JP2015220296A - Contamination evaluation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a contamination evaluation method that can evaluate even a minute amount of metal pollution introduced in a wafer with high sensitivity.SOLUTION: A cleaned silicon wafer is immersed in chemical liquid in which metal impurities are dissolved, and a defect (pit, projection) derived from metal pollution existing on the surface of the wafer are actualized (S3). Subsequently, the defect on the wafer surface is detected by a surface inspection device (S4), and SEM-EDX analysis is executed on the defect (S5, S6). The defect is classified in terms of defect seed on the basis of the analysis result (S7). The cross-sections of the defect is observed by TEM, and the depth or height of the defect is determined on the basis of an observation image of TEM. Subsequently, the volume of the defect is determined on the basis of the surface area of the defect determined from an SEM observation image and the depth or height of the defect determined from the TEM observation image (S9). The metal atom density in the defect is further determined. The metal pollution concentration in the wafer is calculated on the basis of the number of defects classified in terms of a predetermined defect seed and the volume and metal atom density of the defect (S11, S12).

Description

  The present invention relates to a method for evaluating metal contamination introduced into a wafer.

  Conventionally, various methods for evaluating metal contamination introduced into a wafer (for example, a silicon wafer) are known. For example, Patent Document 1 describes that surface analysis is performed by SEM-EDX (scanning electron microscope-energy dispersive X-ray analyzer) or the like in order to grasp metal contamination caused by chemical contact with a wafer. Yes.

  In addition to the method described in Patent Document 1, as a means for assuring and guaranteeing the surface quality of a wafer, there is an ultra-sensitive surface inspection apparatus represented by SP-3 manufactured by KLA-Tencor and MAGICS manufactured by Lasertec. For example, Patent Document 2 uses this ultrasensitive surface inspection device to inspect minute defects such as particles and surface defects.

  Conventionally, wafer quality was evaluated by qualitative and quantitative evaluation of metal contamination above the measurement lower limit by chemical analysis. Further, metal contamination at a concentration lower than the detection limit of chemical analysis is evaluated as a total amount of metal contamination and defects that trap electrons by using a measurement technique such as a wafer lifetime method (WLT method) or an SPV method.

  Further, for example, metal contamination brought in from raw material polysilicon or quartz crucible in single crystal production using the CZ method varies depending on the contamination status of the raw material. Further, in the single crystal pulling step, the amount of contamination at the growth interface changes every moment depending on the convection of the melt, the amount of the pulled crystal and the remaining amount of the raw material remaining in the crucible. Furthermore, the metal moves (diffuses) in the crystal due to various factors such as the growth direction and diameter direction of the pulled single crystal and the temperature gradient of the pulling device, and exists as silicide and precipitates at a stable position. . This metal distribution is very difficult to predict in advance. For this reason, conventionally, quantitative qualitative evaluation of pollutants has been performed, for example, by chemically analyzing a manufactured single crystal block or a product wafer that has undergone a wafer processing step.

JP 2009-139148 A JP 2005-063984 A

  By the way, of the particles and minute defects present on the wafer that have been clarified by the surface inspection using the above-described ultrasensitive surface inspection apparatus, pits and protrusions on the surface (hereinafter referred to as metal-derived defects) due to metal impurities. There are not a few things that form. This metal-derived defect causes a difference in test results depending on the size of the defect (pits, protrusions) and the presence or absence of metal in the electrical property test. In the conventional inspection apparatus, there is no means for classifying the defect of crystal defect type represented by COP and the defect derived from metal.

Further, although the conventional chemical analysis is also a difference depending on the accuracy of the elements (metal impurities) and measuring an apparatus to be detected, for example for the ^{Ni, 1 × 10 10 (atoms} / cm 3) following analysis It is said that it is difficult to do. The WLT method and the SPV method have an analytical sensitivity of 1 × 10 ⁸ (atoms / cm ³ ) when only Fe is viewed, but qualitative quantification of other elements has not been guaranteed so far. In image sensor devices typified by CCD and CIS, it is said that metal contamination at a concentration below the detection lower limit deteriorates device characteristics.

  This invention is made | formed in view of the said situation, and makes it a subject to provide the contamination evaluation method which can evaluate the trace amount metal contamination introduced into the wafer with high sensitivity.

In order to solve the above problems, the contamination evaluation method of the present invention includes a detection step of detecting defects derived from metal contamination present on the surface of the wafer,
An acquisition step of acquiring the metal contamination state of each defect detected in the detection step;
A calculation step of calculating a metal contamination concentration of the wafer based on the number of the defects detected in the detection step and the metal contamination state obtained in the acquisition step;
It is characterized by providing.

  Thus, in the present invention, focusing on the fact that when a metal is introduced into the wafer, defects derived from the metal can be formed on the wafer surface, the defect evaluation (number of defects and metal contamination state) Based on this, the metal contamination concentration of the wafer is calculated. As a result, even a trace amount of metal contamination, for which the difference could not be revealed by conventional chemical analysis, can be evaluated with high sensitivity.

  The detection step includes a chemical solution supply step for supplying a chemical solution in which metal impurities are dissolved on the surface of the wafer, and an inspection step for detecting the defect by inspecting the surface of the wafer after the chemical solution supply step is performed. With.

  As a result, the metal-derived defects present on the surface of the wafer can be made apparent. And since the inspection process is implemented with respect to the exposed wafer surface, the detection sensitivity of a metal origin defect can be improved.

The detection step includes
An analysis step of performing component analysis of the detected defect;
And a classification step of classifying the defects into defect types based on the analysis result of the analysis step. In the analysis step, SEM-EDX analysis is performed on the defect.

  According to this, since a defect on the wafer surface is first detected and component analysis is performed on the detected defect, a metal-derived defect can be detected with high accuracy. In particular, by performing SEM-EDX analysis, it is possible to analyze the metal contamination of the defect while observing an image of the defect by SEM. Furthermore, since defects are classified for each defect type based on the analysis result, the metal contamination concentration can be evaluated for each defect type. In addition, by creating a map of localized positions and distributions of the classified defect types (for example, metal impurity types and defect shape types), various crystal manufacturing process histories such as crystal defects and precipitates and subsequent It is possible to obtain information reflecting thermal history characteristics or crystal position characteristics.

  In the calculation step, the metal contamination concentration of the wafer is calculated based on the number of the defects classified into the predetermined defect type in the classification step and the metal contamination state of the defects. Thereby, the metal contamination concentration belonging to a predetermined defect type can be evaluated.

  Moreover, the acquisition step includes a size acquisition step of acquiring the defect size as the metal contamination state. In this case, the size acquisition step includes a surface area acquisition step of acquiring a surface area of the defect, a depth / height acquisition step of acquiring a depth or height of the defect, and the surface area and the depth or height. And a volume calculating step of calculating the volume of the defect as the size.

  It is considered that the larger the size (volume) of the metal-derived defect, the higher the amount of metal contamination in the metal-derived defect. Therefore, by obtaining the size, a value correlated with the metal contamination concentration of the wafer can be obtained.

  Further, in the depth / height acquisition step, a cross section of the defect may be observed with a TEM, and the depth or height may be acquired based on the observed image, or the wafer having the defect may be acquired. The depth or height may be acquired based on the penetration length of a laser (specifically, for example, a laser of an inspection apparatus that performs surface inspection in an inspection step).

  When acquiring the depth or height by TEM, the cross section of all the defects may be observed by TEM, and the depth or height of each defect may be acquired based on each observation image, One or a plurality of defects are selected, the depth or height of the selected defects is obtained as a representative value based on an observation image by TEM, and the depth or height of the other defects is obtained as the representative value. May be substituted. When observing the cross section of all the defects with a TEM, an accurate depth or height can be obtained. Moreover, when substituting with the depth or height (representative value) of a typical metal origin defect, the depth or height of each defect can be obtained simply.

  Moreover, the acquisition step includes a density acquisition step of acquiring the density of metal impurities in the defect as the metal contamination state. According to this, since the density of the metal impurity in the defect is acquired in the density acquisition step, the metal contamination concentration of the wafer can be easily estimated by using this density.

Further, the detection step includes a condition determination step for determining, by simulation, the conditions of the acceleration voltage of the electron beam in the SEM, which is suitable for detecting the metal impurity to be evaluated.
In the analysis step, SEM-EDX analysis is performed under the acceleration voltage condition determined in the condition determination step.

  Thereby, the metal-derived defect derived from the metal impurity to be evaluated can be detected with high sensitivity. As a result, the metal contamination concentration to be evaluated can be evaluated with high sensitivity in a manner that distinguishes it from other types of metal contamination concentrations.

  The calculation step includes a surface contamination calculation step of calculating a metal contamination concentration on the surface of the wafer based on the number of defects and the metal contamination state. As a result, the metal contamination concentration on the wafer surface can be evaluated with high sensitivity and simplicity.

  In addition, the calculation step calculates the number of bulk defects that are defects derived from metal contamination present in the bulk of the wafer and the metal contamination state of the bulk defects based on the number of defects and the metal contamination state. An estimation step for estimation, and a bulk contamination calculation step for calculating a metal contamination concentration in the bulk of the wafer based on the number and metal contamination state obtained in the estimation step. As a result, the metal contamination concentration in the bulk of the wafer can be evaluated with high sensitivity and ease.

It is the flowchart which showed the procedure of the metal contamination evaluation of a wafer. It is the figure which illustrated the pit which became obvious by chemical | medical solution immersion. It is the figure which illustrated the pit which became obvious by chemical | medical solution immersion. It is the figure which showed the result of having simulated the penetration depth of the electron beam to the silicon | silicone for every acceleration voltage with commercial software. It is the figure which illustrated the observation image of the projection part by SEM. It is the figure which illustrated the observation image of the pit by SEM. It is the figure which showed the EDX analysis result with respect to the projection part of FIG. It is the figure which showed the EDX analysis result with respect to the pit of FIG. It is the figure which illustrated the observation image of cross section TEM with respect to the pit of FIG. It is the figure which illustrated the observation image of cross section TEM with respect to the pit of FIG. It is the figure which showed the result of the TEM-EDX analysis. It is the figure which showed the result of SEM-EDX analysis. It is the figure which showed the defect classification map. It is the figure which showed the calculation result of Ni contamination density | concentration.

  Embodiments of the present invention will be described below with reference to the drawings. In this embodiment, among various metal contaminations introduced in the silicon single crystal pulling process, slicing process, grinding, lapping process, epitaxial growth process, etc., for cleaning and inspection performed up to the manufacturing process and inspection process. A method for evaluating metal contamination that causes pits and protrusions to appear on the wafer surface by pretreatment will be described.

  FIG. 1 is a flowchart showing a procedure for evaluating metal contamination of a wafer in this embodiment. First, a silicon wafer to be evaluated (hereinafter simply referred to as a wafer) is prepared (S1). The wafer prepared here may be prepared at any stage, such as a wafer before the epitaxial growth process or a product wafer after the epitaxial growth process. Here, for example, a wafer (PW: Polished Wafer) that has undergone a mirror finishing process for polishing (POLISH) the surface of the wafer is prepared. Further, the characteristics of the prepared wafer such as diameter, plane orientation, and resistivity may be any characteristics.

  Next, the wafer prepared in S1 is cleaned for the purpose of particle removal (S2). This cleaning is a method for cleaning a semiconductor wafer, such as SC1 cleaning or SC2 cleaning, and is not particularly limited as long as particles can be removed.

  Next, the cleaned wafer is immersed in a chemical solution that dissolves a metal element to be evaluated or a metal and silicon compound (S3). That is, a chemical solution is supplied to the surface of the wafer. When the metal impurity to be evaluated is a precipitate of Ni or Cu, or a compound (silicide) of the metal (Ni, Cu) and Si, for example, HF is used as the chemical solution, but is not limited thereto. Further, the form of supplying the chemical solution to the wafer surface is not limited to immersion, and may be a form such as peristalsis, rest, pouring, and spraying.

  2 and 3 show an example in which defects (pits) formed due to the influence of metal impurities existing on the wafer surface are revealed in the step S3. As shown in FIGS. 2 and 3, by performing step S <b> 3, the metal impurity in the defect is dissolved, and the metal-derived defect can be revealed on the wafer surface. Therefore, a metal-derived defect can be detected with high sensitivity in the next step S4. Note that some metal impurities remain in the defects after the step S3. The step S3 corresponds to the “chemical solution supply step” of the present invention.

  Next, using a highly sensitive surface inspection apparatus, defect inspection is performed on the wafer surface after step S3 (S4). Then, by this defect inspection, the number of defects (pits, protrusions) existing on the wafer surface and the coordinates (position) of each defect are obtained. At this time, the size of the pits and protrusions is often about 50 nm or less. Therefore, it is desirable that the surface inspection apparatus has a sensitivity in accordance with the assumed defect size (about 50 nm or less). Specifically, for example, an ultrasensitive surface inspection apparatus represented by SP-3 manufactured by KLA-Tencor or MAGICS manufactured by Lasertec can be used. The step S4 corresponds to the “inspection step” of the present invention.

  Next, the type of metal impurity to be evaluated is determined, and the metal impurity can be detected with the highest sensitivity by EDX analysis (Energy Dispersive X-ray Spectroscopy). The condition of the acceleration voltage of the electron beam of Scanning Electron Microscope) is obtained by simulation (S5). The simulation method may be any method based on a physical phenomenon, such as commercially available computer simulation software or a simulation using a unique computer.

  FIG. 4 shows the result of simulating the penetration depth of an electron beam into silicon for each acceleration voltage using commercially available software. In the simulation of FIG. 4, information (metal element, thickness of the metal layer) of the metal impurity to be evaluated is reflected. FIG. 4 simulates a cross section of silicon containing metal impurities to be evaluated, and the horizontal line at the top of FIG. 4 indicates the wafer surface. The center of the horizontal line is irradiated with an electron beam, and the energy of the electron beam in the wafer is expressed by color shading. Also, the vertical axis in FIG. 4 indicates the depth from the wafer surface. In FIG. 4, from the left, the acceleration voltage of the electron beam is 0.8 kV, 1.0 kV, 3.0 kV, and 5.0 kV, and the results of 0.8 kV and 1.0 kV are respectively shown in enlarged views below. Yes.

  As shown in FIG. 4, the penetration depth R of the electron beam increases as the acceleration voltage increases. Specifically, in the example of FIG. 4, when the acceleration voltage is 0.8 kV, the penetration depth R = 12 nm, when the acceleration voltage is 1.0 kV, the penetration depth R = 18 nm, and the acceleration voltage is 3.0 kV. In some cases, the penetration depth R = 140 nm, and when the acceleration voltage is 5.0 kV, the penetration depth R = 260 nm.

  The penetration depth of the electron beam with respect to this acceleration voltage varies depending on the type of metal element and the thickness of the metal layer. Therefore, in S5, for example, a simulation is performed for each metal impurity (element, thickness), and an optimum acceleration voltage condition corresponding to the type of metal impurity is obtained. Table 1 below exemplifies the optimum value of the acceleration voltage for each representative metal element and thickness of the metal layer obtained by simulation. In Table 1, Kα and Lα1 indicate the types of characteristic X-rays emitted from the metal element by the electron beam of SEM. EDX analysis is performed with the X-rays of Kα and Lα1.

  The time when the simulation of S5 is performed is not limited to the time of FIG. 1 (the time between S4 and S6), and may be any time before the SEM-EDX analysis of S6. For example, assuming that Ni having a thickness of 150 nm to 250 nm is included in the defects detected in S4, in S5, from Table 1, the metal element is Ni and the acceleration voltage having a thickness of 150 nm to 250 nm is obtained. The condition of 2.0 kV is determined as the current acceleration voltage. In S5, when there are a plurality of types of metal impurities to be evaluated, an acceleration voltage condition (optimum value) is determined for each metal impurity. The step S5 corresponds to the “condition determining step” of the present invention.

  Next, SEM-EDX analysis is performed on the defects detected in S4 under the acceleration voltage conditions determined in S5 (S6). That is, an electron microscope equipped with an EDX analyzer is used to irradiate an electron beam to the defect coordinate position detected in S4 on the wafer surface, and defects (pits, protrusions) are generated by secondary electrons generated by the irradiation. Observe the image. At this time, the defect shape (pit or protrusion) is specified, and the defect size (surface area) is calculated. Simultaneously with SEM observation, EDX analysis of defects is performed based on characteristic X-rays generated by electron beam irradiation. At this time, it is desirable to acquire a secondary electron image as high as possible with a detector that detects secondary electrons. Further, it is desirable to perform EDX analysis by applying an electron beam to a portion having the highest contrast in the secondary electron image obtained by SEM. This makes it possible to perform EDX analysis for defects reliably.

  In S6, when there are a plurality of types (metal elements, thicknesses) of metal impurities to be evaluated, the acceleration voltage is changed according to the types of metal impurities to be detected, and the same defect is detected. Perform multiple EDX analyses. For example, when it is desired to detect all Ni of each thickness, EDX analysis is performed once at 2.0 kV (see Table 1) which is the optimum value of the acceleration voltage for detecting Ni of 250 nm or less, and 250 nm or more is detected. EDX analysis is performed once at 10.0 kV (see Table 1), which is the optimum value of the acceleration voltage for detecting Ni. That is, EDX analysis is performed twice for the same defect. Further, for example, when it is desired to detect all the thicknesses of Ni and each thickness of Mo, the optimum values of acceleration voltages corresponding to Ni and Mo (2.0 kV, 10.0 kV, 3.0 kV, 20 0.04 kV (see Table 1)) for a total of 4 EDX analyses.

  Further, in S6, SEM-EDX analysis may be performed on all the defects detected in S4, or one or more representative defects are selected from the defects detected in S4, and the selected representatives are selected. The SEM-EDX analysis may be performed only on the defect, and the analysis on the other defects may be substituted with the result of the SEM-EDX analysis on the representative defect. When SEM-EDX analysis is performed on all defects, the defect type can be more accurately classified in the later-described step S7, and the metal contamination concentration in the wafer can be more accurately calculated in the steps S11 and S12. it can. When SEM-EDX analysis is performed for only typical defects, defect types can be easily classified, and the metal contamination concentration can be easily calculated.

  Here, FIGS. 5 and 6 are views exemplifying observation images obtained by SEM in S6. FIG. 5 shows an observation image of a protrusion (convex portion), and FIG. 6 shows an observation image of a pit (concave portion). Is shown. 7 and 8 are diagrams illustrating the results of the EDX analysis of S6, FIG. 7 shows the EDX analysis results for the protrusions in FIG. 5, and FIG. 8 shows the EDX analysis results for the pits in FIG. ing. As shown in FIGS. 5 and 6, by observing the defect with the SEM, the shape (whether convex or concave) and size (surface area) of the defect can be easily specified. Further, as shown in FIGS. 7 and 8, the component in the defect can be easily identified by performing the EDX analysis of the defect. For example, in the example of FIG. 7, it can be seen that Ni is contained as a metal impurity in the defect. Moreover, in the example of FIG. 8, it turns out that Cu is contained as a metal impurity in a defect. Further, in the examples of FIGS. 7 and 8, it is also understood that C is included in the defect in addition to the metal impurity. 7 and 8, the largest peak indicates Si. The process of S6 corresponds to the “analysis process” and the “surface area acquisition process” of the present invention.

  Next, based on the result of the SEM-EDX analysis in S6, the individual defects detected in S4 are classified for each defect type (S7). Specifically, for example, the defect is classified according to a metal-derived defect derived from metal contamination or another defect (crystal defect type defect), or the shape of the defect is convex (projection) or concave (pit). Or by the size (surface area) of the defect, or by the metal element detected by EDX analysis. As a result, information on the specific defect type (number of defects, position, etc.) can be easily obtained. In addition, a distribution map of the classified defect types can be prepared, and it is possible to easily evaluate which defect type is distributed on the wafer by this distribution map. The process of S7 corresponds to the “classification process” of the present invention.

  Next, a cross-section observation (cross-sectional TEM observation) is performed on a defect classified as a defect type to be evaluated in the step of S7 (for example, a defect containing Ni) by TEM (Transmission Electron Microscope). (S8). At this time, cross-sectional TEM observation may be performed on all defects classified as the defect type to be evaluated, or only one or more representative defects are selected and only the representative defects are selected. Cross-sectional TEM observation may be performed.

  9 and 10 are diagrams illustrating the observation images of the cross-sectional TEM observation of S8, FIG. 9 shows the observation image of the pits of FIG. 6, and FIG. 10 shows the observation image of the pits of FIG. As shown in FIGS. 9 and 10, by performing cross-sectional TEM observation, the cross section of the pit can be easily observed, and the depth of the pit can be easily estimated based on the observation image of the cross section. Further, by performing cross-sectional TEM observation on the protrusion, the height of the protrusion can be easily estimated. In step S8, the depth of the pit and the height of the protrusion are calculated based on the observation image of the cross-section TEM. At this time, when cross-sectional TEM observation is performed on all the defects, the depth or height of all the defects is calculated based on each observation image. Thereby, a highly accurate depth or height can be obtained for each defect.

  On the other hand, when the cross-sectional TEM observation is performed only for the representative defect, the depth or height is calculated based on the observation image of the cross-sectional TEM for the representative defect. And the depth or height of the defect which is not performing cross-sectional TEM observation substitutes for the depth or height of a typical defect. At this time, when a plurality of representative defects are selected, a plurality of representative depths or heights are obtained. For example, the plurality of depths or heights are averaged, and the average value is used as another value. Substituting the depth or height of the defect. Thereby, since it is not necessary to calculate the depth or height based on the cross-sectional TEM observation for all the defects, the depth or height of each defect can be easily obtained. The process of S8 corresponds to the “depth / height acquisition process” of the present invention.

  Next, the volume of each defect is calculated based on the surface area of the defect obtained in S6 and the depth or height of the defect obtained in S8 (S9). Specifically, for example, the defect volume is calculated by multiplying the defect surface area by the depth or height. At this time, individual volumes may be calculated for all target defects, or individual volumes may be calculated only for representative defects, and the volume of other defects may be substituted with the representative defect volume. You may do it. That is, when the surface area, depth, or height is calculated for all the target defects in S6 and S8, in S9, the individual surface area, depth, or height is individually determined for all defects. Calculate the volume. Thereby, a precise volume can be obtained. On the other hand, when the surface area, depth or height is calculated only for typical defects in S6 and S8, in S9, the volume for typical defects is calculated based on the surface area, depth or height. , And the volume of other defects is substituted with this representative defect volume. Thereby, a volume can be obtained easily.

  It is considered that the defect volume (size) correlates with the degree of metal contamination in the defect. That is, the larger the defect volume, the higher the degree of metal contamination in the defect, that is, the amount of metal contamination. Thus, by acquiring the defect volume, an index correlated with metal contamination can be easily obtained. The step S9 corresponds to the “volume calculation step” of the present invention.

Next, the density of metal impurities (the density of metal atoms) in the target defect is calculated (S10). The density of the metal atoms is calculated with reference to literatures, for example, when the target metal impurity is a pure metal or a compound with silicon. Specifically, an example of a method for calculating the atomic density of Ni will be described assuming that the target metal impurity type (molecular formula) is NiSi ₂ . Based on literatures and the like, the crystal structure, lattice constant, and number of atoms per unit cell of NiSi ₂ are specified. In this case, the crystal structure of NiSi ₂ is a CaF 2 type cubic crystal and the lattice constant a = 5.406Å. Further, in the CaF2-type cubic crystal, the number of Si atoms per unit lattice is 8, and the number of Ni atoms is 4. Therefore, since the Ni atom will be present four in a volume ^{a 3} unit cell, in terms of the density of the Ni atom is Ni atoms per ^{1cm 3 (atoms / cm 3)} , 5.23 × 10 A density of ²² (atoms / cm ³ ) is obtained.

The value of the lattice constant a was obtained from a metal data book (edited by the Japan Institute of Metals). The value of the lattice constant a described in the metal data book is “K. Schbert and H. Pfisterer,
Naturwissenschefeten 37, 1950, 112-113 ”and“ K. Schbert and H. Pfisterer, Z. Metallkunde 41, 1950, 348 ”.

  The density of metal atoms in the defect is considered to correlate with the degree of metal contamination in the defect. That is, the higher the density of metal atoms, the higher the degree of metal contamination in the defect, that is, the amount of metal contamination. Thus, by acquiring the density of metal atoms, it is possible to easily obtain an index that correlates with metal contamination. The step S10 corresponds to the “density acquisition step” of the present invention.

The number of defects classified as the defect type to be evaluated, the defect volume obtained in S9, and the metal atom density obtained in S10 are all indicators that correlate with the metal contamination concentration in the wafer. That is, the number of defects, the defect volume, and the metal atom density are all indexes that increase the metal contamination concentration of the wafer as the value increases. Then, based on these defect number, defect volume, and metal atom density, the concentration of metal impurities on the wafer surface (hereinafter referred to as surface contamination concentration) is calculated (S11). Specifically, when the volume of a typical defect is obtained in S9 and the volume of another defect is substituted with the volume of the typical defect, the surface contamination concentration is calculated by the following equation 1. According to Equation 1, the surface contamination concentration can be easily obtained.
Surface contamination concentration (atoms / cm ² ) = Defect volume × Metal atom density × Defect number / Wafer surface area

On the other hand, in S9, when the volume of all the defects is obtained, the surface contamination concentration is calculated by the following equation 2. In Equation 2, x1, x2,..., Xn indicate individual defect volumes, and n in xn indicates the number of defects. By this equation 2, a highly accurate surface contamination concentration can be obtained. The step S11 corresponds to the “surface contamination calculation step” of the present invention.
Surface contamination concentration (atoms / cm ² ) = (defect volume x1 + defect volume x2... Defect volume xn) × metal atom density / wafer surface area

  Next, based on the number of defects, the defect volume, and the metal atom density, the concentration of metal impurities in the bulk of the wafer (hereinafter referred to as bulk contamination concentration) is calculated (S12). Specifically, it is assumed that defects are distributed in the bulk as in the case of defect distribution on the wafer surface. That is, it is assumed that defects exist in the bulk in terms of the number of defects on the wafer surface, defect volume, and metal atom density. In this case, the number of defects in the bulk is calculated by determining how many units the thickness of the wafer corresponds to, assuming that the defect observation region (cross-sectional observation region by TEM) on the wafer surface is one unit. Then, the number of defects in the bulk is obtained by multiplying the number of defects for one unit (number of defects on the wafer surface) by the number of units corresponding to the wafer thickness. For example, if the observation area of defects on the wafer surface is 1 μm from the wafer surface and the thickness of the wafer is 725 μm, the number of defects in the bulk can be obtained by multiplying the number of defects on the wafer surface by 725 times. .

When a typical defect volume is obtained in S9, the bulk contamination concentration is calculated by the following equation 3. According to Equation 3, the bulk contamination concentration can be easily obtained.
Bulk contamination concentration (atoms / cm ³ ) = Defect volume × Metal atom density × Number of defects in bulk / Wafer volume Equation 3

On the other hand, in S9, when the volume of all defects is obtained, the bulk contamination concentration may be calculated by the following equation 4. In Equation 4, x1, x2,..., Xn respectively indicate individual defect volumes of defects existing on the wafer surface, and n in xn indicates the number of defects on the wafer surface. Further, d is a value (number of units) indicating how many units the wafer thickness corresponds to when the observation area of the defect on the wafer surface is taken as one unit. A value obtained by dividing D1 by the defect observation region D2 on the wafer surface, that is, d = D1 / D2. Moreover, the metal atom density in Formula 4 is the density calculated | required by S10. According to Equation 4, a highly accurate bulk contamination concentration can be obtained. The process of S12 corresponds to the “estimation process” and “bulk contamination calculation process” of the present invention.
Bulk contamination concentration (atoms / cm ³ ) = (defect volume x1 + defect volume x2... Defect volume xn) × d × metal atom density / wafer volume Equation 4

  Note that both steps S11 and S12 may be performed, or only one of them may be performed. In S11 and S12, when the defect to be evaluated is a defect containing a specific metal element (for example, Ni), the contamination concentration of the specific metal element can be obtained. Moreover, when the defect of evaluation object is a defect of a specific size (surface area), the density | concentration of the metal impurity which forms the defect of the specific size can be obtained. Further, when the defect to be evaluated is a pit (concave defect), the concentration of the metal impurity forming the pit can be obtained.

  In the above embodiment, the steps S1 to S7 correspond to the “detection step” of the present invention. Steps S6 and S8 to S10 correspond to the “acquisition step” of the present invention. Steps S11 and S12 correspond to the “calculation step” of the present invention. Steps S6, S8, and S9 correspond to the “size acquisition step” of the present invention.

  As described above, according to the present embodiment, when metal contamination is introduced into the wafer, attention is paid to the fact that pits and protrusions are formed on the wafer surface by the pretreatment for cleaning and inspection. The metal contamination state (defect volume, metal atom density) of metal-derived defects (pits, protrusions) is evaluated. Then, the metal contamination concentration of the wafer is calculated based on the metal contamination state of the metal-derived defects and the number of metal-derived defects. As a result, even a trace amount of metal contamination, for which the difference could not be revealed by conventional chemical analysis, can be evaluated with high sensitivity. Further, in the WLT method and the SPV method, evaluation is performed as a total amount of metal contamination and defects that trap electrons, but according to the present embodiment, the metal contamination concentration for each defect type can be easily evaluated.

  The following experiment was conducted to confirm the effect of the present invention. As an example, three silicon wafers having undergone a mirror polishing process were prepared, and each wafer was immersed in HF 1.0% (chemical solution having a mass percentage of HF of 1.0%) for 180 seconds (in S3 of FIG. 1). Equivalent to the process). Wafer surface defects were inspected by MAGICS manufactured by Lasertec Corporation for each wafer after chemical immersion (corresponding to step S4 in FIG. 1).

  After the inspection, the defects present on the surfaces of the three wafers were observed with an automatic observation SEM manufactured by Hitachi High-Technologies Corporation. At the same time, EDX analysis was performed. The EDX analyzer used was Oxford X-MAX. Examples of this observation analysis are shown in FIGS.

  Next, a representative defect was selected and cross-sectional TEM observation was performed (corresponding to S8 in FIG. 1). The results are shown in FIGS. The defect depth was measured from FIGS. In order to confirm the type of impurities obtained by SEM-EDX analysis, TEM-EDX analysis was also performed at the same time. The analysis results are shown in FIG. FIG. 12 shows SEM-EDX analysis. In the examples of FIGS. 11 and 12, it is shown that both the type of impurity obtained by SEM-EDX analysis and the type of impurity obtained by TEM-EDX analysis are Ni.

  Next, in order to clarify the distribution of metal contamination from the defect coordinates detected by MAGICS, which is a surface inspection apparatus, and the SEM-EDX analysis result, the defect type map of FIG. 13 was prepared. The map in FIG. 13 is a map obtained by superimposing the results of three wafers. In FIG. 13, the positions of defects are represented by plot points, and the types of defects are represented by the types of plot points (◯, Δ, □, ●, ▲, ■). Thus, in the present invention, it has been shown that the distribution of defect types (metal impurity types, differences in pits, protrusions, defect sizes, etc.) can be easily evaluated.

Further, Ni contamination concentration (Ni bulk contamination concentration) was calculated by the following equation 5 with the metal impurity to be evaluated being NiSi ₂ .
Defect volume × Ni atom density × number of defects / wafer volume = Ni contamination concentration (atoms / cm ³ ) Equation 5

At this time, the defect A in FIG. 2 and the defect B in FIG. 3 were selected as representative defects used for calculating the Ni contamination concentration. Further, assuming that the number of defects (the sum of the number of defects A and the number of defects B) is 10, 50, 100, and 1000, the Ni contamination concentration at each defect number was calculated. Further, the Ni contamination concentration was calculated when the ratio of the number of defects A and the number of defects B was changed between 0%: 100%, 50%: 50%, and 100%: 0%. The volume of the defect A was 8.67 × 10 ⁻¹⁶ cm ³ and the volume of the defect B was 3.16 × 10 ⁻¹⁶ cm ³ . Moreover, the above-mentioned 5.23 × 10 ²² (atoms / cm ³ ) was used as the Ni atom density in Equation 5. The number of defects was assumed to exist at the same probability in the wafer thickness direction with the TEM observation region being 1 μm. At this time, the thickness of the wafer was 725 μm.

  Table 2 and FIG. 14 show the results of the Ni contamination concentration calculated by the above calculation. FIG. 14 is a graph showing the results of Table 2.

  As shown in Table 2 and FIG. 14, the Ni contamination concentration increases as the number of defects increases. Further, even if the number of defects is the same, if the ratio of the defects A and B changes, the Ni contamination concentration also changes. Specifically, the Ni contamination concentration increases as the ratio of the defect A having a volume larger than that of the defect B increases. Thus, in the present invention, even if a plurality of defects having different volumes occur, this difference in volume can be reflected in the metal contamination concentration. Moreover, it can be said that the results shown in Table 2 and FIG. 14 indicate the maximum and minimum error ranges of the Ni contamination concentration when there are two types of defects.

Furthermore, in the conventional chemical analysis, it is said that it is difficult to analyze Ni at 1 × 10 ¹⁰ (atoms / cm ³ ) or less, but in the present invention as shown in Table 2 and FIG. Analysis of 1 × 10 ¹⁰ (atoms / cm ³ ) or less is also possible.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has substantially the same configuration as the technical idea described in the claims of the present invention, and any device that exhibits the same function and effect is the present invention. It is included in the technical scope of the invention.

  For example, in the above embodiment, the component analysis of the defect is performed by performing EDX analysis on the defect detected by the surface inspection apparatus. However, the surface analysis method other than EDX analysis, for example, XPS (X-ray photoelectron spectroscopy) Component analysis of defects may be performed by TXRF (total reflection fluorescent X-ray), AES (Auger electron spectrometer), SIMS (secondary ion mass spectrometer), or the like. In the above embodiment, the defect is observed with the SEM. However, the defect may be observed with a microscope other than the SEM, for example, a TEM (transmission electron microscope), an STEM (scanning transmission electron microscope), or the like. .

  In the above embodiment, the depth or height of the defect is obtained by TEM in order to obtain the defect volume. However, the defect volume can be obtained by multiplying the surface area of the defect by a predetermined coefficient. good. According to this, the defect volume can be easily obtained.

  In the above embodiment, the volume of the defect and the metal atom density in the defect are obtained as the degree of metal contamination in the defect (the amount of metal contamination in the defect), and the metal contamination concentration is calculated based on the volume and the metal atom density. However, the degree of metal contamination in the defect may be evaluated based on the peak value of the characteristic X-ray in the EDX analysis. It is considered that the greater the peak value, the greater the degree of metal contamination in the defect.

  Moreover, in the said embodiment, although the depth or height of the defect was calculated | required by TEM, you may obtain | require the depth or height of a defect based on the penetration | invasion length with respect to the wafer of a laser. In this case, for example, it is preferable to determine the depth or height of the defect based on the penetration depth of the laser into the wafer of the surface inspection apparatus used in step S4. According to this, the depth or height of the defect can be obtained in the step of S4. The penetration length may be a value recommended by the manufacturer of the surface inspection apparatus, for example.

  Moreover, although the said embodiment demonstrated the example which applied this invention to the metal contamination evaluation of a silicon wafer, you may apply this invention to semiconductor wafers other than a silicon wafer.

Claims (15)

A detection process for detecting defects derived from metal contamination present on the surface of the wafer;
An acquisition step of acquiring the metal contamination state of each defect detected in the detection step;
A calculation step of calculating a metal contamination concentration of the wafer based on the number of the defects detected in the detection step and the metal contamination state obtained in the acquisition step;
A contamination evaluation method comprising: The detection step includes
A chemical solution supplying step of supplying a chemical solution in which metal impurities are dissolved on the surface of the wafer;
The contamination evaluation method according to claim 1, further comprising: an inspection step of detecting the defect by inspecting a surface of the wafer after performing the chemical solution supply step. The detection step includes
An analysis step of performing component analysis of the detected defect;
The contamination evaluation method according to claim 1, further comprising a classification step of classifying the defects for each defect type based on an analysis result of the analysis step.   The metal contamination concentration of the wafer is calculated in the calculation step based on the number of the defects classified into a predetermined defect type in the classification step and the metal contamination state of the defects. Item 4. The contamination evaluation method according to Item 3.   The contamination evaluation method according to claim 3 or 4, wherein in the analysis step, SEM-EDX analysis is performed on the defect.   The contamination obtaining method according to claim 1, wherein the obtaining step includes a size obtaining step for obtaining the size of the defect as the metal contamination state. The size acquisition step includes
A surface area acquisition step of acquiring the surface area of the defect;
A depth / height acquisition step of acquiring the depth or height of the defect;
The contamination evaluation method according to claim 6, further comprising: a volume calculation step of calculating a volume of the defect as the size based on the surface area and the depth or height.   The contamination evaluation method according to claim 7, wherein in the depth / height acquisition step, a cross section of the defect is observed with a TEM, and the depth or height is acquired based on an observation image.   9. The depth / height acquisition step of observing a cross section of all the defects with a TEM and acquiring the depth or height of each defect based on each observation image. Pollution assessment method.   In the depth / height acquisition step, one or a plurality of the defects are selected, the depth or height of the selected defects is acquired as a representative value based on an observation image by TEM, and the other defects The contamination evaluation method according to claim 8, wherein the representative value is substituted for the depth or height.   The contamination evaluation method according to claim 7, wherein in the depth / height acquisition step, the depth or height is acquired based on a laser penetration length with respect to the wafer having the defect.   The contamination evaluation method according to claim 1, wherein the acquisition step includes a density acquisition step of acquiring a density of metal impurities in the defect as the metal contamination state. The detection step includes a condition determination step for determining, by simulation, the conditions of the acceleration voltage of the electron beam in the SEM, which is suitable for detecting the metal impurity to be evaluated.
The contamination evaluation method according to claim 5, wherein in the analysis step, SEM-EDX analysis is performed under the acceleration voltage condition determined in the condition determination step.   The said calculation process is provided with the surface contamination calculation process of calculating the metal contamination density | concentration in the surface of the said wafer based on the number of the said defects, and the said metal contamination state, The any one of Claims 1-13 characterized by the above-mentioned. The contamination evaluation method described in 1. The calculation step includes
Based on the number of defects and the metal contamination state, an estimation step for estimating the number of bulk defects which are defects derived from metal contamination present in the bulk of the wafer and the metal contamination state of the bulk defects;
15. A bulk contamination calculation step of calculating a metal contamination concentration in the bulk of the wafer based on the number obtained in the estimation step and the metal contamination state. The contamination evaluation method described in the paragraph.