Recent Developments in Chemical Vapour Deposition

Dr. A. Kempster

Diffusion Alloys Limited. Hatfield, England

Development of the Technology

The Chemical Vapour Deposition (CVD) processes are used in many fields of industrial activity to protect components from wear, corrosion or abrasion in engineering environments. These industrial fields are diverse and range from gas turbines to gas cookers and from coinage to nuclear power plants.

The CVD process relies firstly on the generation of a species that is produced by the reaction of the element that is to be deposited with another element that results in the substantial increase in the depositing elements vapour pressure. Secondly, this volatile species is then passed over or allowed to come into contact with the substrate being coated. This substrate is held at an elevated temperature, typically from 800 - 1150°C. Finally, the deposition reaction then usually occurs in the presence of a reducing atmosphere, such as Hydrogen.

Take for example the CVD of chromium, (usually referred to as Chromising), onto a steel substrate. This maybe used to achieve the following properties on the surface of the steel:-

• Corrosion resistance
• Wear resistance
• High temperature oxidation resistance

The vapour pressure of metallic chromium is too low to allow the generation of useful concentrations of its vapour at sensible operating temperatures, (i.e., <1150°C). Hence the chromium must be converted in a volatile species. Many species are possible, but commonly the halides are used and particularly the chloride.

The CVD of chromium can be carried out by using one of the following methods:-

(1) The Pack Cementation Method¹

Here the component to be coated is surrounded by a pack medium that is made up essentially of the following ingredients:-

1. A source of chromium, (usually chromium metal powder or sometimes ferrochromium).
2. A ceramic diluent powder to reduce the sintering of the metallic particles to the component.
3. The halide source, such as ammonium chloride.

The process is carried out under a reducing atmosphere, (for example, mixtures of hydrogen and argon) and the following reactions are thought to occur:-

Possible chromium deposition reactions are as follows:-

In all cases reactions (5), (6) and (7) result in the deposition of a chromium atom on the component surface and under the influence of the high temperature the resulting ad-atoms will then diffuse into the surface of the component.

Fig. 1 illustrates a typical chromising retort that may be used for treating various components.

This method allows the component to be surrounded and supported by the compound and hence the component is coated on all surfaces. Many components may be coated in one load and automatic loading techniques have been developed when large numbers of small components require coating².

Relatively large components can also be coated using this technique. Currently the largest components processed are tubes 43 feet in length and 6" diameter.

After the coating run the components must be separated from the processing compound. Particulate adhesion to the components may require special cleaning techniques.

(2) Out of Pack Chromising Method

In this method the components to be treated are laid out on trays above a layer of powder that generates the volatile chromium species. As the temperature increases the chromium halide vapour is generated and then passes over the components during which deposition of chromium occurs. Fig. 2 shows typical equipment used for out of pack chromising. This method results in a smoother coated surface than that obtained for pack cementation. This may be important, for instance, if the coating is being used as a bearing surface. Also, there is no need to separate the chromising compound from the components at the end of the coating run. In some cases, however, the component may not be coated where it touches the tray. Also, higher temperatures will be used for out of pack in comparison to pack cementation, thus reducing equipment life. The size of components that can be coated (about 2' maximum) are generally smaller for out of pack processing.

(3) Gas Phase Chromising Method

Fig. 3 shows typical equipment for gas phase chromising. Here the chromium halide is generated by the passage of an extemally produced gas (hydrogen chloride) over a chromium source material which is held at an elevated temperature. The ensuing reaction produces the chromium halide which is then allowed to pass over the components to be coated. As with the out of pack method, the components to be CVD coated are placed on wire trays or grids.

In this method there is an almost continual supply of coating species. Therefore, compound depletion effects are avoided and coating runs can be of shorter duration than out of pack runs. As with the out of pack chromising, the coated surface is smooth and there is no compound to separate from the component. In the gas phase method the source material can be chips of metal and not powders, resulting in easier handling.

One disadvantage of this technique is that the handling of halide gases at high temperature is problematical and special resistant materials must be used. As with out of pack processing the component size capable of being handled is relatively small in comparison to that achievable for pack cementation processing.

Although all the above three methods are essentially similar processes the coating produced can have distinctive characteristics³.

The following advantages and disadvantages are generally recognised for the CVD process:-

Advantages of the CVD Process

As is seen above, the process is gas phase in nature and therefore given a uniform temperature within the coating retort and likewise uniform concentrations of the depositing species then the deposition rate will be similar on all surfaces. Therefore, variable shaped surfaces, given reasonable access to the coating powders or gases, such as screw threads, blind holes or channels or recesses, can be coated evenly without build-up on edges.

In some cases, as will be evident later on in this paper, it is possible to form ductile CVD layers, (e.g. chromising of low carbon mild steel). Given that the processing temperature will normally hilly anneal any ferrous substrate that the CVD layer is deposited onto, then it is practical to form, press or bend these components successfully after coating. This is in direct contrast with stainless steel components that work harden significantly during forming or pressing and can cause rapid wearing of the tooling used.

The high temperatures used during CVD results in a considerable amount of diffusion of the coating into the substrate and consequently if thermal expansion coefficients are compatible between the coating and substrate then adhesion will be excellent. In many cases the substrate can be heat treated after CVD coating with no distress to the coating.

The range of CVD coatings are diverse and consequently this generates a wide range of properties as indicated in the following table:-

Disadvantages of the CVD Process

The CVD process is carried out at relatively high temperatures and therefore limitations due to dimensional tolerances are an important consideration. Components that have tight dimensional tolerances will not be amenable to CVD. However, the reduction of distortion during coating can sometimes be controlled by careful stress relieving after rough machining of the component during fabrication. Distortion of large tubular components is inevitable, but the situation may be remedied by careful straightening of the tubes after coating. Sometimes the reheating of the substrate after coating may be a problem, for example in the boronising of high speed steels. The heat treatment to restore the mechanical properties of the substrate, (approaching 1200°ree;C in some cases), will melt the boride surface layer.

The CVD of Chromium onto various materials

The following examples show the fundamental influence of the base material on the final properties of the CVD coating.

1. Low carbon steels (<0.15% C)

   In these materials the chromium has a relatively high diffusion rate due to the low carbon concentration. In some cases special steels have been developed where the carbon that is present is tied up or stabilised as a carbide with, for example, titanium, niobium or tantalum. These steels give excellent results when chromised. The coating formed on these steels will thus be a solid solution or chromium in iron and will vary from 40% chromium at the surface to 11% at the base of the coating. Fig. 4 shows schematically how the concentration gradient of chromium in this type of coating will vary.

   Coatings of this type have good oxidation resistance to 650°C and have the corrosion resistance roughly equivalent to a ferritic stainless steel, (for example Type AISI 430). The coatings are ductile and can withstand bending and forming and consequently sheet steel components will generally be chromised flat and then formed after coating. This reduces the volume in the coating retort, hence the cost of the coating operation. Fig. 5 shows typical examples of such components.

2. Medium Carbon Steels (0.25 - 0.35%C)

   When steels containing c. 0.25% carbon are chromised then a coating with a duplex structure can result. Fig. 6 shows a typical chromised layer on this type of steel.

   The chromium layer at the surface is a carbide with a thickness of approximately 10 µm and a hardness of 1200 Hv (0.1 Kg) and under this is a diffusion zone of chromium in iron. The chromium concentration profile obtained is shown in Fig. 7.

   Coatings of this type are used in light wear / corrosion situations. Here the chromium carbide gives some wear resistance, whilst any breaches in the coating do not corrode rapidly due to the enrichment of chromium beneath the surface of the carbide. Similar duplex coatings may be formed on steels or alloys that contain nitrogen. In this case the outermost coating is chromium nitride.

3. High Carbon Steels and Cast Irons >0.4%C

   Here the carbon in the substrate is so high that the chromium-carbon reaction predominates and hence no significant solid solution chromium-iron alloys are formed. Chromium carbide coatings of up to 20um can be formed on these substrates and heat treatment of the substrate will be commonly carried out after coating to achieve the required mechanical properties. These coatings will be used for more arduous wear resistance than type (ii) coatings and also when applied to high chromium substrates, (i.e. D2), can be used in contact with corrosive material, such as food stuffs containing significant amounts of salt.

   From the above it will be seen that the substrate plays an important role in defining the ultimate properties of the coating, an important consideration when initially designing the component.

Industrial Examples of CVD Technology

1. CVD of Chromium for Coinage Applications

   The attractiveness of using a coating on a cheap substrate for coinage applications is becoming a well established production route. Consequently a very high tonnage of coins are now produced from mild steel and coated by electroplating with nickel, copper or brass. The striking of the coin blank is then carried out on the plated blank. The use of a CVD layer of chromium on mild steel has some potential important advantages over stainless steel.

2. The CVD of chromium is carried out at temperatures in excess of 1000°C and hence the substrate becomes hilly annealed after processing. This enables the striking of the coin blank to give a full relief, overcoming the problems of bulk stainless steel, i.e. poor formability and excessive die wear.

3. The total amount of chromium in a coated mild steel coin is only ~1% by weight, whilst in an equivalent stainless steel it could be 17% by weight.

4. The original punching of the coin blanks from the stainless steel strip will leave a more expensive waste than with mild steel.

A plant has been constructed for the automatic production of chromised mild steel blanks and the technology has been shown to be both reliable and reproducible². Fig. 8 shows typical mild steel coins with a CVD layer of chromium after finishing and striking. This product has been assigned the trade mark of Chrominox^TM.

Electroplated chromium is not a viable alternative as the layer cannot be reliably applied by large scale barrel plating, nor is the chromium plate amenable to deformation during striking. The compounds used for chromising are essentially regenerable after each coating run. Deterioration of the chromising powder does not occur and hence disposal of waste material is not a significant problem.

Apart from the advantages mentioned above, the important objective when producing a chromised coin blank is that it must be competitively priced against stainless steel. This was only achieved when the chromising plant had the necessary level of automation.

Cleaning of Turbine Blades using a CVD Technique (Siclean^TM)

Land based turbines are increasingly being used with more aggressive conditions in the hot or exhaust section of the engine. Consequently, material can suffer from considerable attack from contaminants in the fuel such as NaCI, Na₂SO₄ and V₂0₅ or from high temperature oxidation. Over recent years a variety of protective coatings have been developed for hot sections and these include Vacuum Plasma Sprayed (VPS) MCrAIY's, where M=Co, Ni or Fe, stabilised zirconium dioxide thermal barriers and CVD coatings. During an engine outage it is necessary to remove these old consumed coatings and apply a new coating in its place. This was found to be problematic as unless the corrosion products were effectively removed from the substrate, then the new coating would have poor adhesion. Detachment of the new coating could then occur resulting in a poor life for the refurbished blade. Removal of consumed MCrAIY coating is particularly difficult to carry out with mineral acids as the surface layer of the coating is depleted of aluminium and is essentially rich in nickel. Consequently a method was developed that was based on a CVD technique. Fig. 9 shows the sequences of the process, (commercially known as Siclean^TM). The surface of the used blade was first blasted using a Al₂0₃ grit to remove any substantial surface oxides. The surface was then subjected to a CVD process whereby aluminium is deposited and diffused into the surface of the blade to a depth of 200 - 300 µm. The surface is thereby transformed from a metallic into an intermetallic aluminide.

This aluminide coating is seen to encapsulate the corrosion products, (i.e. oxide, sulphide and nitride particles) in the surface of the superalloy. The coating, (nickel aluminide for a nickel superalloy), changes the surface properties of the nickel alloy significantly. Firstly, it is now very susceptible to acid attack and secondly it is brittle and friable and consequently it can be easily removed using a combination of blasting and acid pickling. The resulting surface is clean and ideal for depositing an overlay such as a VPS MCrAIY layer. Fig. 10 shows a typical turbine vane after service and illustrates the Siclean^TM process at various stages.

Another important advantage of the Siclean^TM process makes use of a benefit of CVD mentioned earlier. One standard method of removing corrosive products from the surface of a used turbine blade is to carry out hand linishing with belt grinders. It is extremely difficult to remove uniformly the corroded surface of a turbine blade, especially around awkward contours, such as the trailing edge of a turbine blade. However, the CVD process produces a uniform depth of aluminide over the whole surface of the turbine blade and consequently the layer that is removed is also uniform and therefore independent of operator skill or part geometry.

Internal Aluminising of Turbine Blades and Vanes

Staying with the subject of land based gas turbines, considerable advances in blade and vane design has occurred over the last 10 years. With the higher gas inlet temperatures referred to earlier, has come the need to apply greater ingenuity to the internal cooling of the turbine blades, notwithstanding this, the internals of the turbine blade can still suffer from oxidation and sometimes corrosion. This attack can reduce the fatigue strength of the base material and protective coatings are now being used on the internal passageways of turbine blades to reduce this attack. The more advanced turbines are using very complex internal cooling passageways and conventional methods of internal aluminising using powder or slurries are not possible. Firstly, even if a powder or slurry can be introduced, it cannot be sure that all surfaces are in contact with the coating media, (and consequentially will become coated), and subsequently that all the used powder or debris will be removed from situ after the coating operation.

Therefore, there was the need for the development of a reliable CVD method to coat the internal surface of turbine blades. In earlier work in this field Restall et. al.5 developed the pulsed aluminising approach in the late 70's. This technique relies on the 'stirring' of the retort gas by periodically altering the pressure within the retort and was principally developed to coat the internal passageways of small aircraft turbine blades. One disadvantage of this technique is that all the external surfaces become coated as well as the internals and this may be problematic if the outside of the blade is already coated or needs to be coated with a MCrAIY layer. Another disadvantage of this technique is that reduced pressure working means that any retort used for the coating operation can only be relatively small before bracing is necessary to provide support against collapse when operating at high temperatures.

As land based turbine components are by definition large, then any plant will be large to accommodate a useful number during a coating cycle. Fig. 11 shows a typical vapour aluminising plant at Hatfield, that is capable of coating up to 32 large rotor blades or nozzle guide vanes in one coating operation.

The CVD technique must be capable of coating the internal passageways with a uniform layer without a significant variation of coating thickness from one end of the blade to the other. Also, there should be a consistency in coating characteristics from one blade to another in the same coating batch.

Consider for example, the internal aluminising of the turbine blade shown in Fig. 12. The blade represented by this Fig. is about 5" in length and has several small longitudinal cooling holes, (1 - 2 mm diameter), going from the root base to the squealer tip at the end of the aerofoil. In cross section at point AA' the hole configuration is as shown in Fig. 13.

Potentially the most difficult holes to coat are those that come from the trailing edge, (0.7 - 1.0 mm diameter), of the aerofoil and join into hole B.

Attempts to coat these components using a powder method always failed as no matter how carefully this was carried out, one or two trailing edge holes always remained uncoated. This was due to the inability to obtain a blade where all the holes were completely filled with aluminising powder. However, by using gas phase CVD process and introducing the gases at point (C) then the following results were obtained

Quality control surveillance is difficult for these components as sectioning is the only way of assessing the complete success of any coating run. Experience has shown that only by using this CVD technique can the complete coating of internal passageways be assured.

CVD methods are now routinely used to coat the internals of land based turbines where the internal serpentine length maybe approaching 1 metre.

4. CVD of Chromium onto Forming Tools for the Food Industry

Tools used for the food industry for the closing of cans filled with, for example pasta or ravioli, were suffering from corrosion, probably due to the high salt content of the food material. Also an anti-pickup coating was required to stop transfer of the can end material onto the work piece during the seaming of the end into position. This pick-up would result in poor seaming and a can with an incomplete seal. It was, therefore, decided to investigate the deposition of chromium carbide onto the tool in question. It was necessary after coating that the bearing surface was also smooth and capable of being highly polished. Therefore it was not possible to use a pack cementation method as the resulting surface would be too rough due to particulate adhesion from the chromising pack. Therefore, it was decided to adopt the gas phase chromising route for this component.

Up to 30 components can be processed in one coating cycle. The resulting chromium carbide coating is up to 15 µm thick and has a hardness of 1500 Hv (0.1 Kg). The tools must be heat treated after coating and this is carried out in a vacuum furnace at 1030°C with a rapid quench after 30 minutes at temperature. The chromium carbide coating shows no distress as a result of this treatment. The resulting coated tool was able to withstand the arduous condition in the food canning plant and acceptable lives are routinely achieved.

Summary

CVD processes are used on a surprisingly wide range of industrial components, from aircraft and land gas turbine blades, timing chain pins for the automotive industry, radiant grills for gas cookers and items of chemical plant, to resist various attack by carbon, oxygen and sulphur. In this paper some essential aspects of the technology of CVD have been discussed and some examples of where CVD has been successfully used to overcome some specific industrial problems are illustrated.

The challenge for the future is to continue the development of CVD to even wider fields of applications, with new coatings on a variety of materials to meet the continuing needs of industry for long lasting solutions.

See Figure 14.