JP2000241319A - Method and device for preparing sample - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily observe or analyze a notable defect out of defects being detected near a surface by providing a mark for identification near a desired detect in the detected defects near the surface. SOLUTION: Defects near the surface of a wafer 5 are detected by a defect detection part 2, and the coordinates information, distribution pattern, or the like of the detected defect near the surface being obtained is stored into a calculation processing part 15. After that, the wafer 5 is moved to a marking part 3 for marking based on the coordinates information of the defect. The marking part 3 is introduced to a vacuum container 17 via a load lock chamber 16 along with a wafer tray 6 where the wafer 5 is placed. In the vacuum container 17, a focusing ion beam(FIB) application part 18 for marking, a sample stage 19 for placing the wafer 5, or the like is provided. Then, a mark for identification is provided near a defect close to a desired surface out of defects near the surface being detected by a defect detection part 2. The calculation processing part 15 controls the FIB application part 18 or the like.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method and an apparatus for preparing a sample for detecting and analyzing defects near the surface of a wafer.

[0002]

2. Description of the Related Art In the manufacture of semiconductor devices, it is required to continue producing non-defective products without stagnation. Due to the large number of products produced, the occurrence of defects in a certain process directly leads to a decrease in product yield, which greatly affects profitability. However, it is rare that a product can be produced without any defective products, and some defective products always occur. For this reason, how to detect defects, foreign matter, and processing defects as early as possible and reduce them is a major issue.

For this reason, in a semiconductor device manufacturing site, a careful inspection is performed after a specific process or after a device is completed, and an effort is made to find out the cause of the defect and eradicate the defective product. Actually, wafers or devices are periodically or periodically extracted in the manufacturing process and inspected for defective portions.
For example, the inspection locations and inspection items in each process are determined in advance, and the inspection locations are constantly monitored for each wafer to detect abnormalities in the manufacturing process. Inspection is performed over a wide range of the wafer, operation confirmation is performed, and if there is an abnormal portion, the device is discarded, and a method of pursuing the cause of the abnormality and taking measures is performed.

However, it is becoming clear that the existence of surface defects of the wafer itself before performing various processes greatly affects the occurrence of such a defect in the semiconductor device. Surface evaluation has become very important.

Defects that occur near the wafer surface include COP
(Crystal Originated Parti
cles) defects and glow-in defects, and particularly, the grown-in defects are concentrated at a depth of 0.1 to 0.3 μm from the surface. It is said that when devices are formed on a wafer having such a defect, defective products frequently occur due to dielectric breakdown of a gate oxide film or excessive junction leakage current. For this reason, it is a wise measure not to increase the number of defective products to measure the number and distribution of defects on the entire surface of the wafer used in the manufacture of the semiconductor device in advance, and to exclude in advance the wafer having an abnormally large number of defects.

However, the density of existing defects is approximately 1 / cm 2 when converted to a surface density from the surface, which is a device formation region, to a depth of about 0.5 μm, and it is difficult to detect defects using measuring means such as an electron microscope. It takes an enormous amount of time. In other words, it is a major problem to detect near-surface defects over a wide area of the wafer in a short time.

[0007] Recently, an optical near-defect measuring device (Optical Shallow Defec) has attracted attention.
t Analysis (hereinafter abbreviated as OSDA) is an apparatus that overcomes the above-mentioned problems, and is capable of non-destructively measuring a defect distribution over a wide range of a wafer in a short time, and is an effective apparatus for wafer evaluation.

FIG. 2 is a diagram for explaining the principle of operation of the OSDA device 30. Laser light 32A, 32B emitted from two types of laser light sources 31A, 31B (for example, the wavelength is 523 nm)
And 810 nm) are obliquely incident on the wafer 5 by the reflecting mirror 33 and the lens 36, and are translated and rotated while the wafer 5 is mounted on the stage 7 while the laser beams 32 A and 32 B
Is scanned spirally with respect to the wafer surface, and pulse-like scattered light 35 generated from the defect 34 due to the scanning is condensed by a lens 36.
7E, the generated coordinate information and the signal intensity of the two wavelengths are recorded in the calculation processing unit 38. Further, the surface shape of the wafer 5 can be recorded by the camera 37A, and a light guide 3 is provided in front of the photomultiplier tube 37E to improve sensitivity.
There is an image intensifier 37C via 7D. Further, a beam collimating lens 36 and a reflecting mirror 33 are provided in the middle of the optical path.

The in-plane distribution of the defect, the depth position distribution of the defect, and the particle size distribution of the defect can be obtained from the data stored in the calculation processing section 38. The measurable range is about 1 μm in depth from the surface, the depth resolution is about 0.1 μm, and the plane resolution is about 10 nm. Therefore, in order to analyze defects near the surface of the wafer, a method of measuring the surface of the wafer in a wide range and in a short time with the OSDA apparatus and analyzing the desired defect of interest in detail using an electron microscope is effective.

For a more detailed description of the OSDA device and a measurement example, see “Applied Physics” (1997).
Vol. 55, March issue, pages 109 to 110 (Known Example 1).

[0011]

The above-described OSDA apparatus can detect the number and distribution of near-surface defects over a wide area of a wafer in a short time, but cannot clarify the types of defects. Further, even if an attempt is made to analyze in detail a desired defect of interest using an electron microscope, it has been difficult to accurately analyze the desired defect due to the following problems.

(1) Problem of Defect Type Identification According to the OSDA apparatus, although information such as the number and distribution of near surface defects can be obtained in a wide area of a wafer in a short time, the type and shape of the defect cannot be identified. for that reason,
There is a problem that it is not possible to pursue the cause of occurrence of a specific defect type among a plurality of types of defects and to take measures for reducing them.

(2) Problems in preparation of a transmission electron microscope (TEM) sample Since the spatial resolution of the measurement by the OSDA device is about 10 μm, even if a defect detected by the OSDA device is analyzed by a high-resolution TEM. However, the position of the near-surface defect of interest having a size of sub μm cannot be specified well. In particular, since the wafer surface evaluated by OSDA is a smooth surface without any mark or pattern, even if the approximate position of the defect is clarified by OSDA, there is no clue to use it as a TEM sample.

In the conventional planar TEM observation method for observing a thin section parallel to the wafer surface with a TEM, the observable area is large, so that the probability that a desired defect portion can be captured in the observation field of view increases. Since the observation is performed only from the direction, it is not possible to observe the near-surface defect at different angles. For example, if the shape can be observed from the cross-sectional direction, the type of defect can be determined accurately and accurately, but there is no method for manufacturing a cross-sectional sample.

That is, according to the prior art, a planar TEM sample including a specific near-surface defect is prepared, and the TEM or ST is prepared.
Although EM observation is an effective method for analyzing near-surface defects, a method of producing a planar TEM sample that accurately captures the defect position, and a cross-sectional observation of a defect obtained by planar TEM observation are also available. There was no method for producing a sample that could be made.

In consideration of such problems, a sample preparation method and a sample preparation apparatus capable of identifying a defect type for a desired defect at a specific position by using defect position information obtained in a short time by the OSDA method are desired. Was rare.

In view of the above-mentioned problems, a first object of the present invention is to provide a sample manufacturing method capable of easily observing or analyzing a defect of interest out of near-surface defects detected by a near-surface defect inspection of a wafer. It is in. Further, a second object is to provide a sample preparation apparatus that achieves the first object.

[0018]

In order to realize the first object of the present invention, a method or apparatus having the following configuration is provided.

(1) A defect detecting step of detecting a near-surface defect of a wafer, and a marking step of providing an identification mark near a desired near-surface defect among the near-surface defects detected by the defect detecting step. Sample preparation method.

(2) Detecting defects near the surface of the wafer,
A defect detection step of storing coordinate information of the near-surface defect, a marking step of applying a mark for identifying a desired near-surface defect based on the coordinate information, and By processing the sample into a test piece suitable for performing any of observation or measurement or analysis on the sample preparation method,
A sample is prepared which can reliably observe a defect of interest among defects near the wafer surface and observe the shape.

(3) In the defect detection step in (1) or (2), the laser beam is scanned over a wide area of the wafer surface and relatively to the wafer, and is generated near the wafer surface. This is the step of receiving the scattered light and detecting the location of the near-surface defect.

(4) In the step (3), a plurality of laser beams having different wavelengths are scanned relative to the wafer, and the near surface defect is generated by scattered light generated from the vicinity of the surface of the wafer. By detecting the existence position of the wafer, a defect near the wafer surface can be more reliably detected.

(5) The marking step in (1) or (2) is performed using at least one of a laser beam, an ion beam, an electron beam, and a mechanical probe.

(6) In particular, in the above (5), processing for forming a concave portion or a damaged region in a part of the surface of the wafer by drawing or processing by irradiation of a focused ion beam or a projected ion beam of the ion beam. Or (7) a process of forming a deposit containing components constituting the atmosphere around the sample on the sample surface by electron beam irradiation, or (8) a mechanical concave portion formed by a sharpened metal probe or diamond. Alternatively, by performing a process including a process of forming a damaged region, a mark for identifying a desired near-surface defect can be reliably provided.

(9) The test piece in the above (2) is processed as a test piece for observation or measurement with a transmission electron microscope or a scanning transmission electron microscope, so that the detected near-surface defect can be reliably identified. it can. (10) In particular, in the above (9), the surface to be observed or measured in the test piece is perpendicular to the surface of the wafer, or (11) The surface to be observed or measured is set to the surface of the wafer. By making them parallel, the defects near the surface of interest can be observed from multiple sides. (12) If the test piece in the above (10) is manufactured by cutting the wafer with a dicing apparatus and processing with a focused ion beam, a desired test piece is obtained.

(13) The processing step in the above (2)
An extraction step of extracting the micro sample containing the desired near-surface defect based on the mark from the wafer by a focused ion beam, a fixing step of transferring the micro sample to an analysis sample holder and fixing the micro sample, Producing a sample piece containing near-surface defects of interest without breaking the wafer by subjecting the sample to the above-described focused ion beam processing and finishing the sample into a shape that is easy to analyze. Can be.

(14) The processing step in the above (2)
A step of cutting or cleaving the wafer based on the mark to form a small piece including the mark, and subjecting the small piece to at least one of polishing or ion thinning from the back surface of the wafer and forming a thin piece substantially parallel to the wafer surface; If the step includes a thinning step of manufacturing a sliced sample having a portion, a flat sampled piece including a near-surface defect of interest can be manufactured.

(15) In the above (14), after the thinning step, the thin section is further observed by a transmission electron microscope or a scanning transmission electron microscope to determine a relative positional relationship between the mark and the desired near-surface defect. The process of seeking
A step of extracting a part including the near-surface defect from the thin section based on the positional relationship and fixing it to a sample holder for analysis, and including the near-surface defect and substantially perpendicular to the wafer surface; By including a cross-section processing step of processing into a thin section,
It is possible to produce a cross-sectional sample piece that definitely includes the near-surface defect of interest.

(16) In the processing step of (2), the wafer is cut or cleaved into small pieces containing the marks with reference to the marks, and the small pieces are polished or ion thinned from the back surface of the wafer. A step of producing a flake sample having a flake section substantially parallel to the wafer surface by applying any of the above, and observing the flake section with a transmission electron microscope or a scanning transmission electron microscope to obtain the mark and the desired surface vicinity defect. A step of obtaining a relative positional relationship, and a step of further performing a second marking step of further performing a second mark for identifying the position of a desired near-surface defect in detail by using a focused ion beam or an electron beam on the slice sample. Thus, it is possible to more accurately produce a sample piece including a near-surface defect of interest.

(17) In the second marking step (16), if at least one of a pulse laser, a focused ion beam, a projection ion beam, and an electron beam is used, marking is surely performed on a minute portion. Can be applied.

(18) If the second mark in the above (16) or (17) is any one of a set of dots or line segments or symbols arranged in a straight line or in a lattice point,
This is suitable for making it easy to identify a defect originally present in the sample.

(19) In any one of the above (16) to (18), the thin sample on which the second mark is provided is
A position determining step of observing with a transmission electron microscope or a scanning transmission electron microscope to clarify a relative positional relationship between the desired near-surface defect and the second mark and to specify an existing position of the near-surface defect. With this, it is possible to attach a mark that allows a desired portion to be accurately identified.

(20) In the above (19), after the above-mentioned position determining step, the above-mentioned near-surface defects including the desired near-surface defects are observed from a cross-sectional direction with a transmission electron microscope or a scanning transmission electron microscope. A sample piece including a cross-sectional thinning step of processing into a cross-sectional thin section sample for the purpose of manufacture can surely manufacture a sample piece including a noticeable near-surface defect.

(21) A method of analyzing near-surface defects in a semiconductor wafer includes the steps of (a) detecting a near-surface defect of a wafer by laser beam irradiation;
(B) a storage step of storing the coordinate information of the near-surface defect, (c) a marking step of marking the vicinity of the near-surface defect of interest by irradiating a focused ion beam (FIB) among the near-surface defects, d) a step of cutting or cutting the wafer into small pieces containing the defect of interest and the mark; (e) thinning the small pieces from the back surface to process them into a thin sample; Sample preparation for identifying near-surface defects in a semiconductor wafer can be realized by a method according to the procedure of an observation step of observing the noted near-surface defects with a transmission electron microscope or a scanning transmission electron microscope based on the marks.

(22) In the above (21),
(A) a second marking step of applying a more detailed second mark with a focused ion beam based on the observation result by the transmission electron microscope or the scanning transmission electron microscope, and (b) the second mark An observation step of observing the applied sample with a transmission electron microscope or a scanning transmission electron microscope; and c) a step of determining an observation area including the defect of interest based on the transmission electron microscope or scanning transmission electron microscope observation. (D) a step of extracting a micro sample including the determined observation region, (e) a cross-section thinning step of processing the micro sample into a cross-section thin sample including the desired near-surface defect by FIB, f) Observing the near-surface defects in the semiconductor wafer by the method of adding a step of observing the cross-sectional slice sample with a transmission electron microscope or a scanning transmission electron microscope. Sample Preparation for can be realized.

(23) In the above (22) or (23), a desired one of the defects obtained in the defect detecting step is transmitted through a transmission electron microscope or a scanning transmission microscope from a direction perpendicular to a surface of the wafer and a sectional direction. By observing with an electron microscope and analyzing the near-surface defect described above to analyze the near-surface defect, the near-surface defect of interest can be observed from multiple angles.

Further, according to the present invention, in order to manufacture a semiconductor device with a good yield by using a non-defective wafer having few defects near the surface, the following wafer management method for selecting non-defective semiconductor wafers in advance. It becomes possible.

(24) At least one of the position, number, distribution, and type of the near-surface defects obtained by detecting the near-surface defects of the semiconductor wafer, the shape observation result of the near-surface defects by electron microscope observation, The correlation between at least one of the occurrence position, the number of occurrences, and the form of the failure of the semiconductor device formed on the wafer is stored in a database, and a defect near the surface of the semiconductor wafer is detected during the manufacture of the semiconductor device. A method of manufacturing a semiconductor device, wherein screening of the semiconductor wafer is performed based on a detection result obtained periodically and the database.

The sample preparation apparatus for realizing the above-described method for preparing a sample or the above-described method for analyzing near-surface defects, the method for managing a wafer or the method for manufacturing a semiconductor device has the following configuration.

(25) A sample preparation apparatus having at least a defect detection unit for detecting a defect near the surface of a wafer, and a marking unit for marking a vicinity of a desired defect among the defects near the surface.

(26) A defect detecting section for detecting a near-surface defect of a wafer, a marking section for marking a desired near-surface defect among the above-described near-surface defects, and a defect detecting section for detecting a defect near the surface. A sample preparation apparatus having at least a sample processing section for extracting a micro sample containing at least a defect of interest and processing it into a shape suitable for mounting on at least one of an analyzer, an observation device, and a measurement device.

(27) A defect detection unit for detecting a defect near the surface of the wafer, a mark is placed near a desired defect among the above-described defects near the surface, and a desired one of the defects near the surface is determined based on the mark. A sample preparation apparatus having a sample processing unit for extracting a micro sample including a near-surface defect and processing the sample into a sample piece having a shape suitable for at least one of an analysis device, an observation device, and a measurement device.

(28) The sample processing section of (26) or (27) has a structure having a sample moving means for moving the extracted micro sample to another member, so that an analyzer such as a transmission electron microscope can be used. It can be reliably mounted on a dedicated sample holder and in a vacuum atmosphere.

(29) In any one of the above (25) to (28), the defect detection section and the marking section, or the defect detection section and the marking section and the sample processing section, or the defect detection section and the above A configuration in which a wafer handler is provided for loading and unloading a wafer between any of the sample processing sections, or (30) from (25) to (2)
8) In any one of the above, the defect detection unit and the marking unit, or the defect detection unit and the marking unit and the sample processing unit, or the wafer between the defect detection unit and the sample processing unit. By providing a transport path through which the wafer moves, it is possible to efficiently carry out from defect inspection to marking or sample preparation.

(31) In any one of the above (25) to (30), by providing at least a calculation processing section for storing coordinate information of a defect detected by the defect detection section, the positioning of the marking section can be performed accurately and accurately. Can be done quickly.

(32) The defect detection unit according to any one of (25) to (31) is a laser beam irradiation means for generating a laser beam and irradiating the wafer, and a light for detecting scattered light from the wafer. (33) a plurality of laser generating means for generating laser beams having different wavelengths from each other and irradiating the wafer, and detecting scattered light from the wafer for each wavelength; In the case of a sample preparation apparatus having at least a plurality of light detection means, the existence position of the near-surface defect can be accurately detected.

(34) The marking portion according to any one of (25) to (32) has at least one of a laser beam irradiation optical system, an ion beam irradiation optical system, an electron beam irradiation optical system, and a mechanical probe. With this configuration, a mark that can be clearly distinguished from a defect can be provided.

(35) By making the ion beam irradiation optical system of the above (34) either a focused ion beam irradiation means or a projected ion beam irradiation means, a mark which can be clearly distinguished from a defect can be provided. .

(36) A near-surface defect detecting device for detecting a near-surface defect of a wafer and storing coordinate information of the detected defect, and a micro-sample piece containing a desired defect is extracted from the wafer and analyzed or observed. A sample preparation device for processing into a sample piece having a shape suitable for at least one of an apparatus or a measurement device, a TEM or STEM for analyzing the shape and structure of a defect, a wafer name, defect coordinates, a defect distribution, a sample piece name, By configuring a near-surface defect analysis system by interconnecting a computer that stores at least one of the defect images with a network,
An apparatus for efficiently analyzing near-surface defects can be configured.

(37) A laser beam irradiating section for generating a laser beam of at least one wavelength and irradiating the wafer, a light detecting section for detecting scattered light from the wafer, and a coordinate information generating the scattered light. A near-surface defect detection apparatus having at least a calculation processing means for storing, wherein the surface having a focused ion beam irradiation means equipped with a liquid metal ion source or an electric field ionization gas ion source for irradiating a focused ion beam to a part of the wafer According to the nearby defect detection device, the vicinity of the defect can be accurately and accurately marked by the near surface defect detection device.

(38) A laser beam irradiating section for generating a laser beam of at least one wavelength and irradiating the wafer, a light detecting section for detecting scattered light from the wafer, and a coordinate information generating the scattered light. It can be realized by a near-surface defect detection device having a calculation processing means for storing and a laser light irradiation means for a laser mark for making a laser mark near the light emitting part based on the coordinate information of the desired scattered light generation part. .

(39) In an ion beam apparatus for irradiating a part of a wafer with an ion beam, a laser beam irradiator for generating a laser beam of at least one wavelength and irradiating the wafer with the laser beam, and scattering light from the wafer. A light detection means for detecting, and a calculation processing means for storing coordinate information where scattered light is generated, at least in the vicinity of a defect near the wafer surface detected by a focused ion beam device having a configuration including a near surface defect detection unit having at least a defect detection unit near the surface. Can be marked.

[0053]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A sample manufacturing method according to the present invention comprises a defect detecting step for detecting a defect near the surface of a wafer, and an identification method for identifying a defect near the desired surface near the surface near defect detected by the defect detecting step. And a marking step of making a mark. Further, an apparatus for realizing this sample manufacturing method has at least a defect detection unit that detects a near-surface defect of a wafer, and a marking unit that marks a vicinity of a desired near-surface defect among the near-surface defects. I do.

<Embodiment 1> In this embodiment, a method for preparing a sample for observing a desired defect from the wafer surface with a TEM, that is, a so-called planar TEM sample, among the defects of the wafer particularly detected by the near-surface defect detection unit. Will be described.

In order to produce a planar TEM sample, it is important to first make a mark for identifying a defect.
Since the wafer surface before processing is a smooth surface without a pattern such as a wiring serving as a mark, it is very difficult to attach an accurate mark for identifying a defect of interest. Therefore, in this method, a sample preparation method including the following marking step is used.

First, a defect near the surface of the wafer is detected by the defect detection unit, and at least coordinate information, a distribution pattern, and the like of the obtained near-surface defect are stored in the calculation processing unit. The origin of the coordinates of the defect at this time is a reference position provided in advance on the wafer. Thereafter, the measured wafer is transferred to a marking section, and marking is performed based on the coordinate information of the previous defect.

The wafer to be subjected to near-surface defect detection is smooth without any pattern before processing in semiconductor manufacturing.
Scanning Electron Image by Scanning Electron Beam
Microscopy: Hereinafter, a SEM image or a secondary electron image obtained by FIB scanning (Scanning Ion Microscop)
(y: hereinafter abbreviated as SIM image), the surface shape of the sample can be observed, but the inside of the sample cannot be seen. Further, it is difficult to specify the position of a defect to be focused on a sample having no peculiar pattern such as a mark on the wafer surface. That is,
SEM when marking near defects near the surface
The exact position cannot be determined from the image or the SIM image. In addition, the size of the defect is as small as about 0.1 μm, and there is no positioning accuracy at all by manual mechanical marking. On the other hand, observation by FIB scanning can be performed at a high magnification and processing can be performed on the spot. Therefore, if defect coordinates are clear, FIB can be used as a marking means.

The marking section has a sample stage on which the wafer is placed, and can move the sample stage so that the designated coordinates can be brought into the field of view with the reference position of the wafer as the origin. The same reference position on the wafer is used as the origin when detecting a defect and when marking. However, even if the method of setting the coordinates at the time of defect detection differs from the method of setting the coordinates of the marking part, mutual By determining a coordinate conversion method for one coordinate system and incorporating the coordinate conversion method in a computer, mutual coordinate information can be shared.

FIG. 3 shows an example in which the surface of the wafer 5 is marked by FIB. FIG. 3A shows an example in which the marks are arranged in a mesh shape, and the interval between the marks 41a is about 5 μm.
The line width was about 0.1 μm, and it was composed of 10 lines vertically and horizontally (Fig. 3
(A) is a part thereof). FIG. 3 (b) shows a dot 4 of a grid point shape.
It is an example of marking by 1b. The marking pattern may be a short line segment 41c as shown in FIG. 3C, or may be a set of appropriate symbols 41d as shown in FIG. 3D.

At the time of SIM image observation, the near-surface defect 42A of interest cannot be observed, but exists inside the area 42 near the mark. This area 42 is an area determined by the spatial resolution of the near-surface defect inspection unit. The mark 41 applied to the sample surface can be clearly distinguished from a defect if the mark 41 has an artificial shape or arrangement.

It should be noted that the mark has a small current FI at a predetermined coordinate.
B. If the current of the FIB is large, the damaged area of the sample becomes large, and when the wafer is subsequently processed into a thin slice, the sample is greatly damaged, and the near-surface defect 42 which should be observed originally.
This is because there is a high possibility that the prototype of A will be damaged. When the area 42 is of the extent shown in this embodiment, the mark 41 and the near-surface defect 42A of interest can be observed in the same field of view by low magnification TEM observation. The mark 41 can be sufficiently identified by TEM observation even if the mark 41 is processed with a low-current FIB and shallow enough to cause slight damage to the wafer 5 surface.

Further, as shown in FIG. 4, a large large mark 43 which can be identified by an optical microscope is formed. FIG. 4 shows an example of the positional relationship of the mark 41 for indicating the relative position between the large mark 43 for the optical microscope and the defect. The large mark 43 serves as a mark when the wafer 5 is cut into small pieces 5A, and may be about several tens μm in size. Further, the large mark 43 itself can be clearly confirmed by a light reflection method with an optical microscope only by forming a concave portion having an appropriate depth in the sample.

Next, the wafer 5 is formed into a square of about 1 mm including the mark 41 and the large mark 43 by using a diamond cutter or a dicing saw so that the middle of the large mark 43 is substantially at the center of the small piece 5A. Disconnect. Next, the small piece 5A obtained by cutting is fixed to a polishing jig with its wafer surface side as an adhesive surface. There are various forms of the above polishing jig.
Using a polishing jig equipped with a device capable of measuring the dimensions of the order, the flatness of the sample is adjusted and pressed against the polishing plate to grind the back surface, thereby producing a polished sample having a thickness of about 10 μm.

Next, the above-mentioned polishing sample removed from the polishing jig is fixed to a single-hole mesh and mounted on an ion thinning apparatus. In ion thinning, only the back surface is obliquely irradiated with argon ions while rotating the single-hole mesh to further thin the polished sample. Ion thinning is
Usually, ion irradiation is performed from the front and back surfaces of the sample. However, in this example, ion irradiation is not performed from the front surface side because the purpose is to observe defects near the surface. Finally the above mark 4
1 is thinned to about 100 nm or less near the center to obtain a planar TEM sample 5B.

The flat TEM sample 5B is mounted on a TEM stage together with a single-hole mesh and subjected to TEM analysis. Plane T
In the observation of the EM sample 5B, a wide area of several tens of μm square can be observed at a low magnification at a time, and the marks 41 previously processed by FIB can be easily found because of their artificial shape and arrangement. The near-surface defect 42A of interest is specified based on the mark 41, and TEM observation can be performed at a high magnification from a planar direction.

As described above, according to the sample preparation method of the present invention, the FI
This is a sample manufacturing method in which a mark is formed by B and a defect to which attention is paid based on the mark can be definitely found.

In this embodiment, an example is described in which the FIB is used as a method for making a mark. However, a marking using ablation of the wafer surface by pulsed laser beam irradiation may be used. Furthermore, an impurity such as carbon contained in the atmosphere near the sample may be deposited on the wafer by electron beam irradiation to form a mark, or may be pressed or scratched by a mechanical metal probe or a diamond probe driven with high precision. The mark may be used as a mark. Either method is preferable because it can be distinguished from a defect by TEM observation. As described above, among the defects obtained by the near-surface defect detection according to the sample manufacturing method of the present invention, the defect of interest can be observed without fail from the planar direction, so that the defect can be efficiently and accurately analyzed. it can.

In the above-described embodiment, an example has been described in which the defect detection unit and the marking unit are separately configured, and the target sample is transferred from the defect detection unit to the marking unit. The present invention is not limited to a configuration in which a part is separated, and marking may be performed at a place where a defect is detected.

<Embodiment 2> This embodiment is a sample manufacturing method for observing a near-surface defect of interest, which is detected by TEM observation of the planar TEM sample manufactured in Embodiment 1, and further TEM observation from the cross-sectional direction. is there. This method will be described as a step continuing from the first embodiment.

FIG. 5A shows the plane T of the plane TEM sample 5B.
It is an EM image 44, which includes a mark 41c and a near-surface defect 42A.
Can be confirmed, and the approximate relative positional relationship between the mark 41c and the near-surface defect 42A can be understood. First, a plurality of areas from the low magnification image to the high magnification image of the planar TEM image 44 including the noted near-surface defect 42A are stored in a computer or left on a recording paper such as a photograph. At this time, the reference mark 41
It is important that c also falls within the same screen as the near-surface defect 42A. Preferably a plurality of marks 41
If c is also included in the image, it is very easy to locate the target near-surface defect 42A.

Next, a planar TEM sample 5B observed by TEM was used.
Is mounted on the mesh and moved to the sample preparation apparatus, and an operation of extracting a micrometer-level micro sample including the noticeable near-surface defect 42A from the planar TEM sample 5B is performed. In order to extract a minute sample including a defect of interest from the planar TEM sample 5B, the position of the target portion must be accurately grasped.

In the mark described above, the interval between the marks is large and the positional relationship with the defect of interest can be accurately grasped. Therefore, in order to accurately identify the position of the defect, as shown in FIG. Further fine marking is performed again based on the planar TEM image (hereinafter, the mark is referred to as a second mark 46).
).

The shape of the second mark 46 may be the same pattern as that illustrated in FIG. 3, but the overall size is further reduced. As a specific example of dimensions, short line segments having a width of about 0.05 μm and a length of 0.3 μm, which are much smaller than the mark 41c, are arranged on four sides at 0.5 μm intervals. In order to make the formed second mark 46 as small as possible, the current of the FIB is made as small as possible and the irradiation time is made as short as possible. Even if the mark does not penetrate the slice, it remains as an artificially arranged damaged area, and can be easily distinguished from a defect because of the artificial shape.

The plane T on which such a second mark 46 is provided
When the EM sample 5B is observed again on a plane with a TEM, the fifth
As shown in (b), the position of the near-surface defect 42A can be found in more detail from the relative positional relationship with the second mark 46.
The TEM image is stored in a computer or left on a recording paper such as a photograph. Based on the TEM image, a minute area including the near-surface defect 42A to be noted in the marking portion is extracted from the planar TEM sample, and the near-surface defect 42A is extracted. The TEM sample for cross-sectional observation can be formed by processing into a vertical thin section containing.

Hereinafter, extraction of the sample 5C including the noticeable surface near-surface defect 42A will be described in detail with reference to FIG.

First, based on a TEM image in which the second mark 46 and the near-surface defect 42A of interest are in the same field of view, a position identification mark for identifying the position of the near-surface defect 42A by surface observation. 61 is applied. Specifically, FIG.
As shown in (b), the FIB 62 processes the arrangement so as to sandwich the near-surface defect 42A. The position specifying mark 61 also has a mark portion 61 'having a small line width in order to realize more precise processing position alignment. In this embodiment, four line segment marks are provided at intervals of 5 μm and 1 μm across the observation region, and the FIB
It was applied by irradiation.

Next, as shown in FIG. 6C, a protective film 63 for protecting the sample is formed so as to cover the line connecting the two position specifying marks 61. The protective film 63 is formed by scanning the FIB 62 in a rectangular shape while supplying a deposition gas to form a metal film in a scan area.
In this embodiment, a tungsten film having a width of about 2 μm, a length of about 14 μm, and a height of about 1 μm was formed. However, the protective film 63 is not limited to tungsten, and may be formed of a carbon film, a platinum film, a silicon film, or the like.

Next, a process of extracting a micro sample is started. FIG.
The FIB 62 surrounds the protective film 63 as shown in FIG.
The cantilever 65 is formed by making a cut in a U-shape. As an example of the dimensions, the width is approximately 4 μm, the length is 18 μm, and the sample thickness is 0.1 μm.

Next, as shown in FIG. 6E, the probe 66 is brought into contact with the cantilever 65. The probe 66 is fixed to the tip of the transferring means of the micro sample, and a driving unit (not shown) of the transferring means is provided with a probe for preventing the cantilever 65 and the probe 66 from being damaged by careless pressing of the probe 66. 66 has a function of stopping the drive in the + Z direction when it comes into contact with the sample. After the contact of the probe 66, in order to fix the probe 66 to the cantilever 65, a deposition film 67 is formed in an area of about 2 μm square including the tip of the probe 66 while the probe 66 is in contact with the cantilever 65. The deposition film 67 is formed by the same method as the protection film 63, and may be the same component. In this way, the probe 66 and the cantilever 65 (sample to be extracted) can be connected (FIG. 6 (f)).

Thereafter, as shown in FIG. 6 (g), the support portion 68 of the cantilever 65 is irradiated with the FIB 62 and subjected to sputtering to cut the cantilever 65. After cutting, probe 6
When the sample 6 is gradually raised, the minute thin sample 5C can be separated and extracted from the original flat TEM sample 5B as shown in FIG. 6 (h).

The processing procedure for forming a cross-sectional TEM sample for viewing the defect 42A from the cross-sectional direction will be described below with reference to FIG.

First, as shown in FIG. 7 (a), the micro-slice sample 5C extracted is stopped in a state where it is raised from the planar TEM sample 5B, and the sample stage is relatively moved so that the sample holder 70 comes directly below the micro-slice sample 5C. Then, as shown in FIG. 7 (b), the probe 66 is lowered to bring the minute thin sample 5C into contact with the sample holder 70. The minute thin sample 5
When C comes into contact with the sample holder 70, the lowering of the probe 66 is automatically stopped.

Next, the minute thin sample 5C is placed in the sample holder 7
In order to fix to 0, the deposition film 71 is formed so as to straddle the micro-slice sample 5C and the sample holder 70. The deposition film 71 in the present embodiment has a size of about 2 μm square at the longitudinal end of the micro flake sample 5C (FIG. 7).
(C)). Thereafter, the deposition film 67 is irradiated with the FIB 62 to separate the probe 66 from the minute thin sample 5C, and the probe 66 is further kept away from the sample holder 70 to prevent the probe 66 from being damaged. Further, in order to securely fix the micro-laminate sample 5C to the sample holder 70,
A deposition film 71 'of about 2 μm square is also formed on the other end in the longitudinal direction of the micro flake sample 70 (FIG. 7D).

Next, the microblade sample 5C fixed to the sample holder 70 is irradiated with the FIB 62, and the side surface is removed together with the sample holder 70 to form the concave portion 72. In this embodiment, the frontage of the recess 72 is about 10 μm, the depth is about 2 μm, and the depth is about 3 μm (FIG. 7E). Further, a concave portion 72 ′ is formed on the opposite side surface so that a thin wall having a thickness of about 0.1 μm remains at the center (FIG. 7F).

FIG. 7 (g) is a cross-sectional sample of the completed thin sample, which is a cross-sectional sample of a minute thin sample 5C including a near-surface defect 42A (not shown) in which the upper portion 74 of the central thin wall portion 73 is noticed. Approximately 0.1 μm in height, 0.1 μm in width and 1 in length
It is a fine line of 0 μm. For reference, the direction of the incident electron beam during TEM observation is indicated by an arrow 75. As described above, the electron beam is incident on the surface of the micro flake sample almost in parallel.

In this manner, a sample for observing the near-surface defect 42A from the cross-sectional direction could be prepared by the procedures shown in FIGS. According to the method of the present embodiment, the position of a defective portion can be located with high accuracy. Moreover, since all of these steps can be performed in a vacuum atmosphere, processing can be performed with almost no unnecessary fine powder attached.

The above-mentioned cross-sectional sample thus prepared was subjected to TEM.
To be introduced. Subsequent TEM observation techniques are well known and will not be described here.

As described above, the plane TEM image and the cross section T
By interpreting the composite with the cross-sectional TEM image obtained by the EM observation, it is possible to estimate the three-dimensional structure of the near-surface defect 42A of interest. For example, in the case of the present embodiment, if the defect image is square in planar TEM observation and rhombic in cross-sectional TEM observation, it can be interpreted that both TEM images are interpreted compositely and that the defect has an octahedral structure. As described above, even if the defect structure is difficult to be interpreted by TEM observation of only a plane or a cross section, the structure can be accurately interpreted by observing the same defect from different directions by 90 °.

The marking method is described in FIB
Alternatively, a plurality of mark patterns may be formed at once by a projected ion beam using a stencil mask provided with a mark-shaped opening pattern in advance. In particular, since a fixed mark pattern can be formed with the projected ion beam regardless of the size of the observation visual field, marking with a similar number and arrangement of marks can be performed even in a high-magnification visual field.

<Embodiment 3> This embodiment is an embodiment relating to a sample preparation apparatus for realizing the above-described sample preparation method.

FIG. 1 is a plan view for explaining an overall schematic configuration of an embodiment of a sample preparation apparatus according to the present invention. The sample manufacturing apparatus 1 includes a defect detection unit 2 and a marking unit 3
Are connected via a wafer handling unit 4. The marking section 3 includes a load lock chamber 16 for introducing the wafer 5 into the vacuum container. The wafer handling section 4 is a mechanism for taking the wafer 5 in and out of the defect detection section 2 and the marking section 3, and includes a telescopic arm 4A and a rotating mechanism 4B.

The movement of the wafer 5 is usually performed in a state where the wafer 5 is set on the wafer tray 6. When a noticeable defect is detected by the defect detection unit 2, the wafer 5 is transferred from the defect detection unit 2 to the marking unit 3, or the wafer 5 is unloaded from the marking unit 3. You can also inspect. The control of the defect detection unit 2 and the marking unit 3 and the collection and processing of data are performed by the calculation processing unit 15 through the signal lines 15A and 1A.
5B.

The detailed configuration of the defect detector 2 will be described with reference to FIG. The defect detection unit 2 measures a near-surface defect on the surface of the wafer 5, and the target wafer 5 can be set on the wafer tray 6 and fixed. The wafer tray 6 is placed on a sample stage 7 that can rotate and translate at high speed.

Laser light 8A, 8B applied to wafer 5
Are emitted from a semiconductor-pumped YAG laser 9A (wavelength 532 nm) and a semiconductor laser 9B (wavelength 810 nm). A beam collimating lens 10 is provided in the path of each laser beam.
A, 10B, reflectors 11A, 11'A, 11B, 11 '
B and the like are arranged, and can be focused and incident on the wafer 5 at an incident angle of 75 °. The laser light 8A, 8B scans the wafer surface in a spiral by rotating the wafer 5. The beam diameter is approximately 5 μm × 25 μm.

If there is a defect near the surface, the condensing lens 13 for collecting the scattered light 12 generated in a pulsed manner from the defect in accordance with the scanning of the laser beam, a reflecting plate 11C,
11'C, photomultiplier tubes 14A and 14B for detecting scattered lights 12A and 12B for each wavelength, a calculation processor 15 for storing the position where the scattered light is generated and the signal intensity of two wavelengths. From the data obtained by such a configuration, the distribution of defects in the wafer surface, the depth position distribution of defects, the particle size distribution of defects, and the like can be obtained.

On the other hand, the marking section 3 has a structure that allows the wafer tray 6 on which the wafer 5 is placed to be introduced into the vacuum vessel 17 via the load lock chamber 16 (see FIG. 1). Observe the wafer 5 and the like in the vacuum container 17 and make a mark F
IB irradiation unit 18 and sample stage 1 on which wafer 5 is placed
9. It includes a secondary particle detector 20 for detecting secondary electrons or secondary ions. The calculation processing unit 15 also serves as the calculation processing unit 15 of the defect inspection unit 2, and controls the FIB irradiation unit 18, the sample stage 19, the secondary particle detector 20, and the like via a signal line 15B.

The FIB irradiator 18 comprises, for example, a liquid metal ion source, a beam limiting aperture, a focusing lens, an objective lens, etc., and forms an ion beam emitted from the liquid metal ion source into a FIB having a diameter of about 10 nm to several 100 nm. I do. The FIB can scan in various shapes depending on the voltage application method of the deflector, and can perform local processing. By detecting secondary electrons or secondary ions generated from the sample at the time of FIB irradiation by the secondary particle detector 20, a SIM image or a secondary ion image of the sample can be displayed on a display, and also serves as a microscope. The SIM image is stored in the calculation processing unit 15 as necessary.

The calculation processing unit 15 can receive and process the information on the defect coordinates obtained by the defect detection unit 2, and among the coordinates of the detected near-surface defects, consider the depth and size of the defect and consider the defect. By designating the coordinates, the sample stage 19 operates, and the noted defect can be stopped in the field of view of the FIB. In addition, by storing the mark shape in the calculation processing unit 15 in advance, by giving a marking command to the calculation processing unit 15, it is possible to instantaneously apply the mark as described in the first embodiment.

By using such a sample preparation apparatus 1, a mark can be made in the vicinity of the defect of interest among the detected defects, and the form of the defect of interest can be analyzed by analyzing the marked sample. The details can be clarified.

In this embodiment, the FIB irradiating section 18 is provided as a means for making a mark.
In place of the above-described method, a pulse laser beam irradiation unit such as an excimer laser or a semiconductor laser beam irradiation system may be provided to perform marking using ablation of the sample surface, or impurities contained in the atmosphere near the sample may be used. An electron beam irradiation system may be provided in order to attach a mark to form a mark. Further, a mechanical metal probe or a diamond probe which can be driven with high precision for marking by mechanical pressing or scratching may be provided.
However, in the case of a mechanical probe, microscopic means is required, and a scanning electron microscope or the like is also used. Also, FIB
A projection ion beam optical system may be used instead of the irradiation optical system. In any of the methods, a defect can be distinguished from an artificial mark by TEM observation, so that the position of the defect can be easily specified, which is preferable.

<Embodiment 4> A micro sample of μm level including a defect of interest is extracted from the defects detected by the defect inspection unit 2 with reference to FIG. 9 and processed into a sample for TEM observation from a cross-sectional direction. The device will be described.

The sample preparation section 23 in FIG. 9 also plays a role of the marking section 3 in FIG. 1, and the sample transfer section 24 for transferring a micro sample to a sample holder in the basic configuration of the marking section 3.
Gas source 2 for supplying a gas for locally forming a deposition film with a configuration in which at least
5 may be mounted. Of course, the sample preparation unit 23 can also mark the sample.

The sample transfer control unit 24 ′ for driving the sample transfer unit 24, the deposition source control unit 25 ′ for driving the deposition gas source 25, and the FIB for operating the FIB irradiation unit 18 for irradiating the FIB 49. The control unit 18 ′ and the stage control unit 19 ′ for controlling the position of the sample stage 19 are controlled by the calculation processing unit 15. Code 22
Indicates the surface state of sample 5. It is a display unit for displaying calculation results of the calculation processor 15, instructions to various control devices, operation states, and the like.

The sample holder (not shown) is a jig for mounting the sample on the sample stage of the TEM, and has a structure detachable from a part of the wafer tray 6. Leave from 6 and T
It can be introduced to EM, SEM, and other analyzers. The dimensions of the sample holder used in this embodiment are 2.6 mm in length and 30 mm in upper width.
μm, lower width 200 μm, height 0.7 mm (silicon wafer thickness), the fixed surface of the extracted micro sample is a silicon wafer surface or a cleavage surface, and the extracted micro sample is fixed to the fixing surface. Even by TEM observation, unevenness of the fixed surface does not hinder electron beam irradiation. Although the shape and dimensions of the sample holder are not limited to this, it is necessary to make the fixed surface a wafer surface or a cleavage surface and make the fixed surface width as thin as possible in order to process a TEM sample in a short time. Is necessary to facilitate TEM observation.

The holder cassette 27 is a jig for mounting the above-mentioned sample holder, and is attached to the wafer tray 6. One or a plurality of sample holders may be mounted on one holder cassette 27. In particular, a plurality of sample holders is convenient for attaching and detaching and managing the sample holder. Further, a plurality of holder cassettes 27 installed on the wafer tray 6 may be provided. For example, assuming that one extracted micro sample is mounted on one sample holder, one micro cassette is stored in one holder cassette 27.
0 sample holders can be mounted.
Assuming that five wafers are mounted on the wafer tray 6, 50 minute samples can be continuously extracted from one wafer, and near-surface defects can be efficiently extracted under various conditions, and defect analysis can be performed. Can be implemented systematically.

The sample transfer section 24 is composed of a coarse movement mechanism section 24A which moves in three axes of XYZ and a fine movement mechanism section 24B which moves finely in the Z direction, and both are operated by a sample transfer control section 24 '. At the tip, a probe 28 for extracting a micro sample
Is connected.

The sample transfer section 24 is provided with a coarse movement mechanism section 24A having a fast moving speed and a large stroke, and a coarse movement section 24A so that the wafer 5 can be quickly sampled from an arbitrary position even if the wafer 5 has a large diameter. It is composed of a fine movement section 24B having a stroke equivalent to the movement resolution and having a high movement resolution, and the entire sample transfer section 24 is installed independently of the sample stage 7, and large movement of the sampling position is shared by movement of the sample stage. I let it.

The coarse movement mechanism 24A is driven in the XYZ directions by a motor, gears, piezoelectric elements, etc., and has a moving resolution of 1 μm or less with a stroke of about several mm. The Z-direction fine movement mechanism section 24B is required to be as compact as possible and to be precisely moved due to mechanical constraints such as the distance between the objective lens and the sample. Moving resolution is achieved.

A probe 28 used for extracting a micro sample is provided at the tip of the fine movement mechanism 24B. However, the configuration of the sample transfer unit 24 is not limited to the configuration shown here, and the fine movement function of the Z axis may be omitted by installing a mechanism that operates with high accuracy for all three axes, A rotation mechanism may be added.

As a conventional technique similar to the sample transfer section 24, there is Japanese Patent Application Laid-Open No. 5-52721 (known example 2).
According to this conventional technique, the transport means for transporting the separated sample is configured with three bimorph piezoelectric elements corresponding to the XYZ axes, and the installation position of the transport means is set on the stage from the above publication. Can be read. When the transfer means is installed on the sample stage as in this conventional example, when the separation point of interest is, for example, at the center of a wafer having a diameter of 300 mm, the moving stroke of the tip of the transfer means moves the sample from the transfer means position to the sample. Since the distance is much smaller than the distance to the desired position, there is a fatal problem that the transport means provided on the sample stage does not reach the separation point of interest. One of the features of the present embodiment is that the sample transfer unit 24 is provided independently of the sample stage in order to avoid the above-described problems.

The deposition gas source 25 has a structure capable of injecting an organic metal gas or the like from a nozzle 29 at the tip, and forms a deposition film of a gas component in a scanning region by performing FIB scanning while injecting the organic metal gas onto a sample. be able to.
The deposition film plays a role of protecting the sample surface and bonding members.

The sample 5 wafer is loaded in the load lock chamber 16.
Can be introduced onto the sample stage 19 without impairing the degree of vacuum in the sample chamber.

By using such a device, the plane TE
A sample for observing a defect of interest in the M sample for TEM observation can be reliably processed, and a defect near the same surface can be observed by TEM from two directions of a plane and a cross section. The morphology of near-surface defects can be analyzed from multiple angles.

In the fourth embodiment, the marking section and the sample preparation section are used in common, but they may be separated from each other.
The marking unit may employ a marking method using an electron beam, and the sample preparation unit may employ a configuration using the above-described FIB. The method for extracting a micro sample using the sample preparation unit described above is described in detail in Example 2.

<Embodiment 5> In Embodiment 3, the defect detector 2
4 shows a sample preparation apparatus including the marking unit 3 and the marking unit 3 together. In the fourth embodiment, the sample preparation unit 23 serving also as a marking unit is shown.
FIG. 10 shows a sample preparation apparatus having the
As described above, the near-surface defect detection device 80 that exclusively detects the near-surface defect of the wafer 5 and the marking device 81 that exclusively marks the wafer 5 may be mechanically independent. In the case of the present embodiment, the coordinate system of the near-surface defect detection device 80 and the marking device 81 is made common, the respective coordinate information is shared, an instruction for controlling the drive system of the sample stage of each device is issued, , And has a configuration in which each device is connected to each device via a network 83.

The target wafer 5 is marked on the marking device 81.
And, by specifying the coordinate information of the defect of interest detected by the near-surface defect detection device 80, the stage of the marking device 81 is driven so that the defect portion of interest can be quickly captured in the field of view. Can be. Further, a configuration in which processing can be performed continuously based on a mark shape registered in the calculation processor 82 in advance may be employed. Further, the marking device 81 may also serve as a sample preparation device. The marking device and the sample preparation device are made independent, and the above three devices are controlled by a common computer.
A configuration for collecting data may be used.

Further, the conventional focused ion beam apparatus can irradiate the wafer with lasers having different wavelengths, and has a structure in which a mechanism for focusing and dispersing the scattered light and sending it to the detector is provided, and it is obtained from the scattered light of each wavelength. The computer that stores the coordinates of the existence of the near-surface defect also serves as a control of the focused ion beam and the control of the sample stage, so that the focused ion beam can be irradiated to the vicinity of the detected near-surface defect.

Further, in the above-described near-surface defect detecting apparatus (for example, OSDA apparatus), the sample stage can be set in a vacuum vessel, and at least the focused ion beam irradiation optical system, the secondary By installing an electron detector and a deposition gas supply source, and installing a control system and a computer processing unit that can move the detected defects among the detected defects into the focused ion beam irradiation area on the sample stage, Can be marked with a focused ion beam in the vicinity of.

<Embodiment 6> In this embodiment, a defect near the surface of a semiconductor wafer is detected at an early stage of manufacturing, and based on the detection result, it is determined whether or not the wafer can be used for manufacturing a semiconductor device. The present invention relates to a semiconductor wafer management method for performing judgment, that is, screening.

First, a database must be created. For a large number of semiconductor wafers, at least one of the number, in-plane and depth distribution, position, shape, and type of near-surface defects is stored in a computer using a surface defect detection device. On the other hand, with respect to the near-surface defect, for the sample prepared by the method of the present invention described above, T
The shape is observed by EM or STEM, and the results are stored in a computer as many conditions as possible, for example, the depth from the wafer surface, the density of the wafer, the periphery and center of the wafer, and the size of the defect. From the many data stored in this way, the result obtained by the surface defect detection device and the shape observation result by TEM or STEM are correlated. For example, a tendency such as "defects distributed at a certain depth have many octahedral vacancies" is made into a database and stored in the calculation processing unit.

On the other hand, before manufacturing a semiconductor device, surface defects are detected on a wafer to be used. At least one of the number, in-plane and depth distribution, position, shape, and type of the near-surface defects obtained at that time is stored in the calculation processing unit together with the wafer identification number. The operation of the finally completed semiconductor device is confirmed by a device operation measuring device such as a prober inspection. At this time, a correlation is obtained between a chip having a good operation and a chip having a bad operation and each evaluation item at a corresponding place obtained by surface defect detection. Further, for a chip that malfunctions, the result of the defect detection, the correlation between the defect and the result of electron microscopic observation, and the cause of the malfunction, such as the withstand voltage failure of the insulating film, are stored in association with each other.

In this way, the correlation between the surface defect detection results for a number of wafers, the electron microscopic observation results of the defects, the results of good and bad operation of the semiconductor device, and the cause of the defective portion is stored in the calculation processing unit as a database. Save it.

On the basis of such a database, at a semiconductor device manufacturing site, surface defects are desirably detected for all the wafers used for manufacturing. The number of near surface defects obtained at this time, in-plane and depth distribution, position,
When at least one of the shape and type data is taken into the computer and compared with the database stored so far, one wafer when the semiconductor device manufacturing process is continued using the measured wafer thereafter. It is possible to roughly estimate the yield of good products per unit. A comparison of the cost associated with the production of a single wafer and the expected profits that can be obtained from the production reveals that if the number and distribution of defects exceeds a certain level, it will be harmful to produce them. Wafer can be screened. This screening improves the final yield of good semiconductor devices.

[0124]

According to the present invention, defects near the surface detected over a wide area of the wafer can be easily marked, and a portion containing the defects near the surface can be easily extracted and thinned. For this reason, the focused surface near defect can be analyzed in detail in a short time.

[Brief description of the drawings]

FIG. 1 is a plan view showing a schematic configuration of a sample manufacturing apparatus according to an embodiment of the present invention.

FIG. 2 is a longitudinal sectional view for explaining the principle of a conventional OSDA device.

FIG. 3 is a partial plan view of a wafer surface showing an embodiment of markking according to the present invention.

FIG. 4 is a partial plan view of the wafer surface showing an embodiment of markking according to the present invention.

FIG. 5 is a partial plan view of the wafer surface showing a step of forming a second mark according to the present invention.

FIG. 6 is a perspective view showing a procedure of a step of extracting a micro sample according to one embodiment of the present invention.

FIG. 7 is a perspective view showing a procedure of a step of forming a cross section of a micro sample according to one embodiment of the present invention.

FIG. 8 is an essential part perspective view showing the configuration of a defect detection unit according to one embodiment of the present invention.

FIG. 9 is an essential part perspective view showing the configuration of a mark detection unit according to one embodiment of the present invention.

FIG. 10 is a block diagram of a sample preparation apparatus according to one embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Sample preparation apparatus, 2 ... Defect detection part, 3 ... Marking part, 4 ... Wafer handling part, 5 ... Wafer, 6 ... Wafer tray, 7 ... Sample stage, 8 ... Laser light, 9A, 9
B: light source, 15: computer, 16: load lock room,
17: vacuum container, 18: FIB irradiation unit, 19: sample stage, 22: display, 23: sample preparation unit, 24:
Sample transfer unit, 27: Holder cassette, 28: Probe,
41: mark, 42A: near surface defect, 44: plane TE
M sample image, 46: second mark, 5A: wafer piece, 5B
... Plane TEM sample, 5C ... micro flake sample, 63 ... protective film, 66 ... probe, 70 ... sample holder, 74 ... cross section sample.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H01L 21/66 H01L 21 / 66H G01N 1/28 FGN 21/88 645A (72) Inventor Satoshi Tomimatsu 1-280 Higashi-Koigakubo, Kokubunji-shi, Tokyo Inside Hitachi, Ltd. Central Research Laboratory (72) Inventor Kyoko Minawa 1-280 Higashi-Koikekubo, Kokubunji-shi, Tokyo F-term in Hitachi, Ltd. Central Research Laboratory 2G001 AA03 BA07 BA11 CA03 FA06 HA09 JA15 KA03 PA03 RA01 RA04 RA06 2G051 AA51 AB07 BA08 BA10 CA02 CA07 CB05 DA08 EA14 EB01 EB02 FA10 4M106 AA01 BA02 BA03 BA05 BA11 CA70 DA01 DA05 DA20 DH01 DH12 DH24 DH32 DH57 DH60 DJ01 DJ03 DJ04 DJ04

Claims (39)

[Claims] 1. A defect detecting step for detecting a defect near the surface of a wafer, and a marking step for marking an identification mark near a desired defect among the surface near defects detected by the defect detecting step. Characteristic sample preparation method. A defect detecting step of detecting a defect near the surface of the wafer and storing coordinate information of the defect; a marking step of making a mark for identifying a desired near surface defect based on the coordinate information; A sample manufacturing method, comprising: a processing step of processing the wafer on a test piece suitable for performing any one of observation, measurement, and analysis of a defect based on the mark. 3. The sample manufacturing method according to claim 1, wherein the defect detecting step scans the laser light over a wide area of the surface of the wafer and relatively to the wafer, and scans the laser light from the wafer. A method for receiving a scattered light to detect an existing position of the near-surface defect. 4. The sample manufacturing method according to claim 1, wherein said defect detecting step includes a step of applying a plurality of laser beams having different wavelengths over a wide area on the surface of said wafer and relative to said wafer. A method of scanning and detecting a position of the near-surface defect using scattered light generated from near the surface of the wafer. 5. The sample preparation method according to claim 1, wherein the marking step is performed using at least one of a laser beam, an ion beam, an electron beam, and a mechanical probe. Production method. 6. A method according to claim 1, wherein said marking step is performed by drawing or processing by irradiation with a focused ion beam or a projected ion beam. A method for producing a sample, comprising processing to form a damaged region. 7. The sample manufacturing method according to claim 1, wherein the marking step includes forming an adhering substance containing components in an atmosphere around the defect portion on the wafer surface by electron beam irradiation. A method for producing a sample, comprising: 8. The method according to claim 1, wherein the marking step includes a step of mechanically forming a concave portion or a damaged region on the wafer with a sharpened metal probe or diamond. A sample processing method comprising: 9. The method according to claim 2, wherein the test piece is a test piece for observation or measurement with a transmission electron microscope or a scanning transmission electron microscope. 10. The method according to claim 9, wherein:
A sample manufacturing method, characterized in that processing is performed such that a surface of the test piece for observing or measuring a defect is perpendicular to a surface of the wafer. 11. The method according to claim 9, wherein:
A sample manufacturing method, wherein a surface of a test piece for observing or measuring a defect is processed so as to be parallel to a surface of the wafer. 12. The method according to claim 10, wherein the test piece is manufactured by cutting the wafer with a dicing apparatus and processing with a focused ion beam. 13. The method according to claim 2, wherein:
The processing step is an extraction step of extracting a micro sample including the desired near-surface defect from the wafer using at least ion beam processing with reference to the mark, and transferring the micro sample to an analysis sample holder. A sample manufacturing method, comprising: a fixing step of fixing; and a finishing step of further processing the micro sample with the ion beam so as to make the shape easy to analyze. 14. The method according to claim 2, wherein:
The processing step is a step of cutting or cleaving the wafer based on the mark to form a small piece including the mark, and a thinning step of processing the small piece into a thin sample having a thin portion substantially parallel to the wafer surface. A method for preparing a sample, comprising: 15. The method for preparing a sample according to claim 14, wherein after the thinning step, the thin section is further inspected with a transmission electron microscope or a scanning transmission electron microscope to determine whether the mark and the desired near-surface defect are present. A step of obtaining a relative positional relationship; a step of extracting a part including the near-surface defect from the thin section based on the above-described positional relationship and fixing the same to a sample holder for analysis; And a cross-section processing step of processing into a thin section substantially perpendicular to the wafer surface. 16. The method according to claim 2, wherein:
The processing step comprises cutting or cleaving the wafer based on the mark to form a small piece including the mark, and producing a thin sample having a thin section substantially parallel to a wafer surface of the small piece; Determining the relative positional relationship between the mark and the desired near-surface defect by observation with a transmission electron microscope or a scanning transmission electron microscope; and A sample manufacturing method, comprising: a second marking step of applying a second mark for identifying a position of a defect in detail. 17. The method according to claim 16, wherein the second marking step is performed using at least one of a pulse laser, a focused ion beam, a projected ion beam, and an electron beam. Sample preparation method. 18. The method according to claim 16, wherein the second mark is at least one of a dot, a line segment, and a set of symbols arranged in a straight line or a lattice point. Sample preparation method. 19. The method for preparing a sample according to claim 16, wherein the thin sample provided with the second mark is observed with a transmission electron microscope or a scanning transmission electron microscope to obtain a desired one. A sample preparation method, comprising: clarifying a relative positional relationship between the near-surface defect and the second mark and specifying a position where the near-surface defect exists. 20. The sample preparation method according to claim 19, wherein after the position determining step, the desired near-surface defect is included in the desired near-surface defect, and the near-surface defect is cross-sectionally viewed with a transmission electron microscope or a scanning transmission electron microscope. A sample preparation method, comprising: a cross-sectional thinning step of processing into a cross-sectional thin sample for observation by the method. 21. (a) a defect detecting step of detecting a defect near the surface of the wafer by laser beam irradiation; (b) a storing step of storing coordinate information of the defect near the surface; and (c).
A marking step of making a mark in the vicinity of the surface near defect of interest out of the surface near defects, (d) cutting the wafer into small pieces containing the defect of interest and the mark, and (e) removing the small piece from the wafer. And (f) an observation step of observing the near-surface defect of interest with reference to the mark using a transmission electron microscope or a scanning transmission electron microscope. A method for analyzing near-surface defects, characterized by the following. 22. The method according to claim 21, further comprising: (a) a second marking step of applying a more detailed second mark to the flake sample based on the observation result in the observation step. (B) an observation step of observing the sample provided with the second mark with a transmission electron microscope or a scanning transmission electron microscope, and (c) an observation result obtained by the transmission electron microscope or the scanning transmission electron microscope. (C) determining an observation region including the defect of interest;
Extracting a micro sample including the determined area,
(E) a cross-sectional thinning process of processing the micro sample into a cross-sectional thin sample containing the desired near-surface defect, and (f) a process of observing the cross-sectional thin sample with a transmission electron microscope or a scanning transmission electron microscope. A method of analyzing near-surface defects. 23. The method for analyzing near-surface defects according to claim 21 or 22, wherein a desired one of the defects obtained in said defect detection step is transmitted from a direction perpendicular to a surface of said wafer and in a cross-sectional direction. A method for analyzing near-surface defects, characterized in that the near-surface defects are analyzed by observation with a microscope or a scanning transmission electron microscope. 24. At least one of the position, number, distribution, and type of near-surface defects obtained by detecting near-surface defects of the semiconductor wafer, and the shape observation result of the near-surface defects by electron microscope observation; The correlation between at least one of the occurrence position, the number of occurrences, and the form of the failure of the semiconductor device formed on the wafer is stored in a database, and a defect near the surface of the semiconductor wafer is periodically detected during the manufacture of the semiconductor device. A method for manufacturing a semiconductor device, wherein the semiconductor wafer is screened on the basis of a detection result obtained by performing the method and the database. 25. A sample preparation apparatus comprising at least a defect detection unit for detecting a near-surface defect of a wafer, and a marking unit for marking a vicinity of a desired defect among the near-surface defects. 26. A defect detection unit for detecting a defect near the surface of a wafer, a marking unit for marking a vicinity of a desired defect among the defects near the surface, and at least one of defects detected by the defect detection unit. A sample processing unit for extracting a micro sample containing a defect to be processed and processing the sample into a shape suitable for mounting on at least one of an analyzer, an observation device, and a measurement device. 27. A defect detector for detecting a defect near the surface of a wafer, a mark placed near a desired defect among the defects near the surface, and a minute sample containing the desired defect based on the mark. A sample processing unit for extracting a sample and processing the sample into a sample piece having a shape suitable for an analysis device, an observation device, or a measurement device. 28. The sample preparation apparatus according to claim 26, wherein the sample processing section has a sample moving means for moving the extracted micro sample to another member. 29. The sample manufacturing apparatus according to claim 25, wherein the defect detecting section and the marking section, or the defect detecting section and the marking section and the sample processing section, or the defect A sample preparation apparatus, comprising: a wafer handler for taking a wafer in and out of a detection unit and any one of the sample processing units. 30. The sample preparation apparatus according to claim 25, wherein the defect detection unit and the marking unit, or the defect detection unit and the marking unit and the sample processing unit, or the defect. A sample preparation apparatus, comprising: a transfer path through which the wafer moves between at least one of a detection unit and the sample processing unit. 31. The sample preparation apparatus according to claim 25, further comprising a calculation processing unit for storing at least coordinate information of the defect detected by said defect detection unit. Production equipment. 32. The sample preparation apparatus according to claim 25, wherein the defect detection unit generates a laser and irradiates the wafer with the laser, and detects scattered light from the wafer. And a light detecting means for performing the measurement. 33. The sample preparation apparatus according to claim 25, wherein said defect detection section generates a plurality of lasers having different wavelengths from each other and irradiates said wafer with said plurality of lasers. A sample preparation apparatus, comprising: at least a plurality of light detection means for detecting scattered light from a wafer for each wavelength. 34. The sample preparation apparatus according to claim 25, wherein the marking section is at least one of a laser beam irradiation optical system, an ion beam irradiation optical system, an electron beam irradiation optical system, and a mechanical probe. A sample preparation apparatus comprising any one of the above. 35. A sample preparation apparatus according to claim 34, wherein said ion beam irradiation optical system is one of a focused ion beam irradiation unit and a projected ion beam irradiation unit. 36. A near-surface defect detecting device for detecting a near-surface defect of a wafer and storing coordinate information of the detected defect, and an analyzing device or an observing device for extracting a small sample piece containing a desired defect from the wafer. Alternatively, a sample preparation device that processes a sample piece having a shape suitable for one of the measurement devices, a transmission electron microscope or a scanning transmission electron microscope, and a wafer name, defect coordinates, a defect distribution, a sample piece name, and a defect image. A near-surface defect analysis system, wherein a computer storing at least one of the data is connected to each other via a network. 37. A laser light irradiator for generating laser light of at least one wavelength and irradiating the wafer with a light, a light detector for detecting scattered light from the wafer, and information on coordinates at which the scattered light is generated. A near-surface defect detection apparatus, comprising: at least a calculation processing means for storing; and a means for irradiating the wafer with an ion beam. 38. A laser light irradiator for generating laser light of at least one wavelength and irradiating the wafer with light, a light detector for detecting scattered light from the wafer, and a coordinate information generating the scattered light. A defect near the surface, comprising: a calculation processing means for storing; and a laser light irradiation means for a laser mark for performing a laser mark near the light emitting part based on the coordinate information of the desired scattered light generation part. Detection device. 39. An ion beam apparatus for irradiating a part of a wafer with an ion beam, a laser beam irradiator for generating a laser beam and irradiating the wafer, and a light detecting means for detecting scattered light from the wafer. A focused ion beam apparatus, comprising: a near-surface defect detection unit having at least a calculation processing means for storing coordinate information that has generated the scattered light.