature corrosion in biomass combustion plants and to study the corrosion mechanisms.

Detailed studies on the first local reactions of alkali chloride particles with iron, chromium and nickel surfaces have been reported.7,13-16 The reaction of KCl particles with iron in a N^sub 2^-20 vol.-%O^sub 2^ atmosphere at 300C showed the deformation and local spreading of the particles, probably by melt formation in contact with the iron. On addition of HCl, a significant increase in the chlorine and oxygen content of the KCl deposited sample surfaces was observed. The increasing chlorine potential of the gas resulted in an increase in iron chloride formation, thus leading to the formation of low melting eutectics in the KCl-FeCl^sub 3^ system. The results showed that especially in the presence of HCl in the combustion gas and at relatively low temperatures, KCl particles led to the formation of a chlorine rich reaction layer on iron, which would promote the adhesion of further salt particles from the flue gas. Furthermore, the reactions on KCl deposited nickel surfaces with HCl led to primarily local formation of chlorine rich nickel oxide, probably NiO/NiCl^sub 2^ mixtures and after longer reaction times NiCl^sub 2^ was found to evaporate significantly. Compared with iron, the nickel surface is more stable, even at 500C. In addition, KCl reactions with chromium showed that the KCl particles reacted at 500C in N^sub 2^-20 vol.-%O^sub 2^ locally with the chromium surface, thus leading to enhanced oxidation. On adding HCl, the reactions led primarily to the formation of a crust like chlorine rich layer and more chloride is formed with increasing reaction temperature. The effect can also be explained by local melt formation between KCl and the chromium chloride to form low melting eutectics of the KCl-CrCl^sub 2^ system (462-475C). Above 600C, the chlorine content of the surface decreases with increasing reaction time owing to oxidation. Thereby, the chloride diffuses through the oxide layer to the chromium/oxide interface, i.e. active oxidation occurs.

1 Phase stability diagram of (Fe, Cr, Ni, Mn, Mo)-O-Cl system at 535C

Biomass includes a large variety of fuels with different chemical composition and combustion characteristics.17 Biomass has a high content of alkali metals and chlorine, and most biomass fuels contain very little sulphur. The content ranges in biomass (dry basis) are: potassium 0.2-1.9 wt-%, chlorine 0.1.1 wt-% and sulphur 0.1-0.2 wt-%.18,19 Each 0.1% of chlorine in the flue gases corresponds to around 100 ppm HCl. In biomass, alkali metals exist primarily as organometallics, sometimes in salt form. The salt form is readily released to the gas phase during combustion. The thermodynamic behaviour of potassium during biomass combustion is primarily influenced by chlorine and sulphur. Chlorine increases the volatility of potassium, which is mainly found as KCl and KOH in the gas phase.6,20

Table 1 Composition of filter and cyclone ashes used as deposits, wt-%

Table 2 Reactions, Glbbs free energy changes and partial pressure for metal chlorides at 535C

Table 3 Composition of sample materials used In exposure tests, wt-%

2 Scheme of experimental setup of high temperature exposure tests

Table 4 Atmospheres of high temperature exposure tests

3 Mass losses of tested materials with 22 vol.-% water (effect of HCl addition and comparison of corrosion rates of materials)

For comparison, Table 1 shows the chemical compositions of untreated filter and cyclone ashes from the biomass combustion plant, identified by inductively coupled plasma atomic absorption spectroscopy (ICP-AAS) and chemical analysis. The C content of filter ash is significantly higher than that of cyclone ash, but the S and Cl contents of filter and cyclone ash are in the same range. The cyclone ash has a higher alkali, Fe and Si content than the filter ash. The behaviour of exposed filter and cyclone ashes was reported in detail.21,22 The results showed that in a HCl containing atmosphere alkali chloride could be formed from alkali carbonate and this accelerated the oxidation process. In contrast, in a HCl free atmosphere, alkali carbonates and hydroxides can be formed, so the enhanced oxidation involving the alkali chloride does not occur.

4 Comparison of mass losses for variation of deposition setting in HCl containing atmospheres

5 Comparison of mass losses for variation of water content and effect of HCl addition

The presence of HCl or Cl^sub 2^ in the gas phase accelerates the corrosion rate of superheater alloys. In oxidizing environments this phenomenon is often referred to as active oxidation. There is a general acceptance that chlorides may cause breakdown of the normally protective surface oxide scales. Cl^sub 2^ is formed from HCl with oxygen giving H2O and this reaction is catalysed by metal oxides. When chlorine reaches the metal surface, it reacts with the metal forming metal chlorides. Gibbs free energies for metal chloride formation are strongly negative, which means that the formation of metal chlorides is thermodynamically favoured.23-26 Table 2 shows the Gibbs free energies for metal chloride formation and partial pressures of metal chlorides for Fe, Cr, Ni and Mo calculated by FactSage at 535C, which is selected as the experimental temperature in the present work.

However, metal oxides are stable in the presence of chlorine, if the partial pressure of oxygen is high enough. According to Grabke et al.24,25 the partial pressure of Cl^sub 2^ below chloride containing deposits is in the range 10^sup -10^ 10^sup -13^ bar. The phase stability diagram of the (Fe, Cr, Ni, Mn, Mo)-O-Cl system at 535C is presented in Fig. 1.

The objectives of this work are to determine the mass losses of eight superheater materials in atmospheres typical of biomass combustion and to investigate the corrosion products by scanning electron microscopy energy dispersive X-ray spectroscopy (SEM-EDX).


Eight superheater materials were exposed for 360 h at 535C in different environments: samples with no deposits, with filter ashes and with cyclone ashes. The test materials were high temperature ferritic steels (10CrMo9 10, X10, X20), which have Ni-Cr contents of

Exposure tests were carried out in atmospher\es corresponding to the composition of flue gas produced from biomass combustion. Eight experiments were conducted in N^sub 2^-5O^sub 2^-13CO^sub 2^ (vol.- %) with variation of the water content (5, 10, 15 and 22 vol.-%) with and without addition of 200 vppm HCl, given in Table 4.

The experimental set up of high temperature exposure tests makes it possible to investigate the corrosion behaviour of materials under well defined corrosive atmospheres. The equipment is composed of two saturators, a furnace and several gas pipelines, containing nitrogen, oxygen, and carbon dioxide (Fig. 2). The gas flow is controlled by capillary cylinders and the gas is dried in a P^sub 2^O^sub 5^ filled pillar. Two N^sub 2^ gas lines are used to add HCl and H2O and pass through saturators containing H2O and HCl respectively. To avoid condensation these gas lines are heated directly after passing through the saturators until they reach the furnace. The third gas line containing nitrogen is used to produce the defined composition of the favoured atmosphere in the balance gas. In front of the furnace the oxygen and carbon dioxide gases are mixed with other gases and allowed to flow into the furnace. The furnace is a horizontal quartz cylinder of 1 m length and can be heated at a heating rate of 5C min^sup -1^ up to 1000C. The exhaust gas flows through a condensation vessel to the gas collector. The corrosive medium flows at a volume flow rate of 2.5 10^sup -3^ s^sup -1^.

6 Corrosion resistances of tested materials in atmospheres of biomass combustion

7 Images (SEM) of corrosion products on 10CrMo9 10

8 Images (SEM) of corrosion products on X10

The samples were cut to approximately 10 10 2 mm and ground with 600 grit SiC paper. Afterwards, the samples were cleaned and degreased with ethanol in an ultrasonic bath and subsequently weighed using a balance with 0.1 mg resolution. During exposure the samples in the furnace had three different compositions: samples with no deposit, with filter ash and with cyclone ash. About 50 mg cm^sup -2^ of ash was used on average and it was subsequently put in three different quartz holders. Four pieces of each material were prepared for determination of mass loss and investigation by SEMEDX. Two samples were deposited with cyclone ashes, one for mass loss and the other for SEM-EDX of the surface and the cross-section. The other two samples with no deposit and with filter ash were only for the determination of the mass loss.

After the exposure experiment, the corrosion products were removed by chemical treatment (pickling in an alkaline KMnO4 solution followed by inhibition using hydrochloric acid) and subsequently the mass loss was determined. The analysis of the corrosion scales was conducted by looking at surfaces and subsequently metallographic cross-sections, which were ground and polished dry to prevent dissolution of the formed chlorides.

9 Images (SEM) of corrosion products on X20

10 Images (SEM) of corrosion products on Esshete 1250

Results and discussion

Mass losses of tested materials in high temperature exposure tests

The mass losses of materials which were tested with 22 vol.-% of water are shown in Fig. 3. As the results show, the HCl addition led to higher mass losses than those without HCl addition. The corrosion rate of materials decreased from ferritic steels to austenitic and Ni based superalloys. It is well known that Ni based materials are more resistant than Fe based materials because chromium and nickel alloying increases the corrosion resistance of steels in chlorine containing oxidising atmosphere.27

The mass losses of materials in HCl containing atmospheres are compared to identify the effect of different settings in Fig. 4. The mass losses with deposition of cyclone ashes are higher than those with deposition of filter ash, followed by those of no deposit. It is reported that corrosion of superheaters is even more severe below chloride containing ash deposits than in chlorine containing atmospheres without deposits.23,24

In Fig. 5, the effect of water variation and the difference of HCl presence are presented. In the HCl free atmospheres (Fig. 5a), the mass losses increased from no deposit to filter ash and cyclone ash for the materials between 10CrMo9 10 and Esshete 1250, but this behaviour is not consistent between TP 347 H and Alloy 625. In contrast, in the HCl containing atmospheres (Fig. 5b) the effect of water increasing from 5 to 22 vol.-% shows clearly that increasing of water content increases the mass losses. A large amount of water vapour in flue gases accelerates corrosion processes at elevated temperatures.28 The effect of different deposits is also evident, i.e. the mass loss depends on the amount of aggressive elements in the deposit.

In summary, the corrosion resistance of the tested materials in biomass combustion atmospheres is compared in Fig. 6. Average rates of mass losses are determined concerning the variation of water content, the HCl addition and the different settings of deposition. The Cr content of 10CrMo9 10, X10 and X20 is 2.3, 8.7 and 10.3 wt-% and the Si content 0.23, 0.38 and 0.23 wt-%. It is known that alloying with chromium27,29 and silicon27,30,31 increases the corrosion resistance of steels, so the corrosion resistance of 10CrMo9 10 was the lowest, followed by those of X10 and X20. The corrosion resistance of TP 347 H was higher than that of Esshete 1250. The NiCr content of TP 347 H is ~4 wt-% higher than that of Esshete 1250, but the Mo content of TP 347 H is zero and that of Esshete 1250 ~1 wt-%. Addition of molybdenum has a positive effect on the corrosion resistance of alloys in chlorine containing environments.31 The compensation of the effects of NiCr and Mo content led to small differences of mass losses between these both materials. The corrosion resistance of Alloy 625 was the highest, followed by Sanicro 28 and AC 66. This can be explained, in addition to the NiCr content, by the Mo content of these superalloys (Alloy 625: 8.9 wt-%, Sanicro 28: 3.3 wt-%, AC 66: 0-1 wt-%). In conclusion, the corrosion resistance of material increases mainly in line with increase in NiCr content.

SEM of corrosion products in high temperature exposure tests

One piece of each sample material deposited with cyclone ash was additionally exposed for SEM-EDX analysis. The corrosion scales were analysed by observation of surfaces and subsequently cross- sections. Figures 7-12 show the corrosion products of each tested material. The composition is described under each figure: main elements are written in bold according to the quantity obtained from EDX analysis and the byproducts are mentioned in bracket.

11 Images (SEM) of corrosion products on TP 347 H

Selected SEM images of corrosion products on 10CrMo9 10 are shown in Fig. 7. Cross-section image (Fig. 7a) of 10CrMo9 10, tested in HCl free atmosphere, shows the simple formation of iron oxide in the outer layer and mixed oxide layer with metal components in the inner layer. Iron oxides with chromium (Fig. 7b) and iron oxides in the form of whiskers (Fig. 7c) were formed on 10CrMo9 10 in the HCl containing atmospheres. The cross-section image (Fig. 7d) shows typical formation of a multi-oxide layer in HCl containing atmospheres. From the outer to the inner layer the formation of iron oxide changed to that of oxide layers mixed with metal components. The chloride is locally formed at the interface owing to active oxidation. In addition, the oxychloride formed with Fe and Mn is shown in Fig. 7e.

The SEM images of corrosion products on X10 are displayed in Fig. 8. In the HCl free atmosphere, oxidation did not lead to severe metal consumption, e.g. simple iron oxide in the form of whiskers is only locally formed on the base material (Fig. 8a). In contrast, the corrosion scales of ferritic steels in HCl containing atmospheres are lamellar and porous,20,31 hence they are spalled completely or partially by the preparation for SEM measurement. In spite of that, the reaction between cyclone ash and metal components was detected, as seen with alkali contents (Fig. 8b).

Corrosion products on X20 are shown in Fig. 9. In HCl free atmospheres, mixed oxides together with resident ash components are formed (Fig. 9a) and the cross-section image shows that the iron oxide is simply formed in the outer layer and mixed metal oxide layer in the inner layer (Fig. 9b). The thickness of the oxide layers was ~20 m and was significantly lower than that of the HCl containing atmosphere (Fig. 9c) at ~150 m, which was formed owing to HCl enhanced oxidation.

12 Images (SEM) of corrosion products on Sanicro 28 and AC 66

Figure 10 shows selected SEM images of corrosion products on Esshete 1250. The cross-section (Fig. 10a) shows that iron oxides mixed with manganese are formed on the outer layer and from outer to inner layer oxides mixed with metal components are observed. The thickness of oxide layers is ~8 m after exposing in HCl free conditions. Figure 10b shows mixed oxides of Fe, Cr, Ni and Mn in the centre of the image (No. 2) and above and around the mixed oxides, iron oxide interspersed with ashes (No. 1). The formed oxides in the HCl containing atmosphere consist of Fe, Cr, Mn and Ni with cyclone ash (Fig. 10c). Furthermore, water vapour is assumed to cause cracking of protective oxide scales on the austenitic steel.20,32 Figure 10d shows the cracked and discontinuously formed oxide with metal and ash components.

The SEM images of TP 347 H are shown in Fig. 11. In Fig. 11a above the mixed oxide scale seen on the left, mixed iron chromium oxides in the form of whiskers are formed on the right (No. 1). In the HCl containing atmosphere on TP 347 H alkali chromate is formed (Fig. 11b). Similar behaviour was observed by Andersson et al.33 in that potassium chromate, iron chromium oxides were formed and at the interface chlorine was detected. In the cross-section image of TP 347 H (\Fig. 11c) the oxide layers, which have a thickness of ~60 m, consisted of mixed Ni, Fe and Cr oxides and at the interface traces of chlorine are detected. In addition to pure iron oxide, the oxides on TP 347 H were mixed, e.g. iron oxide mixed with chromium (Fig. 11d) and chromium oxide mixed with Fe, Ni and Mn (Fig. 11e).

Figure 12 shows the corrosion products of Ni based superalloys (Sanicro 28, AC 66 and Alloy 625). In the HCl free atmospheres, no corrosion layers were generally formed on these three alloys: their corrosion resistances were high enough in these atmospheres (see Fig. 12a as an example). By contrast, in the HCl containing atmospheres corrosion products are variously formed, for instance mixed Cr and Mo oxides with alkalis in the form of mushrooms were detected on Sanicro 28 (Fig. 12b). The cross-section image of AC 66 with thickness of ~8 m (Fig. 12c), shows Fe and Cr mixed oxides with some remanent ash (Ca) on the outer layer and above the interface layer a Fe, Cr and Ni mixed oxide is formed starting from the metal matrix. Mixed oxides (Fig. 12d) and oxychlorides with Ni, Fe and Cr (Fig. 12e) were formed on the surface of AC 66. In addition, the cracked surface was found, again owing to water vapour.

In summarising, cross-sections of corroded materials in HCl free atmospheres showed the formations of simple oxide and no chlorine containing layers. In contrast, the cross-section of HCl containing atmospheres showed multilayers of complex oxides and spinels with traces of chlorine. In general, more multiple oxide layers were found in low NiCr containing ferritic steels than in high NiCr containing austenitic steels and superalloys: the main oxides detected were Fe-O, Fe-Cr-Mn-O and Fe-Cr-Mo-O on ferritic steels, Fe- Cr-(Mn)-O on austenitic steels and Fe-Ni-Cr-O on Ni based superalloys. In addition, the thickness of the corrosion layer corresponds proportionally to the mass losses26 (X20: ~140 m (Fig. 9c), TP 347 H: 60 m (Fig. 11c), AC 66: 8 m (Fig. 12c)).

Summary and conclusion

The high temperature corrosion of eight superheater materials was examined in exposure tests below filter and cyclone ashes from biomass combustion plant in atmospheres (N^sub 2^-5O^sub 2^- 13CO^sub 2^-(5/10/15/22)H2O-(0/ 0.02)HC1 (vol.-%)) characteristic of biomass combustion. The materials were exposed in three different environments (samples with no deposit, with filter ash and with cyclone ash) for 360 h at 535C and subsequently the mass losses of materials were determined and the corrosion products were analysed by SEM-EDX.

The results of high temperature exposure tests can be summarised as follows:

1. The HCl addition led to higher mass losses of materials than those without HCl addition owing to HCl enhanced oxidation.

2. The rate of mass loss is proportional to the aggressiveness of deposition, i.e. cyclone ash is more aggressive than filter ash and specimens with no deposits have the lowest corrosion rates.

3. Increasing water content led to increase of mass losses in the HCl containing atmospheres, caused by water induced corrosion.

4. Corrosion rate decreased from ferritic to austenitic steels and Ni based superalloys.

5. Corrosion products in HCl free atmospheres were simple oxides without Cl traces, but in HCl containing atmospheres they were multilayers of complex oxides and spinels with chlorine.

6. More multiple oxide layers were found in low NiCr containing ferritic steels than in high NiCr containing austenitic steels and superalloys.

7. The thickness of the corrosion layer formed corresponds proportionally to the mass losses.

The corrosion resistances of superheater materials depend on the Ni, Cr, Si and Mo content, which have positive influence on the corrosion behaviour. The corrosion resistance of 10CrMo9 10 was the worst and those of X20 and X10 were relatively similar owing to Cr and Si content. Because of its NiCr and Mo content TP 347 H was more resistant than Esshete 1250. Finally, Alloy 625 has the highest resistance owing to the Mo and NiCr content, followed by Sanicro 28 and AC 66.


The author gratefully acknowledges the financial support of the European Commission for the project 'CORBI' (Mitigation of formation of chlorine rich deposits affecting on superheater corrosion under co-combustion conditions, target action H, ENK5-CT-2001-00532).


1. T. B. Johansson, H. Kelly and A. N. Reddy: 'Renewable energy sources for fuel and electricity', 170; 2000, Beijing, Petroleum Industry Press.

2. D. J. Gielen, M. A. P. C. de Feber and A. J. M. Bos: Energy Policy, 2000, 29, 291.

3. E. D. Larson: Energy Environ., 1993, 8, 567.

4. 'Energy, environment and sustainable development, Directorate- General for Research - biomass, an energy resource for the European Union,' Community research, European Commission, Brussels, Belgium, 2000.

5. G. Chen, J. Andries. H. Spliethoff, M. Fang and P. J. van de Enden: Solar Energy, 2004, 76, 345.

6. H. P. Nielsen, F. J. Frandsen, K. Dam-Johansen and L. L. Baxter: Progr. Energy Combust. Sci., 2000, 26, 283.

7. S.-C. Cha and M. Spiegel: Corrosion, 2005, 61, (8), 743-750.

8. H. P. Michelsen, F. Frandsen, K. Dam-johansen and O. H. Larsen: Fuel Process. Technol., 1998, 54, 95.

9. M. Montgomery and O. H. Larsen: Mater. Corros., 2002, 53, 185.

10. C. Liu, J. Little, P. Henderson and P. Ljung: J. Mater. Sci., 2001, 36, (4), 1015.

11. P. Henderson, A. Karlsson, C. Davies. P. Rademakers. J. Cizner, B. Formanek, K. Giiransson and J. Oakey: 'Materials for advanced power engineering 2002', Lige, Forschungszentrum Julich, 2002.

12. J. Pettersson, C. Pettersson, H. Asteman, J.-E. Svensson and L.-G. Johansson: Mater. Sd. Forum. 2004, 461-464, 965.

13. S.-C. Cha and M. Spiegel: Mater. Sci. Forum, 2004, 461-464, 1055.

14. S.-C. Cha and M. Spiegel: PowerPlant Chem., 2005, 7, (2), 112- 118.

15. S.-C. Cha and M. Spiegel: Mater. Corros., 2006, 57, (2), 159- 164.

16. S.-C. Cha and M. Spiegel: Corros. Eng. Sci. Technol., 2005, 40, (3), 249-254.

17. H. P. Nielsen, L. L. Baxter, G. Sclippab, C. Morey, F. J. Frandsen and K. Dam-Johansen: Fuel, 2000, 79, 131.

18. L. L. Baxter, T. R. Miles, T. R. Jr Miles, B. M. Jenkins, D. Dayton, T. Milne, R. W. Bryers and L. L. Oden: 'The behaviour of inorganic material in biomass-fired power boilers - field and laboratory experiences'. Vol. II; 1996, Golden, CO, National Renewable Energy Laboratory.

19. B. Sander: Biomass Bioenergy, 1997, 122, 177.

20. M. A. Uusitalo, P. M. J. Vuoristo and T. A. Mntyl: Mater. Sci. Eng., 2003, 346, 168.

21. S. Sroda, M. Mkip, S.-C. Cha and M. Spiegel: Mater. Corros., 2006, 57, (2), 176-181.

22. S.-C. Cha: Mater. Corros., 2006, 53, (2), (in press).

23. K. Salmenoja: 'Field and laboratory studies on chlorine- induced corrosion in boilers fired with biofuels', Academic dissertation. bo Akademi, Turku, 2000.

24. H. J. Grabke, E. Reese and M. Spiegel: Corros. Sci., 1995, 37, 1023.

25. H. J. Grabke, E. Reese and M. Spiegel: Molten Salt Forum, 1998, 5-6, 405.

26. M. A. Uusitalo, P. M. J. Vuoristo and T. A. Mntyl: Corrtis. Sci., 2004, 46, 1311.

27. A. Zahs, M. Spiegel and H. J. Grabke: Corros. Sci., 2000, 42, 1093.

28. P. Kofslad: 'High temperature corrosion'; 1988, New York, Elsevier Applied Science.

29. K. Salmenoja and K. Mkel: TAPPI J., 1999, 82. 161.

30. N. Hiramatsu, Y. Uematsu, T. Tanaka and M. Kinugasa: Mater. Sci. Eng., 1989, 120, 319.

31. Y. Kawahara, N. Orita, M. Nakamura. S. Ayukawa and T. Hosoda: Proc. Conf. Corrosion'99, San Antonio, TX, USA, April 1999, NACE International, Paper 91.

32. V. A. C. Haanappel, T. Fransen and P. J. Geilings: High Temp. Mater. Process., 1992, 10, 67.

33. P. Andersson, M. Norell and R. Gautheron: Proc. 6th Conf. on 'High temperature corrosion and protection of materials'. Les Embiez. France, May 2004, CEFRACOR, ORNL, EPRT, Univ Nancy I.

S. C. Cha*

Max Planck Institute for Iron Research (MPIE), Taubenstr. 4, D- 40479 Dsseldorf, Germany

* Corresponding author, email [email protected]

Copyright Institute of Materials Mar 2007

(c) 2007 British Corrosion Journal. Provided by ProQuest Information and Learning. All rights Reserved.