Device and CVD method for preparing the device

Abstract

The present invention relates to a refractory carbide multilayer (10) comprising a first layer (11), a second layer (12) and a third layer (13), wherein the first layer (11) has an average thickness of at least 25 nm and contains at least one refractory carbide, the second layer (12) has an average thickness of at least 25 nm and contains at least one refractory carbide, and the third layer (13) has an average thickness of at least 25 nm and contains at least one refractory carbide, and wherein the first layer (11) has a larger average linear intercept cross-sectional grain size than the second layer (12).

Description

Technical Field

[0001] The present invention relates to refractory carbide multilayers, devices comprising such multilayers, CVD processes for preparing such multilayers, and uses of such multilayers.

Background Art

[0002] J. Lopez-Vidrier et al. (Materials Science and Engineering B 178 (2013) pages 639 - 644) describe silicon-rich silicon carbide (SRC) layers appropriately separated by SiC barrier layers. Crystalline Si or fused quartz was used as the substrate. This multilayer was deposited by plasma-enhanced chemical vapor deposition (PECVD) using SiH4, CH4, and H2 as gas precursors. Each SiC barrier layer had a thickness of 3 nm, and the SRC layer thickness was either 2, 3, or 4 nm. A 20 nm buffer layer of SiC was also deposited between the substrate and the multilayer. The SiH4 flow rate was varied such that the excess Si concentration in the SRC layer was 20, 30, and 50 atomic % respectively.

[0003] EP 0 935 013 A1 describes a CVD-formed SiC product that is less transmissive to light in a specific wavelength range. This CVD-formed SiC product is produced on a substrate by first depositing a SiC coating on the substrate by a CVD process, then removing the substrate, and forming at least one SiC layer having different particle characteristics either within the structure of the formed SiC product or on its surface. This causes scattering, refraction, reflection, etc.

[0004] WO 2019 / 182306 A1 describes a graphite coating method including: a) supporting solid silicon on a porous support; b) placing a graphite substrate for coating and the porous support supporting solid silicon in a vacuum chamber, and then performing heat treatment at 1,400 to 2,000 °C to prepare a first silicon carbide coating layer; and c) depositing a second silicon carbide coating layer on the first silicon carbide coating layer by chemical vapor deposition, and also describes a method for coating a graphite substrate using a hybrid coating method. In this method, research has been carried out to avoid cracks that occur in conventional SiC coatings during the cooling process after vapor deposition.

[0005] EP 3 514 130 A1 describes a silicon carbide coating (or article) including: I-A) a porous graphite substrate; II-A) at least two SiC coating layers with different densities; and optionally III-A) an interface layer located between the graphite substrate and the SiC coating layers, the interface layer including porous graphite and pores filled with SiC in the form of whiskers at least 50 μm long extending from at least one of the SiC coating layers to the porous graphite substrate. This document further describes that in such a multi-layer SiC coated article, at least two SiC coating layers II-A) can be characterized by different crystal diameters.

[0006] U.S. Patent No. 10,294,163 B2 relates to a SiC-coated carbon composite material including a graphite substrate and a CVD-SiC coating that coats the graphite substrate. The SiC impregnated layer extends from the CVD-SiC coating and is included around the core part. The SiC impregnated layer is composed of a plurality of regions arranged such that the Si content gradually decreases in order from the first surface on the CVD-SiC coating side to the second surface on the graphite substrate side.

[0007] WO 2019 / 182306 A1, EP 3 514 130 A1, and US Patent No. 10,294,163 B2 focus on achieving better adhesion of SiC coatings to graphite substrates, but little attention has been paid so far to improving the integrity of the carbide coatings themselves. Coating substrates with ceramics by CVD is a common method for extending the lifespan of graphite substrates, especially when used in high-temperature and corrosive / oxidizing environments. Nevertheless, ceramics are subject to degradation. Degradation pathways include, for example, microcracks in the film that propagate to the substrate over time, thereby exposing the substrate to aggressive chemicals and often causing premature etching and physical breakdown of the substrate. In particular, refractory carbide coatings on graphite are prone to such damage. The use of an intermediate layer between carbide and graphite, already proposed in the prior art, can only mitigate these problems to a certain extent.

Prior Art Documents

Patent Documents

[0008]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Non-Patent Documents

[0009]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0010] The problem underlying the present invention is to provide a constituent material for high-temperature applications that can be used for a long time under conditions involving mechanical (and in some cases mechanical due to temperature gradients) stress and exposure to a thermal etching gas, is very easily available, and has an emissivity that is accurately controlled.

Means for Solving the Problem

[0011] The inventors have discovered that a solution to this problem is provided by designing the structure of the refractory carbide coating itself. This problem is solved by a refractory carbide multilayer including a first layer, a second layer, and a third layer, The first layer has an average thickness of at least 25 nm and contains at least one refractory carbide, The second layer has an average thickness of at least 25 nm and contains at least one refractory carbide, The third layer has an average thickness of at least 25 nm and contains at least one refractory carbide, The first layer has a larger average linear cross-sectional particle size than the second layer, which is solved by the multilayer.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 9A

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Best Mode for Carrying Out the Invention

[0013] The multilayer may be a freestanding multilayer or a part of a multilayer on a part of the surface of the device. The freestanding multilayer does not include a substrate. Neither of the two major surfaces of the freestanding multilayer is attached to a substrate. Thus, the present invention further relates to a device for high-temperature applications including the multilayer of the present invention on the surface of the material of the device. The high temperature described herein generally means a temperature in the range of 200 to 4000 °C, preferably 300 to 3000 °C, for example 400 to 2500 °C.

[0014] The material of the device is not limited. All materials that are stable at high temperatures for high-temperature applications are suitable as the material of the device. Preferably, the material of the device includes graphite or monocrystalline silicon. The material of the device may be, for example, a graphite material, a carbon fiber reinforced material, or monocrystalline silicon. Monocrystalline silicon is preferably in the form of a wafer, as used in the semiconductor industry, for example. The material may have the shape of a support for semiconductor applications. Such a support (wafer carrier) for semiconductor applications often has the shape of a disk or a barrel with inclined sidewalls, both of which are provided with one or more recesses for placing a semiconductor substrate such as a silicon-based wafer to be processed on the support. Such a support is customarily used in an epitaxy reactor or MOCVD reactor for growing layers on a wafer, or a plasma etching reactor such as a focusing ring, or an RTP reactor, that is, an RTP ring, and it is well known in the art that it is processed within the RTP ring. However, the invention of the multilayer on the surface of the material of the device does not exclude the existence of another layer between the graphite material and the multilayer. Such other layers will also form the material of the device.

[0015] Accordingly, the constituent materials referred to in this specification may be the multilayer itself or a device. In this specification, the term "multilayer" means a multilayer regardless of whether it is part of a device. Whenever various specific multilayer features or properties are being addressed, explicit reference is made on the one hand to free-standing multilayers and on the other hand to the multilayers of a device.

[0016] Refractory carbides are well known to those skilled in the art of general surface coating or chemical vapor deposition (CVD). Refractory carbides contain non-carbon carbide-constituent materials and carbon. In silicon carbide, for example, the non-carbon carbide-constituent material is silicon. The present invention is not limited to specific refractory carbides, and depending on the particular application to which the multilayers of the present invention are applied, some refractory carbides may be preferred. The refractory carbide contained in one of the first, second, third (and any other if applicable) layers may be the same as or different from the refractory carbide contained in any other of the first, second, third (and any other if applicable) layers.

[0017] The refractory carbide contained in one layer and the refractory carbide contained in any other layer are considered "the same" when both refractory carbides contain the same non-carbon carbide-constituent material (e.g., silicon). The layers may nevertheless differ, for example, in terms of average linear cross-sectional particle size, porosity or content of non-carbon constituents.

[0018] The refractory carbide contained in one layer and the refractory carbide contained in any other layer are considered "different" when both refractory carbides contain different non-carbon carbide-constituent materials.

[0019] Preferably, at least one refractory carbide of at least one layer comprises silicon carbide, tantalum carbide, niobium carbide, hafnium carbide, tungsten carbide, boron carbide, niobium tantalum carbide, hafnium tantalum carbide, or hafnium niobium carbide.

[0020] More preferably, the first, second, and third layers each contain the same refractory carbide selected from silicon carbide, tantalum carbide, niobium carbide, hafnium carbide, tungsten carbide, boron carbide, niobium tantalum carbide, hafnium tantalum carbide, or hafnium niobium carbide, particularly the same refractory carbide selected from silicon carbide, tantalum carbide, niobium carbide, and hafnium carbide. Most preferably, the first, second, and third layers each contain silicon carbide.

[0021] Irrespective of the specific context, in this specification, the term "refractory carbide" preferably means "silicon carbide". Although specific effects of silicon carbide have been shown in this specification, since the growth of other refractory carbides in CVD follows the same or similar physical principles as those shown in the examples, it is reasonable to presume that the same or similar effects will be observed for other refractory carbides.

[0022] According to the present invention, the refractory carbide multilayer comprises a first layer, a second layer, and a third layer. Each of these layers contains a refractory carbide. As can be seen from the various figures in this specification, the number of layers can be recognized extremely easily from the SEM image. The number of refractory carbide-containing layers can be much larger than 3. The multilayer can include, for example, 3 to 1000 refractory carbide-containing layers, 3 to 100 refractory carbide-containing layers, for example 4 to 50 refractory carbide-containing layers. The number of refractory carbide-containing layers can preferably be 3 to 12, for example 3 to 10 (or particularly 4 to 12, for example 4 to 10).

[0023] The present invention does not exclude other layers, i.e., layers that do not contain refractory carbides. The multilayer can include, for example, one carbon layer or a plurality of carbon layers. In this specification, the "carbon layer" preferably means a "pyrolytic carbon layer". Any carbon layer can be present anywhere in the multilayer. This means that the multilayer can include a pyrolytic carbon layer sandwiched between refractory carbide-containing layers. The pyrolytic carbon layer can be the outermost layer. In the device, the pyrolytic carbon layer can be sandwiched between the innermost refractory carbide-containing layer and the material of the device.

[0024] As used herein, the term "average thickness" means the thickness determined from an SEM cross-sectional image produced by cutting or breaking a multilayer (or the material thereof if the multilayer is part of a device). Layers can be distinguished in the SEM cross-sectional image by their appearance, such as light and dark contrast, particle size, morphology, or porosity. The thickness of an individual layer can be determined by measuring the distance between the lower and upper boundaries of the layer along at least five lines running perpendicular to the multilayer surface. The lines are placed at various positions and the layer thickness should represent a majority of the entire layer. The layer thickness is the average of the measurements at each line.

[0025] According to the present invention, the first layer has an average thickness of at least 25 nm, preferably at least 100 nm, for example at least 1 μm. Since the deposition process can be continued until a very thick layer is formed, the average thickness of the first layer can be very large. The average thickness of the first layer in a freestanding multilayer usually does not exceed 5 mm. In the device of the present invention, the first layer is usually thin. When the multilayer of the present invention is on the surface of a material, even a very thin multilayer - and the individual layers contained therein - are prevented from cracking due to the strength of the material. The (multi)layer is stabilized by the material. The average thickness of the first layer contained in the multilayer of the device according to the present invention can be, for example, in the range of 25 nm to 200 μm, preferably in the range of 100 nm to 180 μm, particularly in the range of 1 μm to 160 μm. In the freestanding multilayer according to the present invention, the first layer is usually much thicker. The average thickness of the first layer contained in the freestanding multilayer according to the present invention can be, for example, in the range of 0.25 μm to 5 mm, preferably in the range of 1 μm to 3 mm, particularly in the range of 10 μm to 2 mm.

[0026] The same considerations apply to the second and third layers.

[0027] The average thickness of the second layer included in the multilayer of the device according to the present invention may be, for example, in the range of 25 nm to 200 μm, preferably in the range of 100 nm to 180 μm, particularly in the range of 1 μm to 160 μm. The average thickness of the second layer included in the freestanding multilayer according to the present invention may be, for example, in the range of 0.25 μm to 5 mm, preferably in the range of 1 μm to 3 mm, particularly in the range of 10 μm to 2 mm.

[0028] The average thickness of the third layer included in the multilayer of the device according to the present invention may be, for example, in the range of 25 nm to 200 μm, preferably in the range of 100 nm to 180 μm, particularly in the range of 1 μm to 160 μm. The average thickness of the third layer included in the freestanding multilayer according to the present invention may be, for example, in the range of 0.25 μm to 5 mm, preferably in the range of 1 μm to 3 mm, particularly in the range of 10 μm to 2 mm.

[0029] The average thickness of all individual layers is preferably in the range of 0.025 μm to 500 μm, most preferably in the range of 0.05 μm to 250 μm, particularly in the range of 0.1 μm to 200 μm, for example, 0.2 μm to 150 μm, for all layers that are layers of the multilayer of the device. The average thickness of all individual layers is preferably in the range of 0.1 μm to 100 mm, most preferably in the range of 0.2 μm to 50 mm, particularly in the range of 0.3 μm to 10 mm, for example, 0.4 μm to 5 mm, for all layers that are layers of the freestanding multilayer. The preferred effect of the layer is more consistent when the average thickness of each layer increases a specific minimum value. The upper part of the individual particles of the lower layer can affect the crystal growth of the outer layer even through a very thin intermediate layer. Exceeding the upper limit of the thickness range makes the coating very time-consuming. As a result, the coating method becomes very expensive, and the material or component on which the layer is deposited may bend due to mechanical stress, which may also be caused by the mismatch in thermal expansion between the layer and the coated material.

[0030] When the multilayer is a multilayer of the device, the average thickness of the multilayer is preferably in the range of 1 μm to 1 mm. Thereby, the reliability of the protection of the material coated with the multilayer is ensured with a reasonable deposition time and thus a reasonable cost. When the multilayer is a freestanding multilayer, the average thickness of the multilayer is preferably in the range of 100 μm to 100 mm.

[0031] According to the present invention, the first layer has a larger average linear cross-sectional particle size than the second layer. This expresses in a more quantitative way that the first layer generally contains larger crystals than the second layer. The average linear cross-sectional particle size of the layer can be determined by counting the number of grain boundaries on a line crossing at least 20 grain boundaries at half the thickness of the layer from the cross-section in the electron micrograph. The average linear cross-sectional particle size is calculated by dividing the length of the line by the number of grain boundaries on that line, which is similar to the definition of "average linear section particle size" in ISO 13383-1:2012 (EN).

[0032] As the inventors have discovered, the crystal diameter can be controlled very flexibly and easily by modifying the CVD coating conditions. Pores that appear to impede the growth of refractory carbide crystals have been found to form under certain conditions. This is explained in connection with Example 3 below. Other conditions seem to assist the deposition of elemental silicon and seem to impede the growth of refractory carbide crystals. Thus, there are various possibilities to ensure that the refractory carbide crystals of one layer grow larger than those of another layer. Thereby, a multilayer is obtained in which the average linear cross-sectional particle size of the first layer is larger than that of the second layer. The present invention does not require any other difference between the first layer and the second layer other than the average linear cross-sectional particle size. However, the present invention does not exclude other differences between the first layer and the second layer other than the average linear cross-sectional particle size, as will be explained below.

[0033] The second layer may contain more non-carbon components (e.g., silicon) of refractory carbide (e.g., silicon carbide) than the first layer. When the refractory carbide contained in the second layer is silicon carbide, the second layer may contain more silicon than, for example, the first layer. Surprisingly, it has been found that this prevents the formation of the deep etch pits shown in FIGS. 2 and 9A, thus providing an additional protection mechanism for the multilayer device. The inventors further speculate that the silicon-rich layer is prone to deformation (decrease in Young's modulus / hardness) and relieves stress, thus increasing the toughness of the multilayer. It also makes it possible to adjust the crystal diameter (i.e., surface particle size) of the outer surface of the multilayer when an outermost layer of defined thickness is deposited on such a second layer, as will be discussed below with reference to FIG. 15.

[0034] Elemental silicon is detectable by XRD in the multilayer. When not in the outermost layer and when the outermost layer is thick, the multilayer coating can be removed step by step and XRD can be applied after each removal. The penetration depth of XRD is about 30 - 50 μm. The Si to C atomic ratio can be determined by EDXS, XPS, or TOF-SIMS measurement and may be in the range of, for example, 1.01:1 to 10:1, preferably 1.03:1 to 4:1, for example 1.1:1 to 2:1.

[0035] The porosity of the second layer may exceed that of the first layer. This causes crack deflection, increases the etching resistance, which ensures the long-term use of the constituent material (i.e., the device or multilayer) in high-temperature and / or corrosive environment applications. The porosity, defined herein as the volume fraction of pores within the layer, can be measured from SEM images. For this purpose, the cross-section of the coating is prepared by cutting and gentle polishing, fracturing and gentle polishing, or the focused ion beam (FIB) method. In the cross-section, pores and solid refractory carbide can be distinguished by light and dark contrast. The area of the cross-section covered by pores is measured for each layer and divided by the total inspection area to obtain the area fraction of pores.

[0036] Assuming that the pores are uniformly distributed throughout the volume of the coating, the area fraction of the pores is equal to the porosity defined above. The porosity of the second layer may be, for example, a porosity of 0.05 to 30%, preferably 0.1 to 20%, and most preferably 0.15 to 10%. When the porosity is larger, the layer density is smaller, and vice versa. As described for the porosity, the layer density is also determined from SEM image analysis.

[0037] In the multilayer, the second layer may be sandwiched between the first layer and the third layer. Further, the third layer may have a larger average linear cross-sectional particle size than the second layer. Thereby, the generally good hardness / density of the two layers (the first and the third) containing larger refractory carbide crystals (or having a larger average linear cross-sectional particle size) is combined with the protection against deep etch pit formation and mechanical damage throughout the coating. Therefore, by selecting the thickness of the outermost layer, it is also possible to adjust the surface crystal diameter (i.e., the surface particle size).

[0038] In the multilayer, the first layer may alternatively be sandwiched between the second layer and the third layer. Further, the first layer may have a larger average linear cross-sectional particle size than the third layer, i.e., the first layer may have a larger average linear cross-sectional particle size than the second layer and the third layer. Thereby, the generally good hardness / density of one (the first) layer containing larger refractory carbide crystals (or having a larger average linear cross-sectional particle size) is combined with the protection against deep etch pit formation and mechanical damage throughout the coating. Therefore, by selecting the thickness of the fourth outermost layer, it is also possible to deposit a fourth outermost layer with an adjusted surface crystal diameter (i.e., the surface particle size).

[0039] In other alternative multilayers, the second layer can contain more of the non-carbon component (e.g., silicon) of the refractory carbide (e.g., silicon carbide) than the first layer, and the porosity of the third layer may exceed the porosity of the first layer. In the case where the refractory carbide is silicon carbide in this other alternative multilayer, the second layer can contain more silicon than the first layer, and the porosity of the third layer may exceed the porosity of the first layer. Thereby, the etching resistance is increased, the mechanical stability is increased, and an adjustable crystal diameter (i.e., surface particle diameter) is obtained.

[0040] In a preferred multilayer, the average thickness of the first layer is greater than the average thickness of the second layer. The inventors do not assume that when the second layer is quite thick, the favorable effects brought about by the second layer increase linearly with the thickness of the second layer. Therefore, for a given total multilayer thickness, it is advantageous to increase the thickness of the first layer, which has a larger average linear cross-sectional particle diameter and is more robust, instead.

[0041] The preferred multilayer has an average linear cross-sectional surface particle diameter in the range of 0.1 to 50 μm, more preferably 0.2 to 40 μm, most preferably 0.3 to 30 μm, particularly preferably 0.4 to 20 μm, for example, 0.5 to 10 μm. The average linear cross-sectional surface particle diameter is measured according to the definition of "average linear cross-sectional particle diameter" in ISO 13383-1:2012 (EN). The technical effect related to the surface particle diameter is that the larger the surface particle diameter, the shorter the overall surface grain boundary length.

[0042] Nevertheless, surprisingly, it has been found that more etching damage is observed.

[0043] Therefore, a smaller surface particle diameter is preferred.

[0044] The present invention further relates to a CVD method for preparing a multilayer according to the present invention, in which a refractory carbide layer containing more or less non-carbon components (e.g., silicon) of a refractory carbide (e.g., silicon carbide) than an adjacent layer and / or a refractory carbide layer having a porosity different from that of the adjacent layer is formed. A refractory carbide layer containing more or less non-carbon components (e.g., silicon) of a refractory carbide (e.g., silicon carbide) than an adjacent layer and / or a refractory carbide layer having a porosity different from that of the adjacent layer is preferably applied in a preceding layer or subsequent layer forming step. atmospheric composition, pressure, and / or the temperature of the CVD coating atmosphere different from the conditions (e.g., condition B) of atmospheric composition, pressure, and / or the temperature of the CVD coating atmosphere and is deposited under the conditions (e.g., condition A). According to a preferred method of the present invention, a multilayer of three or more layers is formed by periodically changing conditions A and B. When a multilayer containing two or more different types of layers is desired, different atmospheric composition, pressure, and / or the temperature of the CVD coating atmosphere conditions (e.g., condition C) allow further layers to be deposited.

[0045] Preferably, the CVD method of the present invention includes thermal CVD. Thermal CVD is different from plasma enhanced chemical vapor deposition (PECVD). Thermal CVD usually requires a much higher temperature than PECVD, but does not require a plasma source. In this specification, the expression "thermal CVD" means CVD at a temperature of at least 800°C. A particularly preferred CVD method of the present invention is a thermal CVD method, and the entire multilayer is deposited at a temperature of 800°C or higher.

[0046] The refractory carbide layer having a porosity different from that of the adjacent layer can be deposited from a CVD coating atmosphere, for example, by changing the content of a carbon source, for example, the content of a carbon source selected from CH4, C2H2, C2H4, C2H6, C3H6, C3H8, C4H8, or C4H 10 in the CVD coating atmosphere by changing the content of the carbon source selected from.

[0047] According to a preferred method of the present invention, - a refractory carbide layer containing more or less non-carbon components (e.g., silicon) of a refractory carbide (e.g., silicon carbide) than the adjacent layer, or - a refractory carbide layer having a porosity different from that of the adjacent layer is formed, and a layer having a larger average linear cross-sectional particle size is deposited thereon. The deposition of the layer having a larger average linear cross-sectional particle size is continued until the outer surface of the layer having a larger average linear cross-sectional particle size reaches a predetermined average linear surface particle size.

[0048] The present invention further relates to a multilayer or device obtained by the method of the present invention.

[0049] Another aspect of the present invention is the use of the multilayer of the present invention as a coating of a support for extending the life of the support in an etching gas-containing atmosphere and / or when exposed to mechanical stress. As described above, the support is preferably a support for semiconductor applications. The etching gas may contain, for example, HCl, NH3, H2, N2, a halogen, gallium, aluminum, arsenic, AI(CH3)3, or silane. The inventors have demonstrated the high resistance of the multilayer of the present invention to etching by HCl, but it is considered that the same or similar effects may be observed with other gases. This is because the exact nature of the boundaries or gaps between the particles seems to control the etching process, and like HCl, gases other than HCl may move into these gaps.

[0050] The present invention will be described in more detail by the following non-limiting figures and examples.

[0051] Figures 1 to 4, 10, and 12 show schematic enlarged cross-sectional views of a device for high-temperature applications. Each figure (and in the case of Figures 1 to 3, each panel arranged in rows and columns) shows only a part of the device, and that part includes a part of part 2 of the device (shown in black at the bottom of each panel) and a part of the refractory carbide layer on the surface of part 2.

[0052] For each figure (and in the case of Figures 1 to 3, each panel arranged in rows and columns), the hatching lines of the refractory carbide layer by parallel lines from the left of the lower boundary to the right of the upper boundary (for example, the shading shown at the top of Figure 1) indicate a larger average linear intercept cross-sectional particle size than any directly adjacent refractory carbide layer. The intersection lines of the horizontal and vertical lines (the shading in the middle of Figure 1) indicate a high porosity exceeding the porosity of any directly adjacent refractory carbide layer. The circles (the indication of the layer in the middle column of Figure 2) indicate a layer rich in non-carbon constituents of the refractory carbide and containing more non-carbon constituents of the refractory carbide than any directly adjacent refractory carbide layer.

[0053] The device illustrated in the three lower panels of Figure 1 includes the refractory carbide multilayer 10 of the present invention on the surface of part 2. This refractory carbide multilayer 10 includes a first layer 11, a second layer 12, a third layer 13, a fourth layer 14, and a fifth layer 15. It cannot be grasped from the schematic view of Figure 1 that each of the layers 11, 12, 13, 14, and 15 may have an average thickness far exceeding 25 nm and contains at least one refractory carbide. In this embodiment, the first, third, and fifth layers 11, 13, and 15 have a larger average linear intercept cross-sectional particle size than the second and fourth layers 12 and 14, and the porosity of layers 12 and 14 exceeds the porosity of the first, third, and fifth layers 11, 13, and 15.

[0054] FIG. 1 further illustrates how the effect of mechanical forces (see upper arrow) and the multi-layers of the present invention serve to mitigate crack propagation by refracting cracks along the plane of the layers, thereby reducing the probability that the cracks reach the substrate. As illustrated in the upper and middle rows, right panel, in a single layer containing large refractory carbide crystals (see upper right panel) or a single layer of highly porous refractory carbide (middle row, right panel), crack propagation can very easily propagate to the surface of portion 2.

[0055] The device illustrated on the right side of FIG. 2 includes the refractory carbide multi-layer 10 of the present invention on the surface of portion 2. The refractory carbide multi-layer 10 includes a first layer 11, a second layer 12, a third layer 13, a fourth layer 14, and a fifth layer 15. Each of the layers 11, 12, 13, 14, and 15 has an average thickness far exceeding 25 nm and contains at least one refractory carbide. In this example, the first, third, and fifth layers 11, 13, and 15 have an average linear cross-sectional particle size larger than that of the second and fourth layers 12 and 14. Layers 12 and 14 are different from layers 12 and 14 shown in FIG. 1. In contrast to the multi-layers shown in FIG. 1, layers 12 and 14 contain more non-carbon constituents of refractory carbide than any directly adjacent refractory carbide layers 11, 13, or 15.

[0056] Figure 2 further illustrates the resistance of the refractory carbide multilayer to hot HCl gas, a common etching gas used in semiconductor applications of refractory carbide-coated graphite. The upper row shows the initial states of different devices with different layers before etching. The arrows between the upper and lower rows indicate HCl etching. The lower row illustrates the states of different devices with different layers after etching. It is known that the outer silicon carbide layer with an atomic ratio Si:C = 1 can withstand etching by hot HCl gas (a common gas used in semiconductor applications of SiC-coated graphite) better than SiC with enhanced silicon. Thus, having one or more layers of stoichiometric SiC with crystals (or larger average linear intercept cross-sectional particle sizes) larger than the usually adjacent non-stoichiometric layers is thought to help extend the lifespan. The inventors surprisingly discovered that, by including layers with low etching resistance (layers containing more of the non-carbon constituents of similar refractory carbides, such as SiC with enhanced silicon), the lifespan of the refractory carbide layer when exposed to the etching gas can still be extended. This layer with low etching resistance is assumed to act as a sacrificial layer that slows down the etching and changes the preferred direction of etching from perpendicular to parallel with respect to the surface. This can be understood from the lower part of Figure 2. In the layer with an atomic ratio Si:C = 1, the expected etching occurs towards the graphite surface (see the deep etch pits in Figure 2, left in the lower row), but the single layer containing excess silicon has overall low resistance to HCl and disappears by HCl (see the middle in the lower row of Figure 2). The multilayer on the right device surprisingly provides the best etching protection for the coated graphite because both the expected etching to the graphite surface and the overall rapid disappearance of the (multi)layer are avoided. By adjusting the atomic ratio (e.g., the atomic ratio Si:C) in the individual layers of the refractory carbide (e.g., silicon carbide) multilayer, it becomes possible to thus reduce chemical etching in several important applications.

[0057] Figure 3 is a schematic diagram of surface particle size control using the multi-layer approach of the present invention. The three panels on the left show the columnar growth of refractory carbide for a thin layer of refractory carbide obtained after a short CVD coating time (upper panel), a slightly thicker layer of refractory carbide obtained after a medium CVD coating (central panel), and a thick layer of refractory carbide obtained after a long CVD coating time (lower panel). When SiC growth of the layer thickness as shown in the lower left panel is carried out in a monolithic manner (i.e., without the multi-layer of the present invention), the CVD growth pattern such as columnar growth continues in a substantially uniform manner, which means that as the growth continues from the initial nucleation site on the graphite surface, the width of the crystal column gradually becomes wider. This means that the surface particle size on the surface becomes larger because it is determined by the diameter of the column itself. As a result, the layer thickness will affect the surface particle size. However, in the application of refractory carbide (e.g., SiC coating) graphite in semiconductor devices, such as epitaxial reactors, etc., the form of SiC plays an important role. There are applications that require a smaller surface particle size, while in others, a larger surface particle size is preferred. The surface particle size also affects the emissivity, and controlling the emissivity is often important for controlling the radiant energy emission from the surface. The multi-layer of the present invention can prevent columnar growth, so the surface particle size on the surface is mainly determined by the thickness of the outermost intermediate layer. This provides a valuable method for adjusting the surface particle size. The lower right panel shows a part of the cross-section of the device of the present invention, where the multi-layer includes a first layer 11 and a third layer 13 with an average linear cross-section particle size larger than that of the second layer 12. The average thickness of the third layer 13, which is the surface layer and determines the surface particle size on the surface, corresponds to the entire thickness of the single layer shown in the upper left panel. Nevertheless, the stability against mechanical force and etching described above with reference to FIGS. 1 and 2 can be further incorporated by an appropriate selection of the layer 12. This results in a (multi)-layer having unprecedented stability and a surface particle size on the precisely conforming outer surface.

[0058] In a preferred embodiment of the present invention, the average linear cross-sectional particle size of at least two layers is smaller than that of other layers. Referring to FIG. 3, the first layer 11 in the lower right panel has a larger average linear cross-sectional particle size compared to the second layer 12 and the third layer 13.

Example

[0059] (Examples 1a, 1b, and 1c) A silicon carbide multilayer in which some silicon carbide layers contain more silicon than other silicon carbide layers Purified (less than 5 ppm impurities by mass) hydrostatic graphite 2 was coated with the silicon carbide multilayer 10 of the present invention including six layers: - The innermost silicon-rich silicon carbide layer 12, - The silicon carbide layer 11 containing large silicon carbide crystals, - The silicon-rich silicon carbide layer 13, - The silicon carbide layer containing large silicon carbide crystals, - The silicon-rich silicon carbide layer, and - The outermost silicon carbide layer containing large silicon carbide crystals.

[0060] The expression "containing large silicon carbide crystals" means that the silicon carbide crystals in the silicon-rich layer were smaller. The expression "silicon-rich" means that the silicon content in the layer was higher than that of the silicon carbide layer containing larger refractory carbide crystals.

[0061] During the CVD process, the hydrostatic graphite substrate was heated and exposed to a constant flow of hydrogen (H2) and methyltrichlorosilane (MTS) gas under reduced pressure. MTS underwent a decomposition reaction to produce solid SiC on the surface of the substrate. The properties of SiC, particularly the Si content, were adjusted by changing the deposition parameters. The above-described six-layer multilayer shown in FIG. 4 was deposited on graphite during one CVD run by periodically changing the parameters shown in the following table:

[0062]

Table 1

[0063] While switching the parameters, the MTS flow was stopped for a short time until the pressure and temperature stabilized. In Example 1a, 6 layers were each deposited within 5 minutes, and the total deposition time was 30 minutes. In Example 1b, the silicon-rich layer was deposited within 2 minutes, and the total deposition time was 21 minutes. In Example 1c, the silicon-rich layer was deposited within just 1 minute, and the total deposition time was just 18 minutes.

[0064] Figure 5 shows a cross-sectional scanning electron microscope (SEM) image of a device of the present invention including a multilayer formed according to Example 1a. The silicon carbide layer containing large silicon carbide crystals appears bright, and the silicon-rich silicon carbide layer appears dark. Some crystallites are highlighted in the lower left panel of Figure 5. It is immediately apparent that the silicon carbide crystals in the layer with a larger average linear cross-sectional particle size are actually larger than those in any of the "silicon-rich silicon carbide layers".

[0065] The layers are sufficiently separated from each other and follow the structure of the graphite substrate. A multilayer coating structure was observed across the entire substrate surface. The silicon carbide crystals in the layer containing large silicon carbide crystals are initially small (less than 1 μm) and become larger (about 2 - 10 μm) as the layer thickens. The crystallites in the silicon-rich layer have a diameter of about 0.2 - 1 μm and do not grow with an increase in layer thickness. The silicon-rich layer stops the crystallization process of the silicon carbide crystals and resumes the crystallization process in the next silicon carbide layer deposited on the outer surface of the silicon-rich layer. Therefore, regardless of the total coating thickness, a silicon-rich intermediate layer can be inserted to adjust the maximum silicon carbide crystal diameter in the layer deposited on that intermediate layer.

[0066] Figure 6 shows SEM images of devices of the present invention having a multilayer including 6 layers formed according to Example 1a (left SEM image), 1b (central SEM image), and 1c (right SEM image). Even when the thickness of the silicon-rich layer is smaller, as discussed above, the crystallization process is interrupted.

[0067] The multilayer structure can be defined in a plurality of ways, for example, by changing the number and order of the layers, or by varying the layer thickness by their deposition times.

[0068] The bottom graph of FIG. 7 shows the X-ray diffraction pattern (XRD) of a sample coated with a single layer of silicon carbide containing large silicon carbide crystals. The top graph of FIG. 7 shows the X-ray diffraction pattern of a sample coated with a single layer of silicon-rich silicon carbide. The second, third, and fourth diffraction patterns from the top are those of Sample 1a, Sample 1b, and Sample 1c, respectively. All the X-ray diffraction patterns in FIG. 7 are plotted on a logarithmic y-axis to improve the visibility of the low-intensity peaks.

[0069] The presence of crystalline elemental silicon was confirmed in samples containing at least one silicon-rich layer (the top four diffraction patterns in FIG. 7 having diffraction peaks characteristic of crystalline silicon at the angle 2θ indicated by the black triangles). The diffraction peaks at the angle 2θ indicated by the black squares are characteristic of β-SiC. In addition, some of the diffraction peaks at the angle 2θ indicated by the black circles are from the graphite substrate.

[0070] The multilayers of Samples 1a, 1b, and 1c showed diffraction peaks for both β-SiC and crystalline silicon. The relative intensity of the diffraction peaks of crystalline silicon decreased with the decrease in the thickness of the silicon-rich layer.

[0071] Using energy-dispersive X-ray spectroscopy (EDXS), it was further confirmed that the silicon-rich layers (such as those indicated by reference numerals 12 and 13 in FIG. 4) actually contained more silicon than the layers with a larger average linear cross-sectional particle size (such as the layer indicated by reference numeral 11 in FIG. 4). EDXS is a spectroscopic method in cooperation with SEM and can be performed on a limited region of the SEM image, for example, on the cross-section of a device including multiple layers. FIG. 8 shows an EDXS line scan of the cross-section of an exemplary multilayer including six layers. The intensities of the Si Kα and C Kα signals were measured along the lines shown on the cross-section of the exemplary six-layer multilayer. The intensities obtained in counts per second of Si Kα (left axis) and C Kα (right axis) are plotted against the position along the line (μm). The Si signal intensity varies depending on the cross-section and shows three maximum values in the region where the silicon-rich layers are arranged (solid line). Thereby, it is confirmed that these layers actually contained more silicon than the layers with a larger average linear cross-sectional particle size. As expected, the generally weak C signal in EDXS is strong only on the graphite substrate.

[0072] The etching resistance of an exemplary multilayer including six layers was investigated by exposure to an etching gas (flow of 80 vol% Ar and 20 vol% hydrogen chloride HCl) at 1300 ° C. and 100 mbar for 3 hours. After etching, the top layer was damaged and partially removed. The underlying silicon-rich silicon carbide layer was partially etched, and the four layers below were all intact. No deep etch pits were formed. This can be seen from FIG. 9 showing cross-sectional SEM images of the multilayer including six layers before (left) and after (right) etching in HCl gas at 1300 ° C. The fundamental problem of deep etch pit formation in a single layer of refractory carbide is also illustrated in the SEM image of FIG. 9A. FIG. 9A shows a single SiC layer deposited on a graphite substrate. This layer contains large crystals. The deep etch pits shown in FIG. 9A were observed after exposure to hot hydrogen chloride gas under the same conditions as described above. In this way, FIGS. 9 and 9A provide experimental evidence of the effects described above with respect to FIG. 2.

[0073] (Example 2) Silicon carbide multilayer containing four silicon carbide layers By changing the deposition order and the respective deposition times, the number and thickness of the individual layers from Examples 1a, 1b, and 1c can be changed. By reducing the number of deposition steps from 6 to 4 from Examples 1a, 1b, and 1c, a multilayer containing only four silicon carbide layers was obtained. In order to double the thickness of the silicon carbide layers containing large silicon carbide crystals, their respective deposition times were doubled. The resulting coating is shown in FIG. 10, which includes a schematic in the upper left part of the figure and a SEM micrograph in the upper right part of the figure. The EDXS line scan shown at the bottom of FIG. 10 supports the excess silicon in the silicon-rich layer.

[0074] The four-layer multilayer of Example 2 was also etched using hot HCl gas as described above for the six-layer multilayer. Most of the samples were etched uniformly from the top (compare with the sample before etching in the upper right panel and upper left panel of FIG. 11). There are some places where deeper etch pits were formed (FIG. 11, central and lower panels). These reached the outer silicon-rich layer and stopped there. This can be mainly explained by the breakdown of the crystal boundaries and the preferred etching along the silicon-rich layer.

[0075] The crystal boundaries can extend from the outer surface of the layer to the lower surface of the layer at some locations. This can occur especially in layers containing large refractory carbide crystals in a columnar shape. The silicon-rich layer seems to break these boundaries very efficiently. Thus, HCl entering the layer along these boundaries spreads horizontally into the lower sacrificial silicon-rich layer. This prevents the HCl from penetrating deeper into the (multi)layer and prevents the formation of deep etch pits.

[0076] (Example 3) Silicon carbide multilayer in which the porosity of some silicon carbide layers exceeds the porosity of other silicon carbide layers The device including the multilayer of Example 3 is illustrated in FIG. 12. This includes the refractory carbide multilayer 10 of the present invention on the surface of part 2. The refractory carbide multilayer 10 includes a first silicon carbide layer 11, a second silicon carbide layer 12, a third silicon carbide layer 13, a fourth silicon carbide layer 14, and a fifth silicon carbide layer 15. The porosity of the silicon carbide layers 12 and 14 exceeds the porosity of the silicon carbide layers 11, 13, and 15. Layers 11, 13, and 15 have a larger average linear section cross-sectional particle size than layers 12 and 14.

[0077] Such a multilayer containing alternating layers can be produced by CVD by changing the growth parameters during the CVD process. The multilayer was deposited on hydrostatic graphite by CVD. Layers 11, 13, and 15 were produced from MTS and H2 gas. The more porous layers 12 and 14 were produced from MTS, C2H4, and H2 gas, and the Si / C molar ratio in this gas phase can be adjusted by the amount of C2H4 added. Growth was not stopped when switching from one layer to another, and only the C2H4 flow was activated / inactivated. The 5-layer multilayer of Example 3 was deposited on graphite during a single CVD run by periodically changing the parameters shown in the following table:

[0078]

Table 2

[0079] FIG. 13 shows a SEM cross-sectional image of the device according to Example 3. The SEM cross-sectional image (FIG. 11) shows the presence of five well-separated layers, three of which are high-density and chevron-shaped, and two of which are porous. The layers grew well on top of each other and formed a fairly distinct interface. During sample preparation (breaking of the coated graphite sample), a part of the upper layer peeled off (arrow in the upper panel of FIG. 13). This indicates that the porous layer has lower toughness, which seems to be beneficial for the crack deflection discussed in FIG. 1. The difference in elemental composition was not distinguishable by EDXS. XRD showed only the presence of silicon carbide and graphite, the latter probably being from the graphite substrate.

[0080] As another example, the etching resistance was investigated in HCl gas. The SEM images show that mainly the upper layer of the coating was etched (Figure 14). When HCl reaches the underlying porous layer, it etches further in a dendritic manner. This seems to be beneficial for refracting and stopping deeper etch pits. Thus, the effect of the porous layer is similar to the effect discussed above for the silicon-rich layers in the silicon carbide multilayer.

[0081] (Example 4) Determination of the average linear intercept surface grain size For the outer surfaces of different devices containing the silicon carbide multilayer of the present invention or the comparative silicon carbide single layer, the average linear intercept surface grain size was determined based on the definition of "average linear intercept grain size" in ISO 13383-1:2012 (EN). In all examples, the outer surfaces from which the average linear intercept surface grain size was derived were deposited under the conditions mentioned in the above table for Examples 1a, 1b, and 1c of the "silicon carbide layer containing large silicon carbide crystals".

[0082] In Figure 15, the average linear intercept surface grain size (vertical axis) was plotted against the total layer thickness (horizontal axis). The data obtained for the comparative silicon carbide single layer are represented by black circles. For the silicon carbide single layer, it was found that the average linear intercept surface grain size increases with the increase in the thickness of the single layer, as can be understood from the dotted line in Figure 15. The data obtained for the silicon carbide multilayer of the present invention are represented by black triangles (samples from Example 1a), black squares (samples from Example 2), and black diamonds (samples from Example 3). As shown by the numbers in the brackets in Figure 15, the outermost layer of one investigated silicon carbide multilayer with a total thickness of about 39 μm had an average thickness of 7.5 μm. The outermost layer of another investigated silicon carbide multilayer with a total thickness of about 42 μm had an average thickness of 14.8 μm. The outermost layer of a third investigated silicon carbide multilayer with a total thickness of about 29 μm had an average thickness of 4.9 μm. As shown by the present invention, when the outermost layer is deposited on a silicon-rich silicon carbide layer or a porous silicon carbide layer, it was found that the average linear intercept grain size is determined by the average thickness of the outermost layer.

[0083] Thus, according to the present invention, not only can the etching resistance and the resistance to mechanical forces be increased. Furthermore, the average linear intercept surface grain size can be very accurately controlled by selecting the thickness of the outermost layer according to the situation. Thus, the multi-layers and devices of the present invention, due to their high resistance, can be used for a long time in a wide variety of different semiconductor applications and can also be used particularly in applications that require small or large surface grain sizes.

[0084] (Example 5) Determination of the porosity in the layer with high porosity The porosity of a single-layer coating deposited under the conditions referred to as "porous layers 12 and 14" in Example 3 was estimated. The coated sample was broken, and the cross-section was gently polished using first a diamond polishing disk and then a diamond slurry. As seen in FIG. 16, the obtained polished cross-section was examined by SEM. Here, the region of the cross-section containing pores can be easily distinguished from the high-density region by light and dark contrast. The porosity was calculated by measuring the area of the cross-section containing pores and dividing this area by the total inspection area. This was done at three locations along the cross-section, and the average porosity and standard deviation were estimated to be 25% ± 4%.

Explanation of reference numerals

[0085] 2C... graphite 10... refractory carbide multi-layer 11... first layer 12... second layer 13... third layer 14... fourth layer 15... fifth layer

Claims

1. A device for high-temperature applications, comprising a refractory carbide multilayer (10) on the surface of the material of the device, wherein the refractory carbide multilayer (10) includes a first layer (11), a second layer (12), and a third layer (13), the first layer (11) has an average thickness of at least 25 nm and contains at least one refractory carbide, the second layer (12) has an average thickness of at least 25 nm and contains at least one refractory carbide, the third layer (13) has an average thickness of at least 25 nm and contains at least one refractory carbide, the first layer (11) has a larger average linear cross-sectional particle size than the second layer (12), the second layer (12) is sandwiched between the first layer (11) and the third layer (13), and the device in which the average linear cross-sectional particle size of the third layer (13) is larger than that of the second layer (12).

2. The device according to claim 1, wherein at least one refractory carbide of at least one layer includes silicon carbide, tantalum carbide, niobium carbide, hafnium carbide, tungsten carbide, boron carbide, niobium tantalum carbide, hafnium tantalum carbide, or hafnium niobium carbide.

3. The device according to claim 1, wherein the second layer (12) contains more non-carbon components (e.g., silicon) of the refractory carbide (e.g., silicon carbide) than the first layer (11).

4. The device according to claim 1, wherein the porosity of the second layer (12) exceeds the porosity of the first layer (11).

5. The device according to claim 1, wherein the average thickness of all individual layers is in the range of 0.025 μm to 500 μm for all layers that are layers of the multilayer (10) of the device.

6. The device according to claim 1, wherein the average linear cross-sectional surface particle size of the outer surface is in the range of 0.1 to 50 μm.

7. The device according to claim 1, wherein the material of the device includes graphite (2) or elemental silicon.

8. - A refractory carbide layer containing more or less non-carbon components (e.g., silicon) of the refractory carbide (e.g., silicon carbide) than the adjacent layer, and / or - A refractory carbide layer having a porosity different from that of the adjacent layer A CVD method for preparing the device according to claim 1, in which a refractory carbide layer corresponding to the second layer described above in claim 1 is formed.

9. A refractory carbide layer containing more or less non-carbon components (e.g., silicon) of a refractory carbide (e.g., silicon carbide) than an adjacent layer, and / or a refractory carbide layer having a porosity different from that of an adjacent layer is applied in a preceding layer formation step or a subsequent layer formation step. Atmospheric composition, Pressure, and / or Temperature of the CVD coating atmosphere different from the conditions of Atmospheric composition, Pressure, and / or Temperature of the CVD coating atmosphere The CVD method according to claim 8, which is deposited under the conditions of.

10. The CVD method according to claim 8, wherein a refractory carbide layer having a porosity different from that of an adjacent layer is deposited from a CVD coating atmosphere by changing the content of a carbon source.

11. - The refractory carbide layer containing more or less non-carbon components (e.g., silicon) of a refractory carbide (e.g., silicon carbide) than an adjacent layer, or - The refractory carbide layer having a porosity different from that of an adjacent layer is formed, and a refractory carbide layer corresponding to the third layer described above in claim 1 having a larger average linear cross-sectional particle size is deposited thereon. The deposition of the refractory carbide layer having a larger average linear cross-sectional particle size is continued until the outer surface of the layer having a larger average linear cross-sectional particle size reaches a predetermined average linear surface particle size. The CVD method according to claim 8.

Images