JPH10132714A - Silicon wafer for evaluating lattice strain, production method thereof, and evaluation method for lattice strain - Google Patents

Abstract

PROBLEM TO BE SOLVED: To evaluate lattice strain in a thin film sample while taking various crystallographic effect into account by provided a wedge-shaped part where the thickness in the direction vertical to the major surface of an Si water for evaluating lattice strain varies continuously in a rage lower than a specified value thereby controlling the thickness at a part to be measured sufficiently. SOLUTION: The Si wafer for evaluation is machined to have a wedge-shaped part using FIB. In the FIB method, sputtering is performed using a focused Ga ion beam while observing an SIM image. The thickness the direction vertical to the major surface of the Si wafer is varied continuously in the rage from 10nm to 10μm. When the thickness exceeds 10μm, an electron beam does not transmit through the wedge-shaped part and when the thickness is less than 10nm, strain is relaxed excessively to cause difficulties in the measurement. Lattice strain is analyzed by observing variation in the HOLZ pattern obtained by CBED method. According to the method, lattice strain due to variation of film thickness can be measured accurately.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to Si for lattice strain evaluation.
The present invention relates to a wafer, a method for manufacturing the same, and a method for evaluating lattice distortion, and more particularly, to a Si wafer having a part where the film thickness changes continuously for evaluating lattice distortion, a method for manufacturing the same, and evaluation of lattice distortion using the part. About the method.

[0002]

2. Description of the Related Art In recent years, with the advance of electron microscope technology, electron microscopes are frequently used for research and development of materials. In observation using an electron microscope, sample conditions such as film thickness and surface properties are closely related to scattering and imaging principles. Therefore, the sample preparation technology is a transmission electron microscope (Transmitted Electron Microsc
ope: TEM) is attracting attention as one of related technologies. When a metal material, particularly iron, was observed, an electrolytic polishing method has been used as a sample slicing method. However, with the progress of the electronics field, if the observation target material covers electronics-related materials, that is, semiconductor materials, multilayer films, and electronic ceramics, the observation method differs depending on the observation target and observation purpose, and the sample is required accordingly. Conditions are different. Therefore, a new thin film forming technology corresponding to these new materials is required. For example, when observing a tissue, generally, a low-magnification observation is performed, so that a thin sample must be prepared in a wide area.
When observing a particle image, it is necessary to prepare a sample as thin as possible to remove the effect of inelastic scattering as much as possible. Among TEM observations, Convergent Beam Electron Diffraction (CBED)
In the method, since it is necessary to obtain a clear HOLZ (High Order Laue Zone) line, a sample slightly thicker than a normal electron microscope observation sample is preferable. As described above, the preferable thickness of the sample varies depending on the observation method. When observing precipitates, interfaces, defects, and the like, they must be included in the thinned region. In particular, with the densification of electronic materials, the necessity of selective polishing (selective thinning) has emerged. However, local and selective thinning is not possible with conventional polishing methods, and we have seen what we see rather than what we want to see.

[0003] The optimum polishing method for a material differs depending on the target material. For example, a material such as bulk Si is preferably manufactured by a chemical polishing method. Even if the material has a low electrical conductivity like Si and has a cleavage, it is better to prepare the material by a cleavage method. In addition, when the material is steel or the like, it may be produced by an extraction replica method or the like and used for observation of fine precipitates. However, a material such as a multi-layered film, which is complicatedly composed of various types of atoms, has a different sputter rate depending on the constituent materials of the film, and it is difficult to achieve a uniform thin film.

Other techniques for thinning a sample include an ion polishing method and a microtome method, and various manufacturing techniques have been applied to various materials. However, it is impossible to process a sample into an arbitrary shape by using these sample manufacturing methods because a thin film cannot be formed at a specific portion in a micrometer region. In addition, it is difficult to experimentally estimate the lattice strain in the bulk state by the conventional method.

[0005]

TEM observation is suitable for measuring local strain at a specific location in a Si wafer. For example, an S wafer grown by using the Czochralski method is manufactured.
The CBED method was used to analyze the strain near the oxygen precipitate in the i-wafer, and quantitative analysis was attempted (Electron Microscopy. Volume 2. EUREM 92, Grana
da, Spain, (1992) Okuyama T. etc., p.157-158). However, the sample is manufactured using the conventional ion polishing method, and the thickness of the portion to be measured cannot be sufficiently controlled. It was difficult to consider the biological effects.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and the thickness of a portion to be measured is sufficiently controlled, and various crystallographic effects in a thin film sample such as the release of distortion due to thinning and the influence of the film thickness are eliminated. S for lattice distortion evaluation that can be considered
It is an object of the present invention to provide an i-wafer, a method for manufacturing the same, and a method for evaluating lattice distortion.

[0007]

In order to achieve the above object, a Si wafer for evaluating lattice strain according to the present invention is a Si wafer for evaluating lattice strain, and has a thickness in a direction perpendicular to a principal surface of the wafer. Is 10 nm or more and 10 μm
It is characterized by having a wedge-shaped portion that changes continuously in the following range.

According to the above-mentioned lattice distortion evaluation Si wafer, since the cross-sectional wedge-shaped portion having a continuously changing thickness is provided, it is possible to measure the change in the strain due to the change in the film thickness. It is possible to estimate the strain in the bulk state in consideration of the release of the strain due to the influence of the thickness. If the thickness of the sample exceeds 10 μm, it becomes difficult for the electron beam to penetrate the sample, making it impossible to measure the strain. On the other hand, if the thickness of the sample is less than 10 nm, the release of the strain due to thinning becomes too large, and It becomes difficult to measure a large distortion.

Further, a method of manufacturing a lattice distortion evaluation Si wafer according to the present invention is characterized in that a focused ion beam is used for forming the wedge-shaped portion in the lattice distortion evaluation Si wafer.

[0010] Focused ion beam (Focused
If Ion Beam (FIB) is used, the wedge-shaped part
It can be easily formed in i-wafer. In the FIB method, the ion beam can be narrowed down to several nanometers, and sputtering can be performed while observing the processing status with a scanning ion microscope (SIM) image. Can be subjected to selective thin film processing.

Further, the method for evaluating lattice distortion according to the present invention is characterized in that a focused electron beam diffraction method is applied to a wedge-shaped portion in the Si wafer for lattice distortion evaluation to detect lattice distortion in the Si wafer. I have.

[0012] For TEM observation, it is necessary to thin the sample, and the strain existing in the thin film sample is released to some extent by the thinning as compared with the strain existing in the bulk state.

In the conventional sample preparation method, the sample is thinned almost uniformly, so that it is difficult to consider the effect of releasing strain. According to the evaluation method of the lattice strain, a wedge-shaped portion having a continuous thickness change is prepared in a sample, and observed by a CBED method.
By measuring a change in strain due to a change in film thickness, it is possible to estimate strain in a bulk state in consideration of the release of strain and the effect of thickness.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of a lattice distortion evaluation Si wafer, a method for manufacturing the same, and a lattice distortion evaluation method according to the present invention will be described below with reference to the drawings.

The strain evaluation Si wafer according to the embodiment is processed to have a wedge-shaped portion using FIB. In the FIB method, Ga focused while observing with a SIM image is used.
Sputtering is performed using an ion beam or the like. In addition, it is possible to process the thin film into various forms by inclining the sample stage, and the inclination angle of the wedge-shaped portion is set by the inclination of the sample stage.

Normally, FIB processing is performed in three stages: rough cutting, intermediate processing, and finishing processing. Table 1 below summarizes the preferable beam current value, sputtering area, and beam diameter range in each stage.

[0017]

[Table 1]

Further, the processing by FIB for a long time may cause a shift of a processing position due to vibration, a shift of a focus due to instability of a beam, and the like, and it is desirable to finish the finishing processing within 30 minutes. . Also,
For a sample that is easily damaged or a sample that requires a resolution of about 0.2 nm, it is desirable to perform sputtering (cleaning) with low-energy Ar ions using an ion milling device after processing by FIB.

If the range of the change in the thickness in the direction perpendicular to the main surface of the Si wafer exceeds 10 μm, the electron beam will not pass through the wedge-shaped portion, while if the range of the change is less than 10 nm, the thickness will be reduced. The effect is so great that the strain release becomes too large, making the strain measurement difficult.

The analysis of the lattice distortion in the formed wedge-shaped portion is performed by the CBED method using H existing in the CBED figure.
This is performed by observing a change in an OLZ (Higher Order Laue Zone) pattern.

A method of analyzing the strain distribution by the CBED method will be described below.

Using the CBED method capable of quantitatively analyzing the local region, the distortion is quantified according to the flowchart shown in FIG. Actually, the distortion is quantified by comparing the observed image and the calculated pattern by using a holographically approximated HOLZ pattern simulation. The parameters necessary for the HOLZ pattern simulation are a lattice constant, an effective acceleration voltage, a crystal system, a lattice type, and an incident direction, as shown in FIG.

(I) Determination of Effective Acceleration Voltage Using the HOLZ pattern simulation, FIG.
As shown in the flowchart of (b), the voltage is minutely changed in the vicinity of the acceleration voltage to be observed, the change in the sensitive HOLZ pattern is profiled, and a calibration curve of the acceleration voltage as shown in FIG. 2 is created (step). 1). Next, the HOLZ pattern is observed in a portion having no distortion (step 2). Next, the test curve (calculated value) created in step 1
Is compared with the HOLZ pattern observed in Step 2 (Step 3), and the effective acceleration voltage is determined (Step 4).

(Ii) Distortion detection As shown in the flowchart of FIG. 1C, the lattice constant of the target orientation is changed using the effective acceleration voltage value and the HOLZ pattern simulation. Next, HO
The displacement of the LZ pattern is profiled to create a distortion test curve as shown in FIG. 3 (step 6, FIG. 1).
(C)). Using the created strain test curve, the corresponding lattice strain is obtained from the deviation from the HOLZ pattern actually observed near the oxygen precipitate (steps 7 and 8).
These are plotted to draw a distortion curve (step 9), and distortion is detected (step 10).

By analyzing the change of the HOLZ pattern existing in the CBED figure using the CBED method described above, it is possible to measure the change in thickness and the minute lattice distortion in a local region with high accuracy.

(Iii) Evaluation of Relationship between Film Thickness and Strain Amount Next, at each of a large number of locations having different thicknesses in the wedge-shaped portion, fine lattice strain is detected by the above-described method, and the relationship between the strain amount and the film thickness is determined. Ask for.

By measuring the change in lattice strain due to the change in thickness, the effect of strain relief can be considered.

According to the method for manufacturing a lattice distortion evaluation Si wafer according to the embodiment, sputtering is performed using a Ga ion beam or the like focused while observing with a SIM image. For accuracy,
In addition, thin film processing can be performed quickly. Further, the formation of the wedge-shaped portion can be easily performed at an arbitrary wedge angle by inclining the sample stage.

Then, by applying the CBED method to the wedge-shaped portion where the film thickness changes continuously, it is possible to accurately measure the lattice distortion accompanying the film thickness change,
Taking into account the effect of the film thickness on the release of the strain, the lattice strain in the bulk state can be estimated.

[0030]

EXAMPLES AND COMPARATIVE EXAMPLES The evaluation of strain near oxygen precipitates in a CZ-Si wafer will be described below as an example of a Si strain for evaluating lattice strain, a method for manufacturing the same, and a method for evaluating lattice strain according to the present invention.

(1) Preparation of a sample having a wedge-shaped part whose film thickness changes continuously The sample is a Si wafer containing oxygen precipitates prepared at 800 ° C. for 700 hours under annealing conditions. The size of the oxygen precipitate generated under these annealing conditions is about 0.5 μm. The wedge-shaped portion containing the oxygen precipitate was produced using FIB.

First, the CZ-Si wafer is polished with a parallel polishing machine to reduce the film thickness to about 20 μm or less. Next, a wedge-shaped portion is formed using FIB. The optimum angle of the wedge differs depending on the observation target. In this embodiment, the angle is set to about 15 degrees. FIG. 4 shows a schematic perspective view near the wedge-shaped portion. FIG. 5 shows a SIM image near the wedge-shaped portion.

In FIG. 4, the length of side A (which corresponds to the depth in FIB processing) is about 5 μm, and the length of side B is about 3 μm.
m and the inclination angle θ were set to about 15 degrees as described above.

Under the conditions shown in Table 2 below, rough cutting, intermediate treatment, and finish treatment were performed.

[0035]

[Table 2]

The line scan in the finishing process was performed by inclining the sample by ± 3 degrees. Thereafter, sputtering (cleaning) using Ar ions was performed using an ion milling device.

(2) Evaluation of Distortion and Film Thickness The film thickness was determined from the contrast interval appearing in the vicinity of [001] under the two-wave condition by the CBED method, and the continuity of the film thickness was evaluated. Micro-strain was also detected by the same method. TEM used was analytical electron microscope JEM-3010 equipped with a LaB ₆ filament (manufactured by JEOL), the electron beam acceleration voltage at the time of TEM observation and 200 kV,
The spot size of the focused electron beam in the CBED method is set to 15
nm.

(A) Parameters for HOLZ pattern drawing Lattice constant: 0.54305 nm Acceleration voltage: 200 kV Crystal system: Diamond Bonding system: Cubic Incident direction: [012] (b) Determination of effective acceleration voltage A calibration curve of the acceleration voltage shown in FIG.
Based on the flowchart shown in FIG.
It was determined to be 9.98 kV.

(C) The lattice constant was changed at an effective acceleration voltage of 199.98 kV, and a strain test curve shown in FIG. 3 was drawn. Then, the HOLZ patterns were observed at measurement points having different thicknesses in the wedge-shaped portion, and distortion at each measurement point was detected based on the flowchart shown in FIG.

(3) Measurement Results FIG. 6 shows the measurement results of the film thickness change.

The vertical axis indicates the film thickness, and the horizontal axis indicates the measurement interval. When an approximate straight line was drawn, each measurement point was almost a straight line, confirming that the continuity of the film was very good. FIG. 7 is a TEM image showing the tip of the wedge-shaped portion. Equal-thickness interference fringes appeared at equal intervals, and good continuity of the film thickness could be visually confirmed.

FIG. 8 shows a bright field image near the oxygen precipitate in the CZ-Si wafer measured this time. The size of the oxygen precipitate was about 0.5 μm. Bend Con around the oxygen precipitate
tour is appearing. The result of measuring four points in the direction [100] perpendicular to the plate-like oxygen precipitate is shown. In the figure, a to g are respectively 200 nm and 300 nm in the vertical direction from the oxygen precipitate.
nm, 380 nm, 560 nm, 150 n in parallel direction
Points at distances of m, 200 nm, and 400 nm are shown.
The recording of these points utilized multiple exposures.

The direction perpendicular to the oxygen precipitate was evaluated.
9A to 9D correspond to the measurement points a to d in FIG.

Looking at the change in the HOLZ pattern in FIG. 9, in the vertical direction with respect to the oxygen precipitates, as (a) goes from (a) to (d) (as it moves away from the oxygen precipitates in the vertical direction), the {1113} line The intersection (arrow) has moved downward.

FIG. 10 shows the result of quantifying the distortion in accordance with the flowchart shown in FIG. 1 (c). The curve is an approximate curve by a Gaussian function. Here, attention was paid to a compression field generated in a direction perpendicular to the oxygen precipitate. Lattice strains were obtained at different places in the wedge-shaped portion at different thicknesses, and the relationship between the amount of strain related to the distance between the oxygen precipitate and the interface of the Si matrix and the film thickness at the measurement point is shown in FIG. The film thickness at the measurement position was determined by the two-wave condition pattern of the CBED method in the same manner as described above. As is apparent from FIG. 11, the distortion is slightly changed with the change in the thickness. Usually, the thickness suitable for the measurement by the CBED method is about 0.3 μm, which is a slightly thicker portion than that in the conventional TEM observation. Considering that point, when the film thickness is 0.3 μm, the lattice strain of the sample used in this example is 100 μm.
× (5.4478-5.4468) /5.43051≒
1.8 ≒ 2, where (5.4305 is the lattice constant of Si,
5.4478 is the lattice constant of Si x 1.0032 (Si lattice constant when strain is not released from the graph),
It was found that 5.4468 was released by about 2% of the lattice constant of Si × 1.030 (the lattice constant of Si in the case of a film thickness of 0.3 μm that is normally observed from the graph)).

[Brief description of the drawings]

1A is a block diagram showing parameters used for drawing a HOLZ pattern, FIG. 1B is a flowchart for determining an effective acceleration voltage, and FIG. 1C is a flowchart for detecting distortion.

FIG. 2 is a graph showing a test curve of an acceleration voltage.

FIG. 3 is a graph showing a distortion test curve.

FIG. 4 is a partially enlarged perspective view showing the vicinity of a wedge-shaped portion of a sample.

FIG. 5 is a scanning ion micrograph showing the vicinity of a wedge-shaped portion of a sample.

FIG. 6 is a graph showing a relationship between a measurement interval and a film thickness in a wedge-shaped portion.

FIG. 7 is an electron micrograph showing a wedge-shaped portion.

FIG. 8 is an electron micrograph showing the vicinity of an oxygen precipitate.

FIGS. 9A to 9D are diagrams showing HOLZ patterns when changed in a direction perpendicular to oxygen precipitates.

FIG. 10 is a graph showing a relationship between a distance between an oxygen precipitate and a Si matrix and lattice distortion.

FIG. 11 is a graph showing the relationship between the film thickness and the amount of lattice distortion.

Claims (3)

[Claims] 1. A Si wafer for evaluating lattice distortion, wherein the Si wafer has a wedge-shaped portion whose thickness in a direction perpendicular to a main surface of the wafer continuously changes in a range of 10 nm or more and 10 μm or less. Si wafer for lattice distortion evaluation. 2. The method according to claim 1, wherein a focused ion beam is used for forming the wedge-shaped portion. 3. A convergent electron beam diffraction method is applied to the wedge-shaped portion of the Si wafer for lattice strain evaluation according to claim 1,
A lattice distortion evaluation method characterized by detecting lattice distortion in an i-wafer.