Sputtering target and method for manufacturing

Abstract

Sputtering target made of a Mo alloy containing at least one metal from group 5 of the periodic table, wherein the average CMan group 5 metal content is 5 to 15 at% and the Mo content is ≥ 80 at%, characterized in that the sputtering target has an average C / O ratio in (at% / at%) of ≥ 1, and that the sputtering target is free of oxides.

Description

[0001] The invention relates to a sputtering target comprising molybdenum (Mo) and at least one metal of group 5 of the periodic table, wherein the average content C M of Group 5 metal 5 to 15 at% and the Mo content ≥ 80 at%.

[0002] Sputtering, also known as cathode sputtering, is a physical process in which atoms are released from a sputtering target by bombardment with high-energy ions and enter the gas phase. Sputtering targets made of Mo containing Group 5 metals are known. EP 0 285 130 A1 describes a sputtering target made of a Mo alloy containing 50 to 85 at% tantalum (Ta). JP 2002 - 327 264 A discloses a sputtering target made of a Mo alloy containing 2 to 50 at% niobium (Nb) and / or vanadium (V), a relative density > 95%, a flexural strength > 300 MPa, and a grain size < 300 µm. The sputtering target has a diffusion phase and at least one pure phase or only a diffusion phase. JP 2005 - 307 226 A discloses a sputtering target made of a Mo alloy containing 0.1 to 50 at% of a transition metal. The sputtering target has a length of ≥ 1 m and a homogeneous density of ≥ 98%.Alternatively, JP 2005 - 307 226 A describes a sputtering target that exhibits composition variations of ≤ 20% over its entire length.

[0003] Mo-Nb and Mo-Ta sputtering targets are used, for example, for the production of electrode layers for thin-film transistors or contact layers for touch panels. Meeting the increasing demands regarding layer quality and homogeneity, even in ever-increasing dimensions, is the goal of numerous development activities. For example, JP 2008 - 280 570 A describes a manufacturing process for a Mo-Nb sputtering target with an Nb content of 0.5 to 50 at%. This process involves first producing a Mo sintered piece, which is then crushed into powder. The resulting Mo powder undergoes a reducing treatment and is mixed with Nb powder. This mixture is then compacted by hot isostatic pressing.While this process allows for a reduction in the oxygen content in the powder, it cannot further reduce the oxygen content in the sputtering target because hot isostatic pressing takes place in a closed container (can). Furthermore, it is not possible to distribute Nb in the Mo with the homogeneity required for many applications.

[0004] JP 2005 - 290 409 A, in turn, describes a sputtering target made of a Mo alloy containing 0.5 to 50 at% of a metal from the group Ti, Zr, V, Nb, and Cr, wherein the oxygen contained in the target is arranged in the form of oxides in the interface region between the Mo-rich phase and the alloying element-rich phase. The preferred manufacturing method for this comprises the steps of mixing Mo powder and powder of the alloying element, sintering, crushing the sintered part into powder, and compacting the resulting powder by hot isostatic pressing in the known state. The oxides adversely affect the homogenization of the sputtering target during hot pressing because the grain boundary diffusion rate is reduced. Furthermore, the oxides adversely affect the sputtering behavior.

[0005] JP 2013 - 083 000 A describes the production of a sputtering target from a Mo alloy containing 0.5 to 60 at% of one or more elements from the group Ti, Nb, and Ta. Mo powder is mixed with a hydride powder of the alloying element, this mixture is degassed at 300°C to 1,000°C, and then densified by hot isostatic pressing. While the hydride powder decomposes into metal powder during degassing, oxygen absorption occurs again during further processing steps through adsorption on the surfaces of the powder particles. This oxygen is not removed during hot isostatic pressing.

[0006] The described sputtering targets do not meet the increasing demands regarding layer homogeneity, homogeneity of sputtering behavior, and the avoidance of unwanted local melting. Local melting is caused, for example, by arc processes (local formation of an electric arc).

[0007] With the described manufacturing technologies, it is not possible to produce sputtering targets that meet the requirements described above for at least one of the following reasons: a) Oxides hinder grain boundary diffusion; b) oxygen depletion during the consolidation process is not possible; c) the consolidation process does not result in sufficient homogenisation of the alloying elements; d) Interface and grain boundary volumes as well as defect density, which are partly responsible for a sufficiently high diffusion rate, are not sufficiently high; (e) the consolidation process leads to an unacceptably high grain coarsening; f) the powder used results in a coarse-grained sputtering target.

[0008] The object of the present invention is to provide a sputtering target which meets the requirements described above and / or does not have the deficiencies described above. In particular, the object of the invention is to provide a sputtering target with which a very homogeneous layer can be produced, both in terms of chemical composition and layer thickness distribution, and which is not prone to local melting due to arc processes. In addition, the sputtering target should have uniform sputtering behavior. Uniform sputtering behavior means that the individual grains or the individual regions of the sputtering target can be ablated at the same rate, so that no relief structure is created in the region of the sputtered surface during the sputtering process.

[0009] A further object of the present invention is to provide a manufacturing method that allows the production of a sputtering target having the aforementioned properties in a simple and process-consistent manner.

[0010] The problem is solved by the independent claims. Particular embodiments are described in the subclaims.

[0011] The sputtering target contains Mo and at least one metal from group 5 of the periodic table. Group 5 metals are Ta, Nb, and V. The average content C Mof Group 5 metal is 5 to 15 at%, the Mo content ≥ 80 at%. The Group 5 metal is preferably completely dissolved in the Mo, which has a positive influence on uniform sputtering behavior. Completely dissolved means that the content of Group 5 metal, which is present in elemental form (as Ta, Nb and / or V grains) or as an oxide, is < 1 vol%. The sputtering target has an average C / O (carbon / oxygen) ratio in (at% / at%) of ≥ 1, preferably ≥ 1.2. To determine the average C / O ratio, three center and three edge samples are taken from the sputtering target, analyzed, and the average value is calculated. The carbon is determined by combustion analysis (CA), the oxygen by carrier gas hot extraction (HE). In the following text, the average C / O ratio is referred to as the C / O ratio.

[0012] Group 5 metals exert a strong solid solution strengthening effect on Mo in the dissolved state. This solid solution strengthening is accompanied by a significant reduction in ductility and formability. While two-phase (Mo-rich phase + Group 5 metal-rich phase) alloys can be processed more easily and with more process consistency by forming because the Group 5 metal-rich phase exerts a ductilizing effect, this was not possible with very homogeneous solid solution alloys to date. A C / O ratio of ≥ 1 now ensures that production can include a forming step, whereas a C / O ratio of < 1 does not sufficiently ensure process-reliable production by forming. The reason for this is probably that a C / O ratio of ≥ 1 leads to an increase in grain boundary strength, which can prevent grain boundary cracks.The positive effect of the forming step on the properties of the sputtering target will be explained in more detail below. With a C / O ratio (at% / at%) of ≥ 1, it is now possible for the first time to combine the positive effects of alloy homogeneity and forming texture in a single product. Surprisingly, a C / O ratio of ≥ 1 not only has a positive effect on formed sputtering targets, but also favorably influences the sputtering behavior of sputtering targets that have only been sintered or sintered and densified by hot isostatic pressing. Hot isostatic pressing is preferably carried out without the use of a can.

[0013] How a C / O ratio of ≥ 1 can be set consistently throughout the process will be described in more detail below. The C / O ratio of ≥ 1 also enables the setting of a low oxygen content in the sputtering target. An oxygen content of ≤ 0.04 at%, preferably ≤ 0.03 at%, particularly preferably ≤ 0.02 at% is feasible. The sputtering target is free of oxides. Undesired arc processes can thus be reliably avoided. Free of oxides in the context of this invention is to be understood as meaning that when examined using a scanning electron microscope at a magnification of 1,000x, the number of detectable oxide particles in a range of 0.01 mm 2 ≤ 1. Preferably in a range of 0.1 mm 2 the number of detectable oxide particles ≤ 1.

[0014] Furthermore, the sputtering target preferably has a forming texture. As the name suggests, a forming texture is created during a forming process. A forming texture is not lost during a subsequent annealing treatment, such as recovery or recrystallization annealing. The sputtering target according to the invention can therefore be in an as-formed, recovered, partially recrystallized, or fully recrystallized state. The forming texture can be attributed, for example, to a rolling, forging, or extrusion process. The forming process creates grains that are largely aligned with the same or similar orientation to the surface of the sputtering target. This ensures uniform sputtering behavior, since the removal rate depends on the orientation of the grains.

[0015] It is also advantageous for uniform sputter removal if the forming texture has the following dominant orientations: a. In forming direction: 110 b. Perpendicular to the forming direction: at least one orientation from group 100 and 111.

[0016] If the direction was changed during forming, as is possible with plate-shaped geometries, the forming direction is defined as the direction in which the deformation occurred more strongly (with a higher degree of deformation). Dominant is defined as the orientation with the highest intensity. Typically, the intensity is greater than 1.5 times, preferably twice, the background intensity.

[0017] The deformation texture is determined using SEM (scanning electron microscope) and EBSD (electron backscatter diffraction). The sample is mounted at an angle of 70°. The incident primary electron beam is inelastically scattered by the sample's atoms. If some electrons strike lattice surfaces in such a way that the Bragg condition is met, constructive interference occurs. This amplification occurs for all lattice surfaces in the crystal, so that the resulting diffraction pattern (electron backscatter pattern, also known as the Kikuchi pattern) includes all angular relationships in the crystal and thus also the crystal symmetry. The measurement is performed under the following conditions: - Acceleration voltage: 20 kV, - Aperture 120 µm, - Working distance 22 mm - High current mode - activated - Scanned area: 1761 x 2643 µm 2 . - Index step size: 3 µm.

[0018] The preferred density of the sputtering target, based on the theoretical density of the respective composition, is > 88% in the as-sintered state, > 96% in the sintered and hot isostatically densified state, and > 99.5%, preferably > 99.9% in the formed state. The high density combined with the low oxygen content also ensures arc-free sputtering.

[0019] Furthermore, it is advantageous if the d 50 and the d 90 Value of the grain size distribution, measured transverse to the last forming direction, which satisfies the following relationship: d 50 / d 50 ≤ 5.

[0020] Preferred is d 90 / d 50 ≤ 3, particularly preferably ≤ 1.5.

[0021] To determine the grain size, a cross-section is taken, and the grain boundaries are visualized using EBSD. The mean and maximum grain size are then evaluated using quantitative metallography. The evaluation is carried out in accordance with ASTM E 2627-10. A grain boundary is defined such that the orientation difference between two adjacent grains is ≥ 5°. The grain size distribution with d 90 and d 50The value is determined by quantitative image analysis. It has been shown that a narrow grain size distribution has a very positive influence on the homogeneity of the sputtering behavior. In contrast to other materials, with Mo Group 5 metal sputtering targets, grains with a larger grain diameter sputter more strongly than grains with a smaller grain diameter. The reason for this is not yet clear, but may be due to different defect densities or a channeling effect (lattice guidance effect - penetration of an ion due to linear regions without lattice atoms). With the previously mentioned d 90 / d 50 ratio, this unfavorable uneven sputtering behavior can be almost eliminated.

[0022] The Group 5 metal is not only completely dissolved but also extremely uniformly distributed in Mo. The standard deviation σ of the Group 5 metal distribution measured by SEM / WDX preferentially satisfies the relationship σ≤CM×0.15, particularly preferably σ≤CM×0.1.

[0023] Since the sputtering rate depends on the respective alloying element content, a sputtering target with a very homogeneous Group 5 metal distribution according to the invention exhibits extremely uniform sputtering behavior. This uniform sputtering behavior ensures, on the one hand, that the produced layers have an extremely homogeneous thickness distribution and, on the other hand, that the sputtering target still exhibits low surface roughness / relief formation even after extended use. This, in turn, is a prerequisite for uniform sputtering behavior over a long period of time.

[0024] Furthermore, the Group 5 metal is preferably Ta and / or Nb. Mo-Ta and Mo-Nb alloys exhibit particularly favorable corrosion and etching behavior. The alloy advantageously consists of Mo and 5 to 15 at% Group 5 metal, along with typical impurities. Typical impurities include impurities that are usually already present in the raw materials or are attributable to the manufacturing process.

[0025] A sputtering target according to the invention is particularly advantageously designed as a tubular target. It has been shown that, under the usual sputtering conditions for tubular targets, microstructural features such as oxides, homogeneity, or the ratio of the average to maximum grain size have a greater influence than is the case with flat targets.

[0026] The sputtering target according to the invention can be produced in a particularly simple and process-consistent manner if the method comprises the following steps: - Preparation of a powder mixture comprising: i. ≥ 80 at% Mo powder; ii. powder of at least one Group 5 metal, wherein the content of Group 5 metal in the powder mixture is 5 to 15 at%; and iii. a C source, the amount of C being chosen such that the total content of C in the powder mixture is ∑ c in At% and the total content of O ∑ o in At% satisfy the following relationship: 0.2≤∑C / ∑O≤1.2; and - Consolidation of the powder mixture.

[0027] By a ∑ c / ∑ oA ratio in the range of 0.2 to 1.2 ensures that a C / O ratio of ≥ 1 can be set in the sputtering target. Oxygen decomposition during further process steps occurs preferentially by reaction of oxygen with carbon and hydrogen.

[0028] The total content ∑ o of oxygen in the powder mixture includes the oxygen content in the Mo powder and the oxygen content in the Group 5 metal. The oxygen is mainly present in adsorbed form on the surface of the powder particles. During normal production and storage, the oxygen content in the Mo powder with a Fisher particle size of 2 to 7 µm is typically 0.1 to 0.4 at%. For Group 5 metals with a Fisher particle size of 4 to 20 µm, the oxygen content is typically 0.3 to 3 at%. The total content ∑ cThe carbon content includes the carbon content in the Mo powder, the carbon content in the Group 5 metal, and the carbon content of the C source. The carbon source can be, for example, carbon black, activated carbon, or graphite powder. However, it can also be a carbon-releasing compound, such as Nb carbide or Mo carbide.

[0029] First, the oxygen and carbon content of the powders used is determined using conventional methods, and then the required amount of powder from the carbon source is determined. The powders are then mixed and consolidated using conventional methods. Consolidation refers to processes that lead to densification. Consolidation is preferably carried out by cold isostatic pressing and sintering. Sintering refers to processes in which densification is due solely to the application of heat and not to pressure (as is the case, for example, with hot isostatic pressing).

[0030] During heat treatment, preferably during the sintering process, the carbon from the carbon source reacts with the oxygen present in the powder to form CO2 and, to a lesser extent, CO. This reaction preferably occurs at temperatures where the sintered part still has open porosity. Compaction processes in which the material to be compacted is contained in a can, as is the case, for example, with hot isostatic pressing, are less suitable for advantageously implementing the inventive method. If hot isostatic pressing is carried out with a can, the inventive powder mixture must be subjected to a separate annealing / degassing treatment.

[0031] Preferably, the total carbon content ∑ c and the total oxygen content ∑ o in powder the following relationship: 0.4≤∑C / ∑O≤1.1, particularly preferably 0.6≤∑C / ∑O≤1.

[0032] This enables a very high level of process reliability to be achieved.

[0033] The pressing process is best carried out at pressures of 100 to 500 MPa. If the pressure is < 100 MPa, sufficient density cannot be achieved during sintering. Pressures > 500 MPa result in the compounds resulting from the reaction of carbon and oxygen not being removed from the sintered part quickly enough during the sintering process because the gas permeability is too low. The preferred sintering temperature is between 1,800 and 2,500°C. Temperatures below 1,800°C lead to very long sintering times and insufficient density and homogeneity. Temperatures above 2,500°C lead to grain growth, which adversely affects the advantageous homogeneity of the grain size distribution.

[0034] The advantageous particle size of the Mo powder is 2 to 7 µm, and that of the Group 5 metal powder is 4 to 20 µm. The particle size is determined using the Fisher method. If the particle size of the Group 5 metal is

[0035] > 20 µm, the alloy has an increased tendency to form Kirkendall pores when a pressureless compaction process is used. If the powder grain size of the Group 5 metal is < 4 µm, the oxygen content (oxygen adsorbed on the surface of the powder particles) is too high, and the advantageous, low oxygen values can only be achieved through costly production steps, such as special degassing steps. If the particle size of the Mo powder exceeds 7 µm, this leads to reduced sintering activity. If the particle size is below 2 µm, the gas permeability in the green compact is significantly impaired. The green compact also begins to sinter at lower temperatures. Both effects lead to impaired oxygen removal during the sintering process.

[0036] Preferably, the powder mixture contains no alloying elements other than Mo, Group 5 metal, and carbon source. Impurities are present to a degree typical for these materials.

[0037] If additional alloying elements are used, their total content must not exceed 15 at%. Alloying elements that do not adversely affect sputtering and etching behavior are suitable. Suitable alloying metals include W and Ti.

[0038] Sintering is advantageously carried out in a vacuum, an inert atmosphere, and / or a reducing atmosphere. An inert atmosphere is understood to mean a gaseous medium that does not react with the alloy components, such as a noble gas. Hydrogen is particularly suitable as a reducing atmosphere. The conversion of C and O to CO2 or CO is advantageously carried out in a vacuum or in an inert atmosphere, for example during the heating process. This allows the resulting reaction products to be efficiently removed. In addition, the formation of hydrides of Group 5 metals is avoided. The final sintering is then preferably carried out at least temporarily in a reducing atmosphere, preferably under hydrogen.

[0039] After consolidation, a forming process preferably follows. Forming can be performed, for example, by rolling for flat targets, or by extrusion or forging for tubular targets. The preferred degree of forming is 45 to 90%. The degree of forming is defined as follows: (Aa−Au) / Aa×100(in %) A a Cross-sectional area before forming A u Cross-sectional area after forming

[0040] At degrees of deformation < 45%, the density and uniformity of the sputtering behavior are adversely affected. Degrees of deformation > 90% have a negative impact on production costs. The forming temperature is preferably between 900°C and 1,500°C, at least temporarily. Temporary means, for example, that the initial forming steps are carried out at this temperature. After that, the forming temperature can also be below 900°C. The forming can be carried out in one or more steps.

[0041] If the sputtering target is designed as a flat target, it is preferably soldered to a backplate. Tubular targets can be connected to a support tube, preferably again through a soldering process, or they can be used as monolithic sputtering targets. Indium or an indium-rich alloy is preferred as the solder material.

[0042] In the following, the invention is explained using a manufacturing example.

[0043] Fig. Figure 1 shows a SEM image with WDX scan of rolled Mo-10At% Nb.

[0044] The following powders were used: - Mo powder with a Fisher particle size of 4.5 µm, an oxygen content of 0.24 at% and a carbon content of 0.03 at% - Nb powder with a Fisher particle size of 8 µm, an oxygen content of 1.26 at% and a carbon content of 0.46 at%

[0045] To get a ∑ c / ∑ oTo achieve a value of 0.7 with a Mo input of 758 kg and a Nb input of 81.6 kg, 0.336 kg of carbon black powder with a Fisher grain size of 0.35 µm was mixed with the Mo and Nb powder in a forced mixer. Four plates were produced from this powder mixture by cold isostatic pressing at a pressure of 180 MPa. The plates were sintered at a temperature of 2,150°C, with the heating process taking place in a vacuum for 3 hours up to a temperature of 1,200°C. H2 was then used as the process gas. The sintered body had a density of 8.9 g / cm3 3 (88.6% of theoretical density), a C content of 0.022 at%, and an O content of 0.018 at%. The C / O ratio was 1.22.

[0046] The sintered piece was subjected to SEM / EDX analysis. Nb and Mo were completely dissolved in each other. No oxides were detected.

[0047] The sintered piece was then rolled at a forming temperature of 1450°C and a deformation degree of 78%. A sample was taken from the rolled plate and ground and polished using standard metallographic techniques. The texture of a longitudinal sample was determined using SEM / EBSD.

[0048] The following settings were used: - Acceleration voltage: 20 KV, - Working distance: 22 mm, - High current mode activated, - Aperture 120 µm - Scanned area 1,761 x 2,643 µm 2 - Index step size 3 µm.

[0049] The analysis of the inverse pole figure revealed 110 as the dominant texture in the longitudinal direction (forming direction) with > 2 times the background. In the normal direction (perpendicular to the forming direction), both the 100 and 111 orientations were measured with > 2 times the background.

[0050] The grain size was determined in a cross-section using EBSD. Grain boundaries were defined as all grain orientation differences of ≥ 5° between two adjacent grains. The grain size distribution was determined using quantitative image analysis. 50 Value in an evaluation range of 20,000 µm 2 was 15 µm, the d 90 Value 35 µm. The d 90 / d 50 The ratio was 2.3. This measurement was determined in an analogous manner at 10 other locations and a mean d 90 / d 50 A ratio of 2.41 was determined. The rolled plate was also examined for the homogeneity of the Nb distribution using SEM / EDX and SEM / WDX. Fig. Figure 1 shows a WDX scan over a distance of 1 mm. The standard deviation of the Nb distribution measured over this distance was 1.02 at%.

[0051] The sputtering behavior of sputtering targets prepared in this way was determined by sputtering tests at Ar (argon) pressures in the range of 2.5 × 10 3 up to 1 × 10 -2 mbar and a power of 400 or 800 watts. Soda-lime glass was used as the substrate material. The sputtering targets could be sputtered without the occurrence of arc processes. The specific electrical resistance of the deposited layers (layer thickness = 200 nm) was low, ranging from 13.7 to 18.5 µΩcm, depending on the sputtering conditions. The layers exhibited compressive stresses in the range of -1,400 to -850 MPa.

Claims

[1] Sputtering target made of a Mo alloy containing at least one metal of group 5 of the periodic table, wherein the average content C M of Group 5 metal 5 to 15 at% and the Mo content ≥ 80 at%, characterized by , that the sputtering target has an average C / O ratio in (At% / At%) of ≥ 1, and that the sputtering target is free of oxides. [2] Sputtering target according to claim 1, characterized by that the group 5 metal is completely dissolved in Mo. [3] Sputtering target according to claim 1 or 2 characterized by a forming texture. [4] Sputtering target according to claim 3, characterized by that the forming texture has the following dominant crystal orientations: a. In forming direction: 110 b. Perpendicular to the forming direction: at least one dominant crystal orientation 100 and 111. [5] Sputtering target according to claim 3 or 4, characterized bythat the d 50 and the d 90 Value of the grain size distribution, measured transverse to the last forming direction, which satisfies the following relationship: d 90 / d 50 ≤ 5. [6] Sputtering target according to one of claims 1 to 5 characterized by an O content ≤ 0.04 at%. [7] Sputtering target according to one of claims 1 to 6, characterized by that the relative density is > 99.5% of the theoretical density. [8] Sputtering target according to one of claims 1 to 7, characterized by that the Group 5 metal is uniformly distributed in solution, where the standard deviation σ of the Group 5 metal distribution satisfies the following relationship: σ≤CM×0.

15. [9] Sputtering target according to one of claims 1 to 8, characterized by that the group 5 metal is Ta or Nb. [10] Sputtering target according to one of claims 1 to 9, characterized bythat it consists of 5 to 15 at% Group 5 metal, balance Mo and typical impurities. [11] Sputtering target according to one of claims 1 to 10, characterized by that this is a tube target. [12] A method for producing a sputtering target according to any one of claims 1 to 11, characterized by that this includes the following steps: a. Preparation of a powder mixture comprising: i. ≥ 80 at% Mo powder, ii. powder of at least one Group 5 metal, wherein the content of Group 5 metal in the powder mixture is 5 to 15 at%; and iii. a C source, the amount of C being chosen such that the total content of C in the powder mixture is ∑ C in At% and the total content of O ∑ O in At% satisfy the following relationship: 0.2≤∑C / ∑O≤1.2; b. Consolidation of the powder mixture. [13] Method according to claim 12, characterized bythat the process includes a forming process. [14] Method according to one of claims 12 to 13, characterized by that the consolidation is carried out by: a. Pressing the powder mixture at 100 to 500 MPa to form a green compact, and b. Sintering of the green body at a temperature T, with 1,800°C < T < 2,500°C. [15] Method according to one of claims 12 to 14, characterized by that the Mo powder has a particle size of 2 to 7 µm measured according to Fisher and the Group 5 metal has a particle size of 4 to 20 µm measured according to Fisher. [16] Method according to one of claims 12 to 15, characterized by that ∑ C and ∑ O satisfy the following relationship: 0.4≤∑C / ∑O≤1.

1. [17] Method according to one of claims 12 to 16, characterized by that the powder mixture does not contain any other alloying elements apart from typical impurities. [18] Method according to one of claims 12 to 17, characterized by that the forming is carried out by rolling, extrusion or forging, with the degree of forming being 45 to 90%. [19] Method according to one of claims 12 to 18, characterized by that the sintering is carried out in at least one atmosphere selected from vacuum, inert atmosphere and reducing atmosphere. [20] Method according to claim 19, characterized by that the sintering is carried out at least temporarily during the heating process in at least one atmosphere selected from vacuum and inert atmosphere and at least temporarily during a holding time at sintering temperature in a reducing atmosphere.

Images