How Surface Prep Prevents CVD Coating Flaking on Graphite Parts

When coatings begin to flake, many parts made from graphite are subject to failure in the Cvd coating process. Sometimes the coating is the problem, however, and sometimes it is due to improper treatment of the surface prior to the coating. The bond is weak if the surface is dirty, oily or too smooth. At the interface little gaps are formed and are increased with the number of heating cycles. Surface preparation can enhance the adhesion for the coating and enhance thermal stability. Even a small defect in roughening and cleaning will cause the coating to fail resulting in expensive downtime of a production line. That's why surface preparation is typically the first thing techs look at.

Surface Roughness Optimization for High-Purity Graphite Substrates

If high purity graphite is to be used, one of the most important parameters to set is surface roughness. If the coating is too smooth, then of course, there's nothing for it to “grab a hold of.” May start out perfect but little spots may start to lift with the start of heating cycles and will turn into flakes. Finally, if the surface is too rough, the coating may not be level and cause weak points or gaps in it, which will also fail. The goal is to have a well-balanced texture. Think about applying a coating and then providing a stable base to the paint with no roughening the paint surface. In many production systems the balance is achieved through controlled grit blasting, and/or very fine machining marks left by the technician. The most important thing is to be consistent. If there are non-uniform stresses after heating and cooling cycles, they could be due to non-uniform scratching or patterns. The real application is a coating line for susceptors in semiconductor tools which are fabricated from graphite. Edge peeling was seen in the parts surface treated only lightly, and not surface treated after 2-3 thermal cycles. The roughening step was adjusted to be controlled and the level of flaking was decreased resulting in higher coatability. It just depended on how well the coating was able to be “locked” into place as it was being deposited. Surface contamination is one component of the problem that is often overlooked, and can be caused by roughening. If it is good texture, but still has power, then there may be some dust or handling oils that will not be removed, and it may be difficult or impossible to bond. Cleaning after surface treatment is also as crucial as the surface treatment process itself and hence cleaning immediately after surface treatment is important. It also aids with maintaining the roughness within a range as opposed to a guess. Many teams will do the microscopic checks and/or simple checks to the surface profile of the product to make sure that the product is consistent from one batch to the next. Once the proper coating range is established for any particular coating system, the maintenance of the range is one of the easiest procedures that could greatly help reduce flaking of the coating over time if the range is maintained.

Cleaning and Pre-Treatment Techniques Before CVD Coating

The cleaning and pre-treatment process have a significant impact on much of the CVD process and the subsequent treatment of parts coated by CVD. It should be clean, dust free and free from any machining debris and should be free from a thin film of oil which forms when the surface is handled otherwise a coating will not stick to the surface. At first it might seem OK but over time, these weak spots will expand and the coating will start to flake or peel off. A simple solvent cleaning is best. Parts are cleaned and wiped generally with approved solvents to remove light contamination and grease from parts. But that's not typically enough. A lot of shops adhere to it and they then use ultrasonic cleaning to dislodge any minute particles which may be trapped in the tiny pores within the graphite. It works particularly well in complicated shapes, where it's not possible to wipe hands. Another process that people tend to rush into is cleaning; this is followed by drying. During coating, it is possible to have bubbles of gas in the coating due to excessive moisture. They are the part of the car that are vulnerable and will be the first to fail. In fact, in some production teams, they even give a soft thermal baking before coating in the actual production. This will eliminate any moisture and volatiles that have been trapped in the graphite. Repeated peeling of the coating was observed in the center area of a susceptor in a semiconductor tool line. It was reviewed and it was decided the process was satisfactory as far as cleaning was concerned, but wasn't satisfactory as far as drying time was concerned. With more careful and extended drying the problem has been reduced greatly. After cleaning, the handling is an important factor as well. It is important to note that contamination easily adheres to bare surfaces of graphite, through skin contact or even clean room air. Hence, time, clean trays and gloves may be a part of the process flow. The key to good pre-treatment is NOT to perform one good cleaning step! It's about layering basic steps to eliminate various types of contamination. If each step is properly completed the coating will have a clean base and will also be stable while in service, reducing flaking in service.

Adhesion Enhancement Mechanisms in CVD SiC Coating

Once the surface has been cleaned and properly prepared, the next hurdle is obtaining good adhesion and stability of the Cvd sic coating when operating. It doesn't until here that adhesion mechanisms are introduced. It's more than just gluing! It is composed of surface adhesion, chemical bonding and type of coating growth at the time of deposition. Mechanical interlocking is one of the important factors. In a controlled texture graphite surface, the SiC coating is allowed to grow in small grooves. It helps to minimize the chances of peeling of the building during expansion and contraction in the temperature. There is also a major factor of chemical bonding. The first step of CVD is the reaction of the silicon atoms with the surface of the graphite. In certain instances, thin carbide transition layers can be obtained at the interface. This layer acts more as a binder that will bind the coating and substrate together rather than acting as 2 layers. A thin interlayer is also sometimes employed prior to full SiC deposition on some production lines. This layer helps to decrease the stress and to create a more uniform coating. Otherwise, there could be a few growth spurts or weak spots that will become peeling spots later in the game. An actual application is a susceptor for an epitaxy reactor used for a graphite. The part had been thermally cycled many times, resulting in edge flaking. After process review, engineers realized that the nucleation stage of the process was too fast, which led to the issue of poor bonding between the bottom layer. The amount of gas flow during the first deposition stage was varied to obtain a more even coating and later on adhesion was improved. Stress control is also a factor. When surface coating is subjected to a high rate of stress, it will try to relieve the stress when temperature is raised, which may result in cracks on the surface coating/connecting area. It's not just one change that will make good adhesion. It is an outcome of controlling the initiation, development and progressive adhesion of the coating.

Impact of Surface Defects on Coating Lifespan

Surface defects in the graphite can appear quite small, but can affect the lifespan of a Cvd sic coating . Very small scratches, pits or remaining particles can become weak points at high heat cycles. A minor fault can gradually develop into cracks, which eventually can cause the coating to flake or delaminate. Common problems are micro scratches caused by machine shop or handling. These scratches can be harmless, but will form irregular growth during coating. The coating is not even in the inside or in the scratch and causes hidden stress zones. As the part is heated, the expansion of the various zones will differ creating separation. Surface pitting is another issue. Graphite has natural porosity, but if the surface is not treated to become smooth or filled, these tiny holes will become a sponge for gas and impurities. As the trapped gas expands during CVD, it can create little blisters under the coating. The blisters tend to break during repeated thermocycling. Contamination is also a silent issue. Fingerprints will cause a thin layer of oil that can prevent bonding in that area. Random edge flaking was persistent on finished coatings in one production line on which graphite susceptor parts were being produced. The cause was found, it was handling marks from the transferring between cleaning and coating steps. The problem was reduced after the handling was tightened and contact was reduced. In particular edge defects are of interest. Rough corners or sharp edges are the areas more likely to be the subject of stress when heated. Usually, these zones have higher degradation rates of the coating, which means that prior to the coating deposition, many processes require an additional “edge rounding.” The primary rule is the coating will not mask surface defects. Looks at the bottom and copies the shape and state of it. The surface defects will lead to failure of the coating. Smoothness, cleanliness and stability of the base surface helps to guarantee the longevity, reliability and stability of the coating.

Standardization Across AIXTRON, Veeco, and LPE Platforms

With the use of CVD across various platforms such as AIXTRON, Veeco, and LPE, one hurdle arises: Each machine has a slightly different operating style, but the coating must perform the same way in both. If the results are not the same when the recipe is the same, it may be due to a lack of standardisation of the tools. Often teams synchronise on a surface first. The cleaning and roughening procedures for the graphite parts that will be put in the depot is the same as outlined above. This includes frequent solvent cleaning, drying time and handling procedures. If one line skips or shortens a step, this can result in a coating life difference in the platforms, which can be seen later. There should also be an alignment of gas flow and temperature profile. The temperature and distribution of gases in the AIXTRON, Veeco and LPE systems can be different, influencing the beginning of the formation of the coating on the surface. By this point, the nucleation can be faster in one system than in another resulting in a change in the coating structure – even if the input recipe is the same. This contrast can manifest itself as density or internal stress differences.