JPH1154579A - Evaluation of semiconductor substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable observation on the entire surface of substrate to a predetermined depth in a wide range and with high sensitivity, and evaluation from the observation, by increasing the size of pits formed by etching and exposing the surface of the substrate, and measuring a pit density by a laser foreign matter detecting device. SOLUTION: Oxygen precipitates l are cleaned with aqueous solution having HF concentration of 20 to 50% for about 30 minutes, thus the oxygen precipitates 1 exposed on the surface of a wafer are dissolved. The HF cleaning forms pits 2 on the wafer surface. The density of pits 2 is measured by the laser foreign matter detecting device. The current minimum detection size of the laser foreign matter detecting device is about 110 nm. If there is a pit of a size smaller than this size, all the pits cannot be measured. Accordingly, cleaning is performed 5 to 10 cycles (10 minutes/cycle) by using alkali cleaning liquid composed of NH4 OH/H2 O2 /H2 O, attaining a comparatively large amount of silicon etching. The cleaning greatly changes the width of pits 3. Next, the laser foreign matter detecting device may detect oxygen precipitates of about several 10 nm, in defective distribution on the entire wafer surface.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an evaluation method for manufacturing a high-quality semiconductor substrate having few crystal defects by evaluating crystal defects in a semiconductor substrate manufactured by using a CZ method and subjected to various heat treatments. The density of the minute oxygen precipitates existing in the bulk near the very surface of the semiconductor substrate is
A semiconductor substrate that can be observed and evaluated over a wide area and with high sensitivity using a laser foreign matter detector by removing oxides by F cleaning to expose pits and expanding the pit size by alkali cleaning. The evaluation method.

[0002]

2. Description of the Related Art The integrity of crystal defects near the surface of a semiconductor substrate manufactured by using the CZ (Czochralski) method is as follows.
Although some improvement is achieved by performing high-temperature treatment (1000 ° C. to 1200 ° C.), it is not perfect, and is incomplete compared to an epitaxial wafer formed epitaxially on the surface.

[0003] In order to manufacture a semiconductor device, a semiconductor substrate is placed on a semiconductor substrate in an area which affects device characteristics, ie, from a surface.
It is desirable that there is no defect in a region having a depth of about 0 μm. However, it is generally difficult to make the region defect-free in a semiconductor substrate manufactured by the CZ method unless an epitaxial wafer is used. It is.

However, a substrate subjected to high-temperature processing is widely used as a substrate for a MOS type LSI process because the cost of the substrate is increased by epitaxial film formation. Also, in the device process,
Since the heat treatment at 500 ° C. to 1200 ° C. is scheduled, the completeness of the crystal defects near the surface is slightly improved by the device process heat treatment without performing the high temperature treatment in advance.
It is not as complete as a substrate subjected to high temperature processing.

[0005] Regarding the type of crystal defects existing near the surface, by performing a heat treatment, the higher the temperature, the more
In the vicinity of the surface of the substrate, a DZ (Deened Zone)
e) Although a defect-free layer called a layer is formed deeply, it is well known that it is not a defect-free layer because 1) vacancy-type grown-in defects and 2) minute oxygen precipitates exist. Have been.

[0006]

As a method of evaluating the density of crystal defects existing near the surface of the semiconductor substrate, a selective etching method, an IR tomography method, an OPP method, and a COP evaluation method for the grown-in method are used. Has been measured. More specifically, the selective etching method is currently widely used mainly as a crystal defect evaluation method. Da
The sh ET method and the Right ET method can be mentioned.
The detection sensitivity for microdefects is low, and it is extremely difficult to judge it as a defect during observation with an optical microscope after ET.
It is impossible to detect a grown-in defect.

E for detecting a grown-in defect
As the T method, a Secco ET method is used, and a defect detected by this method is detected by using an FPD (Flow Pattern).
nDefect), SPD (Secco Pit De
This is called a “fact” and reported by the Japan Society of Applied Physics, but it is difficult to accurately detect the density of pits with an optical microscope during ET due to particles adhering.

In the IR tomography method and the OPP method, it takes several tens of hours to evaluate the entire surface of the evaluation substrate over a wide area, and if only the depth of a certain region from the surface is to be evaluated, there is a problem with the laser diameter. It can be evaluated only in a wide range in 5 μm and 10 μm steps, and is not suitable for evaluation of a narrow specific region. In addition, the detection sensitivity for minute defects may be greatly affected by the state of the front and back surfaces of the evaluation substrate, and the defect density adhered to at that time shows an abnormally high value, making it difficult to distinguish the density of only defects. It is. This evaluation method is suitable for evaluating a grown-in defect of a substrate that has not been heat-treated and for evaluating defect information near and inside the surface of a substrate that has been heat-treated in a relatively wide range. ~ 2μ
It is not possible to evaluate in a narrow range of m.

In the COP evaluation method, the density of the grown-in defects on the substrate can be evaluated over a wide area over the entire surface of the substrate by repeatedly performing alkali cleaning, and the defect density is measured using a laser surface inspection machine. Is to do
When various evaluation heat treatments are performed, COPs are buried in the oxide film to reduce the unevenness of the defects, which makes it difficult to detect defects by a laser surface inspection machine, and that oxides with respect to minute oxygen precipitates Therefore, the irregularities of the defect are small and it is difficult to detect the defect only by alkali cleaning.

[0010] In short, in the IR tomography method or the OPP method, the measurement area is narrow, and it is very difficult to evaluate a wide range over the entire surface of the semiconductor substrate with respect to minute oxygen precipitates with high sensitivity. Further, when trying to obtain information only in a certain area near the surface, the measurement results are obtained in a wide area at intervals of 5 μm and 10 μm due to the problem of the laser diameter, so that it is difficult to perform measurement in fine steps.

According to the present invention, for example, a semiconductor substrate manufactured by using the CZ method is subjected to a high-temperature treatment or a heat treatment (500 ° C. to 1200 ° C.) assuming a device process.
Conventionally, regarding a semiconductor substrate subjected to, for example, a proposal is made in view of the fact that there is no method for easily and easily measuring a minute oxygen precipitate present near the surface of the semiconductor substrate over a wide area of the substrate with high sensitivity. It is an object of the present invention to provide a method for evaluating a semiconductor substrate in which a predetermined depth can be observed and evaluated over a wide area of the entire surface of the substrate with high sensitivity using a simple laser foreign matter detection device.

[0012]

Means for Solving the Problems The inventors of the present invention have developed various methods for evaluating the density of minute oxygen precipitates present in the bulk near the very surface of a heat-treated semiconductor substrate in a wide range and with high sensitivity. As a result of the examination, after performing HF cleaning with an aqueous solution having a HF concentration of 20 to 50%, removing oxides to form concave pits, and then performing alkali cleaning and etching the silicon substrate, detection with a laser foreign substance detection device is performed. By changing the pit size to a possible pit size, it was found that a minute oxygen precipitate existing on the entire surface of the substrate could be evaluated over a wide range and with high sensitivity, and the present invention was completed.

That is, according to the present invention, a semiconductor substrate subjected to a required heat treatment is subjected to a mirror polishing from the surface to a region to be evaluated after removing a thermal oxide film formed by the heat treatment, and then the substrate is subjected to HF cleaning. A method of evaluating a semiconductor substrate in which pits are exposed by removing oxide precipitates on the surface, and the size of the exposed pits is further enlarged by etching the substrate surface by alkali cleaning, and the pit density is measured by a laser foreign matter detection device. It is.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The evaluation method according to the present invention is shown in FIG.
The steps are as follows. Step 1: subjecting the semiconductor substrate to a heat treatment at, for example, 500 ° C. to 1200 ° C. for evaluating crystal defects. Step 2: removing a thermal oxide film formed by the heat treatment. Step 3: existing near the surface of the substrate. In order to measure the density of minute oxygen precipitates, the target area (~ 50μ)
m) mirror polishing is performed, Step 4: dissolving the precipitates exposed on the surface of the semiconductor substrate by HF cleaning, Step 5: removing particles adhered during HF cleaning, Step 6: performing alkali cleaning and HF Enlarging the concave minute pits dissolved by the washing. Step 7: Measure the pit density using a laser foreign matter inspection device.

In the present invention, as a semiconductor substrate whose crystal defect density is to be measured, a wafer which has been subjected to a high-temperature heat treatment or an IG process on a mirror-processed wafer pulled up by the CZ method, or a wafer which has been processed with them. The object in the state is subjected to a crystal defect evaluation heat treatment (500 ° C. to 1200 ° C.) assuming a device process.

Most of the semiconductor substrates subjected to the heat treatment are subjected to heat treatment (5) in both oxidizing and non-oxidizing gas atmospheres.
(00 ° C. to 1200 ° C.), a thermal oxide film is formed. HF cleaning is performed to remove the oxide film,
After removing the thermal oxide film, mirror polishing is performed to a depth at which measurement is performed.

At the time of polishing, as shown in FIG. 2A), the oxygen precipitate 1 exposed on the substrate surface exists. Then H
Perform F cleaning. The HF cleaning is performed when the HF concentration is 20 to 50.
By washing with an aqueous solution of about 30 minutes for about 30 minutes, the oxygen precipitate 1 exposed on the wafer surface is dissolved. Then, HF
Particles adhering at the time of cleaning are removed. Specifically, the cleaning is performed with an alkali cleaning liquid having a smaller silicon etching amount than the alkali cleaning employed in the present invention.

By dissolving the HF cleaning, pits 2 shown in FIG. 2B) are formed on the wafer surface. The density of the formed pits is subsequently measured using a laser foreign matter inspection device. However, since the current detection lower limit size of the laser foreign matter inspection device is about 110 nm, if a pit size smaller than this exists, all Pit cannot be measured.

Therefore, in order to enlarge the pit size to a size that can be measured by a laser foreign matter inspection apparatus, NH ₄ O
As shown in FIG. 2C), the pits 3 were cleaned by using an alkaline cleaning solution composed of H / H ₂ O ₂ / H ₂ O and having a relatively large silicon etching amount and performing cleaning for 5 to 10 cycles at 10 minutes / cycle. The width changes greatly.

Next, by performing measurement using a laser foreign substance inspection apparatus, it is possible to detect a very small oxygen precipitate of several tens of nanometers in the defect distribution over the entire surface of the wafer with high sensitivity.

[0021]

【Example】

Example 1 As an evaluation substrate, a 6-inch silicon p (100) wafer pulled up by the CZ method was used. Initial oxygen concentration is 1
4 × 10 ¹⁷ atoms / cm ³ , and specific resistance is 5 Ω · cm. This sample was heat-treated at 800 ° C. for 2 hours in a non-oxidizing gas atmosphere to form oxygen precipitation nuclei in the bulk, and then heat-treated at 1000 ° C. for 16 hours in an oxidizing gas atmosphere to remove the formed oxygen precipitate. A grown heat-treated wafer was produced.

To ensure that oxygen precipitates are exposed on the wafer surface, mirror polishing is performed after removing the thermal oxide film.
μm. First, selective etching (Wright Etch 1 min) is performed on the wafer substrate in that state,
When the density of the etch pits present on the substrate surface was measured, it was 75 / cm ² . This is the result of oxygen precipitate density measurement by the etching method.

The pit density in the state where the oxygen precipitates were exposed on the substrate surface (Step 3) was measured by using the same heat treatment as that of the sample previously measured by the etching method and the sample lot subjected to 20 μm polishing, and the laser foreign matter was measured. Inspection equipment (Ten
cor Surfscan) as a measurement map in FIG.
As shown in A), it was 5 / cm ² .

Next, washing with an aqueous solution having a HF concentration of 30% for 30 minutes and washing with an alkaline washing solution of NH ₄ OH: H ₂ O ₂ : H ₂ O = 1: 5: 20 for 10 minutes were performed to remove oxygen precipitates. The pit density after melting (step 5) was 90 / cm ² as shown in FIG. 3B) as a measurement map of the laser foreign matter inspection device.

After performing 5 cycles of cleaning at 10 minutes / cycle with an alkaline cleaning solution of NH ₄ OH: H ₂ O ₂ : H ₂ O = 1: 1: 5 (step 7), the pit density can be measured by using a laser particle inspection apparatus. As shown in FIG.
/ Cm ² .

As is apparent from FIG. 3, the pit density is about twice as high as that in the step 5, ie, when the method of the present invention is completed, compared with the pit density after the particle removal is completed after the HF cleaning. I have. Also, the density was detected to be about twice as high as that in the selective etching method.

Next, the pit size of the sample in the above steps 5 and 7 was measured by AFM. FIG. 4 shows the measurement results. In step 5, the width is 50 to 120 nm and the depth is 30 to 80 nm.
２５０250 nm, depth 50-120, and width has changed to about twice the size. This means that a few tens of nanometer oxygen precipitates that could not be detected by the laser foreign matter detection device in step 5 can be detected in step 7.

Example 2 A 6-inch silicon p (100) wafer pulled up by the CZ method was used as an evaluation substrate. Initial oxygen concentration is 15 × 10 ¹⁷ atoms / cm ³ , specific resistance is 10Ω · c
m. This sample was treated with D under the following heat treatment conditions.
High temperature heat treatment (oxidizing and non-oxidizing gas atmosphere) to form a Z layer, and IG (oxygen precipitation nuclei) inside the bulk
Heat treatment (non-oxidizing gas atmosphere) and heat treatment for growing the oxygen precipitation nuclei (oxidizing gas atmosphere)
Was performed to produce an evaluation substrate.

Evaluation substrate A) DZ layer 10 μm Oxygen precipitate density 5 × 10 ⁵
/ Cm ² Heat treatment condition 1100 ° C × 1hr → 800 ° C × 3hr →
1000 ° C. × 16 hr B) DZ layer 20 μm Oxygen precipitate density 4 × 10 ⁵
/ Cm ² Heat treatment condition 1150 ° C × 1hr → 800 ° C × 3.5h
r → 1000 ° C. × 16 hr C) DZ layer 30 μm Enzyme precipitate density 6 × 10 ⁵
/ Cm ² Heat treatment condition 1200 ° C × 1hr → 800 ° C × 4hr →
1000 ℃ × 16hr

The DZ layer and the oxygen precipitate density of the evaluation substrate were determined by selective etching (Wright Etch).
5 minutes) from the cross section of the substrate by the optical microscope after the application, and is a sample of the same lot as the silicon wafer used for the evaluation according to the present invention.

Since a thermal oxide film is formed on the evaluation substrate prepared by performing the heat treatment, the substrate is washed with an aqueous solution having an HF concentration of 10%, and after removing the thermal oxide film, the defect density distribution in the DZ layer is evaluated. To perform 0, 2, 0 on the surface by mirror polishing
Polishing is performed to a depth of 4, 6, 10, and 30 μm, and an HF concentration of 3 is used to dissolve oxygen precipitates protruding from the substrate surface.
Washing was performed with a 0% aqueous solution.

Further, in order to enlarge the pit shape,
After 5 cycles of 10 minutes / cycle with an alkaline cleaning solution of NH ₄ OH: H ₂ O ₂ : H ₂ O = 1: 1: 5, the number of defects was counted by a laser foreign matter inspection device. The measurement result is shown in FIG. The measured density is the density per area detected by the laser foreign matter detection device, and the alkali cleaning 5
This is a density converted into a volume by an etching amount when a cycle is performed.

Further, a substrate of the same lot subjected to the same evaluation thermodynamics was separately manufactured, and a comparative evaluation was performed by OPP to measure a defect density distribution in the DZ layer. The measured defect density is shown in FIG. 5A). As shown in the figure, in the case of OPP, there is no significant difference in the measured density up to 5 μm from the surface, and there is no difference in the defect density near the very surface even if the position of the DZ layer is changed.

Next, the measurement results according to the present invention show that a substrate having a narrower DZ layer has a higher defect density in a shallower surface region, and conversely, a substrate having a deeper DZ layer has a lower defect density in a shallower surface region. It can be clearly measured that the deeper the DZ layer, the lower the defect density near the substrate surface. In addition, since measurement can be performed in small steps, it can be seen that the measurement is optimal for measurement of only a predetermined area.

[0035]

According to the present invention, heat treatment (500 ° C.-
(1200 ° C.) and the like, a micro-oxygen precipitate existing near the surface of the semiconductor substrate can be easily measured with high sensitivity over a wide area over the entire surface of the substrate.
It is also possible to collect information only from the surface in the region where measurement is desired, which is effective as an evaluation means for developing a high-quality semiconductor substrate.

[Brief description of the drawings]

FIG. 1 is a process explanatory view showing an evaluation process according to the present invention.

FIGS. 2A to 2C are explanatory diagrams showing changes in the shape of oxygen precipitates in an evaluation step according to the present invention.

FIGS. 3A to 3C are explanatory diagrams showing maps of the entire surface of the substrate of oxygen precipitates detected by the laser foreign matter detection device.

FIG. 4A is an explanatory diagram showing a pit size,
B) and C) are graphs showing pit sizes by AFM in evaluation step 5 and step 7 according to the present invention.

FIG. 5A is a graph of the depth and the pit density showing the result of the substrate evaluation by the OPP method, and FIG. 5B is a graph of the depth and the pit density showing the result of the substrate evaluation by the present invention.

[Explanation of symbols]

 1 Oxygen precipitate 2, 3 pits

Claims (1)

[Claims] 1. A semiconductor substrate that has been subjected to a required heat treatment, a thermal oxide film formed by the heat treatment is removed, mirror polishing is performed from the surface to a region to be evaluated, and oxidative deposition on the substrate surface is performed by HF cleaning. A method of evaluating a semiconductor substrate, in which a pit is exposed by removing an object, and the size of the exposed pit is enlarged by etching the surface of the substrate by alkali cleaning, and the number of pits is measured by a laser foreign matter detector.