Review of PVD Process and TBC Application For Turbine Blades
Table of Contents
1.4.1. Metallic Bond Coat
1.4.2. Thermally Grown Oxide
1.4.3. Ceramic Topcoat
2.3.2. Electron Beam
2.3.4. Pulsed Laser Deposition
2.4. PVD vs CVD
3.1.1. Vapor Pressure
3.1.2. Gibb’s Phase Rule
3.1.3. Multiple Atoms
3.2.1. Mean Free Path
3.2.2. Glancing Angle Deposition
Thermal barrier coatings (TBCs) are thick films applied to metal components to provide thermal protection against hot combustion gases. TBCs are most often used in gas turbines, which are used to power and move ships and airplanes or to generate electricity such as a land-based gas turbine (generator).[2,3] The coating is sprayed as a heated, or molten, powder or vaporized onto such high-temperature components as the combustor liner, turbine blades, and vanes.[2,4] They have also found use in braking systems, internal combustion engines (i.e. combustion chamber, pistons, and valves), rocket engines, and nuclear reactors.[2,5] Before the introduction of TBCs there was extensive research into structural materials possessing greater temperature performance using textured microstructure, single-crystal blades, improved alloy design, and cooling by airflow through internal channels. Convection cooling offers the most cooling to the component, even more than TBCs. Early research in TBC coatings focused on increased adhesion between the ceramic topcoat and metallic bond coat since the ceramic was prone to spalling. Ensuring chemical and mechanical compatibility between the layers helped to demonstrate the reliability of the coatings, and today, coatings have many improved lifetimes, surviving tens of thousands of hours in service. In the early 1990s, the focus shifted from improving reliability, by preventing spallation, to lowering the thermal conductivity of the ceramic material to impede thermal transport.
Developing a coating to meet the many, often conflicting, goals is achieved by tailoring the chemical composition and coating microstructure.[2,5,6] TBCs need to be resistant to thermal, corrosive, and mechanical wear that is encountered in the harsh environments in which they are used. TBCs should have erosion resistance, that is, resistance against mechanical wear caused by impinging particulate debris or liquid droplets such as fuel or water.[1,2,4] The failure mechanism proposed for a TBC exposed to high temperatures for long duration is thermal fatigue and oxide (TGO) growth.[5,6,8] TBCs should have thermo-cyclic fatigue (TCF) resistance, and thus be tolerant of the strain caused by repeated heating and cooling, as well as, the strain caused by TGO growth.[5,6,8] To lessen the thermal stress mismatch, the ceramic layer can have vertical cracks intentionally introduced. The difference in the thermal expansion (TE) of the airfoil and TBC system layers results in strain, and this induces residual stress in the TBC system. TBCs deposited on high-pressure turbine blades have stresses of 2.4 – 2.8 GPa as-prepared and 2.3 – 3.1 GPa after service, which can vary based on the morphology of the TGO. To reduce the elastic strain energy it is important to control the porosity architecture (i.e. pore volume fraction, geometry, morphology, and distribution) and the thickness of the thermally grown oxide (TGO).[2,3,9] Since the growth of the TGO applies significant strain to the TBC system, controlling TGO thickness ultimately controls the lifetime of the coating.[5,6] TBCs with good durability is achieved at a porosity level of 10% (15%10). These voids are created by entrapped air and a lack of complete overlap of splats (flattened molten particles that occur during impact with the substrate) in plasma spraying (PS). Pores that fully penetrate the ceramic coating help the ceramic and metallic layers remain bonded. Porosity is also important in lowering the thermal conductivity of the ceramic.[2,3,9] Using a porosity forming polymer, which is burned out during heat treatment, results in the highest amount of porosity at 23-27%.
TBCs reduce component temperatures by 190°C (150°C), allowing the engine to run hotter, and produce more work, while keeping parts cooler, thereby improving engine performance, efficiency, and fuel economy (and reducing pollution), as well as, part durability and thus lifetime.[2,6] TBCs protect the substrate from high temperatures, as well as, harsh oxidation and combustion environments.[2,11]
These well-engineered structures are composed of a three-layer case: a metallic bond coat, a TGO, and a ceramic topcoat. To enhance the performance of coatings, improvement can occur in lowering the intrinsic thermal conductivity, controlling the porosity (introducing vertical type cracks), and increasing the thickness of the ceramic topcoat. A low density offers a 15% reduction in thermal conductivity. A large reduction in thermal conductivity and improved TCF lifetime have been found for TBCs deposited at high chamber pressures (1.5 X 10-3 Torr) with an oxygen and inert gas environment. TBC layers are 0.1 – 2 mm thick (300 – 500 μm). However, the failure of TBCs using air-plasma spraying (APS) occurs when the thickness is >1.5 mm. The different coating layers and dimensions are shown in figure 1. The Thornton Structure Zone model shows how varying the temperature and pressure results in considerably different microstructures (figure 2).
1.4.1. Metallic Bond Coat
The metallic layer provides corrosion resistance to the substrate since the ceramic layer provides little or no corrosion resistance due to its extremely porous structure. Oxidation is a specific type of corrosion that occurs in high-temperature applications where the air is present. The metallic bond coat, due to its roughness (Ra = 5 – 10μm), has good adhesion to the ceramic layer. The two types of metallic bond coats are diffusion and overlay.[5,14] Diffusion coatings result in better adhesion, due to increased diffusion, are less expensive, and tend to be thinner. Overlay coatings are affected by minimal elemental interdiffusion but still bond fairly well and offer greater composition flexibility. Both diffusion and overlay coatings are corrosion resistant and result in a spall resistant, slow-growing aluminide (intermetallic, i.e. NiAl, Ni3Al, CoAl) scale to protect against oxidation.[4,15] NiCrSi is another composition option but only protects up to 850°C. The metallic bond coat layer is 130 μm thick (200 μm, 190±25 μm, 120 μm). The bond coat is either a low sulfur platinum aluminide diffusion type or MCrAlY overlay type (where the M represents Ni and/or Co5,11,15,17,18, Ni protects against high-temperature oxidation and Co against hot corrosion).[5,17] The bond coat composition is Ni-22Cr-10Al-1.0Y5,11 and is applied to a substrate of composition Ni-20Cr-20Co-5.9Mo-0.5Al-2.1Ti-0.4Mn-0.3Si-0.06C in wt%. The particle size is 56 – 106 μm.
1.4.2. Thermally Grown Oxide (TGO)
When the bond coat oxidizes a TGO layer is formed. This occurs either before or during ceramic layer deposition. Initially, the TGO layer consists of dense alumina (α & γ phase Al2O3 from α-Al2O3 powder[7,19]). The TGO layer is 1 μm thick, grows to 3-9 μm, and if grown to 10 μm will fail.
1.4.3. Ceramic Topcoat
The topcoat decreases the bond coat surface temperature by 100 – 200°C (110°C,140°C, 100-120°C, 170°C). In one experiment, temperatures on the surface of the topcoat, for two different ceramic thicknesses of 600±50μm and 300±50 μm (applied to a 300±50 μm bond coat and a 5 mm thick substrate) were 1700 – 1900°C and 280°C respectively. The temperature drop across the ceramic layer is a function of the coating’s thickness and its thermal resistance. The insulating layer should have low thermal conductivity and high-temperature capability (sintering resistance), as well as, chemical and mechanical compatibility with the bond coat. The standard refractory used is 6 – 9 wt% yttria-stabilized zirconia (YSZ). The maximum ionic conductivity is obtained at a yttria (Y2O3) concentration of 8 wt% (1000°C). The substitution of some zirconium (Z4+) ions with slightly larger yttrium (Y3+) ions acts to stabilize the zirconia (ZrO2) crystal structure and prevent high-temperature phase transformations.[6,17] Phase transformation of the ceramic causes a volume change and thus increased lattice strain. This occurs when the YSZ exceeds 1200°C (1170°C) in service transforming from monoclinic to tetragonal. This is a slow process and may require temperatures more than 1350°C for over 100 hours before the monoclinic zirconia appears. Upon cooling, the crystal lattice transforms back to monoclinic, resulting in a damaging 4%[7,17] (4-5%) volume expansion. Introduction of oxygen vacancies, which occur when adding Y2O3 to ZrO2, is the other mechanism that helps to slow the rate of decomposition to tetragonal phase: the more stabilizer added, the more oxygen vacancies created. An oxygen vacancy is created when two Z4+ ions (Rion = 0.82 Å), that reside in the zirconia structure, are replaced with a Y3+ ion (Rion = 0.96 Å) and four O2- ions are replaced with only three O2- ions. TBCs containing rare-earth metals other than Y3+ are being developed with even lower thermal conductivity. One dopant alternative is dysprosium-stabilized zirconia (DySZ), whereby substituting the Y3+ ions with larger Dy3+ ions (adding Dy2O3) a lower conductivity is achieved by enhanced phonon scattering (lattice vibration). Similar spraying capability and thermal cycling behavior were observed as compared to conventional YSZ powders. The ceramic, topcoat layer is 300 – 500 μm thick (APS- 600±50 μm, 130 – 380 μm, 230±25μm, 150 μm, 100 μm – 1mm). However, ceramic layers thicker than 300 μm are more likely to spall. The particle size is 11 – 150 μm.[5,11]
In general, PS is used to coat nozzle vanes and EBPVD for turbine airfoils. EBPVD is preferable for the blades because it applies a smooth surface, resulting in better aerodynamics properties, with less interference with cooling holes. Gas turbine engines are a $42B industry worldwide; 65% are used in jet engines and 35% are used in land-based generators. Land-based generators are fueled by natural gas, or liquid fuels, producing 25% of all electricity in the US and 20% of all global electricity (2010). Over the last few decades, alloy and process development has led to blades, without a TBC, that can handle average temperatures of 1050°C with occasional excursions (hot spots) of 1200°C. However, the temperatures in the hot section (combustor and turbine) can reach 1650 – 2200°C if there is advanced cooling and 1550°C if there is no cooling; and in a 747 & DC10 gases reach temperatures above 2,000°C. Modern engine turbines experience temperatures of 1370°C and a jet engine turbine, such as the Snecma M88, is 1590°C. The material commonly used for turbine blades is the superalloy NiCoCrAlY which has a melting point of 1330°C (1200°C is 90% the melting point of the blade material). Therefore the gas temperatures the blades are exposed to are higher than the melting point of the blade material. Using cooling schemes along with TBCs has allowed components to withstand gas temperatures over the melting point of the superalloy (>250°C of the melting point).[4,20] The melting temperature of the TGO (i.e. Al2O3) is higher than the gas temperatures at 2053°C.
The methods used to apply TBCs are physical vapor deposition (PVD) and plasma (thermal) spraying (PS). There is also chemical vapor deposition (CVD) and solution precursor plasma spraying (SPPS) which are under development (2004). PVD coatings are prepared by condensation and CVD coatings by decomposition. Thermal spray coatings are prepared by spraying melted, or heated, material on the substrate, where it solidifies. Heating is performed using a plasma, an arc, or chemical means. Techniques include APS, high-velocity oxy-fuel (HVOF), low-pressure plasma spray (LPPS), high-frequency pulse detonation, and pack cementation.[5,6,17,26] APS uses a plasma jet (electrical method) where a mixture of inert gases (i.e. Ar & H) are fed past an electrode forming the plasma. The powder particles are then fed into the plasma and attain a speed of 350 – 560 m/s. HVOF uses a combustion flame (chemical method) comprised of a liquid (i.e. kerosene, etc) or gaseous (i.e. acetylene, propane, propylene, hydrogen, methane, natural gas, etc) fuel and oxygen to spray powder at 700 m/s. The powder that is fed into the combustion flame is melted on the way to the substrate. The HVOF process produces denser, stronger coatings with an improved bond coat oxidation resistance, and thus performance, and is simpler. However, the coatings in APS are cleaner since HVOF yields impurities from the combustion of fuel. The LPPS method is the most cost-prohibitive but gives a dense coat without oxide formation. The macroscopic properties of a coating depend on which deposition process is used (and its parameters) and properties of the particles such as size (and size distribution) and morphology.
Physical vapor deposition (PVD) was popularized in the 1966 book Vapor Deposition, and in 1838 Michael Faraday created the first glow discharge in a vacuum tube. The vacuum tube, which allows electric current to be controlled in a vacuum, consisted of 2 brass electrodes and a 2 Torr vacuum. Enough electrons were made to make enough ions (and electric neutrals) to create enough electrons (ejection of secondary electrons) to sustain the discharge. As of 1996, the worldwide market for functional coatings is €83 billion with vacuum deposition accounting for €200 million. The distinction between thin and thick film coatings is not based on the thickness of the film, rather, it is on the processing techniques used. Thin films are made by building up individual molecules (i.e. PVD), whereas, thick films are made by depositing particles (i.e. PS).The electrochemical deposition has found widespread commercial application. However, this method has limited capability of depositing ceramics. The process deposits material (anode) onto a conductive surface (cathode) by immersing them in an electrolyte solution containing metal salts and ions which permit the flow of electricity. A power supply oxidizes the anode material causing it to dissolve in the liquid medium, then the electrolyte deposits the ions on the cathode. PVD compared to electroplating (i.e. chrome plating) results in more uniform coatings that typically weigh less. This is important in reducing the centrifugal stresses on turbine blades during its operation.[2,3] Plus, using a chromic acid oxidation process is toxic and carcinogenic and potentially harmful to the environment.
The PVD process takes place in a vacuum chamber and either requires a heat source to evaporate/sublimate the target material or energetic ions to bombard it. The volatile vapor is transported in a vacuum or low-pressure plasma environment to the substrate. The vacuum removes air, which can act as a contaminant and lower the density of the coating, from the system; additionally, it reduces collisions between air and source vapor making deposition less diffuse. Air introduces point defects by
adding extra oxygen atoms disrupting the atomic, and ionic, arrangement in the crystal structure. Typical PVD coating growth is 1-10 nm/sec. A surface modification process can take place to the substrate before and/or after the deposition of the TBC system. Plasma nitriding, or ionitriding, thermally diffuses nitrogen into the substrate by heating the substrate to 500-550°C and placing it in a gaseous atmosphere of nitride (other environments are carbide, boride, etc) for 48 hours.[13,28] Shot peening can modify a sputter-deposited overlay coat making it denser and adding residual stresses.
Thermal evaporation is the most common technique for vaporizing material at temperatures <1500°C. The source material is heated, in a vacuum (>10^-4 Torr), to a temperature at which its vapor pressure is 10^-2 Torr[13,25,28] (0.1-0.3 eV) giving a useful deposition rate.[28,30] PVD is a distilling process so it yields coatings of high purity.[30,31] The vaporized source material, assuming no collisions during transport, travels in a straight line to the substrate, known as a line-of-sight deposition, and the flux is assumed from a point source or cosine emitter (flux distribution from a Knudsen cell, used in MBE, is given by the Knudsen effusion model by Roth and Hirth).[3,28,30] Heating is achieved by conduction and radiation; convection is negligible in a vacuum environment. One method using conduction heating is to resistively evaporate material either directly, using a heated coil (filament), or, more commonly, heating a dimpled boat containing the material (charge) (see figure 4).[28,31] The filament (charge rod) is a refractory of either W, Ta, Mo, C, or BN/TiB2 and melts the source rod. The source rod should be in good thermal contact with the filament for efficient heating and to allow wetting or flow of the molten source over the filament. An advanced form of resistive evaporation is Molecular Beam Epitaxy (MBE) or Vapor Phase Epitaxy (VPE) which uses a high or ultra-high vacuum (>10^-9 Torr) and Knudsen cells for evaporation. The effusion cells allow the temperature and thus evaporation to be carefully controlled. The much higher vacuum and absence of a carrier gas result in the highest achievable purity of the grown films. MBE configuration is such that neutral thermal energy beams (molecular or atomic) impinge on the heated substrate. The TBC ceramic layer should have a low density and this occurs at high chamber pressure and low substrate temperature, therefore, MBE would not be effective in this regard.
2.3.2. Electron Beam
As opposed to radiative heating with an electric filament, the electron beam PVD (EBPVD) process is commonly used when vaporizing material at temperatures >1500°C, and is also useful for evaporating large quantities of material. This method employs a high energy e-beam to deposit refractory materials. Electrons are generated from a thermionic emission source and accelerated using a high voltage (70 kv, 60 kv, 10-20 kV[9,28]). The accelerating electrons are focused and deflected using electromagnetic (EM) fields (see figure 5). E-beam guns usually operate at 50-100 kW, with some operating at 150 kW (similar to plasma guns which operate at 50-120 kW). The methods used to apply the topcoat are either APS or EBPVD (or ion enhanced EBPVD- IE EBPVD).[5,26] EBPVD deposits coatings of uniform thickness (in line-of-sight) on all but inner surfaces of complex geometry, due to high chamber pressure scattering the vapor which creates a ribbon-like structure normal to the plane. The downside to this columnar structure and segmentation cracks is higher thermal conductivity and greater permeability to oxygen and molten salts (i.e. Na and K sulfates).[4,8] The electrochemical reaction, known as hot corrosion, between the substrate and the salts accelerates spallation. When the salt builds up and the temperature is above the melting point of the salt, it wicks through cracks and pores in the TBC and upon freezes causes cracking. However, this structure is more strain tolerant and thus compliant in the direction parallel to the interface resulting in good thermal fatigue resistance.[8,9] EBPVD coatings last 8-10 times longer before spalling compared to APS coatings, therefore, EBPVD coatings are preferred in aerospace applications where extreme thermal cycling occurs. IE EBPVD has the advantage of increased coating adhesion due to ion bombardment that provides in-situ cleaning, as well as, surface modification resulting in a denser microstructure from smoothing of pits, scratches, and evaporated droplets on the surface. Additionally, it can be used to heat the substrate which enhances diffusion and chemical reactions. In EBPVD the vacuum is at least 7.5 X 10-5 Torr to allow passage of electrons and ionized gaseous atoms. Electrons are generated from either a thermionic source (i.e. electron gun), field electron emission (i.e. electrostatic field), or an anodic arc. When using a gaseous arc a vacuum of only 10^-2 Torr is used so that the gas does not get entrapped in the coating. A typical arc is initiated if a high current density (10^4-10^6 A/cm2) (> 1010 A/cm2) and low voltage (>25 V, 10-30 V) is passed through a gas or vapor of the electrode material.[28,33]
Sputtering can apply a wide variety of materials such as ceramics, metals, and alloys, as well as, organic and inorganic compounds. Sputtering is not currently used to apply TBCs due to the slow deposition rate, even though it creates a columnar structure, with elongated grains normal to the interface, which is the ideal microstructure for TBCs. The sputtering process physically erodes solid material from the source by ion-bombardment (plasma: Ar-H or Ar-He, a few % of enthalpy enhancing gas such as H). Surface coverage on complex substrates is achieved due to the higher gas pressure (10^-1-10^-3 Torr4) which scatters the source material.[3,4,29] These ions (few eV- 10’s eV) are created when electrons are accelerated and bump into gas molecules which decreases or increases the number of valence electrons in the gas atoms creating positively or negatively charged ions respectively. These ions which are accelerated to the substrate arrive with a nearer than normal angle-of-incidence (collimation) and can modify the surface by creating defects and reactive sites. Thermal and sputter deposition can be employed in the same system; sputtering the minor constituent for example. The target is bonded to a water-cooled Cu backing plate to prevent melting or outgassing of the target material due to ion-bombardment. Additionally, since the radiative heating is less, compared with thermal PVD, the substrate can be placed closer to the source. The ions could be replaced with neutral fragments of gas molecules or atoms (i.e. free radicals) which have high chemical reactivity, but it is easier to accelerate ions. The plasma is used to clean the substrate (“back-sputtering”) before deposition and modify the film during deposition (“re-sputtering”). The high-speed ions can implant or otherwise damage the substrate. The setup can include a high-voltage DC power supply connected between the sputtering target (cathode) and the workpiece (anode). By biasing the substrate the ions can be accelerated at the substrate, as well as, the source, and if the unreacted gas is not incorporated into the film, it will enhance chemical reactions and provide kinetic energy making for a denser coating. In reactive sputtering, the inert gas is replaced with a reactive gas such as O, carbides, or nitrides. A magnet is usually used to localize the plasma around the area of the target that is to emit material, as opposed to somewhere else on the cathode, and it also prolongs residence time. Sputtering with a magnet (cathode) is known as magnetron sputtering. When coating a non-conductive substrate (i.e. polymer), radio-frequency (RF) sputtering is employed to prevent charge build-up.
2.3.4. Pulsed Laser Deposition
Cathodic arc PVD (Arc – PVD), similar to EBPVD, uses a high current, low voltage arc to strike the surface of the source material. The positively charged cations (i.e. Ar+) are accelerated towards the cathode source (uncommon is anodic arc PVD where the source material is the anode instead of the cathode). This gives rise to a cathode spot; a highly energetic emitting area. The cathode spot, size of a few μm, yields a high velocity 10^2 m/s jet of vaporized cathode material when subjected to a high current density of 10^6 -10^12 A/cm2. The vapor (>85% ionized, 10-100 eV) is guided to an anode which is either the vacuum chamber wall or the surface to be coated. An EM field is used to direct and rapidly move the arc over the entire surface of the cathode target. The plasma beam from the cathodic arc source contains some large clusters of atoms, or molecules (macro-particles), which should be filtered out so that a high-quality thin film can be achieved.
2.4. PVD vs CVD
CVD was employed as early as 1880 in incandescent lamps.[13,25] It involves high-temperature decomposition or reduction of a chemical vapor precursor and differs from PVD during the deposition of the vapor on the substrate.[28,30] CVD uses gases, evaporating liquids, and chemically gasified solids as the source and deposition involve a chemical reaction, not just a phase change. The term vacuum or evaporation (vapor) deposition was coined in an attempt to integrate the two methods. The existence of the hybrid-physical CVD process, HPCVD, illustrates the overlap that can take place. CVD is carried out at much higher temperatures than PVD (>800°C vs 200– 500°C), and higher temperatures than its reactive gases. The increased temperature at the substrate helps to desorb the reacting gas and results in more diffusion bonding, giving the metallic layer better adhesion. The chemical reactions in CVD result in corrosive volatile byproducts and unused reactive gas that must be removed by gas flow from the chamber.
3.1.1. Vapor Pressure
The source material is subjected to a certain pressure and temperature by employing a vacuum and applying heat. By using a high vacuum and lowering the pressure on the source, to below its triple point, then adding heat, the source can be made to sublime instead of evaporating.[29,31]
3.1.2. Gibb’s Phase Rule
Any system under equilibrium conditions conforms to the phase rule; the equality is F = C – P + 2, where F is the number of degrees of freedom (dof), C is the number of components, and P is the number of phases in thermodynamic equilibrium with each other. Phases are similar to the four states of matter; gas, liquid, solid, and plasma, except its possible to have several separable, immiscible, phases that are of the same state of matter.[25,28] Components are the amount of chemically different
species. Dof is the amount of parameters: temperature, pressure, and components which may be varied independently without causing a change in the number of phases. The phase rule applies to all of the deposition processes as long as the system is at equilibrium.
3.1.3. Multiple Atoms
When depositing multiple atoms the term partial pressure is used to describe the vapor pressure of each of the gaseous species. Partial pressure describes the pressure of each constituent as if it was the only one in the system. Adding together the partial pressures for each vapor yields the total pressure of the system; and the vapor pressure of the source determines the lowest pressure the vacuum can achieve. If multiple atoms with different vapor pressures are evaporated from a single source, described by Raoult’s Law, this will result in compositional fluctuations across the TBC thickness.[2,28] This can be avoided by separating the components and using different energy levels for evaporation. Fortunately, the vapor pressure of most rare earth oxides is close enough to that of Zr, that, no major composition problems should result by evaporating from a single source. However, oxides that tend to sublime, such as ceria, magnesia, and silica, require additional processing efforts. To avoid sublimation, an element that normally sublimes can be alloyed with another element to give a liquid melt. Compounds often vaporize, or sublime, as anywhere from atoms to clusters of molecules to dissociated, or partially dissociated, molecules.
Transport consists of mainly three steps: transport reactants to the substrate, chemical reaction at or on the substrate, and transport of products away from the substrate. The first and third steps are interdependent and the middle step involves surface and/or gaseous reaction.
3.2.1. Mean Free Path
The degree of vacuum should be >10^-4 Torr so that the mean free path (MFP) – distance an atom travels before colliding with another atom- is greater than the distance between the source and the target (10-16 inches).
3.2.2. Glancing Angle Deposition
Controlling the angle-of-incidence of deposition results in unique morphologies. This angled, off-normal, deposition results in columnar growth in the direction of the flux. Very oblique deposition angles (GLAD) exacerbate the columnar growth since the valleys do not get any flux, resulting in a more porous structure. Rotating the substrate during GLAD results in corkscrew columnar growth. This “Zig-Zag”, or “Herringbone”, the structure has an increased path length for thermal transport and offers a 40% reduction in thermal conductivity.
If a vapor becomes a solid without first becoming a liquid this is called deposition. Deposition rates vary greatly in PVD, ranging from <1 MLS (<3 Å/s) to >10^4 MLS (>30 μm/s). The further the substrate is moved from the source, the more uniform the coating is but the deposition rate is decreased as 1/r^2.
Atoms that land on a substrate and do not react immediately will have some degree of mobility. These mobile atoms, that lie on the substrate (crystal plane), are known as adatoms. Adatom is a contraction of adsorbed atoms (another term is adparticles). Many arriving atoms are in a thermodynamically unfavorable state and depending on the energy of the atom, the atom-surface interactions (chemical bonding), and the temperature of the surface will dictate the amount of mobility (diffusion) that takes place. It is unlikely that adatoms diffuse into stable lattice positions if the substrate temperature is low and there is reduced ion-bombardment (see figure 6). When the atom is immobilized where it lands this is known as a ballistic deposition, aptly named since the projectile motion of the atom is the only effect on the deposition.
Condensation describes the process of a vapor atom cooling (or being compressed to its saturation limit) to a liquid. Not every atom will condense, some atoms are re-evaporated or reflected, especially if the energy of the vaporized atoms or substrate temperature is too high. The term used for the ratio of the condensing atoms to impinging atoms is called the sticking coefficient. After finding preferential nucleation sites the adatom will condense on the surface, assuming the temperature of the part is less than the adatoms melting temperature, by losing energy through diffusion, bonding to other atoms, and collisions with deposited and incoming atoms. Energy mostly in the form of heat of vaporization (enthalpy change on condensation) or heat of sublimation is given up.
In plasma spraying rapid solidification occurs when liquid particles (some vapor) strike the substrate surface, flatten and cool (10^6 K/sec) to form splat (fine, non-columnar, equiaxed grains). The splats develop a lamellar structure parallel to the interface by directional solidification. In PVD, solidification is described by columnar growth where elongated grains are oriented normal to the substrate.[4,28] A new process known as plasma spray-PVD (PS-PVD) is a combination of LPPS and PVD. There is both molten feedstocks that form splats and vapor transported in a supersonic plasma plume to condense. PS-PVD gives superior coating properties than PS or PVD achieve on their own.
Sintering is the high temperature (heat) treatment that causes particles to diffuse together, fusing the ceramic layer, without the very high melting point of the ceramic being reached and the ceramic liquefying. Sintering resistance during service is important for the ceramic topcoat; a desirable structure is created during the spray process that gives strain tolerance and low conductivity, due to the high volume of pore space. Improvements made in lowering the thermal conductivity of the ceramic layer results in a hotter ceramic surface and thus a greater likelihood of sintering. In addition, a thicker coating increases the chances that sintering will occur because the heat conduction distance is increased which decreases the rate of heat transfer. Thermal conductivity of P-YSZ coatings produced by EBPVD are higher at 1.8–2.0 W/mK (1.5–1.9 W/mK) compared to APS at 0.6–1.2 W/mK (0.8-1.0 W/mK).[2,17] The tradeoff for low values of thermal conductivity, in APS, will be an increase in sintering due to the higher temperatures generated on the surface, and thus an increase in thermal conductivity. Sintering of the microcracks at the splat interfaces is the main cause of the increased thermal conductivity.
Improved thermal surface protection is important for improving engine technology, which continually strives to run engines hotter. Advances in processing resulting in single crystal blades and directional solidification has provided a 120°C temperature improvement for turbine blades.Cooling of the hot structure (blades, disc, and vanes) is the most effective technique at achieving longer part lifetimes but a cooler blade results in lower engine efficiency. Using cooling provides a 370°C (air cooling- 200-300°C) temperature reduction on the blade surface while only using 1-3% of the main airflow.[4,36] Cooling schemes, using either air or liquid, are essential to protect the blade from degradation. However, if the cooling air could instead be used for propulsion it is estimated that this would save 10M gallons of fuel annually in 250 aircraft. TBCs, on the other hand, allow for both blade durability and high efficiency. The advantage of using liquid is its higher specific heat capacity compared with air, however, leaks can occur. A common method, using convection cooling, routes compressor air inside the hollow airfoil and out the many pinholes in its surface. Historically, the focus was on preventing creep. Now the main concern is preventing oxidation and hot corrosion (sulfidation) of the metallic bond coat and thermal-mechanical fatigue (TMF). Elements that provided corrosion resistance are Al, Ta, Zn, Ni, and Cr (and Ca which is being phased out since the dust is carcinogenic). As engine operating temperatures increased the high Cr contents (20%, 25-35%) that prevented hot corrosion were phased out with higher Al content (4-5%, 6%) to provide better oxidation resistance. Resistance to chemical reactions with gas impurities, such as S and V or deposits like Calcium-Magnesium-Alumino-Silicate (CMAS), is important. Recent studies have shown success with other ceramic types that use earth dopant oxides other than yttria.[2,6] Many TBCs containing constituent oxides other than yttria (see figure 7) have been introduced into service or testing with an underlying P-YSZ ceramic layer. This is partly due to the chemical incompatibility of the new ceramic material with the TGO, as found especially for pyrochlores (cerium metals), and partly the new ceramic altered the formation of the TGO layer resulting in reduced lifetimes. Improvements in TBC systems are possible through the use of multi-layer (i.e. micro-layering), or phase mixture ceramics.[2,4,28] One method of layering, at a thickness of 1 µm, is to switch on and off a bias voltage.[2,4,28] This produces layers of different density and yields an impressive 37-45% reduction in thermal conductivity.[2,4] Functionally graded materials (FGM) offer increased part durability by preventing spallation.[2,4,28] One method of blending TBC layers is by vaporizing metal and metal oxides at the same time.[2,4,28] This creates a gradient structure between the bond coat and the ceramic layer that helps to minimize TE mismatch.[2,4,28] This is already performed by vaporizing alloys from a single source, yielding a graded composition as the evaporant is selectively vaporized.[2,4,28] FGM may demand the use of alternative, multiple, and/or both coarse and fine powdered rare-earth metal dopants.[2,4,28] Additionally, improved coating performance is achieved by using ultrafine, or nanopowders, (i.e. gas evaporation) due to the unique microstructure that results.[2,4,28] This structure has greater sintering resistance owing to its nanopores and nano-grains.[2,4] Additionally, using both coarse and fine powders.[2,4] Lastly, if composites are used in the next generation of turbine blades the substrate will possess very different diffusion rates and surface chemistry.
1. D. Clarke, S. Phillpot. 2005. “Thermal Barrier Coating Materials.” Mat’ls Today. Elsevier. 22-9. http://dx.doi.org/10.1016/S1369-7021(05)70934-2
2. U. Schulz, B. Saruhan, K. Fritscher, C. Leyens. 2004.“Review on Advanced EB-PVD Ceramic Topcoats for TBC Applications.” Int. Journal of Applied Ceramic Technology. 1  302-1.
3. T. Rodgers, H. Zhao, and H. Wadley. 2013. “Thermal Barrier Coating Deposition by Rarefied Gas Jet Assisted Processes: Simulations of Deposition on a Stationary Airfoil.” Journal Vacuum Science Technology A. American Vacuum Society. 31  1-14.
4. “Coatings for High-Temperature Structural Materials: Trends and Opportunities.” 1996. Committee on Coatings for High-Temperature Structural Materials, Commission on Engineering and Technical Systems, National Research Council. National Academy Press. 102 p..
5. Z. Lu, S. Myoung, H. Kim, et. al.. 2013.“Microstructure Evolution and Interface Stability of Thermal Barrier Coatings with Vertical Type Cracks in Cyclic Thermal Exposure.” Journal of Thermal Spray Technology. ASM International. 22  671-9.
6. Curry, N. Markocsan, L. Östergren, et. al.. 2013. “Evaluation of the Lifetime and Thermal Conductivity of Dysprosia-Stabilized Thermal Barrier Coating Systems.” Journal of Thermal Spray Technology. ASM International. 22  864-72. http://dx.doi.org/10.1007/s11666-013-9932-9
7. A. Karaoglanli, K. Ogawa, A. Türk, et. al.. 2013. “Thermal Schock and Cycling Behavior of Thermal Barrier Coatings (TBCs) Used in Gas Turbines.” InTech Open. http://dx.doi.org/10.5772/54412
8. R. Miller, W. Brindley, M. Bailey. 1989. “Thermal Barrier Coatings for Gas Turbine and Diesel Engines.” NASA Technical Memorandum.
9. S. Terry, J. Litty, C. Levi. 1999. “Evolution of Porosity and Texture in Thermal Barrier Coatings Grown by EB-PVD.” The Minerals, Metals and Materials. Elevated Temperature Coatings: Science and Technology III Society. 13-26. http://www.materials.ucsb.edu/LINKS/PROFclarke/TBC/24.pdf
10. L. Wang, X. Zhong, Y. Zhao, et. al.. 2014. “Effect of Interface on the Thermal Conductivity of Thermal Barrier Coatings: A Numerical Simulation Study.” International Journal of Heat and Mass Transfer. 79 954-967.
11. R. Ghasemi, R. Shoja-Razavi, R. Mozafarinia, et. al.. 2013. “Comparison of Microstructure and Mechanical Properties of Plasma-sprayed Nanostructured and Conventional Yttria Stabilized
Zirconia Thermal Barrier Coatings.” Ceramics International. Elsevier. 39 8805-13. http://dx.doi.org/10.1016/j.ceramint.2013.04.068.
12. N. Fleck, A. Cocks, S. Lampenscherf. 2014. “Thermal Shock Resistance of Air Plasma Sprayed Thermal Barrier Coatings.” Journal of the European Ceramic Society. Elsevier. 34 2687-94.
13. D. Mattox. 2003. The Foundations of Vacuum Coating Technology. Springer, 164 p. http://www.svc.org/assets/file/HISTORYA.PDF
14. Z. Lu, U. Paik, Y. Jung, et. al.. 2013. “Thermal Fatigue Behavior of Air-Plasma Sprayed Thermal Barrier Coating with Bond Coat Species in Cyclic Thermal Exposure.” Mat’ls.  3387-403. http://dx.doi.org/10.3390/ma6083387
15. H. Bernstein. 1999. “High-Temperature Coatings for Industrial Gas Turbine Users.” Proceedings of the 28th Turbomachinery Symposium. Texas A&M University. 179-88. http://turbolab.tamu.edu/proc/turboproc/T28/Vol28018.pdf.
16. “Progress in Thermal Barrier Coatings.” 2009. The American Ceramics Society. 583 p.. https://books.google.com/books?id=WpcmRsBsKKgC&printsec=frontcover&dq=progress+in+th
17. D. Hass, A. Slifka, H. Wadley. 2001. “Low Thermal Conductivity Vapor Deposited Zirconia Microstructures.” Acta Materialia.  973-83.
18. M. Dorfman, C. Dambra, S. Metco. 2001. “Thermal Barrier Coatings: Improving Thermal Protection.” Sulzer Technical Review.  10-13. http://www.sulzer.com/mr/-/media/Documents/Cross_Division/STR/2001/2001_04_10_dorfman_e.pdf.
19. R. Dutton, R. Wheeler, K. Ravichandran, et. al.. 2000. “Effect of Heat Treatment on the Thermal Conductivity of Plasma-Sprayed Thermal Barrier Coatings.” Journal of Thermal Spray Technology. 9 (2) 204-9. http://dx.doi.org/10.1361/105996300770349935.
20. B. Gleeson. 2006. “Thermal Barrier Coatings for Aeroengine Applications.” Journal of Propulsion and Power. 22 . http://dx.doi.org/10.2514/1.20734.
21. A. Lepeshkin. 2012. “Investigations of Thermal Barrier Coatings for Turbine Parts.” Intech Open. http://cdn.intechopen.com/pdfs-wm/29754.pdf.
22. H. Purwaningsih, L. Noerochim, R. Fajarain, et. al.. 2010. “Phase Transformation on Interface between NiCoCrAlY Bond Coat and Substrate and Study of Thermal Barrier Coating as High-Temperature Material.” The Journal for Technology and Science. 21 (4). http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CCMQFjAB&ur
23. T. Pollock, S. Tin. 2006. “Nickel-Based Superalloys for Advanced Turbine Engines: Chemistry, Microstructure, and Properties.” Journal of Propulsion and Power. 22 . http://deepblue.lib.umich.edu/bitstream/handle/2027.42/77223/AIAA-18239-462.pdf.
24. M. Bäker. 2014. “Influence of Material Models on the Stress State in Thermal Barrier Coating Simulations.” Surface & Coatings Technology. Elsevier. 240 301-10. http://dx.doi.org/10.1016/j.surfcoat.2013.12.045.
25. C. Powell, J. Oxley, J. Blocher Jr.. 1966. Vapor Deposition. John Wiley & Sons, Inc. New York, London, and Sydney. 725 p.
26. “Thermal Barrier Coating System.” 1977. http://www.google.com/patents/US4055705.
27. M. Jarratt. “PVD Coatings: Welcome to the Future.” http://www.pvd-coatings.co.uk/history-pvd-coatings.
28. Mattox, D.M.. 2010. Handbook of Physical Vapor Deposition (PVD) Processing, 2nd Ed.. Elsevier. Oxford. 746 p. http://books.google.com/books?hl=en&lr=&id=aGUxoVTYjA8C&oi=fnd&pg=PP2&dq=handbook+physical+vapor+deposition&ots=bVqZb3dJ1s&sig=spm8qWp9PXik3haZer_Nkxygiqg#v=onepage&q&f=false.
29. D. Smith. 1995. Thin-Film Deposition. McGraw-Hill. 616 p.
30. “M12: Thin Films.” 2006. http://image.sciencenet.cn/olddata/kexue.com.cn/bbs/upload/8181Thin_Films1.pdf.
31. A. Doolittle. “Physical Vapor Deposition: Evaporation and Sputtering.” Georgia Tech. http://users.ece.gatech.edu/~alan/ECE6450/Lectures/ECE6450L12-Physical%20Deposition.pdf.
32. D. Ribeiro. “Thermal barrier coatings (TBCs) – State of the Art.” The University of the Minho. https://repositorium.sdum.uminho.pt/bitstream/1822/8084/1/Chapter%201.pdf.
33. Harsha, K.S.S. 2006. “Principles of Physical Vapor Deposition of Thin Films.” Elsevier. Great Britain. p.400. https://books.google.com/books?id=k8fI2BH1KVEC&pg=PR3&lpg=PR3&dq=Principles+of+Physical+Vapor+Deposition+of+Thin+Films&source=bl&ots=m5dwJYrCOf&sig=2OGtYV8gbVENOmI7QGy17k_H0rA&hl=en&sa=X&ei=FXnGVOPxJ_KLsQTHqYD4Bg&ved=0CEgQ6AEwAw#v=onepage&q&f=false.
34. J. Davis. “ASM Specialty Handbook: Tool Materials.” The Materials Information Society. 505 p.. https://books.google.com/books?id=Kws7x68r_aUC&printsec=frontcover&dq=ASM+Specialty+
35. K. Niesson, M. Gindrat. 2011. “Plasma Spray-PVD: A New Thermal Spray Process to Deposit Out of the Vapor Phase.” Journal of Thermal Spray Technology. 20 (4) 736-43. http://paperity.org/p/6136199/plasma-spray-pvd-a-new-thermal-spray-process-to-deposit-out-of-the-vapor-phase.
36. Yahya. 2010. “Turbines Compressors and Fans.” Mc-Graw Hill. 944 p..
37. “Technology Transfer Network – Air Toxics Web Site.” 2013. EPA. http://www.epa.gov/ttnatw01/hlthef/cadmium.html.
38. N. Tamilselvam, Y. Marcian. “Improving the Thermal Sustainability of Modern Gas Turbine Blade using Newly Proposed Coolant Passage Shapes.” International Journal of Systems, Algorithms, and Applications. http://www.academia.edu/9400370/Improving_the_Thermal_Sustainability_of_Modern_Gas_Turbine_Blade_using_Newly_Proposed_Coolant_Passage_Shapes.
Arc – PVD…….…..cathodic arc physical vapor deposition
C…………………….number of components
dof……………….…degrees of freedom
F…………………….number of dof
FGM…………………….functionally graded materials
GLAD………………glancing angle deposition
HPCVD……….……hybrid-physical chemical vapor deposition
HVOF…………..….high velocity oxy-fuel
IE EBPVD…………ion enhanced electron beam physical vapor deposition
LPCVD………….…low-pressure chemical vapor deposition
LPPS……………….low-pressure plasma spray
MBE……………..…molecular beam epitaxy
MFP………………..mean free path
MLS……………..…monolayer per second
MOCVD……………metalorganic chemical vapor deposition
P…………………….number of phases
PACVD……………..plasma-assisted chemical vapor deposition
PECVD……………..plasma-enhanced chemical vapor deposition
PS-PVD………………….plasma spray-physical vapor deposition
P-T………………..…pressure and temperature
PVD…………………physical vapor deposition
P-YSZ………………partially yttria-stabilized zirconia
TBC…………………thermal barrier coating
TGO…………….….thermally grown oxide
VPE………………..vapor phase epitaxy