18.2.1 General

The need for coating systems in the domain of high temperature arose by the 1960s when improved performance criteria could not be met adequately in spite of new materials with superior physical, mechanical, and metallurgical properties. Operating efficiency and production economy had to be considered and an improvement sought.

In this section, a brief overview of the main categories of coating systems is provided.

18.2.2 Diffusion Coatings

Nowadays the coatings for high temperature applications include mainly diffusion coatings, overlay coatings, and thermal barrier coatings (TBCs). The protective character of diffusion coatings attributes to the protective nature of the Al₂O₃, Cr₂O₃, and SiO₂ scale formed, respectively, on the aluminides, chrominides, or silicides at elevated temperatures. Diffusion coatings have first been developed and still are the most used coating. Aluminide diffusion coatings are proved to be the cost‐effective solution for high temperature oxidation, which are used widely for protecting turbine blades and vanes. The properties of the aluminide coating depend on the process methodologies used to deposit the coating, the substrate composition, and the subsequent treatment. The coating deposition rate and morphology depend on the process temperature and time. Processing temperature influences the rate of diffusion, at which alloy elements may diffuse and the metallurgy of the surface compound may form; thus, it is a critical parameter in the processing and manufacturing of diffusion coatings. The coating time at temperature defines the thickness of the coating formed during the diffusion step. The thickness of the coating is also a main factor of protecting property of the coatings. Two basic mechanisms typify the diffusion coatings, depending on whether the main diffusing species is aluminum diffusing from the coating to substrate of the base metal of the substrate alloy diffusing outward to the coating layer. These coatings are usually produced by pack cementation, out‐of‐pack cementation, and chemical vapor deposition (CVD), which involve the diffusion of a predominant element such as aluminum to form “diffusion coating” layers. It should be noted that coatings are typically fine grained with more grain boundaries compared with substrate materials. Diffusion rates in coatings are therefore expected to be faster than in substrate of similar composition. The three main diffusion coating mechanisms will be analyzed in the next section. Here, the diffusion coating mechanisms will be briefly considered, taking the diffusion aluminide coatings as an example.

Aluminizing is achieved by two different processes that are based on the activity of aluminum in the gas phase and the aluminizing temperature. The low activity high temperature (LAHT) process is a “one‐step” method with outward diffusion of Ni to form β‐NiAl coating. On the other hand, the high activity low temperature (HALT) process concerns the inward diffusion of aluminum, typically giving rise to a δ‐Ni₂Al₃ coating, which requires subsequent heat treatment to convert it to β‐NiAl. This is because δ‐Ni₂Al₃ is a brittle phase that has lower mechanical property than β‐NiAl. As might be expected, the mechanisms by which the coatings grow are also different. For the LAHT process, the aluminum activity is insufficient for it to be the predominant diffusing species, and accordingly, coatings form by diffusion of Ni from the alloy substrate into the region of the coating. For the HALT process, the aluminum activity in the gas phase and at the surface of the coating is high enough to facilitate the inward of Al into the alloy substrate. The growth mode is an important factor when considering coating integrity, since coatings formed by outward Ni diffusion can trap diluent particles (alumina) brought by aluminizing packs. While in the coating β‐NiAl phase layer of LTHA, much unwanted phases can precipitate dispersively, which decrease the mechanical property of the coating. Which coating process is selected depends upon a number of features, e.g. heat treatment specifications for the substrate alloy, applications, nature of packs available, integrity issues, etc.

In LAHT aluminizing (low aluminum content in the pack [1.2–1.5 wt% of Al], 900–1150 °C), the aluminum is deposited on the surface, but at reduced rate, and nickel simultaneously diffuses outward to the surface. Then a β‐NiAl surface layer forms. The interdiffusion coefficient in the β‐NiAl phase of the Al–Ni system varies strongly with composition. At 1050 °C, the minimum value of interdiffusion coefficient is not at the stoichiometric composition, but displaced to the low‐Al side by 1–2 at.% (that is 48–49 at.% of Al). Ni diffuses predominantly in low‐Al β‐NiAl.

The rapid nickel diffusion to the surface, coupled with the low aluminum activity at the surface, effectively holds the surface aluminum content close to 50 at.%. Hence, LAHT coatings are near stoichiometric at the surface and Ni rich within the substrate β‐NiAl phase. This structure ensures outward of Ni transport and the outward growth of the coating microstructure. Another important result is that slowly diffusing elements Ta, W, Re, and Mo from substrate are unable to diffuse to form significant concentration levels in the outwardly growing β‐NiAl. The outer zinc zone of this coating therefore appears much “cleaner” and is free from such precipitates as observed in the HALT coatings. The limited solubility of Cr, as well as Ta, Mo, and W, in the β‐NiAl phase produces precipitates of topologically close‐packed (TCP) phases, forming an interdiffusion zone (IDZ) under the “clean” β‐NiAl phase layer. Thus, the IDZ includes the complicated TCPs and β‐NiAl phases. This coating composition and structure are less dependent on the alloy composition. The further diffusion processes between such aluminide coatings and Ni‐based alloys are determined by the aluminum activity in the β‐NiAl phase. As long as this is lower than the equilibrium activity among β, IDZ, and γ + γ′ matrix, Ni will diffuse out of the base alloy into the β phase. The coating grows further outward.

The kinetics of the formation of aluminide coating on pure Ni by LAHT has also been studied, and this mechanism is illustrated in Figure 18.2.

In respect to the HALT theory interpreted above, if the activity of Al is low, Ni preferentially diffuses out through the coating and combines with Al to form NiAl. The coating grows outward. In this case, Ni diffuses faster than inward diffusion of Al. Pack particles are entrapped in the coating. These particles are assumed to be dragged out by the diffusion of Ni.

As previously mentioned, HALT coatings are processed at lower temperature (700–950 °C) as the first step, with a higher aluminum content (1.7–2.7 wt% of Al). The substrate material (γ‐Ni + γ′‐Ni₃Al) reacts with the depositing aluminum, forming a surface layer of δ‐Ni₂Al₃ over the layer of β‐NiAl. The diffusivity of Al through δ‐Ni₂Al₃ is very high, and substantial amounts of Al can be forced deeper and deeper into the material. In δ‐Ni₂Al₃ phase layer, the diffusivity of Ni is near zero, while Al diffuses rapidly; thus the formation of surface layer of δ‐Ni₂Al₃ results from the inward diffusion of aluminum. Substrate alloying elements such as W, Mo, Ta, and Re are selectively diffusing at the coating/substrate interface. Once the solubilities of alloying elements of W, Mo, Ta, Re, and Cr are saturated, many precipitates rich in these elements will occur for their limited solubility to β‐NiAl phase and distribute in the entire coating structure. After a second step of heat treatment at higher temperature (950–1100 °C), Ni is able to diffuse outward to transform the brittle δ‐Ni₂Al₃ phases into Al‐rich β‐NiAl. This step is also usually combined with the heat treatment required to recover substrate properties.

Now we have preliminary concepts of LAHT and HALT, and then we can primarily establish whether the coating belongs to one process or the other. However, in the actual manufacturing there are all kinds of operations by adding different amounts of Al and at different temperatures. Therefore, the resulting coating could be significantly different and have very complex structures.

Coating–superalloy interdiffusion is principally responsible for the phase transformations, oxidation behavior, and degradation of the mechanical properties of the coating. Moreover, the diffusion behavior in the multicomponent coating layers is very complicated due to the interactions among the components. Thus, diffusion is an important factor to be considered when designing β‐NiAl coatings.

18.2.3 Overlay Coatings

Diffusion‐type coatings, used successfully on early gas turbines, were tied to the substrate composition, microstructure, and design. Later, some changes were introduced: (i) in superalloy composition, such as reduction in Cr and increase in other refractory metals; (ii) in microstructure, by castings with more segregation; and (iii) in design, by air cooling and with thin walls (which introduced higher thermal stresses). These changes required coatings that were much more independent of the substrate. Overlay coatings met this necessity.

Overlay coatings also overcome the process restrictions encountered in diffusion coatings, especially the variants, viz. Cr/Al, Ta + Cr, or the Pt aluminides, all of which give better stability and oxide hot corrosion resistance than Al alone. MCrAlY compositions (M = Ni, Co, Fe alone or in combination; Y represents oxygen‐reactive elements such as Zr, Hf, Si, and Y) are the principals in the series of overlay coatings developed by the electron beam evaporated physical vapor deposition (EBPVD) technique for multiple load use. More than 4 million aerofoils have been successfully processed by EBPVD.

MCrAlY overlay used in gas turbines are usually Ni and/or Co with high Cr, 5–18% Al, and Y addition around less than 1% for stability during cyclic oxidation. They are multiphase alloys with ductile matrix, e.g. gamma Co–Cr, containing a high fraction of brittle phase, e.g. beta CoAl. The Cr provides oxidation hot corrosion resistance, but too much Cr affects substrate phase stability. The success of most overlay coatings is the presence (and perhaps location) of oxygen active elements such as Y and Hf, which promote alumina layer adherence during thermal cycling, giving increased coating protectivity at lower Al levels. Y mostly appears along grain boundaries if a MCrAlY is cast but is homogeneous if plasma sprayed. Thus MCrAlY with 12% Al are more protective than the more brittle diffusion aluminides with 30% Al.

Overlay claddings deposited by hot isostatic pressing (HIP), electron beam evaporation, or sputtering methods are diffusion bonded at the substrate/coating interface, but the intention here is not to convert the whole coating thickness to NiAl or CoAl. There is thus more freedom in coating composition, whose properties can be maximized to the type required. Composition based on NiCr, CoCr, NiCrAl, CoCrAl, NiCrAlY, CoCrAlY, FeCrAlY, and NiCrSi have been successful in gas turbine engines. They are generally alumina formers with only 10% Al unlike the 30% in nickel aluminide coatings. Cr increases the Al activity, which allows this feature. Higher Al levels cause brittleness and a higher DBTT, and the Al levels are generally held below 12% (5–10% preferred). The coatings are also more ductile than NiAl and CoAl and can be rolled and bonded by HIP. In general, NiCrAlY give best results against high temperature oxidation, while CoCrAlY are best for hot corrosion.

18.2.4 Thermal Barrier Coatings

TBCs are ceramic coatings applied over metal substrates to insulate them from high temperatures. They consist of zirconium oxide, ZrO_2, stabilized by about 8 wt% yttria, Y₂O₃, or magnesia, MgO. They are applied over a bond coat that is usually an MCrAlY overlay coating but can be a diffusion coating. The ceramic layer is typically 5–15 mils thick, and the overlay bond coat is usually 3–8 mils thick.

The bond coat improves the adhesion of the ceramic top coat by reducing the oxide buildup underneath the ceramic. (The ceramic is porous to oxygen and will spall off when the oxide formed on the bond coat is sufficiently thick.) The bond coat is applied in the same manner as for overlay coatings.

The ceramic layer is normally plasma sprayed in an air environment. This layer can also be made by electron beam physical vapor deposition (EBPVD), which is used on aircraft engines. The ceramic layer made by EBPVD is columnar grained and is more resistant to spalling than the plasma‐sprayed TBC.

Another type of ceramic layer that has been developed is called a segmented TBC. It has cracks through the thickness of the ceramic, which are perpendicular to the substrate surface. These cracks provide increased resistance to spallation, similar to EBPVD ceramic layers. The segmented TBC is made by a proprietary plasma spray process, and very thick ceramic layers can be made.

The TBC protects the metal by acting as an insulator between the metal and the hot gases. The thermal conductivity of the ceramic is 1–2 orders of magnitude lower than the metal. Although thin, the TBC can significantly reduce the metal temperature provided that the metal component is air cooled. (The air cooling provides a heat sink.) Furthermore, the ceramic has a higher reflectivity than the metal. This means that more of the radiative heat is reflected away, which is important for combustion hardware.

During startup and shutdown, the TBC improves the thermal fatigue life by reducing the magnitude of the temperature transients the metal is exposed to. A 10 mil thick TBC on airfoils in experimental aircraft engines has achieved more than 300 °C reduction of metal temperature. During steady‐state operation, the TBC lowers the temperature of the underlying metal, thereby improving its durability. It also reduces the severity of hot spots.

The principal failure mechanism of TBCs is the spallation of the ceramic layer. Spallation is caused by the synergistic interaction of bond coat oxidation and thermal cycling. The oxidation occurs at the interface between the bond coat and the ceramic. As more oxide forms at this interface, the ceramic layer is more likely to spall.

18.3.1 General

There are three major processes by which the diffusion coatings can be formed: CVD, pack cementation, and out‐of‐pack cementation, which involve the diffusion element such as aluminum (chromium or silicon). Overlay coatings can be applied by CVD, physical vapor deposition (PVD), high velocity oxygen/flame (HVOF), or plasma spray. Plasma spray is one of the most widely used processes today. TBCs are usually deposited by PVD, EBPVD, and plasma spraying techniques. All these processes and a few more advanced techniques for producing high temperature coatings are described in this section.

18.3.2 Chemical Vapor Deposition

CVD is a process wherein a stable solid reaction product nucleates and grows on a substrate in an environment where a vapor‐phase chemical dissociation or chemical reaction occurs. It uses a variety of energy sources, viz. heat, plasma, ultraviolet light, etc., to enable the reaction and operates over a wide range of pressures and temperatures. CVD is a long‐established, economically viable industrial process in the field of extraction and pyrometallurgy. Some of the well‐known processes are:

• The Mond process:
  
  18.1
  
  
• The Van Arkel process:
  
  18.2
  
  
• The Kroll process:
  
  18.3

The method has also been used to form freestanding, simple, and complex shaped articles from metals that are not very amenable to fabrication, e.g. W:

18.4

The technology of CVD took on new dimensions with a fresh view taken at the deposition aspects of the process. This transition of emphasis from extraction to deposition of the CVD phenomenon has made CVD an important sector of coating technology for producing new materials and coatings with improved resistance to wear, erosion and corrosion, good thermal shock resistance, and neutron absorption characteristics. Components from micro‐ to macro‐sizes are now coated by CVD. In the deposition area of CVD, the lamp and electronics industry uses many of the earlier processes to obtain longer service life and better performance for a number of products. The CVD industry took a bigger leap when demands arose for new high temperature materials with greater high temperature corrosion resistance and strength in all the sectors using gas turbines and other propulsion hardware, in the military science, engineering, and aviation fields. In parallel, CVD technology has had rapid application in the semiconductor industry, the many branches of fuel and energy – nuclear, fluidized bed, oil, solar and chemical industries and the tool industry. Solid‐state devices; electrode materials; thermionic devices; erosion‐, corrosion‐, and wear‐resistant components; fission barriers; and many more applications can be listed in CVD technology.

It would be confusing and misleading to list a number of processes as CVD methods in as much as different high temperature coatings cannot be merely termed as diffusion coatings. The minor or major parameters that influence the actual achievement of the coating itself are those that facilitate categorization. Otherwise, conventional CVD, plasma‐assisted CVD (PACVD), laser‐assisted CVD (LACVD), pack, vacuum pack, pack‐slip, vacuum pack‐slip, CCRS, or RS may all be just CVD. Similarly, diffusion coatings can be obtained by first hot dipping, spraying, cladding, fused salt electrodeposition, electrophoresis, plasma, or all PVD, and all the above “CVD” methods. What gives the distinction to any process?

The obvious factor that emerges is that a basic CVD process has a number of limitations, e.g. the types of reactions possible and available, and the minimum substrate temperature that has to be maintained. This results in restricting any eventual substrate/deposit configuration and control. Improved techniques stretch the reactant–product spectrum and can also (i) lower the temperature; (ii) reduce the hazards and mismatch in the morphological aspects, viz. nucleation and growth; (iii) alter the physical–mechanical properties favorably; and (iv) provide better bonding.

CVD technology offers the following favorable aspects to semiconductor industry:

1. A wide range of silicon epitaxy thicknesses and resistivity.
   
2. A low cost, highly productive method of depositing polysilicon, Si₃N₄, and high temperature SiO₂.
   
3. A proven technology for passivating silicon devices with low temperature deposition of silicon oxide and silicon nitride.

A wide range of metallic and inorganic coatings can be produced by CVD. In many instances CVD coatings exhibit unique properties, which are difficult to produce by other methods.

A CVD process was seen as a thermochemical process in which the substrate is regarded as a collector of the deposit, and the substrate/deposit interaction is both undesirable and unnecessary for deposit growth. However, for a CVD deposit, the post‐deposition state is a mandatory parameter to meet the service requirements; substrate/deposit interaction to some degree is expected. Further, heat treatment, and therefore diffusion, is often a step following the primary CVD process (or even during deposition, since the substrate is heated). Thus, substrate/deposit interaction always occurs and is also one of the ways deposit consolidation can be achieved.

A CVD reactor is the apparatus‐equipment configuration that carries out the CVD process. CVD reactors are categorized on the basis of the temperature and pressure under which they operate. A CVD reactor can have a cold wall or a hot wall depending on the experimental optimization parameters. When the reactor wall surrounding the heated substrate is comparatively cool relative to the temperature of the substrate, it is a cold‐wall reactor. The substrate is heated directly, and product deposition occurs mostly on the heated component. In a hot‐wall reactor, the deposition product can nucleate both on the reactor wall and the heated substrate, though to different degrees, if a temperature difference prevails across the wall substrate area and zone. The substrate may be independently heated or not, but the reactor itself is heated in order to facilitate the process.

The CVD reactor assembly consists of three components: (i) the reactant supply system, (ii) the deposition systems, and (iii) the reactant–product retrieval, recycle, and disposal system. The entire reactant supply (using solid and/or gaseous reactants) has to be carefully balanced before introduction into the reactor system, predicted to reach the required partial pressures at the temperatures of reaction and deposition. The primary features to consider concerning the substrate are the substrate/deposit adherence and growth. CVD reactions can be exothermic but mostly are endothermic. This could result in substrate cooling. On a production scale, large volumes of gases are handled, which can induce convective coating. Both these situations have to be compensated. Substrate heating optimization involves achievement of the correct temperature profile and coating exposure time. The substrate/reactor parameters in a deposition system should be designed to achieve a laminar flow. The input can be parallel, perpendicular, or angular to the substrate depending on the substrate holder design. The movement (rotation, lateral, etc.) of the substrate within this region also compensates to some extent any nonuniformity in the deposition.

CVD occurs through a chemical reaction; therefore, the thermodynamics of the system that can drive the chemical reaction, the basic chemical kinetics to provide information on the reaction rate, and mass and heat transfer profiles that are influenced by the reactor and substrate size and design, require consideration.

Mass transport of a CVD system links diffusion control, chemical kinetics, and fluid‐flow dynamics. Diffusion growth can be ascertained from chemical kinetics by rate constant of growth = A exp (−E/kT), where E is the activation energy, k is the Boltzmann’s constant, and T is the temperature. With increase in substrate temperature, the growth rate changes at a transition temperature from a predominant kinetics to a diffusion control, the latter being mostly under mass transport influence (Figure 18.3).

Many models of CVD reactors and their kinetics have appeared in the recent literature, but space restrictions impose our departure to the next coating process.

18.3.3 Pack Cementation

Pack cementation is an extended CVD method in which the substrate to be coated is introduced in a retort and surrounded by a powder mixture, which is composed of the source alloy(s), an activator, and an inert filler. The retort is then heated to the desired coating temperature usually under inert or reducing environment. In these conditions, the source element reacts with the activator, forming a gaseous transporting species. This diffuses through the pack to the substrate surface where it decomposes, allowing the metallic element to be deposited to penetrate into the substrate by solid‐state diffusion.

As for all diffusion coatings, the microstructure is highly dependent on the substrate, which is coated. For this reason, process conditions have to be optimized for each material.

The first “cementation process” was presented by Allison and Hawkins in 1914, who deposited Al on iron and on steel. But it is only since the 1960s that this process stimulated interest because of the development of coatings for the protection of gas turbine blades, especially those made of Ni‐based superalloys. The process has then been extended to Co and Fe‐based alloys. Concerning the elements deposited, these are mainly Al followed by Cr and Si. Nevertheless, codeposition of two or three elements became possible with the theoretical understanding of the process. The extensive works on Cr–Al codeposition published by Rapp on steels (1989, 1994, 1997) and by Young et al. on nickel‐based alloys (da Costa et al. 1994a,b, 1996; Gleeson et al. 1993) are the major examples, but development of Cr–Si (Rapp 1996), Al–Si (Wynns and Bayer 1999), Ti–Al (Weber 2004), and Ti–Si–Al (Rosado and Schütze 2003) coatings can also be mentioned. Eventually, more recent investigations seem to focus on the addition of reactive elements such as Y, Ce, or Hf by codiffusion.

The principles of the pack cementation technique changed remarkably little since 1914, but the composition and the quality of the substrate–deposit systems have been optimized in reply to the requirements of a large range of service applications.

From an industrial point of view, the pack cementation technique has been used very extensively for nearly 60 years for coating gas turbine blades (Mévrel et al. 1986). Nowadays it is estimated that on first‐stage gas turbine blades, more than 80% of all coated airfoils are coated by pack cementation (Goward 1998).

The coating of larger pieces has been the subject of some research investigations and even leads to the development of a few patents, showing that coating even inside a tube is possible by this technique (Baldi 1980; Wynns and Bayer 1998). During the 1990s the US company Alon developed several industrial processes involving the pack cementation technique. The purpose was to protect ethylene cracker furnace components from carburization and coke formation (Kurlekar and Bayer 2001). With Alon’s facilities, 15 m long furnace tubes can be coated with Al, Cr, and/or Si on an industrial scale, producing 100 000 linear meters of tubes per year (Alon). After one year service in an ethane cracking furnace, the coating reduced the coke formation by a factor of 2, and post‐exposure analysis revealed that the coating microstructure was essentially unchanged (Ganser et al. 1999).

Typically the powder pack mixture consists of three components:

• A metallic source: A fine powder of the element(s) to be deposited on the surface of the substrate (Al, Cr, Si, Ti, or combinations of these elements).
  
• A halide salt activator as, for example, NaCl, NH₄Cl, AlCl₃, NaF, NH₄F.
  
• An inert filler powder (e.g. Al₂O₃) preventing the powder mixture from sintering at high temperature.

Table 18.1 lists different packs that can coat various elements with the corresponding temperatures and heat treatment durations.

Figure 18.4 illustrates the major chemical reactions involved in the general pack cementation process. In order to allow a better understanding of the following paragraph, the reactions are written for the particular example of aluminizing of iron activated by NH₄Cl.

The whole process relies on the formation of gaseous halides according to the general reaction:

18.5

AlCl₃, AlCl₂, AlCl, Al₂Cl₆, and Al₂Cl₄ are the chlorides involved in a chloride‐activated aluminizing process.

The partial pressures of each of the gaseous halides formed are established by their thermodynamic stability, which varies with the process conditions: composition of the pack, type of activator, temperature, pressure, and type of inert or reducing environment.

In the case of aluminum deposition, AlCl₃ is the major halide formed at low temperature, whereas at higher temperature the activity of AlCl becomes higher.

Once they are formed, the halide molecules diffuse through the gas phase to the substrate (e.g. iron) surface, where they adsorb and decompose owing to the general reactions:

• Disproportionation:
  
  18.6
  
  
• Displacement/exchange:
  
  18.7
  
  
• Direct reduction:
  
  18.8
  
  
• Dissociation:
  
  18.9

By using alkaline or earth alkaline halide activators as, for instance, NaCl, KCl, or CaCl₂, the precipitation of the activator salt has also to be considered (Ravi et al. 1989):

• Precipitation:
  
  18.10

The aluminum formed at the surface of the substrate of Reactions 18.6–18.10 can then diffuse into the solid substrate, forming the desired coating.

The predominance of the Reactions 18.6–18.10 depends, first of all, on the stability of the gaseous halides involved. The deposition particularly occurs by disproportionation Reaction 18.6 when, first, the vapor pressure of the substrate halide is low and when, secondly, the coating element has higher and lower halides of comparable vapor pressures in order to set atomic aluminum‐free. When the vapor pressure of the substrate halide becomes comparable to the gaseous species of the coating element, the contribution of the exchange Reaction 18.7 becomes important. The latter reaction even gets undesirable if the vapor pressure of the substrate halide becomes higher than the source supplier halide, as it would lead to a significant metal loss and to porous coatings. In a NaCl‐activated pack, this phenomenon can occur during chromizing or siliconizing of iron, whereas Reaction 18.7 stays minimal during aluminizing (Kung and Rapp 1989).

The reader may, at this point, easily catch the compromise that has to be found. Because halide molecules must diffuse through the gas phase from the pack to the substrate, the coating composition depends on the gaseous halide activity and their stability or their ability to decompose at the substrate surface. Hence, the formation and the decomposition of the gaseous halide have to be optimized at the same time in the pack and at the substrate surface, respectively.

Furthermore, it is a prerequisite that the thermodynamic activity of the incorporated element is always lower at the surface than in the pack (Gupta and Seigle 1980). This activity gradient drives the gas‐phase diffusion of the halide molecules from the pack to the substrate surface. As a consequence, a desired coating composition cannot be obtained by simply using a master alloy of the same composition (Rapp 1989).

Moreover, the concept of “major depositing species” has been defined. This corresponds to the gaseous species that is responsible for the major part of the deposition. In the case of a Cr–Al codeposition process by a chloride‐activated pack, Rapp (1989) and da Costa et al. (1994a,b) showed that although the vapor pressure of AlCl₃ is several orders of magnitude higher than that of Cr halides and other Al halides, the codeposition is possible by optimizing the process conditions so as to get comparable vapor pressures of AlCl and CrCl₂. Indeed, AlCl₃ is too stable and does not decompose enough at the substrate surface. The Al transport occurs via AlCl. This is thus considered as the major transporting species for Al, whereas CrCl₂ is the major transporting species for the deposition of Cr (Rapp 1989).

The formation of the coating can be described in the light of diffusion couples (Wachtell 1974). The coating actually corresponds to the IDZ between the substrate element(s) and the deposited element(s). The driving force for the solid‐state diffusion is the activity gradient between the pack/coating and the coating/substrate interface.

The aluminizing of iron can thus be treated as a binary diffusion couple, involving pure Al and pure Fe. As a consequence, the phases formed in the coating should follow the FeAl binary phase diagram (Figure 18.5).

As can be seen from this diagram, the aluminizing of iron can form several intermetallic phases, FeAl₃, Fe₂Al₅, FeAl₂, FeAl, or Fe₃Al as well as solid solution of Al in Fe. The effective formation of these phases is however controlled by the powder composition, the temperature, and the duration of the process. These parameters especially control the diffusion fluxes, which determine the final coating structure.

Goward and Boone (1971) distinguished the low and the high activity pack structures (Figure 18.6), as already discussed in Section 18.2.2. Their investigations considered the aluminizing of a nickel‐based superalloy. Two powder pack mixtures differing in the nature of the Al source (pure Al or Ni₂Al₃) were investigated at about 800 and 1100 °C, respectively. The observations lead the authors to consider that the high activity pack (with the powder containing pure Al) structure is characterized by a high aluminum content at the surface with the coating phase Ni₂Al₃. The underlying phases down to the coating–substrate interface followed the Ni–Al phase diagram with decreasing Al content. For the low activity pack (with the powder containing Ni₂Al₃), the coating is single‐phased β‐NiAl, which is directly in contact with the substrate.

The difference was attributed to the diffusion fluxes:

• In the high activity pack conditions, the coating formation was attributed to exclusive inward Al diffusion, the diffusion of Ni being minimal at low temperature.
  
• In low activity packs, the coating formation was interpreted as single Ni outward diffusion. This last mechanism also requires higher temperatures to allow diffusion of Ni.

One example of nickel aluminide diffusion coating produced by LAHT pack cementation process is illustrated in Figure 18.7. The coating consisted of a well‐defined to “clean” layer with NiAl as its major phase and a large diffusion zone underneath.

An example of SEM‐BSE (backscatter electron) image of LTHA coating on Ni‐based superalloy CMSX4 (this is a typical second generation Ni‐based single‐crystal superalloy characterized by the replacement of most of the Ti with Ta, by relatively high Co and low Mo content) is shown in Figure 18.8. This coating was deposited using an aluminizing pack containing 2 wt% Al at 900 °C and was heat‐treated for 2 hours at 1120 °C and then for 24 hours at 845 °C. The β‐NiAl phase layer is uniform in thickness, and many precipitates (white small particles) distributed dispersively in the entire β‐NiAl phase layer. Squillace et al. (1999) has also analyzed clearly the coating structure after the first step of LTHA. Three layers are visible: the inner layer has a striated appearance, the outer layer is equiaxed with many inclusions, and the middle layer is featureless. The coating layer has deposited very high amount of Al. In the book entitled High Temperature Coatings by Bose (2007), it was presumed that higher inward diffusion of Al with lower outward Ni diffusion has avoided the formation of Kirkendall porosity and also eliminated the embedded pack particles.

Although mixed types of mechanisms can occur as a result of varying pack activities, substrate compositions, and temperature, during practical application of the coating process, it is still of great use nowadays to relate the observed structures to two archetypes (Figure 18.6).

The aluminizing of iron shows the same kind of behavior. At 900 °C, a three‐layer structure with Fe₂Al₅, FeAl₂, and FeAl is formed in high activity packs, and a single‐phased FeAl coating in low activity packs (Levin et al. 1998).

With the help of thermodynamic calculations performed with a computer software, the activities of the gaseous halides can be determined considering all pack reactions together. Such considerations allow the optimization of the pack concentration and process conditions needed to form the desired coating phase. More precisely, the activities of the gaseous halide depend on the reactants involved in the halide formation and thus on the type of metallic source mixed to the pack powder. Therefore, the direct dependence of the coating structure with the master alloy determines whether the coating deposition occurs in a high or low activity process.

The activator is also a determining reactant involved in Reaction 18.5 too, from which it can be said that an unstable activator results in high halide partial pressures. Moreover, concerning the choice of cation for the activator, some researchers recommend the use of ammonium salts, because of their decomposition at high temperature. Indeed, the NH₃ produced in Reaction 18.5 is not stable above 350 °C decomposing by the following reaction:

18.11