JPWO2014156918A1 - Niobium sputtering target - Google Patents

Abstract

A niobium sputtering target containing tungsten of 5 wtppm or more and 100 wtppm or less, having a purity of 99.995% or more excluding tungsten, tantalum and gas components, and an average crystal grain size of 150 μm or less. It is an object of the present invention to provide a niobium sputtering target that has a uniform and fine structure, is stable in plasma, and can improve film thickness uniformity (uniformity). [Selection figure] None

Description

  The present invention relates to a high-purity niobium sputtering target having a uniform and fine structure, stable plasma during sputtering, and excellent film uniformity (uniformity). The high-purity niobium of the present invention contains (adds) tungsten. However, since the content of these elements is small, it will be collectively referred to as “high-purity niobium” in the present specification.

  Sputtering methods are used to form coatings made of metal or ceramic materials in various fields such as electronics, corrosion-resistant materials, decorative fields, catalysts, cutting / polishing materials, and wear-resistant materials. Yes. The sputtering method itself is a well-known method in the above-mentioned field, but recently, particularly in the field of electronics, a sputtering target suitable for forming a complex-shaped film, forming a circuit, and the like is required.

  Among these, a niobium film is attracting attention as a liner film for improving the reflow property of the Al wiring film, and a niobium sputtering target having excellent sputtering characteristics is required to form the niobium film. In general, this niobium target is processed into a target by repeatedly forging and annealing (heat treatment) of an ingot or billet obtained by melting and casting a niobium raw material, and further rolling and finishing (mechanical and polishing).

  In such a manufacturing process, forging of an ingot or billet destroys the cast structure, diffuses and disappears pores and segregation, and further recrystallizes by annealing, thereby increasing the density and strength of the structure. The target is manufactured. In general, a melt-cast ingot or billet has a crystal grain size of 50 mm or more. By forging and recrystallizing the ingot or billet, the cast structure is destroyed, and the average grain size varies depending on the size. Can obtain crystal grains of about 200 μm.

  When sputtering is performed using the target thus manufactured, the finer and uniform the recrystallized structure of the target is, the more uniform the film can be formed, and less arcing and particles are generated. It is said that a film having excellent thickness uniformity (uniformity) can be obtained. Therefore, various measures are adopted in the target manufacturing process from the viewpoint of uniformizing the recrystallized structure.

  For example, Patent Document 1 and Patent Document 2 are sputtering targets for forming an Nb film used as a liner material for an Al wiring film in an integrated circuit wiring technique, and the average crystal grain size of Nb is 100 μm or less. In addition, each crystal grain has a grain size in the range of 0.1 to 10 times the average crystal grain size, and the ratio of the grain size of adjacent crystal grains is in the range of 0.1 to 10, In addition, a target having a variation in the ratio of particle size to size within ± 30% is disclosed.

  Patent Document 3 is a high-purity niobium for producing a sputter target, a capacitor, a resistance film, a wire, etc., having a purity of 99.99% and a metal niobium having an average particle diameter of about 150 μm or less. Disclosure. In addition, the amount of tantalum, molybdenum, and tungsten as impurities in metallic niobium is less than about 100 ppm, and by reducing the amount of impurities in the metal, the properties and performance of the manufactured product should be good. Is disclosed.

  Patent Document 4 discloses a sputtering target made of an Nb material for sputtering an Nb oxide film suitable for an optical thin film, and the hydrogen content in the sputtering target made of the Nb material is set to 1 to 1000 ppm. Thus, a technique for suppressing the variation in the film thickness associated with the change in the film forming speed by making the variation in the hydrogen content within 20% is disclosed.

  However, even if it sees patent documents in any case, inclusion of a specific element does not refine the structure, thereby stabilizing the sputter film formation and making the film thickness uniform. In addition, elements other than niobium are usually recognized as impurities, and as disclosed in Patent Document 3 and Patent Document 4, these elements are removed as much as possible.

JP 2000-104164 A JP 2009-149996 A JP 2004-511651 A Japanese Patent Laying-Open No. 2005-097696

  The present invention maintains a high purity of niobium, and by adding a specific element, it has a uniform fine structure, the plasma during sputtering is stable, and the film thickness is uniform (uniformity). An object is to provide an excellent high-purity niobium sputtering target.

In order to solve the above problems, the present invention maintains a high purity of niobium, and by adding a specific element, has a uniform and fine structure, stable plasma during sputtering, and a film We obtained the knowledge that high purity niobium sputter rig targets with excellent thickness uniformity (uniformity) can be obtained. Based on such knowledge, the present invention
1) A niobium sputtering target containing 5 wtppm or more and 100 wtppm or less of tungsten, having a purity excluding tungsten, tantalum and gas components of 99.995% or more and an average crystal grain size of 150 μm or less,
2) The niobium sputtering target according to 1) above, wherein the variation in tungsten content is within 30%;
3) The niobium sputtering target according to 1) or 2) above, wherein the variation in average crystal grain size is within 20%.

  The present invention maintains the purity of niobium at a high purity, and by adding tungsten as an essential component, it has a uniform and fine structure, the plasma at the time of sputtering becomes stable, and the film thickness uniformity (uniformity) is excellent. Further, it has an excellent effect that a high-purity niobium sputtering target can be provided. And even in the initial stage of sputtering, since the plasma is stabilized, the burn-in time can be shortened.

It is a figure explaining the sampling position for the characteristic investigation of a target.

The niobium sputtering target of the present invention is usually produced by the following steps.
First, niobium having a purity of 5N (99.999%) or more is prepared as a niobium raw material. Although there is no restriction | limiting in particular in purity, It is preferable to use a thing with few impurities beforehand.

  Next, this is melted and purified by electron beam melting or the like to increase the purity, and this is cast to produce an ingot or billet. An appropriate amount of tungsten can be added during this dissolution. There is no particular limitation on the purity of tungsten, but it is preferable to use a material having few impurities in advance.

  Next, the niobium sputtering target of the present invention can be produced by subjecting the ingot or billet to heat treatment, forging, rolling, finishing (machining, polishing), or the like.

  Specifically, ingot: heat treatment at 1500 ° C. to 1800 ° C.—forging at room temperature (forging) —heat treatment at 1150 ° C. to 1250 ° C.—cold rolling—heat treatment at 1050 ° C. to 1150 ° C.—finishing (machining, polishing) By performing processing or the like, the sputtering target of the present invention can be obtained.

  By forging or rolling, the cast structure can be destroyed, and the pores and segregation can be diffused or disappeared. Furthermore, this can be recrystallized by heat treatment, and by repeating this forging or rolling and heat treatment (recrystallization annealing), It is possible to increase the density, refinement and strength of the tissue.

  Usually, the niobium sputtering target is manufactured by the above manufacturing process. However, this manufacturing process shows an example, and the present invention does not invent this manufacturing process. Of course, the present invention includes all of them.

  In order to take advantage of the characteristics of the niobium sputtering target, a high-purity material is often used, but there is a drawback that the crystal grains of the target are apt to be coarsened during the heat treatment. In the production process of such a high-purity target, the present inventors usually have a portion where tungsten having a content of about 7 to 20 wtppm is segregated to about 50 wtppm, and the crystal grain size is local. I found that it was miniaturized.

  From this, we obtained a hint that the addition of tungsten is effective for miniaturization of the niobium sputtering target, and this led to the present invention. That is, in the niobium sputtering target of the present invention, what is important is that niobium having a purity of 99.995% or more excluding tungsten, tantalum and gas components contains 5 wtppm or more and 100 wtppm or less of tungsten as an essential component. .

  The lower limit value of 5 wtppm of the tungsten content is a numerical value for exerting the effect, and the upper limit value of 100 wtppm of the tungsten content is a numerical value for maintaining the effect of the present invention. When the upper limit is exceeded, segregation of tungsten occurs, and a part of the niobium is not recrystallized. As a result, the upper limit value of tungsten is set to 100 wtppm in order to reduce the uniformity of the film.

  As described above, the addition of tungsten can be performed when the niobium raw material is melted by electron beam. However, since the niobium raw material usually contains a small amount of tungsten, the impurity concentration of the purified niobium is reduced. If it is analyzed and the tungsten content specified in the present invention is satisfied, it can be used.

  In the niobium sputtering target of the present invention, its purity needs to be 99.995% or more excluding tungsten, tantalum and gas components. In general, gas components are difficult to remove unless they are special methods, and it is difficult to remove them during purification in a normal production process. Therefore, the gas components of oxygen, hydrogen, carbon, and nitrogen are the same as those of niobium of the present invention. Exclude from purity. Further, tantalum is an element belonging to the same group as niobium, and its influence on the film characteristics when forming a film is small. Further, when tantalum is separated, the cost is greatly increased, so it is excluded from the purity.

  As described above, the inclusion of tungsten leads to a uniform and fine structure of niobium, but the inclusion of other metal components, non-metal components, oxides, nitrides, carbides and other ceramics is harmful and can be tolerated. Can not. This is because these impurities are considered to have a function of suppressing the action of tungsten. Further, these impurities are clearly different from tungsten, and it is difficult to finish the crystal grain size of the niobium sputtering target uniformly, and do not contribute to stabilization of the sputtering characteristics.

  In the niobium sputtering target of the present invention, it is desirable that the variation in the tungsten content in the target is within 30%. As long as the appropriate tungsten content has a function (property) for forming a uniform and fine structure of the niobium sputtering target, the more uniform dispersion of tungsten can greatly contribute to the uniform refinement of the target structure.

  In the present invention, it is of course important to keep in mind that the variation in the content of tungsten in the target is within 20%, and it is important to have a clear intention, and the method for controlling the variation is not particularly limited. . The distribution of tungsten content in the ingot varies depending on the processing conditions of the electron beam melting, but it is possible to uniformly diffuse tungsten by performing a heat treatment at 1500 ° C. to 1800 ° C. after the ingot is manufactured.

  Regarding the variation in the tungsten content in the target, for example, as shown in FIG. 1, in the case of a disk-shaped target, three points (center point) are arranged on eight equal lines starting from the center of the disk. , ½ point of radius, outer periphery or its neighboring points) and analyze the total tungsten content of 17 points {16 points + center point (one point because the center point is common)}. Then, the variation can be calculated based on the formula {(maximum value−minimum value) / (maximum value + minimum value)} × 100. In addition, also in the below-mentioned Examples and Comparative Examples, the variation in tungsten content is obtained using this analysis technique.

  The niobium sputtering target of the present invention preferably further has an average crystal grain size of 150 μm or less. With appropriate addition of tungsten, the crystal grain size can be made finer, but it is important to note that the average crystal grain size is 150 μm or less and that the intention is clear. Since the purpose of the present invention is to refine the crystal grain size, the lower limit is not particularly limited, but from the viewpoint of controlling the crystal grain size, it is preferably 30 μm or more, more preferably 50 μm or more. Further, it is more desirable that the variation of the average crystal grain size is within 20%.

  Regarding the variation in the crystal grain size in this target, for example, as shown in FIG. 1, in a disk-shaped target, three points (center point, Taking the half point of the radius, the outer periphery or the vicinity thereof, and measuring the crystal grain size of niobium of a total of 17 points {16 points + center point (one point because the center point is common)}. The variation in crystal grain size can be calculated based on the formula {(maximum value−minimum value) / (maximum value + minimum value)} × 100. It should be noted that also in examples and comparative examples described later, the variation in crystal grain size is obtained using this analysis technique.

  Thus, the inclusion of tungsten forms a uniform fine structure of the target, thereby stabilizing the plasma during sputtering, thereby improving the film thickness uniformity (uniformity) of the film formed by sputtering. It becomes possible. Further, since the plasma at the time of sputtering is stable even at the initial stage of sputtering, the burn-in time can be shortened.

  Hereinafter, description will be made based on Examples and Comparative Examples. In addition, a present Example is an example to the last, and is not restrict | limited at all by this example. In other words, the present invention is limited only by the scope of the claims, and includes various modifications other than the examples included in the present invention.

Example 1
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 7 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was heat-treated at 1600 ° C. for 5 hours, forged at room temperature, and then heat-treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  When the particle size of the target obtained by the above steps was measured using a scanning electron microscope (SEM), as shown in Table 1, the average crystal particle size was 132 μm, and the variation was 19%. . The crystal grain size was measured for only one part of the target using the cutting method described in JIS / H0501. The variation in the content of tungsten was 29%, and the purity excluding tungsten, tantalum and gas components was 99.999%. The size of each part of the target subjected to component analysis was 2.5 mm × 2.5 mm × 20 mm.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 1.1%, and the film thickness distribution fluctuation was small. The sheet resistance was measured using OMNIMAP RS75 manufactured by KLA-Tencor. The uniformity of the film was evaluated as “good” when the sheet resistance distribution variation was 3 or less, and “bad” when 3 or more.

(Example 2)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 15 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was heat-treated at 1600 ° C. for 5 hours, forged at room temperature, and then heat-treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal grain size was 129 μm, and the variation was 9%. The variation in the content of tungsten was 16%, and the purity excluding tungsten, tantalum and gas components was 99.998%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 1.1%, and the film thickness distribution fluctuation was small.

(Example 3)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 55 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was heat-treated at 1600 ° C. for 5 hours, forged at room temperature, and then heat-treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

As a result of measuring the target obtained by the above steps using the same method as in Example 1,
As shown in Table 1, the average crystal grain size was 107 μm, and the variation was 11%. The variation in the content of tungsten was 16%, and the purity excluding tungsten, tantalum and gas components was 99.996%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 1.2%, and the film thickness distribution fluctuation was small.

Example 4
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 88 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was heat-treated at 1600 ° C. for 5 hours, forged at room temperature, and then heat-treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the target particle size obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 106 μm, and the variation was 10%. The variation in the tungsten content was 10%, and the purity excluding tungsten, tantalum and gas components was 99.995%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 1.0%, and the film thickness distribution fluctuation was small.

(Example 5)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 93 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was heat-treated at 1600 ° C. for 5 hours, forged at room temperature, and then heat-treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the particle size of the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 98 μm, and the variation was 14%. The variation in the content of tungsten was 14%, and the purity excluding tungsten, tantalum and gas components was 99.995%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 1.4%, and the film thickness distribution fluctuation was small.

(Comparative Example 1)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 2 wtppm was melted by electron beam, and this was cast into an ingot. Next, this ingot was forged at room temperature and then heat treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the particle size of the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 230 μm, and the variation was 39%. The variation in the content of tungsten was 60%, and the purity excluding tungsten, tantalum and gas components was 99.999%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as is apparent from Table 1, in this comparative example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 5.2%, and the film thickness distribution fluctuation was large.

(Comparative Example 2)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 4 wtppm was melted by electron beam, and this was cast into an ingot. Next, this ingot was forged at room temperature and then heat treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the particle size of the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 221 μm, and the variation was 36%. The variation in the content of tungsten was 50%, and the purity excluding tungsten, tantalum and gas components was 99.999%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as is apparent from Table 1, in this comparative example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 4.9%, and the film thickness distribution fluctuation was large.

(Comparative Example 3)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 7 wtppm was melted by electron beam, and this was cast into an ingot. Next, this ingot was forged at room temperature and then heat treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the particle size of the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 127 μm, and the variation was 34%. The variation in the content of tungsten was 38%, and the purity excluding tungsten, tantalum and gas components was 99.998%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this comparative example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 4.8%, and the film thickness distribution fluctuation was large.

(Comparative Example 4)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 130 wtppm was melted by electron beam, and this was cast into an ingot. Next, this ingot was forged at room temperature and then heat treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the target particle size obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 103 μm, and the variation was 29%. The variation in the content of tungsten was 31%, and the purity excluding tungsten, tantalum and gas components was 99.994%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as is apparent from Table 1, in this comparative example, the sheet resistance distribution fluctuation (standard deviation / average value × 100) was 4.9%, and the film thickness distribution fluctuation was large.

(Comparative Example 5)
A raw material in which a predetermined amount of tungsten was added to niobium having a purity of 99.999% so that the tungsten content in the target was 143 wtppm was melted by electron beam, and this was cast into an ingot. Next, the ingot was forged at room temperature and then heat treated at 1200 ° C. for 2 hours. Next, after rolling this at room temperature, heat treatment was performed at 1100 ° C. for 1 hour and finishing was performed to obtain a target material having a diameter of 450 mm.

  As a result of measuring the particle size of the target obtained by the above steps using the same method as in Example 1, as shown in Table 1, the average crystal particle size was 108 μm and the variation was 37%. The variation in the content of tungsten was 41%, and the purity excluding tungsten, tantalum and gas components was 99.993%.

  Next, since the film thickness depends on the sheet resistance, the distribution of the sheet resistance in the wafer (12 inches) was measured, thereby examining the distribution of the film thickness. Specifically, the sheet resistance at 49 points on the wafer was measured, and the average value and standard deviation were calculated. As a result, as apparent from Table 1, in this comparative example, the sheet resistance distribution variation (standard deviation / average value × 100) was 5.5%, and the film thickness distribution variation was large.

  The present invention includes a niobium sputtering target containing tungsten of 5 wtppm or more and 100 wtppm or less and having a purity of 99.995% or more excluding tungsten, tantalum and gas components, thereby providing a uniform and fine structure and stable plasma. Yes, it has an excellent effect on the uniformity of film thickness (uniformity). Also, since the plasma during sputtering can be stabilized even in the initial stage, it has the effect of shortening the burn-in time. Therefore, the liner film for improving the reflowability of the field of electronics, particularly the Al wiring film It is useful as a target suitable for forming a film having another complicated shape.

Claims (3)

  A niobium sputtering target containing tungsten of 5 wtppm or more and 100 wtppm or less, having a purity of 99.995% or more excluding tungsten, tantalum and gas components, and an average crystal grain size of 150 μm or less.   The niobium sputtering target according to claim 1, wherein variation in tungsten content is within 30%. The niobium sputtering target according to claim 1 or 2, wherein the variation of the average crystal grain size is within 20%.