Oxid Met (2011) 76:23–42 DOI 10.1007/s11085-011-9248-4 REVIEW

Metal Dusting Protective Coatings. A Literature Review A. Agu¨ero • M. Gutie´rrez • L. Korcakova • T. T. M. Nguyen • B. Hinnemann • S. Saadi

Received: 1 February 2011 / Revised: 9 March 2011 / Published online: 20 March 2011  Springer Science+Business Media, LLC 2011

Abstract Metal dusting is a catastrophic form of carburization attack that takes place in carbon-supersaturated gaseous atmospheres, and is most commonly encountered in steam reforming processes such as the production of hydrogen or syngas for ammonia, Fischer–Tropsch and methanol applications. The consequence of metal dusting can be a severe loss of metal from the process units, leading to high-cost maintenance and serious safety issues. The present literature review discusses the latest developments within metal dusting protection of alloys with special emphasis on protective coatings. In the first part of the paper, an overview of the main theories for metal dusting of alloys as well as fundamental studies is provided. In the second part, the paper focuses on the different methods to prevent metal dusting, including surface poisoning, alloying, chemical, mechanical and laser treatments as well as coatings. Particular focus is given to coatings and their composition, and fabrication methods, and a critical analysis of the different materials’ behaviours and the suitability perspectives of deposition techniques are provided. A. Agu¨ero  M. Gutie´rrez (&) Instituto Nacional de Te´cnica Aerospacial, Torrejo´n de Ardoz, 28850 Madrid, Spain e-mail: [email protected] A. Agu¨ero e-mail: [email protected] L. Korcakova  T. T. M. Nguyen  B. Hinnemann  S. Saadi Haldor Topsøe A/S, Nymøllevej 55, 2800 Kgs Lyngby, Denmark e-mail: [email protected] T. T. M. Nguyen e-mail: [email protected] B. Hinnemann e-mail: [email protected] S. Saadi e-mail: [email protected]

123

24

Keywords

Oxid Met (2011) 76:23–42

Metal dusting  Inhibition  Protection  Coating

Introduction Metal dusting (MD) is a type of corrosive disintegration of metals and alloys into fine particles. Materials prone to this form of corrosion attack are iron, nickel, cobalt and their alloys. MD only occurs at high temperatures (approximately 400–800 C) and in carbon-supersaturated gaseous atmospheres [1]. Below 400 C, the thermodynamic potential for MD is high but the slow kinetics prevents it from happening. Above 800 C, the thermodynamic potential for MD is low, hence MD does not occur. The common gases causing MD are carbon-monoxide (CO) and hydrocarbons. Therefore, MD is most commonly encountered in steam reforming processes such as the production of hydrogen or syngas for ammonia, Fischer–Tropsch and methanol applications. At high temperatures, CO and hydrocarbons tend to dissociate on metal surfaces and form carbon [2]. The carbon is then transferred to the solid phase, draws the susceptible metals out of their homogeneous solid matrix, which leads to pitting, general attack, and finally to the breakdown of the materials. The consequence of this corrosion attack can be a severe loss of metal from the process units, leading to high-cost maintenance and serious safety issues. Various industrial failures due to metal dusting have been reported, including among others the Mossgas plant converting natural gas to synthetic transportation fuels [3] and DSM Fertilisers plant producing ammonia [4]. The former failure involved the disintegration of the burner liner in the secondary reformer due to metal dusting, leading to the meltdown of a part of the reformer. The latter failure, albeit less catastrophic, still caused a severe loss of material, which required many repairs and replacements of various parts of the units. The main problem with metal dusting is that unless appropriate protection measures are taken, such as using a sufficiently resistant material or modifying the process conditions, the process equipment will continue to dust, as shown in Fig. 1. This may result in continuous and lengthy downtime, hence great expenses, for the plant in question. This paper reviews the open literature including patents of the methods to protect materials from metal dusting, in particular coatings. The paper begins with a brief introduction to the mechanisms for MD, theoretical studies of carbon formation, MD on pure Ni, and on Ni-based binary systems using density functional theory (DFT) as well as a general discussion on the underlying electronic material properties that determine their susceptibility to MD. Subsequently, the paper focuses on the different methods to prevent metal dusting including surface poisoning, alloying, chemical, mechanical and laser treatments as well as coatings. Finally, a critical analysis of the different alternatives to prevent or retard MD with respect to both the materials behaviour and the suitability perspectives of deposition techniques is provided. Metal Dusting Mechanisms The precursor for MD is C formation from the C sources, which are often CO and CH4. The main reactions for C formation from CO are:

123

Oxid Met (2011) 76:23–42

25

Fig. 1 An example of metal dusting attack on ferrules in a waste heat boiler in a petrochemical plant

1.

The Boudouard reaction: 2CO ¼ C þ CO2

2.

DH1 ¼ 172:4 kJ

ð1Þ

The reduction reaction: CO þ H2 ¼ C þ H2 O

DH2 ¼ 131:3 kJ

ð2Þ

The main reaction for C formation from CH4 is: CH4 ¼ C þ 2 H2

DH3 ¼ 74:6 kJ

ð3Þ

The thermodynamic potential for the above reactions to proceed to the right, i.e., forming C, is represented by the so-called carbon activity, aC, which is calculated for these three reactions as: P2CO PCO2

ð4Þ

PCO PH2 PH 2 O

ð5Þ

PCH4 P2H2

ð6Þ

aC1 ¼ Kp1 aC2 ¼ Kp2

aC3 ¼ Kp3

where Kpi is the equilibrium constant of the corresponding reaction and Pi is the partial pressure of the corresponding gas. When aC is greater than unity, carbon has a potential to form via the corresponding reaction, although the extent of carbon formation may be limited by the kinetics of the process. When aC is smaller than unity, thermodynamics state that graphite should not form. It is noted from the above equations that aC is a

123

26

Oxid Met (2011) 76:23–42

function of the temperature and the partial pressures of the gases involved [5]. In other words, aC is a function of the temperature, the gas compositions and in some cases the absolute pressure of the gas mixture. Several mechanisms have been proposed for MD of Fe and Ni based materials in the literature. Grabke Mechanisms Grabke and co-workers proposed two separate MD mechanisms; one explains how MD occurs in pure iron and one in pure nickel [6]. Even though other mechanisms for MD have been suggested, the Grabke mechanisms are still the most widely accepted mechanisms for MD corrosion to date. In pure Fe, it is proposed that MD occurs in the following five steps [7, 8]: 1. 2.

3. 4.

5.

C is transferred from the gas phase to the solid phase by one or several of the three reactions (1)–(3) mentioned earlier. Fe reacts with the C in the solid phase to form iron carbides, such as Fe3C (cementite), mainly at the surface and to some extent in the grain boundaries. The C diffusivity is low in cementite, thus this layer acts as a barrier for further C intrusion. Graphite therefore starts to grow at the surface of the newly formed carbide layer. The C in the graphite layer is unity by definition. The carbide in contact with the graphite thus becomes unstable at this low a C and begins to decompose back into Fe and C. The freshly released Fe particles are an excellent catalyst for further C nucleation and the process repeats from the first step at much higher rates.

Fig. 2 SEM image of C whiskers with metallic particles, which are visible as white spots at the ends of the C whiskers. Secondary electron (SE) image

123

Oxid Met (2011) 76:23–42

27

In pure Ni, the mechanism for MD is similar to the one described above, the only exception is that nickel carbides are unstable by nature and do not form. The process thus involves direct C growth into the Ni matrix. Figure 2 shows a scanning electron microscope (SEM) image of C whiskers with metallic particles in MD pit of Ni alloys. Competitive Oxidation and Carburisation Mechanism (J. Albertsen) This mechanism was proposed based on a study of MD in two Ni-alloys, i.e. alloy 602CA and Inconel 693, and may be applicable to other Ni-alloys [9]. The study showed that MD pits formed in these alloys all exhibited an annular structure containing oxides and carbides at the bottom of the pits, and a zebra structure of alternating chromia and alumina layers with graphite, Ni and Fe layers inside the pits, as shown in Fig. 3. The author proposed the following steps for MD corrosion: 1. 2.

3.

Carbon is transferred from the gas phase to the material, forming carbides. The newly formed carbides are oxidised, especially at the interface with graphite since the gas there becomes more oxidising when the equilibrium to obtain aC of unity is allowed. The porous oxide layer formed allows further C penetration and results in the disintegration of the material. This leads to the formation of various phases such as a-Cr and c-Ni.

Fig. 3 Image of MD pit with a zebra like structure in Inconel 690. The SEM micrograph was taken with backscattered electron (BSE) detector

123

28

Oxid Met (2011) 76:23–42

Carbon Crystallinity Mechanism (Z. Zeng and K. Natesan) This mechanism was proposed for MD in Ni-based materials [10] and Fe-based materials [11]. Using Raman and X-ray diffraction, the authors showed that the carbon formed on metal surfaces, including Ni, Fe and cementite, exhibits good crystallinity. Coke could not crystallise well from the gaseous phase at low temperatures without catalytic activation. It was then suggested that the driving force for recrystallization of coke and MD is the free energy difference between the cokes with good and poor crystallinities, with Ni being the catalyst. The accumulation of C in the Ni matrix leads to the segregation of Ni particles from the main bulk and thus MD corrosion.

Theoretical Investigations of Metal Dusting and Carbon Formation As a macroscopic phenomenon, MD has not received much focus in theoretical density functional theory (DFT) investigations. However, it has been demonstrated [10, 12, 13] that MD is closely related to carbon growth and that these processes are essentially the same on the molecular level. Therefore, it is instructive to review the latest experimental and theoretical work on carbon formation on metallic surfaces. As carbon growth causes breakdown of industrial steam reforming catalysts, mechanisms for carbon formation on Ni and other transition metals have been studied extensively [2, 14–16]. This is far from being the case for MD, but some of the carbon growth studies have recently been extended to consider MD as they are very relevant, even though the corresponding experiments and calculations were carried out with a different application in mind. Some years ago [16], it was found that carbon deposition on Ni depends on the facet orientation with the (111) facet being the preferred one, and recently [17, 18], it was also demonstrated in a MD experiment that carbon forms preferably in alignment with the (111) facets. In 2004, Helveg et al. [19] presented in situ transmission electron microscopy (TEM) experiments, in which carbon nanofiber growth on a Ni catalyst could be observed on the atomic scale. They observed that carbon nanotubes grow via an interfacial mechanism and with the Ni nanoparticle positioned at the top of the nanofiber. Parallel theoretical work [19, 20] mapped out in great detail possible growth mechanisms and pointed out that while surface and subsurface carbon transport is likely to occur, Ni transport through the bulk crystal is very unfavourable. Furthermore, the special role of the Ni (211) steps, i.e. the sites where the (111) and (100) facets meet, as graphene nucleation sites and the role of the epitaxial match between this edge and graphene were pointed out as essential prerequisite for carbon growth. In a metal dusting context, these observations can be turned around so that blocking of these step sites or preventing epitaxial match between the metal substrate and graphene can prevent metal dusting [21].

123

Oxid Met (2011) 76:23–42

29

Methods for the Elimination or Reduction of Metal Dusting MD can be stopped or slowed down by changing the process conditions; by surface poisoning; by controlling the alloy composition; by modification of the surface that can be done mechanically, chemically and thermally; or by deposition of a coating on the alloy surface. Changes in Process Conditions As already mentioned, MD may take place if the carbon activity aC [ 1 and the process temperature is in the critical range between 400–800 C. Thus, in order to suppress MD, the compositions of the carbon sources can be reduced, the composition of the oxidising gases (steam and CO2) can be increased, or the process temperature can be taken outside the MD critical range. Unfortunately, in many cases such changes of process conditions are not an option due to the process requirements. Surface Poisoning Another possibility of inhibiting MD is surface poisoning. This can be done with elements such as sulphur [22–24] or phosphorus [25], that can be adsorbed on metal surfaces according to the following equations: •

•

S adsorption equation:

H2 S ¼ H2 þ Sads :

ð7Þ

PH3 ¼ 3=2H2 þ Pads :

ð8Þ

P adsorption equation:

Suitable poisoning elements have to adsorb onto the surface more strongly than carbon. Theoretical studies on sulphur adsorption on Ni surface [26] have shown that both carbon and sulphur preferentially adsorb at steps and defects with sulphur binding stronger than carbon. This explains why sulphur is so effective in blocking MD on the surface. Sulphur adsorbed on metal surfaces retards gas carburization by site blocking in two ways: (i) prevents initial carbon transfer from the gas phase to the surface, thereby stopping the entire MD sequence, and (ii) prevents graphite from nucleating on cementite surface, hence allows continuous growth of iron carbide and stops the disintegration of the solid phase [27]. When S protection is no longer needed, a sulphur removal unit will have to be installed in order to prevent catalyst poisoning and sulphidation attack of downstream equipment. An alternative to S poisoning is P poisoning, which has a minimal influence on the downstream processes [25].

123

30

Oxid Met (2011) 76:23–42

Alloying The effects of some elements on metal dusting is described in this paragraph. Some elements are present in the commercially employed alloys, such as Ni and Fe whereas most of the others are used to fabricate model alloys, simply to study their influence on metal dusting, even when it is known that they can have detrimental effects on the mechanical properties of the alloy. Some of these elements are also used to form a type of coating and will be repeated in ‘‘Coatings’’ section. Ni and Fe The Ni/Fe ratio is very important for MD resistance. In many cases, alloys with higher Ni contents show increased MD resistances. Previous experiments at Argonne National Laboratory have shown that high Fe contents in the alloy lead to the formation of Cr, Fe spinels when Cr is present, which are not as protective against MD as Cr oxide [28]. On the other hand the presence of Fe is crucial for the manufacturability, forgeability [29] and weldability [30] of the alloys. Zhang and Young have also shown that the changes in the alloy composition alter the reaction mechanism and moreover, changing composition within the austenite phase alters the rate of graphite deposition [31]. Sn The effect of Sn on Ni has been demonstrated both in catalytic [32] and in metal dusting studies [33]. Regarding the protection mechanism of Sn, inspiration from catalysis can be gained. From X-ray photoelectron spectroscopy (XPS) measurements [34] it was suggested that Sn segregates to the surface and poisons graphene nucleation sites. Also recent experimental and theoretical investigations [35, 36] of Sn additions to Ni demonstrated that the presence of Sn increases the barrier for C– C bond formation. A recent theoretical study [37] focused on the Ni3Sn alloy and investigated the effect of adsorbate-induced segregation in the presence of CO, C and OH species and the resulting surface composition. Special attention is given to the stepped surface, since these are the carbon nucleation sites. It was found that under MD conditions, the surface is predominantly covered by C and CO species. Furthermore, it was found that in the presence of CO and carbon, Ni segregates to the surface, especially at the steps. However, there is a limit as to how much Sn can be accommodated in subsurface layers, and therefore the segregation of Ni to the surface may be limited by the transport of Sn away from the surface. This might imply that in spite of ongoing Ni segregation, the Ni–Sn alloy can retain some of its protection. Protective Oxide Formers For NiAl and FeAl-based materials, the protection mechanism is a very different one, since Al, in sufficient contents, forms a stable, continuous oxide and the Ni–Al material thus, in many situations, is protected by an oxide layer. The material must

123

Oxid Met (2011) 76:23–42

31

therefore be capable of forming an oxide scale on the surface. Moreover, the material should also act as a reservoir of the oxide forming element, in order to immediately regenerate the protective scale if it spalls during operation conditions. The protective oxides on the alloy surface prevent C diffusion into the metal. There are several studies of the NiAl system for MD, but also numerous investigations of Ni–Al materials in an oxidizing environment due to their use as high temperature oxidation protective coatings, for instance in gas and aeronautic turbines. In this context, the interface between NiAl and an Al2O3 oxide layer has been investigated by DFT calculations [38], and it was pointed out that S weakens this interface, whereas Hf and other early transition metal additions strengthen it. It has been demonstrated that for NiAl-based materials to be protective against MD, a certain aluminium concentration in the top layer needs to be kept so that broken oxide can be re-formed. Strauss and collaborators [39] studied the metal dusting resistance of aluminide intermetallics Fe3Al, (Fe, Ni)3Al and Ni3Al. They concluded that Fe3Al and (Fe, Ni)3Al are susceptible to metal dusting whereas Ni3Al is very resistant. They also observed that alloying with Cr reduced the attack in Fe3Al. The effect of other oxide former elements has also been studied. Bayer [40] pointed out that alumina is more stable than chromia in a high carbon activity environment at low oxygen partial pressure. Rahmel and co-workers [41] indicated that the conversion of oxides into carbides is the principal mechanism of degradation by carburization. They indicated that alumina and silica are stable at the operating conditions and its conversion to the corresponding carbides occurs at significantly higher temperatures than that of chromia. Nevertheless, MD tests performed by Special Metals [41] show that the alloys with the highest Cr and Cr ? Al concentrations have the best resistance against MD. In the case of silica, volatile SiO may form at the very low oxygen partial pressures associated with MD [42]. Carbide Formers Strauss and Grabke have shown that the addition of carbide formers, e.g. Mo, W or Nb, prolongs the incubation period for MD [43]. These elements have greater affinity to C than Cr, hence Mo, W or Nb carbides are formed first, leaving Cr in solid solution. Therefore, more Cr is available for formation of protective Cr oxide and the MD attack is then postponed. Cu Recently Cu has been suggested as an alloying element which helps improve the MD resistance of Ni alloys. Cu reduces the catalytic activity of Ni for CO decomposition and hence retards carbon deposition and MD. It is not difficult to alloy Ni alloys with Cu, since Cu and Ni are fully soluble at all concentrations. MD tests on Ni–Cu alloys with varying Cu contents show little coking when the Cu content is above 10 wt% [44]. Zhang and co-workers [45] found that for Cu contents above 20 wt%, the rate of MD was very low and not sensitive to the alloy

123

32

Oxid Met (2011) 76:23–42

composition whereas for Cu contents below 10 wt%, the rate decreased significantly with increasing the Cu wt%. This behaviour was attributed to Cu interference with sites for graphite nucleation involving several Ni atoms [46]. Copper alloys have recently been reported in the patent literature to be resistant or immune to carburization, metal dusting and coking [47, 48]. Similar results were obtained by Nishiyama et al. [49] who proposed that Cu suppresses CO dissociation (‘‘Surfactant Mediated Suppression’’), one of the fundamental steps in the mechanism of MD as already explained. They also exposed pure Cu, Pt and Ag to metal dusting conditions at 650 C and all exhibited good resistance. Moreover, the same author has observed that Ni30Cr alloys with critical amounts of both Cu and Si showed excellent resistance to metal dusting [50]. These alloys formed a protective M2O3 scale, M being mainly Cr with some Ni, over a discontinuous layer of amorphous SiO2. The role of Cu in the protective mechanism does not seem to be clear. On the other hand, Zhang and collaborators [51] concluded that addition of Cu to 304 stainless steels did not provide benefits due to the limited solubility of Cu in this steel. In steels containing higher Ni levels (310SS and 800H), internal carburization was decreased (higher Cu solubility) and MD was completely suppressed for the alloys with 5 and 10 wt% in Cu. However, for the 20 wt% Cu alloys, Cu rich phases developed causing internal carburization at the Cu-c phase interphases [44]. Recently, Zhang and Young [52] confirmed that Cu does not affect the C permeability in Ni as it was earlier proposed [53] but rather interferes with the nucleation of graphite. Precious Metals It has been shown for Ni based steam reforming catalysts that the addition of Au or Ag decreases the carbon formation rate [54, 55]. Combined experimental and theoretical studies show that Au preferentially adsorbs and diffuses into steps and defects in Ni, instead of flat surfaces. Furthermore, in order for graphite to form, several adjacent Ni atoms are required. Adding Au destroys this periodicity, and as Au is preferentially located at the surface, small additions have a large effect [54]. Mechanical and Chemical Surface Treatments Cold working has a positive effect in increasing the resistance to metal dusting of chromia scale formers, as it results in grain refinement and thus provides easy diffusion paths for Cr. Moreover, for steels subjected to treatments that introduce near-surface deformation, such as grinding, sand blasting, shot-peening and machining, Cr diffusion increases due to dislocation pipes. Grinding is better than cold rolling which in turn is better than electro-polishing. On the other hand, pickling should be avoided as it can strongly accelerate MD by attacking the grain boundaries [56]. The establishment of a dense, well adhered Cr2O3 layer by controlled preoxidation at low Po2 is advantageous [57]. In an alternative approach, Zeng and Natasen proposed a method to mitigate MD by blocking already existing MD pits by

123

Oxid Met (2011) 76:23–42

33

employing an intermediate oxidation treatment [58, 59]. For instance, when prepitted Incoloy 800 was treated by gas mixtures containing steam and H2 at 760 C for 6 h, the incubation time was greatly extended. The main oxide phase found under these conditions is Cr2O3. This mitigating approach can be applied, for instance, in hydrogen reformer plants during schedule maintenance. Thermal Treatments Laser surface melting was employed by Voisey and collaborators to improve the MD resistance of Incoloy 800H [60]. A refined microstructure was created which resulted in an increase in the density of rapid Cr diffusion paths, therefore promoting the formation and maintenance of a protective scale. The samples were treated with a 2 kW CW CO2 slab laser and were subsequently ground with 600 grit SiC paper to remove laser induced oxides, before exposing them to a 20% H2:80% CO2 atmosphere at 650 C. The degree of carbon filament growth on these samples was lower than that observed on non-treated Incoloy 800H. The extent of the microstructural refinement can be controlled by varying the laser parameters. Coatings In general, the protection mechanism provided by coatings is based on the formation of a thin, stable, protective and adherent Al, Cr or Si oxide scale as already mentioned. Depending on the temperature and the chemistry of the corrosive environment, the degradation mechanism may be caused by coating spalling, coating attack because the protective oxide does not form or it is not sufficiently stable, or may be essentially based on the loss of the oxide forming element as it reforms the protective scale, or due to interdiffusion with the substrate. Two types of coatings are usually considered for MD protection: • •

Diffusion coatings. Overlay coatings.

Diffusion Coatings These coatings are formed when an alloy is coated with a metal or mixture of metals, at sufficiently high temperatures to cause interdiffusion between the coating and the substrate. This results in a metallurgical bond with the substrate material, and the coating then becomes an integrated part of the substrate material. Diffusion coatings are obtained in general by thermochemical methods such as pack cementation or chemical vapour deposition (CVD). Both methods involved the production of volatile species which, when in contact with the hot component, react to form atoms that will in turn react and/or interdiffuse with the substrate components. In pack cementation the component is ‘‘buried’’ and therefore in contact with a mixed powder bed, whereas in CVD the component is only exposed to gas phase precursors. Other methods, such as electro-deposition, physical vapour

123

34

Oxid Met (2011) 76:23–42

deposition (PVD) or even thermal spray, can also be used to generate a diffusion coating in a two step process: (1) a substrate independent layer (overlay) is deposited by any of these deposition techniques and (2) the coated component is subjected to a diffusion heat treatment. Aluminium diffusion coatings are currently employed for high temperature oxidation and corrosion protection in the chemical industry and in the hot section of power generation and aeronautic gas turbines as already mentioned. When modified with Cr deposited by pack cementation or Pt deposited by electro-deposition, the lifetime and the operating temperature of the coated alloy can be increased considerably [61]. This type of coating has also shown very good behaviour in protecting steels as well as Ni base alloys from MD. –

–

–

Baumert et al. [62] studied the behaviour of aluminide diffusion coating on stainless steel employed in a methanol reforming unit at temperatures within 565–1150 C. In general, 120 lm aluminide coatings provided good resistance for up to 14 years in service with only occasional and minor pitting observed. Coatings with a 60 lm thickness were penetrated within 13 years. Other studies also concluded that aluminized steel offered good protection in metal dusting environments for several years of service [63]. Winns and Bayer [64] developed a two step pack cementation diffusion coating incorporating Al, Cr and Si in an optimized combination, to limit carburization, metal dusting and the harmful effects of catalytic coke build-up on austenitic FeCrNi alloys. The coating must have a thickness greater than 200 lm. Si stabilizes the aluminide phase and retards Al loss by inwards diffusion and an inner Cr–Si enriched phase ([40 wt% Cr) provides a stabilizing diffusion barrier. They also pointed out that Si aids in lowering the diffusivity of carbon from the process environment into the base alloy. This coating was applied to HK-40 and HP–Nb micro-alloyed tubes placed in an ethylene furnace. The coated tubes could be welded and the coating withstood 27 months in service at temperatures up to 1149 C, providing substantial improvement in carburization resistance when compared to un-coated tubes exposed to similar conditions [65]. The microstructure of the as-deposited coating exhibits a 400 lm layer with two distinct bands, a top Al rich layer with 36 wt% Al and an inner diffusion zone enriched in Cr (30 wt%) and Si (over 2 wt%) which, according to the authors, enhances the stability of the Al rich outer layer. The microstructure of the exposed samples showed thicker layers, indicating that interdiffusion continued, and the Al content at the top had decreased to 16–20 wt%. The authors claim that these coatings could provide protection for a longer period of time. However, in zones where overheating took place (1177–1204 C) the coating reached 650 lm and the top zone had an Al content of only 5 wt%, indicating that the coating would not be protective much longer. The behaviour of siliconized 310 steel was studied in a 1.5 atm 10/1 CO/CO2 atmosphere at 700, 750 and 800 C [66]. The 100 lm coating was applied by pack cementation and exhibited a top layer (10 lm) with 11 wt% Si separated by pores from a thicker layer with 4–5 wt% in Si. Before exposure to the test

123

Oxid Met (2011) 76:23–42

–

–

–

–

35

atmosphere the top layer was removed. For a total of 168 h at 800 C the siliconized specimen experienced significantly less weight gain than the uncoated 310 stainless steel. However, small, unidentified, spherical particles could be observed on the surface of the siliconized specimens. At Argonne National Laboratories [67], Al and Cr/Si coatings were deposited at temperatures ranging from 700 to 1000 C on Fe alloys. Ni based alloys were only aluminised as the authors claimed that it was not feasible to enrich them with Cr/Si. The authors indicate that the 200 lm coating on Incoloy 800 consists of a top Al enriched layer over Ni3Al. Before testing at 593 C in MD atmospheres, the coated specimens were first oxidised in air at 900 C and later in wet air (23 vol% H2O) at 593 C. In general, the pre-oxidised surfaces were intact after exposure to metal dusting (52 H2, 18 CO, 5.6 CO2, 1.1 CH4 and 23 H20 in vol%) for approx. 300 h. Rosado and co-workers [68] studied Al, Cr and Si single element diffusion coatings; a two step chromizing-aluminizing process; and Si–Al, Ti–Al and Ti–Si–Al co-diffusion coatings on ferritic and austenitic steels, including Incoloy 800. They concluded that the properties of the coatings depend on the microstructure and the composition of the substrates and that the best coatings were different for each substrate. In general, plain aluminide coatings were protective for all alloys when tested at temperatures up to 700 C and for at least 2000 h in some cases. However, pores and cracks developed within these coatings. In case of the two step coating, the chromizing step caused the formation of an increased Cr content barrier in the metal-coating interface similar to the results obtained by Wimms [64]. Again, it was shown that this barrier retards the loss of Al by substrate-coating interdiffusion. The pure Si diffusion coatings exhibited poor behaviour at 400 C as a continuous protective oxide could not be formed. Si aluminides were better than the pure aluminide except at 620 C for two of the alloys. A detrimental effect attributed to the formation of a poorly adhered mixture of Al and Si oxide was observed instead of a stable, protective layer. However, for Incoloy 800 (richest in Ni from all studied alloys), Si was beneficial as pores were suppressed and a diffusion barrier rich in Cr–Si was formed under the aluminide layer. TiAl was good only for Incoloy 800 as TiC is formed in other alloys containing C in the substrate, which turned out not to be resistant to metal dusting. TiSiAl was good for all alloys tested. The authors indicate that Ti helps by promoting the formation of a continuous alumina scale even at a temperature range where it does not easily form. However its beneficial effect is limited to low C or high Cr materials. HTAS [69] has patented a method to protect Ni, Fe and Cr alloys from carburization and metal dusting by applying thin layers of a noble or precious metal or of an element from group IV such as Sn or Pb by any method such as PVD, CVD, dipping, spraying, electroplating, etc. and subsequently subjecting the coated component to a heat treatment under inert atmosphere to obtain a uniform distribution of the metal. Schu¨tze and collaborators [70] have produced Sn–Ni coatings by a two stage process including electro-deposition of Ni followed by pack cementation on

123

36

Oxid Met (2011) 76:23–42

steels or by simple pack cementation of Sn on Ni base alloys. Pack cementation was carried out at 650 C and 800 C. The Sn–Ni coating is stable up to 1200 C against coking and metal dusting. However, oxygen can diffuse through this phase and if there are oxide formers in the alloy (Al or Cr), adhesion can be weakened by the formation of oxide layers. Overlay Coatings Overlay coatings have little chemical interaction with the substrate except in areas close to the coating-substrate interface, in which some degree of interdiffusion may occur. Therefore the coating composition is not significantly affected by the substrate composition. Overlay coatings are in general alloys or intermetallics and can be deposited by a number of techniques including thermal spray, PVD, CVD, electro-deposition, etc. For instance, MCrAlY coatings deposited by plasma spray, high velocity oxy-fuel thermal spray (HVOF) or PVD are also being used to protect nickel superalloys in gas turbines from oxidation and hot corrosion. Their lifetime is at least 10 times that of the aluminide coatings previously discussed, and 3–4 times longer that the Pt aluminides. However their cost is significantly higher. Overlay coatings used to protect alloys from metal dusting have been reported in several studies: –

–

–

–

Ni50Cr coatings were applied by arc spray and by HVOF on waste heat boiler inlet tube sheet in the Mossgas gas-to-liquid fuel plant (South Africa) [71]. The arc-sprayed coatings exhibited localized spalling and detachment in many areas after several months in operation but the HVOF deposited coating was very protective for up to at least 17 months with only very minor and isolated delamination in sharply machined parts such as the edges of drill holes due to poor coating coverage. These areas could be repaired by local grit blasting and re-spraying. The authors also point out that the hole edges can be machined to a curved finish instead of a corner so they can be better coated. c-TiAl was deposited by HVOF by Rosado and co-workers [68] on two different steels. The coating was extremely resistant against metal dusting when deposited over a ferritic steel but on an austenitic steel, it failed due to thermal expansion coefficient mismatch. Al and Ni5Al layers were deposited by plasma spray and exposed to a carburizing atmosphere containing a mixture of H2 and hexane (ac & 3) at 950 C [72]. Prior to testing the coatings were subject to a pre-oxidation and diffusion heat treatment at 950 C for 10 h. The Ni5Al coating was totally oxidised after only 10 h whereas the Al coating formed an aluminide coating similar to those described previously and did not show evidence of carburization or oxidation. Wang and collaborators [73] deposited NiAl by HVOF with and without small amounts of CeO2 and/or Cr. The coatings were characterized by X-ray diffraction confirming NiAl as the sole intermetallic phase after deposition, except a 3–30 lm NiAl layer which also exhibited minor amounts of Ni3Al. The coatings were tested under a 2 vol% CH4 atmosphere in H2 at 1100 C (ac = 3).

123

Oxid Met (2011) 76:23–42

–

–

–

–

–

37

No evidence of carburization was found after 100 h for coatings with addition of Cr and CeO2. However, the pure NiAl layer was ineffective in blocking the ingress of C due to its poor integrity (and probably to the lack of sufficient thickness—only half the thickness compared to coatings with Cr and CeO2). Cr and CeO2 seem to be beneficial in forming coatings with less cracks and pores. A moderate to high temperature process, which is used to deposit SiO2 by means of CVD and employs alkosilanes as a gas phase precursor, was patented by BP [74]. The coating reduces coking in a steam-ethane atmosphere within the 890–920 C temperature range. Phillips Petroleum [75] also patented a CVD process to deposit SiO2 but in this case from halogen containing silanes, disilanes or siloxanes in liquid phase, at lower temperatures and for shorter periods of time. Coated specimens were exposed to a steam, ethylene and N2 mixture for 3 cycles of 20 h each, at 900 C and exhibited significantly less coking that bare Incoloy 800. A mixture of SiO2, BaO, CaO and Al2O3 was: (1) melted at 1420 C, (2) ground into a powder, (3) applied as a slurry to steel and (4) sintered at 1200 C for 5 min [76]. The resulting glass coating was exposed to a cracked propane (1600:25 and 1200:50 propane: N2 ml/min ratios) atmosphere at 800 and 850 C for 30 min. The defect and crack free coating exhibited a 40 lm enamel layer composed of Si, Ba, Al and Ca oxides on top of a ‘‘transition’’ layer on which some Fe and Cr from the substrate could be observed. This layer prevented catalytic coking and there was no evidence of carbon permeation through the coating. CoNiCrAlY has been deposited by HVOF on Inconel 601 and alloy 602 CA and tested in metal dusting conditions (ac [ 1) at 700 C for 850 h [77]. The 200 lm coating is composed mainly of the phase c-Ni/c‘-Ni3Al as a matrix in which bNiAl and b-CoAl are also present. On exposure to metal dusting conditions, the coating developed a protective aluminium oxide scale. Plasma sprayed Ni50Cr and NiCrAlY on Incoloy 800H and Inconel 600 were tested in a metal dusting environment at 650 C (20% H2–80% CO) for 2000 h [78]. The Ni50Cr coating exhibited extensive metal dusting already after 579 h of exposure whereas the nominally 150 lm NiCrAlY failed due to cracking and coating spallation. Both coatings were subject to laser re-melting in order to eliminate the interconnecting porosity typical for thermally sprayed coatings and to improve the coating substrate adhesion. The NiCrAlY coating was characterized by XRD and exhibits mostly a bcc a structure as well as the expected c and b phases (previously defined). After having been subject to laser re-melting, the diffraction pattern remained similar except for the appearance of peaks corresponding to a-Al2O3. The laser treated NiCrAlY improved the MD resistance of the alloys but its performance was limited by defects attributed to incomplete melting and the presence of iron spinel oxides within the laser induced oxide. Moreover, the laser treated NiCr coating performed better than the untreated layer but showed extensive MD, in particular at the specimen edges, and it was attributed again to Fe spinel oxides due to deeper melt at the edges and therefore higher Fe content in the coating.

123

38

–

–

–

Oxid Met (2011) 76:23–42

Liu and Chen deposited Ni3Al and CeO2 with dispersed Ni3Al on Fe35Ni25Cr by electro-deposition and determined their resistance to metal dusting and carburization [79]. Both coatings had an initial thickness of 200 lm and were less resistant to metal dusting than the substrate at 650 C in 73.3% H2, 24.4% CO and 2.3% H20 atmosphere (ac(CO reduction) = 12.2). According to the authors, the test temperature of 650 C is too low to form a protective alumina scale, which explains the coating failure. The behaviour of the coating containing dispersed CeO2 was even worse and was attributed to the formation of NiO/Ni particles on the coating surface or at the grain boundaries, promoted by CeO2. However, under cyclic carburization conditions (2% CH4–H2 atmosphere) and at 1050 C, after a long exposure, the pure Ni3Al coating was attacked whereas the CeO2-dispersed coating exhibited a better behaviour. The authors explain that at 1050 C (ac = 3.21) a-Al2O3 forms, and under cyclic conditions it cracks, allowing the migration of corrosive gases, whereas the presence CeO2 reduced the cracking and therefore improved the coatings behaviour [80]. Moreover, the authors observed that if the Ni3Al coating containing dispersed CeO2 is pre-oxidised at 1100 C in order to form protective Al2O3, less coke formation is observed at 650 C in CO– H2–H2O (ac = 12.2) and the coating was almost free of pitting attack. The effectiveness of this pre-treatment strongly depends on the mechanical integrity of the scale [81]. Al2O3, and Cr2O3 deposited by reactive magnetron sputtering (Al and Cr sputtering in an atmosphere containing O2 and Ar respectively) retarded corrosion of HK-40 in carburizing atmospheres [82–84]. The coatings require an inner metallic layer (Al or Cr) to improve adhesion and its thickness varies from 6 to 10 lm. Metal dusting resistance studies were carried out in a thermobalance with uncoated HK-40 as well as coated specimens (CH4/Ar: 2.86 in Vol) for 50 h at 800 C. The Al/Al2O3 coating was protective initially but developed cracks that allowed the carburizing gas to reach the substrate. In case of the Cr/ Cr2O3 system, if the composition is graded from pure Cr to pure Cr2O3, the coatings mechanical properties are improved and crack formation is avoided. As a result, these coatings with less fractures exhibited less corrosion products. However, even the graded coating, which exhibits a dense columnar morphology, was still permeable to carbon. Schu¨tze and coworkers [70] undertook a project with the goal of developing coatings to poison the metal surface, which acts as a catalyst for coking and metal dusting processes, and to thereby prevent MD. Only a very thin layer of coating is needed for this (in the idealized case, one monolayer would be enough). The goal of the first stage of the project was to identify suitable phases for this approach. In particular, the following phases were PVD sputtered on different alloys: Cr–Al, Nb–Al, Cu–Al, Nb–Sn, Sn–Ti, Ni–Sn. The studied substrates were ferritic alloys 13Cr–Mo4-4, 10CrMo9-10 and P91, austenitic Incoloy 800, and nickel-based Inconel 690. The resulted coatings were heat treated and then exposed to a metal dusting gas atmosphere (CO 24%–H2O 2%– H2 bal). In general, the coatings tended to spall off when exposed to MD conditions, and it could be seen that no strong chemical bonds between the coating and the underlying alloy had been formed. The Cu–Al layer behaved

123

Oxid Met (2011) 76:23–42

39

differently from all the others, as it did not remain mixed, but Cu diffused into the underlying alloy, while Al formed an oxide layer. The conclusions were that Cu and Sn, which are stable towards oxidation, can also be used for protection against metal dusting. Investigations showed that even a thin layer of 2–5 lm could improve the corrosion resistance of Incoloy 800 significantly.

Summary There is a strong drive to operate deeper into the metal dusting region in order to create more energy efficient processes. Therefore, there is significant interest in protecting metal parts from metal dusting. The key to suppressing metal dusting is to stop the dissociation of the carbon source or subsequent carbon diffusion into the susceptible materials. There are several methods to do this, which have been discussed in the present review: modifying the process conditions, surface poisoning, alloying, chemical, mechanical and laser treatments as well as coatings. Each method has advantages and drawbacks. Changes of process conditions are not an option due to the process requirements in most cases and the use of poisoning chemicals is still limited due to its interference with downstream processes. Newly developed alloys are sometimes not sufficiently resistant and expensive due to the large quantities involved. Mechanical and chemical surface treatments are not always applicable, and the lifetime of coatings is often limited by adhesion problems. There is thus a need to develop all these methods in order to challenge the metal dusting limits. The protection for both Ni and Fe based alloys against MD by surface modification can be achieved in two ways. Firstly, the growth of a protective oxide (Cr2O3 or Al2O3) is enhanced. Secondly, the CO dissociation is minimized by specific metals present in high enough concentration, such as Cu ([20 wt%), Sn and possibly Pt and Ag. Surface mechanical and laser treatments improve metal dusting resistance by increasing the rate of formation of the protective oxide by creating fast diffusion paths. Chemical treatments enhance the oxidation of the substrate surface, and thereby create a protective scale prior to exposure to the MD atmosphere. Using coatings can provide many advantages: (1) no modification of processing conditions is required, (2) the surface treatment alternatives may not work for long exposures and are not applicable to all type of components, (3) existing alloys with good mechanical properties can be used and therefore the existing component manufacturing processes do not require modifications, (4) may allow the deposition of any suitable coating composition independently from those of the substrates and (5) can be repaired/reapplied in many cases. The various coating materials and deposition techniques that have been explored for this application, have advantages and disadvantages, and a number of factors must be taken into consideration when selecting the most appropriate coating and deposition method for a specific component. Some of these factors are: coating type and composition, substrate material, component size/weight and geometry, coating

123

40

Oxid Met (2011) 76:23–42

temperature, capacity for coating ‘‘in situ’’ and cost among others. The coatings that have shown the best performance are those based on NiAl intermetallics, that behave as an Al source in order to develop protective Al2O3. In general, coatings must be dense and free of cracks and porosity. Aluminide diffusion coatings do not have adhesion problems, and fulfil these criteria at a rather low cost. On the other hand, this type of coatings degrade by coating substrate interdiffusion, which results in a reduction of the subscale Al content and therefore the coating lifetime will be decreased. The best alternatives for this type of coatings so far seem to be diffusion aluminides with some Cr, which helps to retard Al loss by inwards diffusion. Regarding the effect of Si on aluminides, there is no general trend, it is beneficial for some alloys, but detrimental for some others. Clearly, a better understanding of the Si effect is required. Recent work carried out with Sn diffusion coatings have also resulted in promising protective behaviour. However, more data and longer exposures are required. Regarding overlay coatings employing HVOF or plasma spray, there is potential for components of simple geometries (without inner surfaces that require to be coated, such as tubes and pipes) provided that high density and adhesion are achieved. MCrAlYs have exhibited good behaviour but are very expensive. CeO2 dispersed NiAl deposited by electrodeposition also have good potential provided that the process is capable of uniformly coating the inner surfaces of long tubes. Glass coatings have also shown good protective behaviour but a very high temperature (1200 C) is required for sintering and there is not enough data regarding long exposure behaviour. As already demonstrated with gas and aeronautic turbines, the use of coatings offer significant advantages for all industries in which MD is an issue. Energy, as well as cost savings are directly and indirectly ensured: less expensive base materials can be used, the lifetime of components is significantly increased, less production interruptions due to component replacement can be achieved, among others.

References 1. H. J. Grabke, R. Krajak, and E. M. Mu¨ller-Lorenz, Materials and Corrosion 44, 89 (1993). 2. J. Rostrup-Nielsen and D. L. Trimm, Journal of Catalysis 48, 155 (1977). 3. H. J. de Bruyn and M. L. Holland, Materials Experience in a Methane Reforming Plant. Corrosion 98 (1998). 4. R. J. Gommans and T. L. Huurdeman, Experience with metal dusting in waste heat boilers. Vol. 35 AIChE Symposium (1994). 5. J. Zhang and D. Young, Corrosion Science 48, 1496 (2006). 6. H. J. Grabke, Materials and Corrosion 54, 736 (2003). 7. R. F. Hochman, in Proceedings of the 4th Interntional Congress on Metallic Corrosion, ed. N. E. Hamner (NACE, 1972), p. 258. 8. H. J. Grabke, R. Krajak, and J. C. Nava Paz, Corrosion Science 35, 1141 (1993). 9. J. Z. Albertsen, Ø. Grong, J. C. Walmsley, R. H. Mathiesen, and W. Van, Metallurgical and Materials Transactions A 39, 1258 (2007). 10. Z. Zeng and K. Natesan, Chemistry of Materials 15, 872 (2003). 11. Z. Zeng, K. Natesan, and V. A. Maroni, Oxidation of Metals 58, 147 (2002). 12. C. M. Chun, T. A. Ramanarayanan, and J. D. Mumford, Materials Corrosion 50, 634 (1999).

123

Oxid Met (2011) 76:23–42 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54.

41

Z. Zeng and K. Natesan, Chemistry of Materials 17, 3794 (2005). R. T. K. Baker, R. S. Harris, R. B. Thomas, and R. J. Waite, Journal of Catalysis 26, 51 (1972). T. Baird, J. R. Fryer, and B. Grant, Carbon 12, 591 (1974). C. Bernardo and D. L. Trimm, Carbon 14, 225 (1976). J. Zhang, P. Munroe, and D. J. Young, Acta Materialia 56, 68 (2008). J. Zhang, C. Kong, and D. J. Young, Materials at High Temperature 26, 45 (2009). S. Helveg, C. Lo´pez-Cartes, J. Sehested, P. L. Hansen, B. S. Clausen, J. R. Rostrup-Nielsen, F. AbildPedersen, and J. K. Nørskov, Nature (London) 427, 426 (2004). F. Abild-Pedersen, J. K. Nørskov, J. R. Rostrup-Nielsen, J. Sehested, and S. Helveg, Physical Review B 73, 115419 (2006). S. Saadi, F. Abild-Pedersen, S. Helveg, J. Sehested, B. Hinnemann, C. C. Appel, and J. K. Nørskov, Journal of Physical Chemistry C 114, 11221 (2010). J. R. Rostrup-Nielsen, Journal of Catalysis 85, 31 (1984). A. Schneider, H. Viefhaus, and G. Inden, Materials and Corrosion 51, 338 (2000). A. Schneider, G. Inden, H. J. Grabke, Q. Wei, E. Pippel, and J. Woltersdorf, Materials Technology Steel Research 71, 179 (2000). P. Abbott, J. Edwards, J. B. Crewdson, and M. Fowles. Metal passivation in a heat exchange reformer. Patent, WO 03/051771 (2003). F. Abild-Pedersen, O. Lytken, J. Negbæk, G. Nielsen, I. Chorkendorff, and J. K. Nørskov, Surface Science 590, 127 (2005). H. J. Grabke and E. M. Mu¨ller-Lorenz, Steel Research 66, 254 (1995). Z. Zeng, K. Natesan, and M. Grimsditch, Corrosion 60, 632 (2004). C. T. Sims, N. S. Stoloff, and W. C. Hagel, Superalloys II, (John Wiley & Sons, Inc, New York, 1987), p. 169. M. J. L. Oestergaard, L. Korcakova, W. R. Watkins, J. J. Hoffman, E. Slevolden, and E. Ford, Advanced metal dusting resistant alloy—a Materials Technology Institute project, to be published at Stainless Steel World 2010. J. Z. Zhang and D. J. Young, Oxidation of Metals 70, 189 (2008). D. L. Trimm, Catalysis Today 49, 3 (1999). I. Alstrup and I. Chorkendorff, US Patent 6,602,355. C. Padeste, D. L. Trimm, and R. N. Lamb, Catalysis Letters 17, 333 (1992). E. Nikolla, A. Holewinski, J. Schwank, and S. Linic, Journal of the American Chemical Society 128, 11354 (2006). E. Nikolla, A. Holewinski, and S. Linic, Journal of Catalysis 250, 85 (2007). S. Saadi, B. Hinnemann, S. Helveg, C. C. Appel, F. Abild-Pedersen, and J. K. Nørskov, Surface Science 603, 762 (2009). K. M. Carling and E. A. Carter, Acta Materiala 55, 2791 (2007). S. Strauss, R. Krajak, M. Palm, and H. J. Grabke, Materials and Corrosion 47, 701 (1996). G. Bayer, Corrosion 2001, NACE International, paper No. 01387 (2001). A. Rahmel, H. Grabke, and W. Steinkush, European Federation of Corrosion Workshop, Frankfurt, November 1997. B. A. Baker, G. D. Smith, V. W. Hartmann, L. E. Shoemaker, and S. A. McCoy, Corrosion NACE 2002, paper No. 02394 (2002). S. Strauss and H. J. Grabke, Materials and Corrosion 49, 321 (1998). V. B. Trindade, U. Krupp, E.-G. Wagenhuber, and H.-J. Christ, Materials and Corrosion 56, 785 (2005). J. Zhang, D. M. I. Cole, and D. J. Young, Materials and Corrosion 56, 756 (2005). J. Zhang and D. J. Young, Corrosion Science 49, 1450 (2007). T. A. Ramanarayanan, C. M. Chun, and J. D. Munford, US Patent No. 20030029528A1. P. Szakalos, M. Lundberg, and J. Hermblom, US Patent No. 2004005239A1. Y. Nishiyama, K. Moriguchi, N. Otsuka, and T. Kudo, Materials and Corrosion 56, 806 (2005). Y. Nishiyama and N. Otsuka, Materials Science Forum 522–523, 581 (2006). J. Zhang and D. J. Young, ECS Transactions 3, 27 (2007). J. Zhang, M. Safarzadeh, and D. J. Young, Oxidation of Metals 70, 15 (2008). DKI German Copper Institute Booklet, Copper-Nickel Alloys: Properties, Processing, Application. http://www.copper.org/applicaitons/cuni/txt_DKI.html. F. Besenbacher, I. Chorkendorff, B. S. Clausen, B. Hammer, A. M. Molenbroek, J. K. Nørskov, and I. Stensgaard, Science 279, 1913 (1998).

123

42

Oxid Met (2011) 76:23–42

55. H. S. Bengaard, J. K. Nørskov, J. Sehested, B. S. Clausen, L. P. Nielsen, A. M. Molenbroek, and J. R. Rostrup-Nielsen, Journal of Catalysis 209, 365 (2002). 56. H. J. Grabke and M. Schu¨tze, Corrosion by Carbon and Nitrogen—Metal Dusting, Carburization and Nitridation (European Federation of Corrosion Publications, No. 41, WP Cambridge, England, 2007). 57. H. J. Grabke, E. M. Mu¨ller-Lorenz, S. Strauss, E. Pippel, and J. Woltersdorf, Oxidation of Metals 50, 241 (1998). 58. Z. Zeng and K. Natesan, Oxidation of Metals 66, 1 (2006). 59. Z. Zeng and K. Natesan, International Journal of Hydrogen Energy 32, 3640 (2007). 60. K. T. Voisey, Z. Liu, and F. H. Stoot, Applied Surface Science 252, 3658 (2006). 61. N. S. Cherevu, K. S. Chan, and R. Viswanathan, Energy Materials 1, 33 (2006). 62. K. L. Baumert and J. J. Hoffman, Corrosion 1997, NACE International, paper No. 493 (1997). 63. B. Ganser, K. A. Wynns, and A. Kurlekar, Materials and Corrosion 50, 700 (1999). 64. K. A. Wynns and G. T. Bayer, US patent No. 5,873,951 (1999). 65. A. R. Kurkelar, G. T. Bayer, and G. E. Bayer, Corrosion 2001, NACE International, paper No. 01388 (2001). 66. W. Tsu and W. T. Tsai, Materials Chemistry and Physics 64, 147 (2000). 67. K. Natesan, Z. Cheng, W. K. Soppet, and D. L. Rink, Proceedings of the International Symposium on Metal Dusting, September (2001). 68. C. Rosado and M. Shu¨tze, Materials and Corrosion 54, 831 (2003). 69. I. Alstrup, EP Patent No. 0903424B1 (2002). 70. http://kwi.dechema.de/kwi_media/Downloads/hochtemperaturwerkstoffe/ Projekte/HTW Schlussbericht_F468.pdf (2009). 71. M. L. Holland, Corrosion 2001, NACE International, paper No. 01385 (2001). 72. R. R. De Souza, A. S. Takimi, M. D. Lima, and C. P. Bergmann, 58 Congresso Anual da Associacao Brasileira de Metalurgia e Materiais, Rio de Janeiro (2003), p. 490. 73. Y. Wang and W. Chen, Surface and Coatings Technology 183, 18 (2004). 74. D. E. Brown, K. Clarck, J. M. Carroll, and M. L. Sims, US Patent No. 4,099,990. 75. S. Kukes, J. R. Harris, R. A. Porter, and L. E. Reed, US Patent No. 4,410,418 (1983). 76. C. S. Li and Y. S. Yang, Surface and Coatings Technology 185, 68 (2004). 77. W. Brandl, G. Marginean, N. Marginean, V. Chirila, and D. Utu, Corrosion Science 49, 3765 (2007). 78. K. T. Voisey, Z. Liu, and F. H. Stott, Surface and Coatings Technology 201, 637 (2006). 79. H. Liu and W. Chen, Corrosion Science 49, 4134 (2007). 80. H. Liu and W. Chen, Oxidation of Metals 67, 129 (2007). 81. H. Liu and W. Chen, Materials and Corrosion 59, 311 (2008). 82. J. Alvarez, D. Melo, O. Salas, J. Oseguera, and V. Lo´pez, Surface and Coatings Technology 203, 779 (2009). 83. J. Alvarez, D. Melo, O. Salas, J. Oseguera, and V. Lo´pez, Surface and Coatings Technology 203, 422 (2008). 84. D. Melo, D. Sigu¨enza, O. Salas, J. Oseguera, R. Reichelt, and V. M. Lo´pez, Surface and Coatings Technology 204, 788 (2009).

123