JP2007266258A - Bmd density evaluation method of silicon wafer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a BMD density evaluation method of a silicon wafer, in which BMD density is calculated, based on a correlational relationship between a recombination center density, other than Fe and the BMD density, thereby accurately evaluating the BMD density using a simple method. <P>SOLUTION: In the BMD (Bulk Micro Defects) density evaluation method of the silicon wafer, the BMD density evaluation method of the silicon wafer comprises a step of deciding the correlational relationship between the recombination center density excluding Fe and the BMD density; a step of measuring the recombination center density, other than Fe of the silicon wafer by a SPV (Surface Photovoltage) method; and a step of calculating the BMD density of the silicon wafer, based on the measured recombination center density, other than Fe and the correlational relationship between the predetermined recombination center density, other than Fe and the BMD density. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a BMD (Bulk Micro Defects) density evaluation method for a silicon wafer, and more particularly to a BMD density evaluation method for a silicon wafer that calculates the BMD density based on the correlation between the recombination center density other than Fe and the BMD density.

As a wafer for producing a device such as a semiconductor integrated circuit, a silicon single crystal wafer (hereinafter also referred to as “silicon wafer”) grown mainly by a CZ (chocolate ski) method is used.
From the viewpoint of improving the device characteristics on the silicon wafer, it is desirable to make the vicinity of the silicon wafer surface as defect-free as possible. This is because the presence of defects becomes a factor that degrades device characteristics such as an increase in junction leakage, a deterioration in oxide film breakdown voltage, or a decrease in carrier lifetime.
On the other hand, it is usually desirable to form high density defects in the bulk of the wafer. This is because defects in the bulk can capture (getter) heavy metal contamination that degrades device characteristics.

From the above requirements, by heat-treating a silicon wafer produced by the CZ method before the device production process, making the vicinity of the surface a defect-free layer and generating BMD consisting of oxygen precipitates in the bulk In general, a heavy defect gettering site is formed by forming a high defect density layer. This technique is called intrinsic gettering (IG).
Generally, the capability of intrinsic gettering increases as the BMD density of a silicon wafer increases. Therefore, BMD density evaluation of silicon wafers is extremely important for improving and stabilizing device characteristics.

As a conventional BMD evaluation method, an etching method or an infrared tomography method is known. The etching method is a method in which a silicon wafer is divided along a cleavage plane, a cross section is etched, and then observed using a microscope or the like, and the number of BMDs is counted visually. In addition, the infrared tomography method is a method in which a silicon wafer is divided along a cleavage plane, then infrared rays are projected, and scattered light from BMD is counted from the cleavage plane.
Patent Document 1 discloses a method of evaluating the BMD density by intentionally contaminating a predetermined concentration of Fe and then measuring the remaining Fe concentration not captured by the BMD.
Further, Patent Document 2 discloses a method for evaluating the BMD density based on the correlation between the minority carrier diffusion length measured by the SPV method and the BMD density.
JP 2003-257882 A JP-A-8-62122

However, since the etching method is a visual observation, there is a problem that a measurement error is likely to occur, and there is a large variation in measurement values among observers.
Moreover, since the silicon wafer is divided along the cleavage plane in the etching method and the infrared tomography method, there is a problem that it is almost impossible to evaluate the BMD distribution in the wafer surface.
The BMD density evaluation method of Patent Document 1 has a problem that the evaluation process becomes complicated, such as intentional contamination of Fe and high-temperature heat treatment for capturing Fe in BMD.
Furthermore, the technique of Patent Document 2 has a problem that a measurement error due to the influence of the Fe concentration in the wafer bulk becomes large.

  The present invention has been made in consideration of the above circumstances. The object of the present invention is to calculate the BMD density based on the correlation between the recombination center density other than Fe and the BMD density by a simple method. An object of the present invention is to provide a BMD density evaluation method for a silicon wafer that can accurately evaluate the BMD density.

The method for evaluating the BMD density of a silicon wafer according to one aspect of the present invention is as follows.
A BMD (Bulk Micro Defects) density evaluation method for a silicon wafer,
Determining a correlation between recombination center density other than Fe and BMD density;
Measuring a recombination center density other than Fe of the silicon wafer by an SPV (Surface Photovoltage) method;
Calculating the BMD density of the silicon wafer based on the correlation between the measured recombination center density other than Fe and the predetermined correlation between the recombination center density other than Fe and the BMD density. Features.

Here, the Fe concentration in the bulk of the silicon wafer is desirably 1E11 atoms / cm ³ or less.

  The silicon wafer preferably has a resistivity of 5 Ωcm or more.

  Further, in the step of determining the correlation between the recombination center density other than Fe and the BMD density, the correlation is obtained by measuring the recombination center density other than Fe by a SPV method for silicon wafers having different BMD densities. And it is desirable to be determined by measuring the BMD density by infrared tomography.

  According to the present invention, by calculating the BMD density based on the correlation between the recombination center density other than Fe and the BMD density, a BMD density evaluation method for a silicon wafer that can accurately evaluate the BMD density by a simple method is provided. It becomes possible to do.

First, the principle of a general SPV (Surface Photovoltage) method used in the present invention, and measurement of density and Fe concentration of recombination centers (hereinafter referred to as “ORC”) other than general Fe using the same A method will be described.
In the SPV method, by irradiating the surface of the silicon wafer with light, excess minority carriers generated inside the silicon wafer move to the wafer surface due to the electric field of the depletion layer existing on the surface of the wafer, and the surface photovoltage generated as a result is Measure. By measuring the surface photovoltage, the minority carrier diffusion length of the wafer is obtained.

A general method for measuring the ORC density and Fe concentration using the SPV method is as follows.
First, the silicon wafer to be measured is subjected to light irradiation or heat treatment at about 200 ° C. to activate (excite) Fe and ORC. The activation in the case of Fe specifically means dissociating the bond between Fe and B (boron) (Fe-B) into interstitial Fe and B. And the diffusion length of the minority carrier before and behind activation is calculated | required by SPV method, and ORC density and Fe density | concentration are calculated | required by the following formula | equation (1) and Formula (2) from the difference, respectively.

_{N R = D n / [C} R × (P-1)] × (P / L 2 bef -1 / L 2 aft) ··· (1)
_{N Fe = D n / (C} Fe -C FeB) × (1 / L 2 aft -1 / L 2 bef) ··· (2)
However,
N _R : ORC density (/ cm ³ )
N _Fe : Fe concentration (atoms / cm ³ )
Dn: electron diffusion coefficient C _R : ORC electron capture rate C _Fe : Fe electron capture rate C _FeB : Fe-B electron capture rate P: P = C _Fe / C 2 _FeB
L _bef : diffusion length before activation (um)
L _aft : diffusion length after activation (um)

Here, the Dn, C _R , and P values are known. Also, the value of D _n / (C _Fe —C _FeB ) is approximately 1.1E16 as an empirical value from the DLTS (Deep Level Transient Spectroscopy) method. Therefore, the ORC density and Fe concentration can be calculated from the measurement results of the diffusion length before activation and the diffusion length after activation by the SPV method.

  Next, an embodiment of the BMD density measuring method according to the present invention will be described with reference to the accompanying drawings.

Prior to the measurement of the silicon wafer to be measured, the correlation (correlation diagram) between the ORC density and the BMD density is determined.
In determining the correlation, first, a plurality of silicon wafers having different BMD densities are prepared. This is created, for example, by changing the IG heat treatment conditions for the silicon wafer. That is, the silicon wafer is heat-treated at, for example, 780 ° C. for 3 hours to generate oxygen precipitation nuclei. Then, for example, by performing heat treatment at 1000 ° C. for 1 to 16 hours, BMD is grown from the oxygen precipitation nuclei, and a plurality of silicon wafers having different BMD densities are produced.
Next, using the ORC density measurement method using the SPV method described above, each ORC density is measured for a plurality of prepared silicon wafers having different BMD densities.
Further, the BMD density is measured for the same silicon wafer. At this time, any method can be applied to the BMD density measuring method used so far as it is a conventionally used method. However, it is desirable to use the infrared tomography method from the viewpoint of measurement accuracy and simplicity.

FIG. 1 is an example of a correlation between an ORC density and a BMD density created by obtaining a BMD density by an infrared tomograph. As is clear from the figure, there is a positive correlation between the ORC density and the BMD density.
The ORC density reflects the influence of impurities and crystal characteristics other than Fe in the silicon wafer. Therefore, it is considered that the positive correlation between the ORC density and the BMD density confirmed here reflects carrier trapping by BMD.

After determining the correlation between the ORC density and the BMD density, a silicon wafer to be measured is prepared.
For example, the silicon wafer is doped with B (boron) at the stage of growing the silicon single crystal by the CZ method, and BMD is formed inside the silicon wafer by a predetermined heat treatment (IG). The doping amount of B (boron) is, for example, 2.8E17 to 1.3E14 atoms / cm ³ . The predetermined heat treatment includes, for example, a first stage heat treatment for generating oxygen precipitation nuclei for holding a silicon wafer at a temperature of 600 ° C. to 1000 ° C. for about 4 hours in an oxygen atmosphere, and after the first stage heat treatment. And a second stage heat treatment for growing the BMD by holding the silicon wafer at a temperature of 900 ° C. to 1000 ° C. for 16 hours in an oxygen atmosphere. By this heat treatment, BMD having a density of 5E3 / cm ^{3 to} 5E6 / cm ³ is formed in the silicon wafer.

Next, the ORC density in the silicon wafer is measured using the SPV method. Specifically, surface charge treatment is first performed on a silicon wafer on which BMD is formed. The surface charge treatment is performed, for example, by first immersing the silicon wafer in 0.5 to 50% HF for 1 to 10 minutes, and then washing the silicon wafer with pure water and then spin drying. Then, after the surface charge treatment, the silicon wafer is measured by the SPV method with the surface being stabilized, whereby the diffusion length (L _bef ) before activation is obtained. When surface charge treatment is performed on a silicon wafer as described above, the surface of the p-type silicon wafer doped with B (boron) is positively charged. Since the layer can be secured, highly accurate measurement is possible.

Next, the activation process of Fe and ORC in the silicon wafer is performed. For example, the surface of the silicon wafer is intermittently irradiated with monochromatic light having a forbidden band energy of 1.1 eV or more on the surface of the silicon wafer, or the silicon wafer is held at 200 ° C. or higher for 5 to 15 minutes. It is carried out by quenching at a temperature drop rate of 0.1 to 3.0 ° C./second. Then, the diffusion length (L _aft ) after activation is obtained by measuring this silicon wafer by the SPV method.

Then, the ORC density is calculated by substituting the obtained diffusion length before activation (L _bef ) and the diffusion length after activation (L _aft ) into the above equation (1). Then, the BMD density of the silicon wafer to be measured is obtained from the correlation between the ORC density and the BMD density determined in advance.

In the BMD evaluation method of the present invention, since the ORC excluding Fe is used as an index instead of the minority carrier diffusion length that is directly affected by Fe, the influence on the calculated BMD density due to the presence of Fe can be reduced. .
However, as is clear from the above equation (1), the diffusion lengths L _bef and L _{aft including} the influence of Fe are used in the ORC calculation expression. Therefore, the influence of Fe is completely determined from the ORC density. It is difficult to eliminate. Therefore, the inventors calculated the dependence of the ORC density on the Fe concentration in order to clarify the influence of the Fe concentration on the ORC density. Specifically, the diffusion length (L _bef ) before activation is obtained by substituting it into the above equation (2) using the Fe concentration and the diffusion length (L _aft ) after activation as variables. The ORC density is calculated by substituting the diffusion length after activation (L _aft ) and the diffusion length before activation (L _bef ) into the above equation (1). Then, the relationship between the ORC density and the diffusion length (L _bef ) before activation of the wafer was determined. Here, the Fe concentration was 1E9 to 1E14 atoms / cm ³ .

The result is shown in FIG. As is apparent from the figure, when the Fe concentration is 1E12 atoms / cm ³ or more, the influence of the Fe concentration on the diffusion length increases, so that the ORC density obtained by using the diffusion length is also affected. Thus, in the case of a silicon wafer having a high Fe concentration, an error occurs in the ORC density calculation, and as a result, the BMD density calculation error also increases. Therefore, it is desirable to use a silicon wafer in which the influence of Fe concentration does not appear in the ORC density calculation and the BMD density calculation accuracy is high, that is, a silicon wafer having a Fe concentration in the bulk of 1E11 atoms / cm ³ or less.

  Furthermore, it is desirable that the silicon wafer to which the BMD density measuring method of the present embodiment is applied has a resistivity of 5 Ωcm or more from the viewpoint of improving the accuracy of SPV measurement.

As described above, according to the present embodiment, the BMD density can be easily evaluated because it can be performed non-destructively as compared with the conventional destructive measurement. Moreover, it is also simple in the point which does not require intentional contamination of Fe and high temperature heat processing.
Furthermore, since the ORC excluding Fe is used as an index instead of the minority carrier diffusion length that is directly affected by Fe, the influence on the calculated BMD density due to the presence of Fe can be reduced, and accurate measurement is possible. Become. In addition, the accuracy can be ensured even in the point that there is no variation in the measured value among the observers.
Since this measurement method can be performed non-destructively, it is possible to measure the entire wafer surface, which has been difficult in the past. Therefore, it is possible to obtain the ORC density in a map shape, and it is possible to verify the in-plane non-uniformity of the BMD generated by the silicon wafer pulling or heat treatment process, etc., and to improve and improve these silicon wafer related processes. Can be expected.

  The embodiments of the present invention have been described above with reference to specific examples. In the description of the embodiments, the description of the BMD evaluation method and the like that is not directly necessary for the description of the present invention is omitted, but the elements related to the required BMD evaluation method and the like are appropriately selected and used. be able to.

  In addition, all BMD evaluation methods that include elements of the present invention and can be appropriately modified by those skilled in the art are included in the scope of the present invention.

It is an example of the correlation of the ORC density created by calculating | requiring BMD density by an infrared tomograph, and BMD density. It is a figure which shows the influence of Fe concentration with respect to ORC density calculation.

Claims (4)

A BMD (Bulk Micro Defects) density evaluation method for a silicon wafer,
Determining a correlation between recombination center density other than Fe and BMD density;
Measuring a recombination center density other than Fe of the silicon wafer by an SPV (Surface Photovoltage) method;
Calculating the BMD density of the silicon wafer based on the correlation between the measured recombination center density other than Fe and the predetermined correlation between the recombination center density other than Fe and the BMD density. A method for evaluating BMD density of a silicon wafer. The method for evaluating the BMD density of a silicon wafer according to claim 1, wherein the Fe concentration in the bulk of the silicon wafer is 1E11 atoms / cm ³ or less.   3. The BMD density evaluation method for a silicon wafer according to claim 1, wherein the resistivity of the silicon wafer is 5 Ωcm or more. In the step of determining the correlation between the recombination center density other than Fe and the BMD density, the correlation is performed by measuring the recombination center density other than Fe by the SPV method for silicon wafers having different BMD densities; and 4. The method for evaluating a BMD density of a silicon wafer according to claim 1, wherein the BMD density is determined by measuring the BMD density by an infrared tomography method.

Images