JPH09145641A - Method and device for evaluating crystallinity of single crystal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quantitatively evaluate a difference between a uniform part and a specific part in crystallinity by measuring a locking curve one-dimensionally on the line where a part having a specific crystallinity such as crystal defect observed by a topograph crosses and a part having uniform crystallinity are included. SOLUTION: A first crystal 3 and a second crystal 4 (a crystal to be tested, specimen, wafer) are fitted on a scanning stage 6. The stage 6 is scanned parallely to the linear direction, and the stage 4 is rotated relatively against the stage 6 by means of a specimen rotating table 8. A film 5 is provided in a direction where a diffracted light is emitted from the crystal 4, in order to pick up a topograph thereof, while the film 5 is removed to measure a locking curve, and the intensity of diffracted X-ray is measured by acintilation counter 7. By scanning the stage 6, the positions of X ray hitting the crystals 3 and 4 varies, and a topograph including a band part is exposed to a photographic plate, so that the entire topograph can be obtained with time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a single crystal evaluation method and evaluation apparatus using X-rays. The single crystal means a crystal having a common crystal orientation in any part of the sample and having a fixed crystal orientation as a whole. It is the opposite concept of the words polycrystal and amorphous. A polycrystal is composed of a large number of crystal grains having different orientations and a grain boundary which is a boundary between them. The present invention aims to evaluate the crystallinity of a single crystal.

A single crystal may have defects.
In that case, even a single crystal cannot be used to produce a device of good quality. Here, single crystal means
Of course, it includes all semiconductors, insulators, and good conductors. However, the present invention is particularly focused on the quality evaluation of semiconductor single crystals.

The crystal quality of a semiconductor wafer greatly affects the characteristics of semiconductor devices or the processing yield. for that reason,
It is necessary to evaluate the crystal quality of the semiconductor wafer in advance. In particular, it is strongly desired to accurately grasp the distribution of crystal defects within the wafer surface and to quantitatively evaluate the crystal defects. Evaluation of crystallinity is performed by various methods. There are destructive inspection and non-destructive inspection. There is an X-ray diffraction method as an effective method for nondestructive inspection of crystallinity.

There are various types of X-ray methods. It is possible to determine whether the sample is single crystal, polycrystal, or amorphous by X-ray diffraction. The X-ray may be white (including many wavelength components) or may be monochromatic (single wavelength). Sometimes it is narrowed down to a narrow beam, and sometimes it is wide. Various X-ray diffraction methods are used depending on the purpose and object. Since X-rays are diffracted by the crystal lattice when irradiated to the crystal, the crystal structure is obtained from the direction and intensity of the diffracted X-rays. It is always Bragg's condition that gives the diffraction direction

Λ = 2 dsin θ (1)

[0006] Where λ is the wavelength of the X-rays, d is the lattice spacing, and θ is the angle between the incident X-rays and the diffracted X-rays with the lattice plane. θ is called the Bragg angle. When used for structural analysis of a polycrystalline sample, for example, monochromatic X-rays are irradiated and a powder X-ray photograph is taken. X-rays are diffracted in many directions around the sample. From the Bragg angles θ of many diffraction lines, the distance d between the lattice planes that caused the diffraction is obtained. From the combination of d, it is possible to understand what kind of crystal structure the polycrystalline sample has and what the lattice constant a is.

Alternatively, a single crystal sample is irradiated with white X-rays,
Take a two-dimensional diffraction photo. This corresponds to the reciprocal lattice of the crystal and is sometimes called Laue photograph. When a clear point cloud appears in the photograph, it can be seen that it is a single crystal. The crystal structure is known from the position of the diffraction point. The directions of the incident X-ray beam and the diffracted X-ray beam are symmetric with respect to the normal line standing on the plane that causes diffraction, and the plane spacing is fixed. Therefore, the Bragg condition is defined as not including X-rays of various wavelengths. I'm not satisfied. So in this case we use white X-rays.

These methods do not rotate the sample with respect to the x-ray beam. The incident direction of the X-ray beam on the sample is fixed. Both are extremely effective for structural analysis of a sample having an unknown structure.

The problem of the present invention is not such an X-ray structural analysis method for unknown samples. This is an X-ray device for evaluating the crystallinity when the sample is a known substance, the structure is also known, and it is a single crystal. A device using X-rays only detects Bragg diffraction λ = 2d sin θ similarly, but there are various modes. What is described above is one in which the sample crystal is directly irradiated with X-rays. This applies X-rays to only one crystal.

The subject of the invention is an X-ray device for the evaluation of the crystallinity of known single crystals. Polycrystals are not the subject of this invention. The purpose is to evaluate crystallinity rather than structural analysis.
The present invention has a known crystal structure, a single crystal with a known orientation,
Since it has a local structure and disorder of orientation, it is intended to detect this. The disorder of the structure and orientation is also called a crystal defect. This can be called a crystal evaluation because it detects the type, distribution, and number of defects. It must be distinguished from X-ray diffraction as a structural analysis method.

Among the X-ray diffraction methods, there is a method called a double crystal method as a highly accurate method. The X-rays emitted from the monochromatic X-ray source are first applied to the first crystal having good crystallinity, and the X-rays diffracted therefrom are applied to the sample crystal (referred to as the second crystal). This is called the two-crystal method because two single crystals are used.
The first crystal is not a sample. As the first crystal, a single crystal having excellent crystallinity such as silicon or germanium is used.
By once diffracting the X-ray generated from the X-ray source by the first crystal, the monochromaticity and parallelism of the X-ray are improved.

Even a monochromatic X-ray source may include X-ray components (white X-rays) other than a predetermined wavelength. Since the first crystal is an excellent single crystal, X-rays other than the predetermined wavelength are diffracted in different angle directions and do not enter the sample. Even X-rays having a predetermined wavelength may contain non-parallel components. First
When the X-ray is diffracted by the crystal, the X-ray other than the predetermined direction cannot satisfy the Bragg condition and is not diffracted.

Therefore, when diffracted by the first crystal, monochromaticity and parallelism are further improved. Of course, since there are components diffracted in other directions, the X-ray power is reduced by using the first crystal. However, even with such a loss, the benefit of using X-rays with improved monochromaticity and parallelism is greater.

Which surface of the first crystal diffracts the incident X-ray is a parameter that can be freely determined. In the Bragg equation, since the wavelength λ and the surface spacing d are determined, the diffractive surface can be selected depending on the inclination angle θ. First
When the surface plane orientation of the crystal and the diffraction plane orientation match, the angles of the incident X-ray and the diffracted X-ray with respect to the surface are the same. This is sometimes called symmetrical reflection. The surface plane orientation of the first crystal may be different from the diffraction plane orientation. This can be called asymmetric reflection.

As the X-ray source, a characteristic X-ray of metal such as CuKα1 ray is usually used. In some cases, synchrotron radiation may be used. This is because it is possible to obtain an X-ray beam having excellent parallelism, monochromaticity, and intensity.

The diffraction plane of the first crystal and the diffraction plane of the second crystal (crystal to be inspected) are selected according to the purpose. For example, when selecting the diffraction plane of the inspected crystal whose crystal surface is (100), it is preferable to use the asymmetric reflection of the (422) plane for the first crystal in order to check the crystallinity in the vicinity of the surface. Further, when it is desired to relatively investigate the state up to the inside of the crystal, it is appropriate to use symmetrical reflection such as the (400) plane. Here, the diffraction represented by the surface index will be described.

FIG. 16 shows symmetrical reflection. The symmetry means that the incident beam and the outgoing beam are symmetrical with respect to the plane. The cross-sectional area of the incident beam and the cross-sectional area of the reflected beam are the same. This is a single crystal (400) with a (100) plane
Shows diffraction. It is assumed that the crystal is a cubic crystal and the lattice constant is a. The surface spacing d of the surface having the surface index (klm) is

D = a (k ² + l ² + m ² ) ^-1/2 (2)

Is given by The Bragg angle θ can be obtained by substituting the Bragg condition.

Sin θ = (λ / 2a) (k ² + l ² + m ² ) ^1/2 (3)

Thus, the Bragg angle θ from the crystal plane having an arbitrary plane orientation can be calculated. The range of possible plane orientations is determined by the wavelength of X-rays. For example, the characteristic X of CuKα1
If a line (λ = 0.154 nm) is used and Si is used for the first crystal (a = 0.543 nm),

Sin θ = 0.1418 (k ² + l ² + m ² ) ^1/2 (4)

It becomes For (400) diffraction of (100) Si, substituting k = 4, l = 0 and m = 0,

Φ ₁ = Φ ₂ = 34.5 ° (5)

FIG. 17 shows asymmetric reflection. The (422) plane is inclined by 35.26 ° with respect to the (100) plane.
Since θ = 44.00 ° in the diffraction with respect to the (422) diffraction, the incident tilt angle θ ₁ = 8.74 °. The inclination angle Θ ₂ on the exit side is 79.26 °. Beam expansion rate is s
in Θ ₂ / sin Θ ₁ = 6.5. In other words, this asymmetrical reflection spreads the beam size 6.5 times. When Ge (a = 0.565 nm) is used as the first crystal, the same thing can be said although the angle is slightly different.

[0026]

2. Description of the Related Art There are still two methods in the two-crystal method as a method for evaluating crystallinity. One is rocking curve measurement. Another is topograph photography. It is a completely contrasting method and is used according to the purpose.

[Topographic Imaging] Topograph is an overall method and is a method of projecting a diffraction image by applying X-rays diffracted by a crystal to be inspected (second crystal) to an X-ray film, an imaging plate or the like. The entire sample is irradiated with a large area of monochromatic and uniform direction X-rays. Since the sample is a single crystal, X-rays should be diffracted in the same direction from the same crystal plane throughout the plane. A photographic plate is placed in the direction of diffraction from a particular crystal plane to expose it to diffracted X-rays. Since it is a single crystal and has a fixed orientation, the diffraction direction from a specific plane is uniquely determined, and the dry plate is placed there.

If the crystal has no defects, the diffraction X
Since the directions of the lines are the same and the intensities are also the same, X-rays of uniform intensity which spread the size of the sample should be incident on the photographic plate. This causes the dry plate to be exposed. A uniform white figure appears. This is an image of the sample. The term "image" is strange because X-rays do not form a condensing optical system in this case, but since it corresponds one-to-one to the sample spatially, what is reflected on the dry plate is called "sample image". I can do it.

The quality of incident X-rays is important in X-ray topography. Must be monochromatic and parallel. This is a strong condition. If the irradiated X-rays are not parallel, even if there is a defect in the crystal structure on the surface of the sample, the diffraction lines will be non-parallel, resulting in the result as if there were no defect. Therefore, as mentioned above, the two-crystal method is used. Although photography requires processing such as development, it is useful for photography because of its high resolution. Imaging plates have low resolution,
It has the advantage of being able to observe in real time.

Since the X-ray source emits a parallel large-area beam,
The X-ray image on the photographic dry plate is the result of exposure to parallel beams. An arbitrary point S on the dry plate has a one-to-one correspondence with a specific point T on the sample surface. The points on the sample are numbered and referred to as points S _j . Let T _{j be} the corresponding point on the plate.

It is assumed that a point S _j on the sample is defective. Here, the defect means that the crystal plane is locally inclined. Therefore, in the defect, the diffraction direction from the predetermined crystal plane is deviated. Since X-rays are not diffracted by T _j on the dry plate, the amount of X-rays incident on this portion is small. X
The part that does not hit the line becomes dark, so this part appears black. That is, when there is a local defect at the point S _j on the sample surface, a black portion appears at T _j of the photographic plate. In this way, it is possible to project a good portion in white and a defective portion in black in the entire crystal. There are several types of crystal defects. The type of defect can also be determined by topography.

This is the X-ray topography. The location of defects in the sample surface can be clarified at once. Topograph is an overall, intuitive, spatial evaluation method for single crystals.

[Rocking curve measurement] On the other hand, the rocking curve detects the intensity of X-rays diffracted by the crystal to be inspected by an X-ray counter or the like. The inspected crystal is rotated in the range of several tens to several hundreds of seconds around the diffraction angle, and the detected value of the X-ray counter is monitored. The spectrum showing the relationship between the rotation angle and the X-ray detection value is called a rocking curve. The angle that gives the peak, the X-ray intensity at the peak, and the full width at half maximum (FWHM) of the peak are the observed quantities.

The peak angle (θ) is determined by the diffractive surface. The fact that it fluctuates means that the crystal plane is tilted, which is directly evidence of crystal disorder. The peak intensity indicates the intensity of diffraction on that surface. The full width at half maximum represents the fluctuation of the angle of the diffractive surface. Therefore, the narrower the half-width is, the better the crystal plane is.

As described above, the crystallinity can be quantitatively evaluated from the parameters of the rocking curve, peak angle, peak intensity, FWHM and the like. In the rocking curve measurement, if the measurement area is wide, diffraction from all in the area is collected, so only the average property of the area is known. As for crystallinity, only the average crystallinity in that region is known. Therefore, the measurement area is often narrow and limited. This is a local measurement and evaluation method. Many measurement points consisting of a narrow area are determined, and rocking curve measurement is performed at these measurement points.

[0036]

Although the topography has an advantage that crystal defects can be evaluated with high spatial resolution, it has a problem that crystal defects cannot be quantitatively evaluated. Further, the topography has a problem that an image of the entire crystal is not recorded when the crystal under inspection is warped.

A conventional topograph photographing method is shown in FIGS.
It will be explained by. In FIG. 1, the parallel monochromatic X-ray beam emitted from the X-ray source 1 passes through the elongated slit 2 and becomes a band-shaped parallel beam. This is incident on the first crystal 3 at a small irradiation angle and is diffracted from the first crystal 3 so as to form a large angle. It is a parallel beam with a wide cross section. This strikes the inspected crystal 4 and is diffracted again. The width of X-rays is expanded by asymmetrical diffraction by the first crystal, and X is spread over the entire surface of the inspected crystal at once.
Irradiate the line. In this case, the whole image can be obtained by one shot. It takes less time to shoot. In this case, both the first crystal and the second crystal remain stationary.

In FIG. 2, the first crystal 3 is irradiated with X-rays at a large angle and is symmetrically reflected. The X-ray beam has a narrow strip-shaped cross section. The entire surface of the inspected crystal cannot be irradiated at once. Therefore, the stage 6 on which the first crystal 3, the second crystal 4 and the film 5 are entirely mounted is moved in parallel (scanning) to scan the X-ray irradiation portion of the inspected crystal, and The entire diffraction image is taken on film.

The angle of the crystal does not change in both the collective photographing (non-scanning) of FIG. 1 and the scanning photographing method of FIG. The scanning of FIG. 2 also translates the stage. That is, the angle of incidence of X-rays on the wafer (crystal to be inspected) is constant in any method. If the wafer (crystal to be inspected) has a warp, the incident angle of the X-ray with respect to the plane deviates due to the warp and the Bragg condition is not satisfied, so that there is no diffracted beam. That part gets dark. FIG. 4 shows such a topographic photograph.

The vertical strip-shaped portion in the central portion has a proper incident angle, satisfies the Bragg condition, and diffracted X-rays are generated, so that the image is shown in white. Bragg diffraction does not occur on both sides and it is black because of warpage. This is not due to a defect. Since the area is large, it can be seen that it is due to warpage, not due to defects. However, the distribution of defects in the darkened area is completely unknown.

Jenichen et al. Have proposed a device that enables topography imaging even if the crystal to be inspected has a warp. B. Jenichen, R. Kohler and W. Mohling, "RT
K 2-a double-crystal x-ray topographic camera ap
plying new principles ", J. Phys. E: Sci. Instrum.
21 (1988) 1062-1066.

Jenichen points out the problem that the X-ray topograph is an excellent method that can reveal all the crystal defects in the wafer surface at once, but is not routinely used as an inspection method. . This is because the wafer is distorted, so only a part of it can be photographed by the topography and the defect inspection of the entire surface cannot be performed. It is shown in FIG. X-rays hit the flat first crystal 12 from the X-ray source 11 and become a parallel large-area beam. This is the second warp
Hit crystal 13. For the crystal 13 to be inspected which is distorted in a convex shape, only the narrow band-shaped portion 15 in the central portion satisfies the Bragg condition.

On both sides 16, 17 the X-ray is not diffracted because the Bragg condition is not satisfied. In order to solve this, Jenichen used a curved first crystal 14 as shown in FIG.
Is used. The diffracted light is not a parallel beam but a slightly converging beam. This beam has an incident angle of θ _B everywhere in the second crystal. Therefore, X-rays are diffracted on the entire surface of the second crystal.

According to Jenichen's method, the first crystal is curved so that the incident angle with respect to the second crystal is always at a predetermined angle. It's a very clever method. In this method, the deflection of the first crystal must exactly correspond to the strain of the crystal to be inspected. Therefore, the first crystal is attached to a thin base material, and the base material is bent to distort the first crystal. FIG. 15 shows this. Bars L, L are provided at both ends of the base material CU, and the first crystal is distorted by pushing the bars in and out and pushing them in.

However, the strain of the inspected crystal is diverse,
It is not easy to prepare a strained first crystal that matches this. Jenichen should mechanically bend the first crystal
However, the bending mode obtained by holding both ends of the elastic body to distort it is extremely simple and does not always match the bending of the crystal to be inspected. Moreover, the degree of bending is limited. It cannot be applied to inspected crystals with complicated strain.

Even if the X-rays are incident on the surface of the first crystal barely even if the distortion is simple, even if the first crystal is distorted in a round shape, the distribution of the bending angle of the reflected X-rays becomes non-linear. Even if the second crystal having linear distortion is irradiated, the distortion cannot be corrected. Further, as shown in FIG. 14, since the beam is contracted or expanded due to the curvature of the first crystal,
In order to irradiate the beam to the entire first crystal, it is necessary to oversize the first crystal.

Topographic imaging must be applicable to wafers (samples) with any warp. At present, topographs cannot be applied to warped samples. What has been described above is a drawback of the topographic method.

Next, the difficulties of the current rocking curve method will be described. This is to measure the diffraction intensity, FWHM, and diffraction angle by applying a narrow beam to many measurement points. If one measurement area is made too narrow (the beam is made thin), crystal measurement may be skipped and mapping measurement may be performed. To avoid missing defects, increase the number of measurement points. However, if the measurement interval is too narrow, the measurement will take too long. Conventionally, since only the rocking curve was measured, the position where defects exist was not known in advance, and measurement points were measured in the vertical and horizontal directions on the entire wafer to observe the same density at all points.

However, it is not necessary to measure the rocking curve in the portion where no defect exists. What is needed is a rocking curve near the defect. It is best to make the measurement point density high only near the defect. However, since in advance where the defect do not know what is distributed, it can not be such a measure.

Further, only by analyzing numerical data such as peak full width at half maximum (FWHM), peak angle and peak intensity,
It is difficult to determine whether the numerical abnormality is caused by a crystal defect or by something unrelated to the crystal defect (for example, warpage of the crystal). As described above, the conventional rocking curve measurement has problems that the measurement time cannot be shortened by previously distributing the measurement points only to the site where the measurement is necessary, and it cannot be determined whether the measurement is a defect or a warp.

[0051]

[Means for Solving the Problems] As described above, the overall evaluation by the topography and the quantitative evaluation by the rocking curve have problems. Therefore, the present inventor has devised a method for solving these problems and quantitatively evaluating crystallinity with high accuracy. It does not make much sense to examine the uniformity of crystallinity over the entire surface. What is important is to quantitatively understand the crystallographic difference between a uniform crystal part and a peculiar part (non-uniform part).

Therefore, the present inventor has devised a crystallinity evaluation method which combines a topography and a rocking curve and can be applied to a warped crystal. According to the present invention, in the same X-ray measuring apparatus, the sample is held as it is, and the topography and rocking curve measurement are continuously performed. That is, the present invention has two features. One is a topograph that can be applied to samples with warpage. The other is the integration of the topograph and the rocking curve.

First, the position of the crystal defect is specified by topographic photography applicable to a warped crystal. The rocking curve measurement location is determined based on the topographic image. Then, a one-dimensional rocking curve mapping measurement is performed on a straight line that includes a portion that has been confirmed to have uniform crystallinity and that crosses a peculiar portion such as a crystal defect observed by a topography. By doing so, it is possible to quantitatively and accurately evaluate the difference between the part having uniform crystallinity and the part having peculiarity.

[0054]

BEST MODE FOR CARRYING OUT THE INVENTION At this time, the X-ray irradiation area is determined based on the size of the unique portion observed by the topography. If the irradiation area is made smaller than the size of the peculiar portion, the difference between the peculiar portion and the uniform portion can be evaluated more accurately than in the case where it is not. Further, the interval between the measurement positions of the mapping measurement in the vicinity of the singular point is set to be the length of the X-ray irradiation range or less. This allows the singular points to be reliably captured and evaluated without jumping over the singular points.

Next, a case will be described in which crystal defects are linearly extended in the topography, and a portion in which the contrast of the image changes at the defect line is observed. Conventionally, in such a region where the light and shade changes discontinuously, two topographs are taken by changing the incident angle of X-rays until the light and shade of the topograph reverse to those of the first one.
The deviation of the diffraction angles of the two regions was calculated. However, this is not a quantitative evaluation because it is based on a qualitative judgment that the shades will be reversed.

According to the present invention, two lines having different shades crossing the singular line are used.
The one-dimensional distribution measurement of the rocking curve is performed on one straight line extending over two areas. In this method, the edge of the singular line can be measured so as not to include the singular line. It becomes possible to quantitatively and accurately evaluate the discontinuity of light and shade on the singular line.

Even if the measurement position of the rocking curve is determined based on the image obtained by the topography, there is no point unless the rocking curve can be measured by irradiating an actually aimed place with X-rays. According to the present invention, X-rays can be accurately applied to the intended part. To this end, the invention does this for the topography and the rocking curve in succession in the same device.
Moreover, the crystal arrangement remains the same, and after the topograph is taken, the rocking curve mapping measurement is performed at that position.

Further, according to the following procedure, the rocking curve measurement position and the position of the crystal defect seen in the topographic image can be accurately matched. 1. The whole image of the crystal is photographed on a piece of film using a linear slit and topography (while correcting the warp).

2. Change the slit from linear to a pinhole slit. And the X-ray reflected from the crystal to be inspected in a pinhole shape on the film that photographed the whole image above
Doubly expose the line. At this time, in order to know the relative positional relationship between the pinhole and the crystal to be inspected, at least two points (X1, H1) at which the mechanical wafer moving position (X) and the pinhole height (H) are changed, The X-ray is exposed with (X2, H2). 4

3. The relative positional relationship between the pinhole and the crystal is calculated from the coordinate data of (X1, H1) and (X2, H2) and the positional relationship between these two points on the film and the image of the entire crystal.

This makes it easy to spatially associate the topography with the rocking curve measurement. If the overall defect distribution is known from the topography, rocking curve measurement can be performed at the defect position to further clarify the defect state. The wholeness of the topography and the locality of the rocking curve work in a complementary manner, so that the wafer can be evaluated more quickly and accurately.

Another feature of the present invention is that topography can be performed on a sample having a warp. As already mentioned, the conventional topograph is effective only for flat samples, and the one with a warp is
As shown in FIG. 4, only the image of the portion satisfying the Bragg condition can be obtained. If the whole picture is not determined by topography, there is a risk of missing crystal defects. There is no point in connecting the topograph and the rocking curve.

Topographic imaging effective for a warped sample is indispensable for the present invention. Therefore, the present invention changes the conventional idea of fixing the X-ray incident angle so that the entire topography can be photographed by changing the incident angle using a device capable of rotating the sample with respect to the stage as shown in FIG. did.

In FIG. 3, the first crystal 3 and the second crystal 4
The (crystal to be inspected, sample, wafer) is attached to the stage 6. The film 5 is provided in the direction in which the diffracted light of the second crystal is emitted, and the topograph is imaged. In the case of rocking curve measurement, the film is removed and the diffracted X-ray intensity is measured by the scintillation counter 7. The stage 6 can be translated (scanned) in a linear direction. By scanning the stage, the positions where X-rays hit the first crystal and the second crystal change, and the topography of the strip-shaped portion can be exposed on the photographic plate. Find the whole topographic graph over time. Up to this point, it is the same as the scanning topography of FIG.

However, in FIG. 3, the sample 4 can be rotated separately from other elements. For this purpose, a sample rotating table 8 is provided, and by rotating this, the second crystal can be rotated relative to the stage.

This is important. If the wafer has a warp, the sample is rotated by the warp angle so that the incident angle ω is always the same with respect to the surface. Then, since the Bragg diffraction condition is satisfied on the crystal surface, the diffracted light hits the film and an image of that portion can be formed. In addition, the sample turntable 8
Enables rocking curve measurements. If the sample cannot be rotated, the rocking curve cannot be obtained. In the present invention, since the sample is rotatable, a rocking curve can be obtained.

In topographic imaging, scanning of the stage and rotation of the sample turntable are performed simultaneously. When the stage is scanned, the portion of the second crystal that the X-ray strikes moves. The incident direction of X-rays must be appropriate for this part. That is, topograph imaging is performed while finely adjusting the angle of the crystal to be inspected so that the band-shaped area being imaged is instantaneously and instantaneously at the correct angle with respect to the incident X-ray.

Here, in order to realize a fine angle adjustment corresponding to the warp of the crystal to be inspected, it is preferable that the width of the X-ray irradiated to the crystal to be inspected is narrow. When asymmetric reflection is used for diffraction of the first crystal, the width of X-rays is widened. In this case, the width of the X-ray can be narrowed by providing a slit between the first crystal and the second crystal. However, a new slit makes the device more complicated. This is not very desirable. On the other hand, if symmetrical reflection is used, the width of the X-ray will not be widened. You can get a narrow beam as it is. In this case, it is not necessary to provide a slit.

Needless to say, the first crystal must have good crystallinity. 1st for these reasons
As the crystal, a single crystal with good crystallinity such as silicon or germanium is preferably used. Of course, the selection of the first crystal also depends on the wavelength of X-rays. CuKα1 (0.154n
In the case of m), Si or Ge can be used. When the beam needs to be made thin, the plane orientation of the first crystal of Si or Ge is (100), and the diffraction plane is (400) symmetric reflection.

Since the existence and position of crystal defects can be known by topography, the rocking curve of the portion containing crystal defects is measured. The rocking curve measurement requires rotating the sample with respect to the X-ray. The device of FIG.
Since the sample rotating table 8 is provided so that only the sample can be rotated, it can be used as it is for measuring the rocking curve. In other words, making the sample rotatable has two meanings. One is to perform the rocking curve with this same device. As shown in FIG. 2, the rocking curve cannot be measured when the sample is not rotating. The other is to rotate the sample relatively and to correct the warp, and to realize topographic imaging of the entire surface.

[0071]

EXAMPLE FIG. 3 shows a block diagram of an apparatus used for evaluation. In this device, not only the second crystal but also the first crystal is mounted on the scanning stage. For the first crystal, a 4-inch φ Si wafer having a thickness of 3 mm and having a (100) plane orientation was used.

[Example 1 (quantitative evaluation of a portion having peculiar crystallinity)] The evaluation in the case where there are isolated defects (singular points) will be described. Using the apparatus shown in FIG. 3, a 2-inch φS (sulfur) -doped InP (100) wafer grown by the LEC method was evaluated. The entire topography was performed while scanning the stage in front of the X-ray source through the elongated slit. FIG. 8 shows the photographing results of the two-crystal topograph. X
Since the line film was set diagonally with respect to the crystal to be inspected, the image is an elliptical image that is vertically distorted. A unique circular contrast is seen in the center of the wafer. Diameter is 2m
It was on the order of m to 3 mm.

Therefore, the peculiar part near the center was quantitatively evaluated by the rocking curve. As a pre-stage of the rocking curve measurement, first, the topographic image of the whole image was photographed again. Slit 0.5mmφ
It was replaced with a pinhole. The position along the scanning line of the scanning stage is X, and the height of the pinhole slit is H. The relative position of the beam with respect to the sample can be defined by the two-dimensional coordinates (X, H) in which the scanning line-shaped position X and the height H of the slit are combined.

When the small-diameter beam was at (X1, H1), the beam was applied to one point of the InP sample, and the diffracted beam was superimposed on the dry plate and exposed. This point is 1. This is indicated by 1 in the topographic photograph of FIG. Similarly, when the small-diameter beam was at (X2, H2), the beam was applied to the sample, and the diffracted beam was superimposed and exposed on the same dry plate. This creates an exposure point in the portion indicated by 2 in FIG.

Points 1 and 2 on the sample correspond to the positions (X1, H1) and (X2, H2) of the stage and the pinhole. 2
Since the positions of the points are determined, the positions on the wafer correspond to the positions of the stage and the pinholes in a one-to-one correspondence.

Next, based on this data, the height was adjusted so that the pinhole would pass through the center of the wafer. The detector is a scintillation counter. Then, the rocking curve was mapped while moving the scanning stage little by little on a straight line including a portion where the crystallinity can be judged to be normal in the topographic image across the center of the abnormal circular cotrust in the central portion. Further, the same rocking curve was measured on this straight line and a parallel line 3 mm apart to the left and right.

At this time, since the diameter of the circular cotrust observed in the topography was 2 mm to 3 mm, the measurement area of the rocking curve was 0.5 mm, which was smaller than that.
The area was mm. The distance between the measurement points was 0.2 mm near the center of the wafer. FIG. 9 shows the result of the rocking curve along the left straight line. FIG. 10 shows the measurement result of the rocking curve along the straight line passing through the center. FIG. 11 shows the results of a rocking curve along the right straight line. The horizontal axis represents the diffraction angle and the vertical axis represents the diffraction intensity. The rocking curves in FIGS. 9 and 11 all have the same shape, and the FWHM is about 1
5 seconds. The narrow FWHM means that it is a good quality crystal.

However, a slight abnormality is recognized in the rocking curve of FIG. 10 along the center line. The peak position is slightly off. This means that the (400) plane is tilted in this part. The amount of change in peak angle is only 1.
5 seconds. In other words, the (400) plane is 1.5 in this part.
It is tilted by the second.

As described above, the present invention can accurately find a small abnormality in an extremely narrow area. It was possible to quantitatively evaluate the slightest abnormality, not only the abnormal portion but also the linear and detailed rocking curve measurement including the portion with normal crystallinity to pursue the difference between the normal portion and the abnormal portion. This is because the evaluation was carried out based on the idea of For comparison, FIGS. 9 and 11 show rocking curves measured along a straight line that does not pass through the center. No abnormal change in rocking curve was observed in these.

[Embodiment 2 (Evaluation of discontinuity of topograph density on a singular line)] In a topographic image, crystal defects may extend linearly and discontinuity of density may occur at the defect line. . Next, the evaluation of such a linear defect according to the present invention will be described.

The sample (crystal to be inspected) is a 3 inch φ non-doped GaAs (100) wafer grown by the LEC method. The diffraction surface of the first crystal (Si) is set to (400), and the sample G
The diffraction surface of aAs was set to (422).

While scanning the stage, the GaAs warp was corrected by changing the incident angle (ω) of X-rays on GaAs in conjunction with the scanning. That is, the rotation angle ψ of the sample turntable is changed as a function of the scanning amount x. Microscopically, the incident angle Θ of the X-ray with respect to the crystal surface is constant, but since there is a warp angle θ, the incident angle ω with respect to the base line of the sample is changed so that ω = Θ + θ. The warp angle θ is a function of the scanning amount x. That is, the wafer is attached to the sample rotary table so that the warp direction is parallel to the scanning direction. Then, it can be set to θ (x).

When the scanning amount is x, the sample is rotated, and the rotation angle ψ of the sample is determined in the direction in which diffraction occurs. This is ψ =
c + θ (c is a constant). In this way, topographical imaging of the entire GaAs wafer can be performed. In the present invention, the correction of the warp is simple, and it is also effective for the warp of complicated modes. This skillfully utilizes the structural feature of the present invention that the sample is minutely rotated even in topographic imaging. The rotatability of the sample is indispensable for the rocking curve, but the problem of warpage is vividly solved by bringing this into the topograph.

Subsequently, with the film on which the whole image was taken set on the sample rotary table, the slit was 0.5 mm.
It was replaced with a φ pinhole. The height (H) of the pinhole remains constant, and at two points (X1, H) and (X2, H) at different positions (X) of the scanning stage,
Dotted X-ray beam diffraction images were double exposed to the same film.

The results are shown in FIG. The points marked with 1 and 2 are the points exposed by X-rays passing through the pinholes.
There is a black part in the upper left of the sample. This linear part is abnormal. I would like to investigate this part by the rocking curve. There is an abnormal area above the line segment 12. Line segment 1
Examining the crystallinity along line 2 reveals the details of the anomaly.

Next, a 0.5 mmφ pinhole slit was used as it was, and a scintillation counter was installed instead of the X-ray film. Then, while the scanning stage was gradually moved, many measurement points were taken on the line segment 12 and the rocking curve at each point was measured. Then, one-dimensional mapping of the rocking curve on the line segment 12 was performed.

The results are shown in Table 7. The horizontal axis is the position coordinates of the measurement points taken between the line segments 12. The vertical axis represents the deviation of the peak angle (second). Since the diffraction by the (422) plane is measured, if this plane is strictly parallel, the peak angle will be a constant angle. However, since the (422) plane is slightly displaced, the peak angle slightly changes. This result shows that there is a jump of 24 seconds on the way. This corresponds to the defect line, and means that the (422) plane is inclined by 24 seconds before and after the defect line.

This measurement can measure the local crystal defects quantitatively only by measuring the line segment 12. Since the topograph is taken at the beginning, it is possible to omit the rocking curve measurement of the part where there is no defect, and the time required for the rocking curve measurement can be greatly shortened. Nevertheless, the evaluation of crystal defects can be performed completely.

[Embodiment 3 (Method for photographing topography of entire surface even if crystal has warp)] L was measured using the apparatus shown in FIG.
3 inch φ undoped Ga grown by EC method
An As (100) wafer was evaluated.

For comparison, first, topographic imaging of the entire surface with a wide X-ray beam as shown in FIG. 1 was performed. The X-ray beam is broadened by using the (422) asymmetric reflection surface, and then the sample is irradiated. The sample GaAs itself also uses (422) diffraction. FIG. 4 shows a topographic photograph. Since the GaAs wafer is warped, the entire image is not visible. Only the narrow strip in the center is visible.

For comparison, the topography of the same GaAs wafer was obtained using a scanning type topography device as shown in FIG. The first crystal forms a narrow beam using (400) symmetrical diffraction and irradiates GaAs. GaAs is diffracted by (422). GaAs does not rotate and the incident angle (ω) is constant. The result was similar to that of FIG. Only the central strip does not show the whole image.

Next, the same GaAs sample was topographically photographed by the method of the present invention in which the sample was rotated. First
The diffraction surface of crystalline Si (100) is (400), and the diffraction surface of the sample is (422). The stage is scanned, and the sample turntable is rotated little by little in conjunction with this. This corrected the incident angle of the X-ray entering the GaAs crystal. The scanning stage was reciprocated at a constant speed in a section where the entire surface of the crystal could be photographed.

FIG. 12 shows how the moving position (X) of the scanning stage and the incident angle (ω) of the X-ray change when the image is actually taken. Take three points along the scan line. X1 is the left end of the crystal, X2 is the central part of the crystal, and X3 is the right end of the crystal. At these three points, the incident angle of the X-ray is adjusted in advance. As a result, the optimum incident angle ω at these three points
1, ω2, ω3 are obtained. The optimum incident angle here means that the (422) diffraction intensity of GaAs is maximized.

When X2 is 0 °, X3 is +22 seconds, X
In No. 1, it is changing to -43 seconds. Then, for the three intermediate test points, the incident angle was determined by interpolation. This graph gives the incident angle ω as a function of the scanning distance X, but since ω and the rotation angle ψ of the sample turntable are the same, ψ = c + ω (c is a constant) ψ can be obtained as a function of X, ψ (X).

As described above, the topographic image of the entire sample can be obtained by the photographing method in which the sample is rotated with the scanning. FIG. 5 is a topographic image taken by the method of the present invention. By adjusting the angle at three points, the warp of the wafer was corrected and the topograph of the entire crystal could be obtained.

If the warp condition is more complicated, the angle adjustment is performed at more points along the scanning line. The optimum angles may be obtained at 4 points or 5 points, and the incident angle between them may be determined by a straight line connecting the angles at the adjacent adjustment points, or may be given by a curve determined by the least square method. Even if the sample has a complicated warp, the topography of the entire sample can be obtained by changing the X-ray incident angle in association with scanning.

Since the state of the entire defect can be seen from the topography, the rocking curve is measured along the line segment including the region where the defect exists. This point is the same as the above-mentioned embodiment.

[0098]

INDUSTRIAL APPLICABILITY The present invention provides a method for quantitatively evaluating abnormalities of crystallinity represented by crystal defects with high accuracy by using a two-crystal X-ray diffraction method. Since the rocking curve is measured after the area such as a defect that needs to be quantitatively evaluated is obtained in advance by the topography, the evaluation efficiency is high.

It is possible to easily and accurately make one-to-one correspondence between the position to be quantitatively evaluated such as a crystal defect observed by the topography and the measurement position of the rocking curve. Therefore, the quantitative evaluation can be performed with high accuracy, and no omission of the rocking curve measurement point occurs. Further, although the whole image of the warped sample could not be photographed by the topography in the related art, the present invention improves the topograph imaging method and can grasp the whole image of the complex warped crystal by the topography.

When a crystal is evaluated using a two-crystal topography, the in-plane distribution of crystal defects and lattice strain can be easily known in a crystal curved due to warpage or a crystal having a tilt in the lattice plane. In addition, the position of the crystal defect or lattice strain observed with the two-crystal topograph and the measurement position of the rocking curve are set to 1: 1.
Can be dealt with. The crystal defects and lattice strains captured as images in the two-crystal topography can be easily quantitatively evaluated by using rocking curves.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing a non-scanning topographic imaging system according to a conventional example.

FIG. 2 is a schematic configuration diagram showing a scanning topograph imaging system according to a conventional example.

[Fig. 3] Topography imaging by rotating the sample with scanning,
The schematic block diagram of the X-ray evaluation apparatus of this invention which made it possible to measure a rocking curve with the same apparatus.

FIG. 4 is a topography photograph of a warped GaAs wafer taken by the non-scanning topography photography shown in FIG.

5 is a photograph of a topography of a warped GaAs wafer by the evaluation apparatus of the present invention shown in FIG.

FIG. 6 is a topographic photograph for explaining that a GaAs wafer is topographically photographed by the method of the present invention, two images of two points are double-exposed by a pinhole slit, and the positions of the slit and the topographic photograph are associated with each other.

7 is a line segment 1 in the GaAs wafer-sample of FIG.
The graph which shows the fluctuation | variation by the position of the peak angle of a diffraction angle by performing the rocking curve measurement along 2. The peak angle changes discontinuously for 24 seconds before and after the defect.

FIG. 8 is a topographical photograph of a (100) InP wafer according to the method of the present invention. There is a small circular defect in the center. A diameter Q passing through the central portion, a string P parallel to this 3 mm to the left, and a string R parallel to this to the right of the diameter Q.

FIG. 9 is 3 to the left from the straight line passing through the center in the photograph of FIG.
The graph which shows the rocking curve measurement result in the measurement point along the chord P which separated mm. The horizontal axis represents the rotation angle of the crystal to be inspected, and the vertical axis represents the intensity of the diffracted X-ray.

10 is a graph showing rocking curve measurement results at measurement points along a straight line Q passing through the center of the photograph of FIG. The horizontal axis represents the rotation angle of the crystal to be inspected, and the vertical axis represents the intensity of the diffracted X-ray. The peak position is off by about 1.5 seconds at the measurement point in the center.

FIG. 11 is a graph showing a rocking curve measurement result at a measurement point along the chord R 3 mm to the right from the straight line passing through the center in the photograph of FIG. The horizontal axis represents the rotation angle of the crystal to be inspected, and the vertical axis represents the intensity of the diffracted X-ray.

FIG. 12 is a graph showing the rotation angle of the sample rotated with scanning to take the topograph photograph of FIG.
Optimal rotation angles are obtained at the three points X1, X2, and X3, and at other points, the rotation angles are determined by interpolation.

FIG. 13 is an explanatory view showing that when Jenichen et al. Has a warp in a crystal to be inspected (second crystal), X-rays are only partially diffracted and only a partial topographic image is obtained.

FIG. 14 is an explanatory view for showing Jenichen's proposal that it is possible to take a general topographical photograph of a warped sample by using a curved first crystal.

FIG. 15 is an explanatory view of an apparatus for bending the first crystal proposed by Jenichen.

FIG. 16 shows (40) in a single crystal having a (100) plane.
0) Schematic diagram for explaining symmetrical reflection.

FIG. 17 shows (42) in a single crystal having a (100) plane.
2) Schematic diagram for explaining asymmetric reflection.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 X-ray source 2 Slit 3 1st crystal 4 Sample (2nd crystal, to-be-tested crystal) 5 Film 6 Scanning stage 7 Scintillation counter 8 Sample rotary stage 11 X-ray source 12 1st crystal 13 2nd crystal 14 Curved 1st Crystal 15 Central portion where X-ray diffraction occurs 16 Right side portion where Bragg conditions are not satisfied and X-ray diffraction does not occur 17 Left side portion where X-ray diffraction does not occur and Bragg conditions are not satisfied

Claims (10)

[Claims] 1. In the evaluation of a single crystal using the X-ray two-crystal method, the crystallinity of the entire sample is evaluated by topography, and the crystallinity which is observed in the topograph such as a crystal defect is crossed and crystallized. The crystallinity of a single crystal characterized by performing a one-dimensional measurement of a rocking curve on a line that includes a portion with uniform crystallinity and quantitatively evaluating the difference between the portion with uniform crystallinity and the specific portion. Evaluation method. 2. When performing topography by using the X-ray two-crystal method, the sample is rotated according to the warpage of the sample crystal as the sample and the first crystal are scanned, and the incident angle of the X-ray with respect to the sample is changed. The method for evaluating crystallinity of a single crystal according to claim 1, wherein the topography is adjusted so that the angle of the X-ray with respect to a predetermined diffraction surface is constant. 3. When performing a one-dimensional mapping measurement of a rocking curve on a line that crosses a peculiar part of a crystal such as a crystal defect observed in a topograph and includes a part having a uniform crystallinity, X for rocking curve measurement is measured. The method for evaluating the crystallinity of a single crystal according to claim 1 or 2, wherein the irradiation region of the line is determined based on the size of the peculiar portion observed by the topography. 4. When performing a one-dimensional mapping measurement of a rocking curve on a line that crosses a peculiar part of a crystal such as a crystal defect observed by a topography and includes a part with uniform crystallinity, mapping in the vicinity of the peculiar part is performed. The crystallinity evaluation method for a single crystal according to any one of claims 1 to 3, wherein the interval between the measurement points of the measurement is set to be equal to or less than the length of the X-ray irradiation range. 5. When the rocking curve measurement is performed on a straight line that crosses the singular line and includes a portion with uniform crystallinity, when the crystalline discontinuity observed on the topography is divided by a line, the singular line 5. The difference is quantitatively evaluated by measuring the vicinity of the two areas partitioned by the line so as not to include a singular line.
The method for evaluating crystallinity of a single crystal according to any one of 1. 6. A peculiar portion such as a crystal defect is exposed by superimposing and exposing a topograph for recording a part of a sample crystal and a topograph for recording an entire image of a sample by controlling an X-ray irradiation position. The method for evaluating crystallinity of a single crystal according to claim 1, wherein the method for evaluating single crystallinity is performed with high accuracy. 7. An apparatus for evaluating crystallinity of a single crystal, characterized in that the topography of a single crystal and the rocking curve measurement, which are carried out by using the X-ray two-crystal method, are carried out in the same apparatus with the same crystal arrangement. 8. The crystallinity evaluation apparatus for a single crystal according to claim 7, wherein the topographic recording medium is an X-ray film, a nuclear dry plate or an imaging plate. 9. The silicon or germanium having a plane orientation (100) is used as the first crystal.
Alternatively, the single crystal crystallinity evaluation device according to item 8. 10. The single crystal crystallinity evaluation apparatus according to claim 7, wherein a (400) plane is used as a diffraction plane of the first crystal.