The applied thermal barrier coatings (TBCs) on turbine blades structurally consist of a metallic bond coat formed from special heat-resistant alloys of the MCrAlY system and an outer ceramic layer, most often made of yttria-stabilized zirconia (ZrO₂·Y₂O₃). The most advanced method for applying the outer ceramic layer is electron-beam physical vapor deposition (EB-PVD) in vacuum, which imparts to the deposited ceramic a specific columnar structure (fig. 1). As the experience of the authors of [1][2] and of other researchers [3–6] shows, the microstructure, thickness, and growth direction of the columns depend significantly on the technological parameters of the ceramic deposition process.

During electron‑beam evaporation of the material in a high vacuum of 10−⁴…10−⁶ mm Hg, the spatial density of the vapor flow above the evaporator follows a cosine distribution law, according to which the maximum density is observed in the direction of normal line towards the evaporation surface (angle = 0). The heating of the evaporated material is carried out by automatic scanning of the electron beam over the surface of the rod. Over time, due to imperfect scanning or the emergence of certain differences in local absorption/reflection conditions, protruding areas form on the surface; the evaporation rate on these areas tends to decrease, even to a complete stop, while their height naturally increases [7][8]. This undesirable phenomenon leads to non‑uniform evaporation of the ceramic rod (fig. 2) and affects the shape of the vapor cloud [8][9].

Therefore, subjected to the electron beam, a zone with very high vapor pressure may form above the molten material, causing the melt surface to become concave (fig. 3). As a result, the vapor flow density decreases more sharply with increasing angle φ (the vapor cloud becomes focused). The deepening of the evaporation surface also leads to this same effect. If the cold crucible walls hinder the distribution of the vapor flow (fig. 3, c), their influence on the evaporator characteristic is called the “chimney effect”. Both of these factors change the evaporator characteristic so that it becomes similar to the spatial distribution of the vapor flow density issuing from a discharge nozzle.

Stable maintenance of the most even shape of the melt pool (without protrusions or low spots) and constancy of its level relative to the crucible are among the main technological conditions for ensuring repeatability of the coating deposition process on parts from melt to melt. To achieve this, the operator must concentrate the beam scanning on the protruding parts of the rod that have formed and melt them flush with the pool surface, and set the melt level by changing the feed rate of the rod into the crucible. At high evaporation rates, a zone (1, in fig. 3, d) [7] with high vapor density may form above the crucible surface; vapor particles outside this zone distribute rectilinearly from its surface (the vapor cloud is substantially scattered).

To improve evaporation stability in modern electron-beam coating deposition facilities, the most advanced automated control systems are used, such as EBCS (Electron Beam Control System) for precise positioning and control of the electron beam in equipment manufactured by ALD Vacuum Technologies GmbH [10][11]. These systems make it possible to achieve a nearly flat surface of the evaporating ceramic rod; however, they cannot completely eliminate the formation of protrusions, and operator involvement in the process cannot be fully automated.

As it can be seen from the schematic of the electron‑beam evaporation process (fig. 1, a), in order to obtain the required coating thickness at each point of the blade, it is necessary to create conditions for the ceramic particles evaporated from the crucible to reach that point. For this purpose, a high vacuum is maintained in the installation to prevent collisions between ceramic molecules and their scattering [10][11]; in this case, the “line-of-sight” condition applies, i.e., only surfaces that are in direct line of sight from the crucible are coated. However, the cosine distribution of the vapor flow (fig. 3) also leads to significant deposition of the evaporated material on the walls of the electron‑beam facility installation (fig. 4), which considerably increases its consumption. Furthermore, in order to obtain a uniform coating thickness at every point along the blade airfoil profile during its rotation in the facility, it is necessary to select a coating zone (working zone) with a relatively uniform vapor flow density [12][13].

It can be seen that in order to reduce the consumption of the evaporated material, it is necessary to control the vapor flow in a given direction. Improving the ceramic layer of the TBC by densifying it, with the aim of reducing gas permeability and porosity – which will reduce the penetration of fuel combustion products into the inter‑columnar space – is one of the directions for improving TBCs.

In the electron‑beam method of coating deposition on parts, the material, for example ZrO₂·Y₂O₃, is evaporated in a vacuum installation from a crucible by heating with an electron beam and then condenses on the part surface (fig. 1, a). Evaporation of oxides is generally accompanied by a change in the form of the original molecules. For most refractory compounds used in protective coating deposition, dissociation of the original molecules with the formation of gaseous products is characteristic. However, in works [12][14][15], ZrO₂·Y₂O₃ ceramic is cited as an example of a material that evaporates from water‑cooled copper crucibles under a focused electron beam with virtually no change in its original composition. The dissociation process is discussed in works [4][16], where the degree of dissociation of ZrO₂ is given (tab. 1) and it is indicated that the oxygen present in the dissociation products is partially removed from the evaporation zone and carried away into the vacuum system. To compensate for this loss, a method is used in which a certain amount of oxygen is supplied to the crucible zone, resulting in the formation of a white layer of ZrO₂·Y₂O₃ of stoichiometric composition.

The degree of dissociation depends on many process parameters, including the temperature of the melt pool of the evaporated ceramic, the evaporation rate, and the vacuum level. Evidence of the ceramic dissociation process is the black color (fig. 2) of the surface of the ceramic rod after it has cooled [4][16].

In order to investigate the process of formation of thicknesses and microstructures of coatings deposited in a cosine vapor flow, the authors conducted the following experiment. To eliminate the influence of the blade shape (curvature), flat specimens (plates) were fabricated and welded to the locks of a test blade (fig. 5). The plates were divided into sections for the study of the formed coatings.

The coating structure was investigated by X‑ray diffractometry and scanning electron microscopy. The coating thickness was evaluated using scanning electron microscopy and optical metallography (at magnifications of up to 1000×). Thickness measurements for each of the specimens were performed at least 3 times.

The formation of coatings begins with the deposition of a thin film of the condensing material on the substrate (part surface). An atom (molecule) striking the surface enters the field of attractive forces of the surface atoms (molecules). The following situations are possible after this: (1) the atom is adsorbed and finally “sticks” to the surface; (2) some time after adsorption, the atom leaves the surface again; (3) the impinging atom is instantly reflected from the surface [17][18]. Usually, the first two possibilities occur. The probability of capture of an incident atom is characterized by the condensation coefficient αк, equal to the ratio of the number of atoms condensed on the surface to the total number of incident atoms. The value of αк decreases with increasing substrate temperature and with increasing size mismatch (difference in crystal lattice parameters) between the condensing substance and the substrate [12]. It should be emphasized that in electron‑beam evaporation, the vapor flow consists almost entirely of neutral atoms; the degree of ionization does not exceed 0.05–0.1% [12].

The material distribution during deposition from a point source is determined by the formula

dm/ds ~ (m/h²) ‧ cosφ,

where dm is the mass of material deposited on the surface element dS, m is the total amount of evaporated material, h is the distance from the evaporator to the condensation surface, and φ is the angle between the vapor flow direction and the normal to the surface element. Typically, the following values are adopted: h = 250 ÷ 300 mm, and the length (diameter) of the condensation surface l = 150 ÷ 180 mm [12]; in fact, these are limiting values.

As follows from the material distribution formula for evaporation in vacuum using electron beams, the zone of uniform deposition occupies a small area on a hemisphere with a large radius, i.e., the efficiency of the process (facility productivity) is low and apparently cannot be significantly increased.

During the experiment in the UE-175 unit, it was found that to deposit a ceramic layer of ~100 μm thickness on the blades, about 450….500 grams of ceramic are evaporated during the process cycle. At the same time, about 1.5 grams are deposited on one blade, and on 14 blades – 21 grams of ceramic. Thus, the utilization factor of the evaporated material is only 4%. Most of the evaporated material overflies the parts and deposits on the vacuum installation walls in all directions. Screens are installed to protect them. Therefore, it seems reasonable to evaluate the possibility of using special screens to direct (concentrate) the vapor cloud toward the parts to be coated.

It was found that the vapor flow density distribution (fig.  6) changes cosinusoidally. Actually, the vapor flow density decreases more sharply with increasing angle φ (the angle of incidence of the vapor flow on the substrate), which is why it is recommended to set the radius R of the substrate placement relatively large.

The nature of the variation of the vapor flow profile was analyzed based on the results of thickness measurements (tab. 2) and microstructures (fig. 7) of coatings on different sections of the plates, depending on their position above the crucible.

As it can be seen (fig. 7), the inclination of the crystallites and the thickness of the ceramic layer are consistent with the variation in the vapor flow density profile of the deposited ceramic. It can be observed that in the central part of the plates (fig. 7, a) the coating thickness is 10–15% greater than away from the central part (fig. 7, b, c, d), and the inclination angle of the ceramic columns reaches 10–13 (fig.  7, c, d). Consequently, the performance characteristics of the formed coating will be lower at the periphery of the working zone.

The same pattern of thickness variation is also observed on other specimens – the plates (fig. 5, 7).

In view of the foregoing, as one possible option for concentrating the vapor cloud of the evaporator, it is proposed to install a special cone above the crucible (fig. 8), either with cooling of its walls or uncooled (this will have to be determined in experiments, as well as the degree of condensation of the material evaporated from the crucible on the walls of this cone). In the electron-beam installation, the electron beam is moved by scanning at an angle of ~50° to the crucible surface. As a result, the electron beam will impinge on the crucible through an open aperture in the protective cone near its base. An observation port for viewing the surface of the evaporated material is also required. Therefore, in this case, a part of the vapor flow will immediately escape into the vacuum chamber.

The modification of the electron-beam facility proposed by the authors is intended to reduce the consumption of the evaporated material.