JP4928694B2 - Method for identifying the structure of a thin film polycrystal - Google Patents

Description

【００９０】
次に、本発明の膜構造シミュレーション法を用いて、この薄膜多結晶の膜構造をシミュレーションした。
【００９１】
初期結晶粒の成長結晶面は（１００）面と（１１１）面とが共存するとした。また初期結晶粒の基板に対する面方位は、（１１１）面、（２２０）面、（３１１）面、（４００）面からあらかじめ与えた確率分布に従って決定した。
【００９２】
初期パラメータは、シミュレーション時間ｔ＝１００秒、単位時間ステップΔｔ＝０．７２秒、成長速度Ｖ_iは（１００）面＝４ｎｍ／ｓおよび（１１１）面＝４ｎｍ／ｓ、単位セル長ΔＬ＝５ｍｍ、セル空間＝５０×５０×５０、初期核密度ρ_1st＝１０^-3／ｎｍ²である。
【００９３】
そして、膜構造シミュレーションによって計算した積分強度比と半値幅とが表１に示した測定結果に一致するように初期核密度ρ_1stを変化させて、最終的に薄膜多結晶の膜構造を同定した。
【００９４】
図１４（ａ）に膜構造の同定結果を示す。同図は、結晶粒の形状とＸ線回折に基づく面方位の配向性とのシミュレーション結果を、ソフトウェアで可視化したものである。面方位を逆極図で示した図１４（ｂ）に基づいて、個々の結晶粒を面方位で分類してマッピングしている。図１４（ａ）に示されるように、薄膜多結晶の表面形状、結晶粒形状、粒界などの３次元膜構造が、十分な精度で同定されていることが確認された。
【００９５】
（実施例２）
ＣＶＤ法で製膜した別の薄膜多結晶シリコンの膜構造を、双晶の発生を考慮して同定した。
【００９６】
まず、この薄膜多結晶のＸ線回折スペクトルを測定した。下表２に、各ピークについての積分強度比と半値幅との測定結果を示す。
【００９７】
【表２】

【００９８】
次に、本発明の双晶の発生を考慮した膜構造シミュレーション法を用いて、この膜構造をシミュレーションした。
【００９９】
初期結晶粒の成長結晶面は（１００）面と（１１１）面とが共存するとし、初期結晶粒の基板に対する面方位は、（１１１）面、（２２０）面、（３１１）面、（４００）面からあらかじめ与えた確率分布に従って決定した。
【０１００】
本実施例における初期パラメータは、シミュレーション時間ｔ＝１００秒、単位時間ステップΔｔ＝０．７２秒、成長速度Ｖ_iは（１００）面＝４ｎｍ／ｓおよび（１１１）面＝４ｎｍ／ｓ、単位セル長ΔＬ＝５ｍｍ、セル空間＝５０×５０×５０、初期核密度ρ_1st＝１０^-3／ｎｍ^{2 、}双晶発生確率Ｒ_2nd＝２．７３×１０^-4／ｎｍ²／ｓである。
【０１０１】
そして、膜構造シミュレーションによって計算した積分強度比と半値幅とが表２に示した測定結果に一致するように初期核密度ρ_1stと双晶発生確率Ｒ_2ndとを変化させて、最終的に薄膜多結晶の膜構造を同定した。
【０１０２】
図１５に膜構造の同定結果を示す。同図は、実施例１の図１４（ａ）と同様に、シミュレーション結果をソフトウェアで可視化したものであり、図１４（ｂ）に示した面方位に基づいて、個々の結晶粒を面方位で分類している。図１５に示されるように、やはり薄膜多結晶の表面形状、結晶粒形状、粒界などの３次元膜構造が、十分な精度で同定されていることが確認された。
【０１０３】
【発明の効果】
以上のように、本発明においては、薄膜多結晶を構成する個々の結晶粒の成長と結晶粒間の相互作用とを、任意の精度で表現することが可能である。また、薄膜多結晶の製膜条件から各結晶表面の成長速度を計算した結果を組み合せることで、あらゆる薄膜多結晶の成長をシミュレーションすることができる。そのため、あらゆる物質の薄膜多結晶構造の同定に適用可能である。そして同定した薄膜多結晶の膜構造から所望の物性値を計算することで、ある製膜条件における薄膜多結晶の性質を、製膜・計測試験することなく計算から求めることができる。
【０１０４】
さらに、今までの分子動力学法や分子軌道法等では、計算可能な空間・時間スケールが実際の製膜時間と比べてはるかに小さかったため、製膜メカニズムの全貌を扱うことはできず、膜構造の同定は事実上不可能であった。しかし、本発明においては、気相・表面反応を分子動力学法や分子軌道法等で計算してパラメータ化した結果（結晶表面の成長速度など（表面上の結晶面の成長速度の値と、初期核密度の大まかなオーダ等））を用いることで、実用的な計算時間と空間スケールで膜構造の同定を行なうことができるため、製膜条件の最適化を行なうことができる。
【０１０５】
また本発明では、２次核の主因である双晶の発生メカニズムを成長シミュレーションに導入することで、薄膜多結晶の膜構造を正確に同定することができる。さらに、双晶のもつ対称性に由来する結晶粒間の成長競合の影響を成長シミュレーションに考慮できるため、薄膜多結晶の膜構造に対する双晶の影響を定量的に評価することができる。
【図面の簡単な説明】
【図１】本発明に係る薄膜多結晶の膜構造同定方法の一例を示す流れ図。
【図２】本発明に係る薄膜多結晶の膜構造同定方法で用いる膜構造の成長シミュレーションモデルの原理を説明するための図。
【図３】本発明に係る膜構造シミュレーション方法の一例を示す流れ図。
【図４】本発明に係る初期結晶粒の成長形の決め方の一例を説明するための図。
【図５】本発明に係る初期結晶粒の成長形の位置決めの仕方の一例を説明するための図。
【図６】本発明に係るセル操作の手順の一例を示す流れ図。
【図７】本発明に係る結晶粒の物性値の計算方法の一例を説明するための図。
【図８】本発明に係るシミュレーション結果について計算および計測スペクトルの一致の判定および再計算の方法の一例を示す図。
【図９】本発明に係る製膜途中に発生する２次核を含む薄膜多結晶の膜構造同定方法の一例を示す流れ図。
【図１０】本発明に係る双晶の発生モデルを組み込んだ膜構造シミュレーション方法の一例を示す流れ図。
【図１１】本発明に係る双晶の発生手順の一例を示す流れ図。
【図１２】本発明に係る双晶の面方位の設定方法の一例を説明する図。
【図１３】本発明に係る双晶の仮想結晶粒の形状の一例を示す図。
【図１４】本発明の実施例で得られた膜構造の同定結果の一例を示す中間調画像。
【図１５】本発明の実施例で得られた膜構造の同定結果の他の例を示す中間調画像。
【図１６】従来技術における膜構造のシミュレーション結果を示す図。
【符号の説明】
１１…基板
１２、１３、１００…初期結晶粒
１４、１５、４１０、４４０…仮想結晶粒
１６…仮想結晶粒の増分
１７…実際の薄膜多結晶の表面
１８…粒界
１９、２０…実際の結晶粒の形状
２１、２２、２３、２４、４５０…３次元セル
１３０、１４０…結晶面の成長速度
２００、３００…成長形
４００…母結晶
４２０…結晶表面
４３０…双晶[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for identifying a film structure of a thin film polycrystal.
[0002]
[Prior art]
In order to optimize the characteristics of the thin film that constitutes a solar cell or the like, it is desirable to be able to predict characteristics such as electrical conduction resulting from the film structure, and film properties related to the characteristics. For these predictions, when the thin film is a thin film polycrystal composed of a polycrystal, it is necessary to handle individual crystal grains.
[0003]
In recent years, a CVD (chemical vapor deposition) method and a PVD (physical vapor deposition) method are widely used for forming a thin film polycrystal. When using the CVD method or the PVD method, it is necessary to identify the characteristics and film structure of the thin film polycrystal according to the film forming conditions. This identification can be performed by mapping the orientation, grain boundary, and the like of each crystal grain constituting the thin film by measurement using electron backscatter diffraction imaging (EBSP). However, a scanning electron microscope (SEM) to which an EBSP apparatus usually used for this measurement is added has a limit on the electron beam diameter. Therefore, it is impossible to measure each crystal grain separately for a thin film polycrystal composed of a small crystal grain size. In addition, with a transmission electron microscope (TEM), measurement can be performed for each crystal grain, but it is difficult to evaluate a large number of samples because preparation of a measurement sample requires a lot of labor.
[0004]
For this reason, in order to save the trouble of measurement, a method has been developed for obtaining the orientation distribution of crystal grains constituting a polycrystal using a measurement technique such as X-ray diffraction.
[0005]
In the apparatus disclosed in Japanese Patent Laid-Open No. 55-78233, mapping is performed by scanning an X-ray beam with respect to a sample. However, since there is a limit to the minimum value of the X-ray beam diameter, it is impossible to separate and evaluate each crystal grain of a thin film polycrystal having a small crystal grain size.
[0006]
The polycrystal pole figure measured by X-ray diffraction used in JP-A-5-256799 and JP-A-9-178675, and the rocking curve by θ scanning used in JP-A-9-281061 are as follows: It is the average data of the whole polycrystal. Therefore, although the frequency of the orientation in which the crystal grains are oriented can be grasped, information on the grain boundaries inside the polycrystal cannot be obtained. Therefore, this result alone is insufficient to grasp the film structure of the thin film polycrystal.
[0007]
Therefore, in order to identify in detail the film structure including grain boundaries, etc., a process for analyzing measurement results such as X-ray diffraction using an appropriate model for the film structure of a thin-film polycrystal should be introduced. is required.
[0008]
Examples of the characteristics of the thin film polycrystal used for identifying the film structure and the model of the film structure include the following.
[0009]
In the first model, thin film polycrystals are handled as a homogeneous thin film by obtaining average characteristics of individual crystal grains. In this model, the difference in individual crystal grains does not affect the entire thin film. However, a separate model is required for determining the average characteristics of individual grains, and variations from the average characteristics cannot be handled.
[0010]
As shown in FIG. 16, the second model is a specific crystal grain having the fastest growth rate in the normal direction of the substrate when the crystal grains collide with each other as the individual crystal grains grow from the start of film formation. This is a model in which thin film polycrystals are grown and the film structure is determined based on the “geometric selection” mechanism in which only the film grows preferentially (the source of FIG. 16 is A. Van der Drift, Philips Res. Repts. Vol. .22 (1967) p.267 and the “geometric selection” mechanism is described in AA Chernov, Modern Crystallography III: Crystal Growth, Springer-Verlag, Berlin (1984) p.284). Examples of simulations using this model include A. Van der Drift, Philips Res. Repts. Vol. 22 (1967) p. 267, and C. Wild et al., Diamond and Related Materials Vol. 2 (1993) p. There are .158. However, in these examples, since crystal grains are handled in a two-dimensional space, it is not possible to accurately model the plane orientation of the crystal grains that can be expressed only by handling them in a three-dimensional space.
[0011]
[Problems to be solved by the invention]
The present invention uses an interaction model between crystal grains based on the above-mentioned “geometric selection” mechanism accompanying the growth of a thin film polycrystal, but does not exist in the first and second models. The purpose is to provide a method for identifying the film structure of a thin film polycrystal from the measurement of crystal analysis such as X-ray diffraction by using a growth simulation using a film structure model of a thin film polycrystal that realizes spatial representation. To do.
[0012]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, according to the present invention, the process of measuring crystal analysis data for a target thin film polycrystal and the film structure of the thin film polycrystal are simulated based on the film forming conditions of the thin film polycrystal. A method for identifying a film structure of a thin film polycrystal, comprising adjusting a parameter for designating the film structure so as to match the crystal analysis data, thereby estimating the film structure of the thin film polycrystal Provided.
[0013]
In the present invention, the thin film polycrystalline film structure simulation includes a step of preparing a simulation space divided by a plurality of cells, and a plurality of growth forms assigned to the substrate in accordance with the film forming conditions in the space. The initial crystal grains are arranged, the growth form of each initial crystal grain is similarly expanded for each unit growth step in the space, and the increment of the enlarged crystal grains is included in each step. If a cell is not shared with an increase in another crystal grain, the cell is determined to be a constituent of the crystal grain, and if it is shared with an increase in another crystal grain, the cell Preferably includes a step of determining that the grain boundary is between the crystal grains.
[0014]
Further, in the present invention, for each initial crystal grain arranged on the substrate, the plane orientation of the plane parallel to the substrate is assigned according to the film forming conditions, and the cell determined to be a component of the crystal grain, It is preferable to give the plane orientation assigned to the crystal grains.
[0015]
Further, in the present invention, based on the cell arrangement of the cell group to which the plane orientation is assigned, the shape of each crystal grain is obtained to calculate the film structure of the thin film polycrystal, and the plane orientation, volume, grain of each crystal grain It is preferable to calculate the crystal analysis data of the thin film polycrystal by obtaining the diameter.
[0016]
Furthermore, in the present invention, twins serving as secondary nuclei are generated according to the probability of occurrence on the growth surface of each crystal grain at each unit growth step, and the plane orientation of the twins is the same as that of the crystal grains where twins are generated. It is preferable to determine from the surface orientation by a mirroring operation in which the generation surface is a mirror surface.
[0017]
In the present invention, it is preferable that the film structure designation parameter adjusted so as to coincide with the crystal analysis data is a density of initial crystal grains arranged on the substrate.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
In the present invention, the film structure of the thin film polycrystal is identified by using the measurement of the film structure of the thin film polycrystal by crystal analysis using X-ray diffraction or neutron diffraction and the simulation of the film structure of the thin film polycrystal. . Specifically, the parameters of the simulation are adjusted so that the measurement result of the X-ray diffraction spectrum or the like matches the calculation result of the spectrum obtained from the simulation, and the film structure of the thin film polycrystal is identified.
[0019]
In the film structure simulation of the present invention, the crystal grains are virtually enlarged while maintaining the growth form from the initial crystal grains arranged on the substrate in the three-dimensional simulation space divided by the cells. Then, a crystal grain attribute (such as a plane orientation of a plane parallel to the substrate) is given to the cell included in the enlarged virtual crystal grain, and a grain boundary attribute is given to the cell where the virtual crystal grains collide with each other. From the three-dimensional cell group having the crystal grain and grain boundary attributes obtained in this way, the shape and plane orientation of the actually grown crystal grain are determined, and the X-ray diffraction spectrum and the like are calculated. The film structure simulation of the present invention can also take into account the effect that twins, which are secondary nuclei for crystal growth, occur during film formation. Note that in this specification, the plane orientation with respect to a crystal grain or twin crystal substrate refers to the plane orientation of a crystal plane parallel to the substrate among crystal planes of the crystal grain or twin crystal.
[0020]
First, a first embodiment of the present invention will be described.
[0021]
FIG. 1 shows a flowchart of a method for identifying a thin film polycrystalline film structure according to the present invention. This method includes the following first to sixth procedures.
[0022]
In the first procedure, a thin film polycrystal for which the film structure is to be identified is subjected to crystal analysis by X-ray diffraction or neutron diffraction, for example, X-ray diffraction spectrum measurement using the θ-2θ method or the like.
[0023]
In the second procedure, initial setting of parameters for a film structure model (described later) used in the film structure simulation is performed. The parameters to be set are, for example, film formation time t, initial nuclear density ρ generated at the start of film formation_1st(Density of initial crystal grains on the substrate). Note that parameters handled as film forming conditions, such as the substrate temperature during film formation, the pressure in the film forming chamber, and the like are also set.
[0024]
In the third procedure, as described above, the film structure simulation is performed using the film structure model, and the film structure of the thin film polycrystal under the parameter conditions given in the second procedure is calculated.
[0025]
In the fourth procedure, an X-ray diffraction spectrum is calculated for the target thin film polycrystal using the thin film polycrystal film structure obtained by the simulation in the third procedure.
[0026]
In the fifth procedure, it is determined whether or not the simulation time has reached t set in the second procedure. When the simulation time t is reached, the simulation is terminated. If the time t has not yet been reached, the time is further advanced by the unit time step Δt of the simulation, and the third and fourth procedures are repeated.
[0027]
In the sixth procedure, it is determined whether or not the calculation result of the X-ray diffraction spectrum obtained in the above-described film structure simulation matches the measurement spectrum obtained in the first procedure. If they do not match, the initial nuclear density ρ_1stAnd recalculate from the third procedure.
[0028]
Next, a film structure model used in the film structure simulation according to the present invention will be described. FIG. 2 is a schematic diagram for explaining the principle of the film structure model.
[0029]
FIG. 2A shows a method for expressing an interaction model between crystal grains based on a “geometric selection” mechanism, which expresses an interaction between crystal grains accompanying the growth of a thin film polycrystal, in order to apply to the present invention. This is an improved version.
[0030]
In FIG. 2 (a), for the sake of simplicity, initial crystal grains a and b (12, 13) are generated on the substrate 11 at the initial stage of the growth of the thin film polycrystal, and the planes parallel to the respective substrates have arbitrary plane orientations. Is shown. These crystal grains grow while maintaining an arbitrary shape called a growth shape determined from the growth rate of each crystal surface under the same film-forming conditions. However, if the film-forming conditions differ locally and temporally, they grow differently. Take shape.
[0031]
In the “geometric selection” mechanism, each crystal grain keeps growing in its own growth form, and when the crystal grains become sufficiently large and collide with each other, the growth at that location is performed with the collision position as the grain boundary. To stop. At this time, depending on the plane orientation and position of each crystal grain, a growth competing action occurs in which the crystal grain continues to grow further or another crystal grain stops growing at the grain boundary. By this growth competition action, the growth form that the initial crystal grains had cannot be maintained, but a part of the growth form is maintained as it is in a portion where no collision with other crystal grains occurs. As a result, as shown in FIG. 2A, the crystal grain a (19) in which the initial crystal grain a (12) has grown and the crystal grain b (20) in which the initial crystal grain b (13) has grown A thin film polycrystal surface 17 is formed through the grain boundaries 18.
[0032]
In the thin film polycrystal growth simulation of the present invention, each of the crystal grains a, b (19, 20) is virtually compared to the actual individual crystal grains a, b (19, 20) constituting the thin film polycrystal. Assume hypothetical crystal grains a and b (14, 15) when the initial crystal grain growth form is maintained. Assuming that a part of each virtual crystal grain increment 16 collides with each other while continuing the growth is regarded as a grain boundary 18, it is assumed that the part does not grow beyond the grain boundary 18. Represents the “selection” mechanism.
[0033]
FIG. 2B is a diagram for explaining a representation of a three-dimensional space using cell elements in order to apply the “geometric selection” mechanism to the film structure simulation method of the present invention.
In FIG. 2B, the virtual crystal grains a and b (14, 15) are enlarged while maintaining the growth shape according to the growth time. And among the spaces taken in with the expansion of the virtual crystal grains, the space that does not collide with other virtual crystal grains retains the attributes of the crystal grains to which the virtual crystal grains belong, The collision space retains the grain boundary attributes. Holding such information requires an enormous storage area. However, in practice, it is often unnecessary to handle a region smaller than the lattice constant of the target polycrystalline material or the size of the crystal grain handled in film formation.
[0034]
Therefore, in the present invention, the three-dimensional space is divided by the cell elements 21 as shown in FIG.
Then, in each growth step, for example, if the cell taken in the increment 16 of the virtual grain b (15) is not included in any other virtual grain increment (for example, the cell 23 in the figure), The cell 23 is given the attribute of the crystal grain to which the virtual crystal grain b (15) belongs. Specifically, as the attributes of the crystal grains, information such as the plane orientation of the crystal grains with respect to the substrate and the exposed crystal plane when the cell 23 is exposed on the surface of the thin film polycrystal is given to the cell 23. Similarly, the attribute of the crystal grain to which the virtual crystal grain a (14) belongs is given to the cell 22 taken in the increment of the virtual crystal grain a (14).
[0035]
On the contrary, if the cell taken into the increment 16 of the virtual crystal grain b (15) is also included in the increment of the other virtual crystal grain a (14) (cell 24 in the figure), the cell 24 has a grain. Gives the attributes of the world. Specifically, information such as the plane orientation of the crystal grains to which the two virtual crystal grains a and b belong to the substrate is given to the cell 24.
[0036]
In this way, the growth model of the thin film polycrystal including the interaction between crystal grains by the “geometric selection” mechanism shown in FIG. 2A can be expressed with necessary accuracy. Here, the contents of the cell 23 or 24 once given the attribute of the crystal grain or grain boundary are basically kept unchanged. However, when the X-ray spectrum calculation result obtained from the film structure simulation is compared with the measured spectrum, the crystal grains are considered to have shrunk due to the influence of etching by flying particles during crystal growth. Changes such as canceling the attribute of the crystal grain given to the cell.
[0037]
FIG. 3 is a flowchart illustrating the details of the second to fifth procedures related to the film structure simulation among the procedures shown in FIG.
[0038]
In the second procedure of FIG. 3, as described above, initial setting of parameters necessary for the film structure simulation is performed.
First, as an initial setting, a growth simulation time (film formation time) t and a growth unit time step Δt are given.
Next, the crystal surface (growth surface) constituting the initial crystal grains, and the growth rate V of each crystal surface i_igive.
The crystal surface constituting the crystal grain is determined crystallographically depending on the type of crystal to be grown. In general, a crystal face with a high index has a high growth rate and cannot constantly exist on the surface. For this reason, crystallography reveals that crystal planes that grow under various conditions are low-index crystal planes such as {100} planes. However, since it is not known which low index crystal plane is exposed under each film forming condition, all the low index crystal planes observed so far are taken into consideration for the target substance. For example, when a silicon crystal growth simulation is performed, a crystal surface of {100}, {111}, etc. is given to each initial crystal grain. Note that different crystal surfaces may be provided for each crystal grain, and different types of crystal surfaces may coexist on the same crystal grain.
[0039]
Growth rate V of each crystal surface i_iIs obtained by calculation using a molecular dynamics method or the like based on the film forming conditions. If the growth conditions are locally different due to diffusion of the film-forming species even during the growth, the corresponding growth rate V_iAs appropriate.
Subsequently, the unit cell length ΔL of the three-dimensional cell is given, and the cell space is set by giving the number of cells in each of the three directions. Finally, an array of virtual grains is set. This array stores the center position of the virtual grain, the radius of the virtual grain, and a flag indicating whether the virtual grain can be grown, and information on the virtual crystal grain (virtual grain) used when the crystal grain growth is performed in the third procedure. Is to hold.
[0040]
Next, initial crystal grains are set.
First, the number of initial crystal grains is determined. This is the initial nuclear density ρ described above._1stFrom (initial nuclear density ρ_1st) × (substrate area) = (number of initial crystal grains). Next, the diameter and position of each initial crystal grain are determined. Here, the diameter is the diameter (or radius) of the initial crystal grain in a direction parallel to the substrate, and is determined using a random number based on a predetermined probability distribution. Also, the position of each initial crystal grain is determined by the diameter of each initial crystal grain determined as described above so that the initial crystal grains do not overlap each other and the interval between the initial crystal grains based on the probability distribution given in advance. Decide to make it happen.
[0041]
Next, the surface orientation of each initial crystal grain with respect to the substrate is strongly dependent on the film formation conditions (specifically, the surface orientation of the substrate used for film formation, the surface condition such as surface cleanliness, etc. However, the surface It may vary depending on the composition of nearby chemical species) and is determined based on the probability distribution given in advance. For example, when performing a silicon crystal growth simulation, plane orientations such as {111}, {220}, {100} are given. The probability distribution is set in consideration of the fact that the plane orientation of the initial crystal grains with respect to the substrate is affected by the plane orientation of the substrate.
[0042]
Next, for each initial crystal grain, the growth rate V of the crystal surface and each crystal surface i given by the above-mentioned initial setting._iThen, determine the shape of the growth shape.
[0043]
For example, as described above, consider a case where a {100} or {111} plane is given as a growth plane. As is well known, {100} has six equivalent surfaces, and {111} has eight equivalent surfaces. Equivalent surface growth rate V_iTherefore, the growth form of a crystal grain given a {100} crystal surface is a regular hexahedron having six faces, and the growth form of a crystal grain given a {111} crystal surface is It is a regular octahedron consisting of eight faces.
[0044]
FIG. 4 shows the growth rate V depending on the crystal surface._iIt is a schematic diagram for demonstrating how to determine the growth form in case where is different. In FIG. 4, it is shown in two dimensions for simplicity.
FIG. 4A shows an example in which the initial crystal grains have six crystal surfaces and form a cube (a square 100 having four sides in the figure). Growth speed V of two opposite faces (two sides 110 in the figure) among these six faces._i130 is the growth rate V of the other four surfaces (two sides 120 in the figure)._iSuppose that it is twice 140. In this case, when the initial crystal grain 100 continues to grow, the vertical direction connecting the two surfaces 110 finally becomes a rectangular parallelepiped (rectangle 150 in the drawing) in which the direction connecting the two surfaces 120 is doubled. Accordingly, in this case, as a growth form of the initial crystal grains 100, a rectangular parallelepiped (rectangle 200 in FIG. 4B) whose vertical direction is twice as large as the other direction is determined as shown in FIG. 4B. . In this way, the growth rate V depends on the crystal surface._iThe growth forms of initial crystal grains having different values are determined.
[0045]
Next, the shape of the initial crystal grain is determined by positioning the growth form given to the initial crystal grain as described above so that the plane orientation of the initial crystal grain determined as described above with respect to the substrate is realized. .
[0046]
FIG. 5 is a schematic diagram for explaining a method of positioning the growth shape. As shown in FIG. 5A, consider a case where the growth form is a regular octahedron 300 composed of eight equivalent crystal surfaces {111}. Further, as shown in FIG. 5B, a case where the plane orientation of the initial crystal grain 300 with respect to the substrate is {100} is considered. In such a case, as shown in FIG. 5B, the regular octahedron 300 is positioned so that the {100} plane inside the growth shape is parallel to the substrate 11. As a result, as shown in the figure, the shape of the initial crystal grain is a shape in which the pyramid is exposed on the surface of the substrate 11.
[0047]
Returning to FIG. 3, in the second procedure, the attributes (such as the plane orientation of the plane parallel to the substrate) of the initial crystal grains are assigned to the cells included in the initial crystal grains represented by the virtual crystal grains as described above. Given, the initial crystal grain is expressed by a cell.
Finally, parameters (substrate temperature at the time of film formation, pressure in the film formation chamber, etc.) handled as film formation conditions are set.
[0048]
In the third procedure of FIG. 3, crystal grain growth is performed for each unit time step Δt. That is, the virtual crystal grains are enlarged (similar enlargement) from the initial crystal grains while maintaining the growth-like similarity at every unit time step Δt. As shown in FIG. 2 (b) described above, the cell operation is performed using the virtual crystal grains a and b (14, 15) to express the growth of the actual crystal grain, and the virtual crystal in which the cell operation cannot be performed. It is determined that the grains have disappeared from the surface of the thin film polycrystal.
[0049]
FIG. 6 shows the flow of a specific procedure for this cell operation. First, each virtual crystal grain is enlarged every unit time step Δt. And every time it expands, among the cells included in the increment (increment 16 in FIG. 2B) which is the volume increase of each virtual crystal grain, there is an empty cell (a cell having no crystal grain attribute). Determine whether or not.
[0050]
If the empty cell is included in the increment of the virtual crystal grain, it is determined whether or not the empty cell is also included in the increment of the other virtual crystal grain. That is, it is determined whether the number of virtual crystal grains including the empty cell in the increment is one (no empty cell is shared) or two or more (an empty cell is shared). When there is one kind of virtual crystal grain including the empty cell in the increment, the attribute (plane orientation, etc.) of the crystal grain to which the virtual crystal grain belongs is given to the empty cell. When there are two or more types of virtual crystal grains including an empty cell in an increment, two or more types of virtual crystal grains collide with each other in the empty cell region, and a grain boundary is formed. Gives the grain boundary attribute. Thus, every unit time step, each empty cell in the three-dimensional space is filled with a crystal grain attribute or grain boundary attribute.
[0051]
Further, when no empty cell is included in the increment of the virtual crystal grains, the virtual crystal grains are stopped without being enlarged. This means that in the previous growth step, the crystal grains to which the virtual crystal grains belong were surrounded by other crystal grains (that is, grain boundaries) and thus could not be grown in this step. The crystal grains whose growth has been stopped in this way are not finally exposed on the surface of the thin film polycrystal, that is, disappeared from the surface of the thin film polycrystal.
[0052]
Returning to FIG. 3 again, in the fourth procedure after the third procedure in which crystal grains are grown as described above, the attributes of each three-dimensional cell to which the attributes of crystal grains or grain boundaries are given are referred to. Thus, the actual shape and plane orientation of each crystal grain are estimated. (Here, the shape of the crystal grain is estimated, and the plane orientation of the plane parallel to the substrate, which is an attribute of the cell included in the shape, is estimated as the plane orientation of the crystal grain.) Then, for example, the X-ray diffraction θ-2θ method is calculated, such as the grain size, volume, X-ray integral intensity, half-width of the X-ray peak, and distribution of the crystal surface exposed on the thin film polycrystal surface. .
[0053]
FIG. 7 is a diagram for explaining the calculation method. As shown in the figure, the crystal grain size, the crystal grain volume, and the X-ray integral intensity of the crystal grain are calculated using the following equations (1), (2), and (3).
Crystal grain size = average height of the cell group constituting the crystal grain from the substrate (1)
Volume of crystal grains = total number of cells constituting crystal grains × unit cell volume (2)
X-ray integrated intensity of crystal grains = constant dependent on crystal plane orientation with respect to substrate x crystal grain volume ...・ (3)
Note that the X-ray integrated intensity of the above equation (3) is a value for each crystal grain. Therefore, in order to compare with the actual X-ray measurement spectrum result, a value obtained by summing up the integrated intensity of the above equation (3) for all crystal grains having the same plane orientation is calculated.
Further, the half width of the X-ray peak derived from the plane orientation of the crystal grain is calculated using the grain size of the crystal grain. Specifically, from each crystal grain, the Gaussian relationship in which the height of the X-ray peak ∝ the volume of the crystal grain × grain size × a constant specific to the diffraction plane orientation, the half width of the X-ray peak ∝ 1 / grain size Since an X-ray peak can be expressed by a curve, the half-value width of the curve obtained by taking the sum of Gaussian curves for all crystal grains is taken as the half-value width of the X-ray peak.
[0054]
In this way, the calculated X-ray integrated intensity and half-value width for each peak of the X-ray diffraction spectrum are compared with the measured spectrum, as will be described later.
[0055]
Returning to FIG. 3, in the fifth procedure after the fourth procedure in which the actual shape of each crystal grain, the X-ray integrated intensity, etc. are calculated as described above, the simulation time is set in the second procedure described above. It is determined whether or not t has been reached. When the simulation time t is reached, the simulation is terminated and the sixth procedure is started. If not yet reached, the time is further advanced by step Δt to enlarge the virtual crystal grains, and the third and fourth procedures are repeated.
[0056]
In the sixth procedure, as shown in FIG. 1, is the calculation result (X-ray integrated intensity and half-value width) of the X-ray diffraction spectrum obtained by the simulation consistent with the measurement spectrum obtained in the first procedure? Determine if. When the error between the calculation result and the measurement result falls within a predetermined error range, it is determined that the two match. If they do not match, the initial nuclear density ρ_1stAnd recalculate from the third procedure described above. Initial nuclear density ρ_1stThe change is performed so that the error between the calculation result of the X-ray diffraction spectrum and the measurement spectrum becomes small, and finally the spectrum matches between the calculation and the measurement.
[0057]
FIG. 8 shows a method for determining and recalculating the coincidence of the calculated and measured spectra. FIG. 8A is a flowchart of determination, and FIG. 8B is a schematic diagram showing a difference between calculation and measurement spectrum for each peak constituting the spectrum.
[0058]
First, in (i) of FIG. 8 (a), as shown in FIG. 8 (b), the calculated spectrum and the measured spectrum are compared for the entire spectrum, and the difference is calculated for each peak constituting the spectrum. That is, as described above, in each of the calculated spectrum and the measured spectrum, the X-ray integrated intensity (peak area) and half-value width (peak width) of each peak are calculated, and the difference between the calculated spectrum and the measured spectrum is calculated. put out. In practice, the ratio of the integrated intensity of each peak to the integrated intensity of a specific peak is used instead of the integrated intensity of each peak.
[0059]
Next, in (ii), each peak is oriented in a specific orientation so that the difference between the calculation and measurement obtained in (i) (difference between the integrated intensity ratio and the half width) is reduced. A correction amount to be changed such as a grain size and a volume of a crystal grain group having the above-described plane orientation (plane orientation of a plane parallel to the substrate) is calculated. The integrated intensity of the diffraction peak is proportional to the volume of the crystal grain in the plane orientation from which the peak is derived, and the half width of the peak is almost inversely proportional to the grain size of the crystal grain.
[0060]
Finally, in (iii), the initial nuclear density ρ, which is a parameter, is satisfied so as to satisfy the correction amount such as the grain size and volume of the crystal grain group obtained in (ii)._1stThe amount of change is predicted and recalculated from the third procedure of FIGS. That is, initial nuclear density ρ_1stA number of candidates that are slightly changed with respect to the direction and amount of change are prepared, and each candidate is recalculated from the third procedure, and a candidate that minimizes the error between the integrated intensity ratio and the half-value width is selected.
[0061]
Initial nuclear density ρ_1stWhen is reduced, the number of initial crystal grains on the substrate is reduced, so that it is less likely that crystal grains collide with each other during growth simulation and hinder growth. Therefore, each crystal grain grows large, and the grain size (grain size in the height direction) and volume of each crystal grain become large. Therefore, as described above, the calculated value of the integrated intensity of the peak derived from the plane orientation of each crystal grain is increased, and the half-value width is decreased. The initial nuclear density ρ_1stWhen is increased, contrary to the above, the integrated intensity of each peak is reduced and the half-value width is increased.
[0062]
As described above, when the calculation result of the X-ray diffraction spectrum and the measured spectrum coincide with each other, from the actual shape and plane orientation of the individual crystal grains obtained by the fourth procedure in FIGS. The film structure of a thin film polycrystal is identified.
[0063]
As described above in detail, in the thin film polycrystalline film structure identification method according to the present invention, a thin film polycrystalline film based on film forming conditions using measurement data from X-ray diffraction or the like, for example, an X-ray diffraction spectrum is used. Identify the structure. In performing the growth simulation necessary for identifying the film structure, the growth rate of each crystal surface calculated from the film forming conditions is used. If information such as a database on the growth rate of the crystal surface obtained from various film forming conditions can be used, the calculation time can be reduced. Further, by adjusting the unit cell length ΔL, a result can be obtained with arbitrary accuracy and calculation time. That is, if ΔL is reduced, the accuracy is increased, and if ΔL is increased, the calculation time is shortened.
[0064]
Since the main point of the thin film polycrystal film structure identification method according to the present invention is the thin film polycrystal growth simulation, the effect of the simulation on the identification will be described. The characteristics of the individual crystal grains constituting the thin film polycrystal can be considered in the growth simulation including not only the average value but also the variation. Therefore, accurate film structure identification can be performed with less calculation time. Specifically, considering the distribution according to the position and diameter of the initial crystal grains and the distribution according to the plane orientation of the plane parallel to the substrate, the influence on the film structure of the thin-film polycrystal can be obtained on the growth simulation. it can. Therefore, in addition to the identification of the film structure after completion of the film formation, it is also possible to identify the structure in the initial and intermediate stages of film formation.
[0065]
The crystal grains constituting the thin film polycrystal are represented by virtual crystal grains and a three-dimensional cell, and the model shown in FIG. 2B is used. Therefore, the complex surface shape and grain boundary structure given as a result of the growth of the thin film polycrystal and the accompanying interaction between crystal grains can be simulated with the required accuracy. In a so-called cellular automaton (CA) model, an error derived from approximation by a cell propagates to the surroundings. However, in the model shown in FIG. 2B of the present invention, only the virtual crystal grains have an influence by giving the crystal grain attributes to the cell, and the cell error propagates to the surroundings as compared with the CA model. Hateful. Therefore, in the present invention, the error in identifying the film structure is reduced.
[0066]
In the present invention, since the calculation results are stored in a three-dimensional cell, the shape of individual crystal grains constituting the thin film polycrystal, the orientation of the plane orientation based on X-ray diffraction, etc. can be easily obtained from the stored data. Can be calculated.
[0067]
Next, a second embodiment of the present invention will be described.
[0068]
In thin-film polycrystalline film formation, not only the initial crystal grains generated in the initial stage of film formation grow to form a thin film, but also separate crystal grains (secondary nuclei) are newly generated in the middle of the film formation to produce the crystals. Grain may grow. Since the generation of secondary nuclei affects the film structure, it is necessary to incorporate a secondary nucleation model in the identification of the film structure of a thin-film polycrystal.
[0069]
There is twinning as a cause of particularly effective secondary nuclei. Twins are a type of defect that occurs during crystal growth, occur only on specific crystal planes of crystal grains, and easily occur on {111} planes in silicon and the like. Therefore, it is necessary to incorporate the twinning generation model, which is one of the main causes of the secondary nuclei, to identify the film structure of the thin film polycrystal in general.
[0070]
FIG. 9 is a flowchart of the film structure identification method for a thin film polycrystal including secondary nuclei due to twins generated during film formation according to the present invention. This figure is a modification of FIG. 1 described above.
In the first procedure of FIG. 9, as in the first procedure of FIG. 1, measurement of an X-ray diffraction spectrum or the like is performed on a thin film polycrystal for which a film structure is to be identified.
In the second procedure, initial settings of parameters used in the film structure simulation set in the second procedure of FIG. 1 (film formation time t, initial nuclear density ρ_1st) In addition to the twin generation probability R per unit area on the crystal plane where twins are generated._2ndSet.
In the third procedure, a thin film polycrystal growth simulation is performed using a film structure model in which a twinning generation model is incorporated into the film structure model used in the third procedure of FIG.
[0071]
In the fourth procedure, similarly to the fourth procedure in FIG. 1, an X-ray diffraction spectrum is calculated for the target thin film polycrystal using the film structure obtained in the third procedure.
In the fifth procedure, as in the fifth procedure of FIG. 1, it is determined whether or not the simulation time has reached t set in the second procedure.
In the sixth procedure, as in the sixth procedure of FIG. 1, it is determined whether the calculation result of the X-ray diffraction spectrum obtained by the above-described film structure simulation matches the measurement spectrum. However, unlike the case of FIG. 1, if the two do not match, the initial nuclear density ρ_1stAnd the probability of twin formation R_2ndAnd recalculate from the third procedure.
[0072]
FIG. 10 is a flowchart illustrating the details of the second to fifth procedures related to the film structure simulation incorporating the twinning generation model among the procedures described above, and is a modification of the flowchart of FIG. 3 described above. .
[0073]
In the second procedure of FIG. 10, as in the second procedure of FIG. 3, initial setting of parameters necessary for the film structure simulation is performed. However, in addition to the simulation time t and unit time step Δt set in the second procedure of FIG. 3, the twin generation probability R per unit area_2ndAlso give in advance. Further, based on the film forming conditions, a crystal surface capable of generating twins is identified. The unit cell length ΔL is set so that the maximum number of twins generated on the unit cell is one. More specifically, the generation probability R of twins per unit area described above is added to the area of one face of the unit cell including the crystal surface of the virtual crystal grains capable of generating twins._2ndThe unit cell length is set so that the value multiplied by 1 is 1 or less. Furthermore, since the minimum size of a thermodynamically stable twin crystal can be found crystallographically, the initial diameter of the twin crystal along the generated crystal surface is set using this size.
[0074]
Similarly to the second procedure of FIG. 3, the growth rate of each twin crystal surface is obtained from the film forming conditions, and the twin growth form is obtained from this growth rate. The kind of crystal surface given to twins is the same as the kind of crystal surface given to initial crystal grains described above. The shape of the twin virtual crystal grains is defined from the twin growth form thus obtained.
[0075]
In the third procedure of FIG. 10, crystal grain growth is performed for each unit time step Δt, as in the third procedure of FIG. However, the cell operation for enlarging the virtual crystal grains while maintaining the growth form is performed not only on the initial crystal grains but also on twins generated during the growth, as will be described later. The twin generation probability R per unit area described above._2ndThus, twins are generated on the crystal surface where twins can be generated.
[0076]
FIG. 11 shows the details of the procedure for generating twins.
First, for each growth step Δt for enlarging the virtual crystal grain, in the three-dimensional cell space, among the empty cells included in the increment of the virtual crystal grain, an empty cell including the crystal surface of the virtual crystal grain capable of generating a twin crystal is obtained. look for.
[0077]
Next, the probability of twin formation R in the area of one face of such an empty cell_2ndWhether or not to generate twins on the corresponding cell is determined using a random number based on the value multiplied by. When twins are not generated, the process proceeds to the fourth procedure in FIG.
When twins are generated, the plane orientation of the twin crystal substrate is set as follows according to the plane orientation of the crystal grain (mother crystal) to which the cell in which the twin crystal occurs belongs.
[0078]
FIG. 12 shows an example of a method for setting the plane orientation of twins. This figure shows an example in which twins are generated on the silicon {111} plane. If a mirroring operation is performed with the silicon {111} plane as a mirror surface with respect to the vector representing the plane orientation of the mother crystal, a vector representing the plane orientation of the twin crystal can be obtained. That is, since both have the symmetry that they are mirror images of the {111} plane, the plane orientation of the twin crystal is set from the plane orientation of the mother crystal using this property. For example, as shown in FIG. 12, the plane orientation vector of the mother crystal is a^ρIf [001], the twin plane orientation vector is b^ρ[001] and the mother crystal vector is a^ρIf [100], the twin vector is b^ρ[100]. For other materials, the plane orientation of twins is set using the symmetry of twins.
[0079]
Returning to FIG. 11, the virtual crystal grains of twins are set based on the plane orientation given to the twins as described above and the twin growth forms set in the second procedure of FIG. .
Next, of the cells contained in the twin virtual crystal grains, twin attributes (such as plane orientation with respect to the twin substrate) are given to empty cells that do not have other crystal grain attributes.
[0080]
FIG. 13 shows an example of the shape of twin virtual crystal grains represented by a three-dimensional cell. As shown in the figure, the virtual crystal grains 440 of twins 430 are generated on the crystal surface 420 capable of generating twins among the crystal surfaces of the virtual crystal grains 410 of the mother crystal 400. The shape of the virtual crystal grain 440 is set based on the plane orientation of the twin 430 obtained from the plane orientation of the mother crystal 400 and the growth form of the twin crystal 400 as described above. The twin 430 attribute is given to the three-dimensional cell 450 included in the virtual crystal grain 440 of the twin 430.
[0081]
In FIG. 11, finally, the twin virtual crystal grains expressed by the cells as described above are put into a group of other virtual crystal grains that continue to grow from the initial crystal grains. Thus, the twin crystal is grown by performing the crystal grain growth for each unit time step Δt shown in the third procedure of FIG. 10 also for the twin virtual crystal grains.
[0082]
Returning to FIG. 10, the fourth procedure and the fifth procedure after the third procedure in which the crystal grains are grown as described above are the same as the fourth and fifth procedures in FIG. That is, in the fourth procedure, the actual shape and plane orientation of each crystal grain are estimated by referring to the attributes of each three-dimensional cell given the attributes of crystal grains or grain boundaries (as described above, the crystal grains The shape of the crystal grain is estimated from the plane orientation of the plane parallel to the substrate, which is an attribute of the cell included in the shape, and the crystal grain size, volume, X Calculation of line integral intensity, half width of X-ray peak, etc. is performed. In the fifth procedure, it is determined whether or not the simulation time has reached the set value t. If the simulation time has reached t, the simulation is terminated and the sixth procedure is started. If not, the time is further advanced by step Δt, and the third and fourth procedures are repeated.
[0083]
In the sixth procedure, as shown in FIG. 9, the X-ray integrated intensity and the half-value width of each peak are calculated for the calculation result of the X-ray diffraction spectrum obtained by the simulation and the measurement spectrum obtained by the first procedure. Determine whether they match. If they do not match, the initial nuclear density ρ_1stAnd the probability of twin formation R_2ndAnd recalculate from the third procedure.
[0084]
As described above, when the calculation result of the X-ray diffraction spectrum agrees with the measurement spectrum, the actual thin film polycrystal of the target thin film polycrystal is obtained from the actual shape and plane orientation of the individual crystal grains obtained in the fourth procedure. Identify the membrane structure.
[0085]
As described above in detail, by incorporating the twinning generation mechanism into the film structure identification method of the first embodiment, it is possible to perform thin film polycrystal growth simulations under various film forming conditions. As a result, the film structure of the thin film polycrystal can be identified more accurately.
[0086]
Further, by introducing the generation mechanism of twins, which is the main cause of secondary nuclei, into the growth simulation of the thin film polycrystal, the film structure of the thin film polycrystal can be identified more accurately. Furthermore, since the plane orientation of twins is accurately set based on the plane orientation of the mother crystal and the effects of growth competition between grains derived from the symmetry of twins can be considered in the growth simulation, The degree of twin effects on the film structure can be quantitatively evaluated.
[0087]
【Example】
Example 1
The film structure of thin-film polycrystalline silicon formed by CVD was identified.
[0088]
First, the X-ray diffraction spectrum of this thin film polycrystal was measured. Table 1 below shows the measurement results of the integrated intensity ratio and the half width for each peak.
[0089]
[Table 1]

[0090]
Next, the film structure of this thin film polycrystal was simulated using the film structure simulation method of the present invention.
[0091]
It is assumed that the (100) plane and the (111) plane coexist as the growth crystal plane of the initial crystal grains. The plane orientation of the initial crystal grains with respect to the substrate was determined according to a probability distribution given in advance from the (111) plane, the (220) plane, the (311) plane, and the (400) plane.
[0092]
The initial parameters are simulation time t = 100 seconds, unit time step Δt = 0.72 seconds, growth rate V_i(100) plane = 4 nm / s and (111) plane = 4 nm / s, unit cell length ΔL = 5 mm, cell space = 50 × 50 × 50, initial nuclear density ρ_1st= 10^-3/ Nm²It is.
[0093]
Then, the initial nuclear density ρ is set so that the integrated intensity ratio and the half width calculated by the film structure simulation agree with the measurement results shown in Table 1._1stFinally, the film structure of the thin polycrystalline film was identified.
[0094]
FIG. 14A shows the identification result of the film structure. The figure visualizes the simulation results of the crystal grain shape and the orientation of the plane orientation based on X-ray diffraction by software. Based on FIG. 14B in which the plane orientation is shown as an inverse polar diagram, individual crystal grains are classified and mapped by the plane orientation. As shown in FIG. 14A, it was confirmed that the three-dimensional film structure such as the surface shape, crystal grain shape, and grain boundary of the thin film polycrystal was identified with sufficient accuracy.
[0095]
(Example 2)
The film structure of another thin film polycrystalline silicon formed by the CVD method was identified in consideration of the generation of twins.
[0096]
First, the X-ray diffraction spectrum of this thin film polycrystal was measured. Table 2 below shows the measurement results of the integrated intensity ratio and half width for each peak.
[0097]
[Table 2]

[0098]
Next, this film structure was simulated using the film structure simulation method considering the generation of twins according to the present invention.
[0099]
It is assumed that the (100) plane and the (111) plane coexist in the growth crystal plane of the initial crystal grains, and the plane orientations of the initial crystal grains with respect to the substrate are (111) plane, (220) plane, (311) plane, (400 ) Determined according to the probability distribution given in advance from the surface.
[0100]
The initial parameters in this example are: simulation time t = 100 seconds, unit time step Δt = 0.72 seconds, growth rate V_i(100) plane = 4 nm / s and (111) plane = 4 nm / s, unit cell length ΔL = 5 mm, cell space = 50 × 50 × 50, initial nuclear density ρ_1st= 10^-3/ Nm^{2 ,}Twin generation probability R_2nd= 2.73 × 10^-Four/ Nm²/ S.
[0101]
Then, the initial nuclear density ρ is set so that the integrated intensity ratio and the half width calculated by the film structure simulation agree with the measurement results shown in Table 2._1stAnd twin generation probability R_2ndAnd finally the film structure of the thin film polycrystal was identified.
[0102]
FIG. 15 shows the identification result of the film structure. This figure visualizes the simulation result with software, as in FIG. 14A of Example 1, and based on the plane orientation shown in FIG. Classification. As shown in FIG. 15, it was confirmed that the three-dimensional film structure such as the surface shape, crystal grain shape, and grain boundary of the thin film polycrystal was identified with sufficient accuracy.
[0103]
【The invention's effect】
As described above, in the present invention, the growth of individual crystal grains constituting the thin film polycrystal and the interaction between the crystal grains can be expressed with arbitrary accuracy. Moreover, the growth of all thin film polycrystals can be simulated by combining the results of calculating the growth rate of each crystal surface from the thin film polycrystal film forming conditions. Therefore, it can be applied to the identification of the thin film polycrystalline structure of any substance. Then, by calculating desired physical property values from the film structure of the identified thin film polycrystal, the properties of the thin film polycrystal under certain film forming conditions can be obtained from the calculation without performing a film forming / measurement test.
[0104]
Furthermore, in the conventional molecular dynamics method and molecular orbital method, the space and time scale that can be calculated is much smaller than the actual film formation time. The identification of the structure was virtually impossible. However, in the present invention, the results of parameterization by calculating the gas phase / surface reaction by the molecular dynamics method or the molecular orbital method (such as the growth rate of the crystal surface (the value of the growth rate of the crystal plane on the surface, By using a rough order of the initial nuclear density, etc.), it is possible to identify the film structure on a practical calculation time and space scale, so that the film forming conditions can be optimized.
[0105]
Moreover, in the present invention, the film structure of a thin film polycrystal can be accurately identified by introducing the generation mechanism of twins, which is the main cause of secondary nuclei, into the growth simulation. Furthermore, since the influence of the growth competition between crystal grains derived from the symmetry of twins can be considered in the growth simulation, the influence of twins on the film structure of the thin film polycrystal can be quantitatively evaluated.
[Brief description of the drawings]
FIG. 1 is a flowchart showing an example of a method for identifying the structure of a thin-film polycrystal according to the present invention.
FIG. 2 is a view for explaining the principle of a growth simulation model of a film structure used in the thin film polycrystalline film structure identification method according to the present invention.
FIG. 3 is a flowchart showing an example of a film structure simulation method according to the present invention.
FIG. 4 is a view for explaining an example of how to determine the growth form of initial crystal grains according to the present invention.
FIG. 5 is a view for explaining an example of a method of positioning a growth form of initial crystal grains according to the present invention.
FIG. 6 is a flowchart showing an example of a procedure for cell operation according to the present invention.
FIG. 7 is a view for explaining an example of a method for calculating physical properties of crystal grains according to the present invention.
FIG. 8 is a diagram showing an example of a calculation and measurement spectrum coincidence determination and recalculation method for simulation results according to the present invention.
FIG. 9 is a flowchart showing an example of a method for identifying a film structure of a thin film polycrystal including secondary nuclei generated during film formation according to the present invention.
FIG. 10 is a flowchart showing an example of a film structure simulation method incorporating a twinning generation model according to the present invention.
FIG. 11 is a flowchart showing an example of a twin generation procedure according to the present invention.
FIG. 12 is a diagram for explaining an example of a twin plane orientation setting method according to the present invention.
FIG. 13 is a diagram showing an example of the shape of twin virtual crystal grains according to the present invention.
FIG. 14 is a halftone image showing an example of a film structure identification result obtained in an example of the present invention.
FIG. 15 is a halftone image showing another example of the identification result of the film structure obtained in the example of the present invention.
FIG. 16 is a diagram showing a simulation result of a film structure in the prior art.
[Explanation of symbols]
11 ... Board
12, 13, 100 ... initial crystal grains
14, 15, 410, 440 ... virtual crystal grains
16 ... virtual grain increment
17 ... Actual thin film polycrystal surface
18 ... Grain boundary
19, 20 ... Actual crystal grain shape
21, 22, 23, 24, 450 ... 3D cell
130, 140 ... crystal plane growth rate
200, 300 ... Growth type
400 ... Mother crystal
420 ... crystal surface
430 ... Twins

Claims (5)

Measure the crystal analysis data for the target thin film polycrystal, and simulate the film structure of the thin film polycrystal based on the film forming conditions of the thin film polycrystal, and specify the film structure to match the crystal analysis data by adjusting the parameters, see contains a step of estimating the thin polycrystalline film structure,
Preparing a simulation space in which the thin film polycrystalline film structure simulation is divided by a plurality of cells;
In the space, arranging a plurality of initial crystal grains to which a growth shape is assigned according to a film forming condition on a substrate;
A step of enlarging the growth form of each initial crystal grain for each unit growth step in the space; and
At each step, if the cell included in the enlarged crystal grain increment is not shared with the other crystal grain increment, it is determined that the cell is a component of the crystal grain, and the other Determining that the cell is a grain boundary between the grains, if shared with the increase in grain size of
Film structure identification process of a thin film polycrystalline characterized by containing Mukoto a. For each initial crystal grain arranged on the substrate, the plane orientation of the plane parallel to the substrate is assigned according to the film forming conditions, and the crystal grain is assigned to the cell determined to be a component of the crystal grain. film structure identification method of a thin film polycrystalline according to claim 1, characterized in providing a plane orientation was. Based on the cell arrangement of the cell population to which the plane orientation is assigned, the shape of each crystal grain is obtained to calculate the film structure of the thin film polycrystal, and the plane orientation, volume, and grain size of each crystal grain are obtained to obtain the thin film polycrystal. 3. The method for identifying a film structure of a thin film polycrystal according to claim 2 , wherein crystal analysis data of the crystal is calculated. At each unit growth step, twins serving as secondary nuclei are generated according to the probability of occurrence on the growth surface of each crystal grain, and the plane orientation of the twin is determined from the plane orientation of the crystal grain where the twin is generated. claims 1 to 3 film structure identification method of a thin film polycrystalline any one of claims, characterized in that determined by the mirror image operation with a mirror surface. The film structure specified parameters be adjusted to match with the crystallography data, we claim 1, characterized in that the density of the initial crystal grains disposed on the substrate 4 to any one of claims thin polycrystalline Membrane structure identification method.

Images