JP2010278363A - Crystal defect detecting method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To facilitate detecting crystal defects with high accuracy, even to a semiconductor material with a large area. <P>SOLUTION: Anisotropic etching of a high selection ratio is performed to crystal defects contained in the semiconductor material used as a crystal defect detection object by etching gas containing at least Cl<SB>2</SB>and O<SB>2</SB>. On an etching bottom face, there is exposed a semiconductor material protrusion formed using the crystal defects in the semiconductor material as a vertex, and the crystal defects are actualized, based on this protrusion for detection. To a Cl<SB>2</SB>flow rate 9 to an etching chamber, an O<SB>2</SB>flow rate is set up in a from about 0.8 to about 1. SF<SB>6</SB>may be contained. A flow rate of ≥0 and ≤0.4 is set up to the Cl<SB>2</SB>flow rate 9 or a flow rate of ≤1/2 is set up to the O<SB>2</SB>flow rate. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method for detecting and evaluating minute crystal defects existing in a semiconductor substrate or the like.

  In an LSI using a semiconductor material such as Si single crystal as a substrate, crystal defects become a problem at various stages of the LSI manufacturing process and affect the characteristics and reliability of the element. In particular, defect control is important and researched for improving reliability and yield in an LSI process where miniaturization advances.

  Conventionally, selective etching with a chemical solution has been adopted as the simplest evaluation method for oxygen precipitation defects, which are a kind of crystal defects. However, this selective etching with a chemical solution does not manifest as pits unless the defect size is about 1.0 μm.

Methods for evaluating finer oxygen precipitates include a method using an infrared laser (infrared laser bright-field interferometry; OPP: Optical Precipitate Profiler, infrared tomography method), a transmission electron microscope (TEM), and the like. . However, in the method using an infrared laser, it takes about 2 hours to evaluate one wafer. Further, in the method using an infrared laser, in principle, the resolution in the depth direction is only about 2 μm. A transmission electron microscope (TEM) can detect a very small defect of about several nanometers, but it takes a lot of time to prepare a sample, and observation is difficult unless the defect density is 10 ¹⁰ cm ⁻³ or more.

  The present applicant proposes in Patent Document 1 a method capable of detecting minute oxygen precipitation defects that cannot be detected by selective etching using a chemical solution or a method using an infrared laser, and capable of easily detecting such defects. is doing.

JP 2000-58509 A

  In Patent Document 1, a crystal defect can be detected with high sensitivity and easily by forming a projection having the crystal defect at the apex by etching with a high selectivity to the crystal defect. However, the area of semiconductor wafers has been increasing, and it is desired to perform defect inspection more efficiently for larger-area semiconductor wafers and the like.

  In addition, while further demands for cost reduction of semiconductor devices are progressing, it is desired that defect inspection is also simpler and requires fewer steps. For example, it is desired to reduce the number of etching gas removal treatments used for inspection as much as possible.

  The present invention evaluates minute crystal defects simply and with high accuracy even for a semiconductor material having a large area.

In order to achieve the above object, a crystal defect detection method according to the present invention is highly selective for crystal defects contained in a semiconductor material to be crystal defect detected by an etching gas containing at least Cl ₂ and O _2. By performing anisotropic etching of the ratio, the semiconductor material protrusion formed with the crystal defect in the semiconductor material as the apex is exposed on the bottom surface of the etching, and the crystal defect is revealed based on the semiconductor material protrusion and detected. To do.

In another aspect of the present invention, in the crystal defect detection method, the O ₂ flow rate is set in a range of about 0.6 to about 1 with respect to the Cl ₂ flow rate 7 to the etching chamber.

In another aspect of the present invention, in the above-mentioned crystal defect detecting method, the etching gas further comprises a SF _6, the flow rate of the SF ₆ into the etching chamber, from less than half of the flow rate of the O ₂ 0 It is set to a larger range.

In another aspect of the present invention, in the crystal defect detection method, the etching gas further includes SF ₆ , and the flow rate of O ₂ is about 0.6 relative to the flow rate 7 of Cl ₂ to the etching chamber. The range of about 1.0 or less and the flow rate of the SF ₆ are set in the range of more than 0 and about 0.4 or less.

  In another aspect of the present invention, in the crystal defect detection method, the pressure in the etching chamber for performing the etching is set to 20 mTorr or more and 30 mTorr or less.

  In another aspect of the present invention, in the crystal defect detection method, a temperature of the cathode electrode disposed in the etching chamber in which the etching is performed is set to 50 ° C. or higher and lower than 90 ° C.

  In another aspect of the present invention, in the crystal defect detection method, the semiconductor material is subjected to heat treatment for 3 hours or more and 16 hours or less before performing the etching.

As described above, in the crystal defect detection method of the present invention, the semiconductor material is etched using an etching gas containing at least Cl ₂ and O ₂ , so that the crystal defect has a high selectivity. In addition, while maintaining the projection shape having the crystal defect at the top, the cone-shaped projection due to the crystal defect can be accurately formed on the etching bottom surface. For this reason, it is possible to detect a crystal defect even on a semiconductor material having a large area based on the obtained protrusion.

It is also possible to add SF ₆ further etching gas. It is possible to adjust the etching rate and detection sensitivity of the oxide of the semiconductor material that is an oxygen precipitation defect or an etching reaction product.

It is a figure which shows typically the procedure of the crystal defect manifestation used in embodiment of this invention. It is a figure which shows the crystal defect evaluation method used in embodiment of this invention. It is a figure explaining the method of calculating | requiring the position of the depth direction of a crystal defect from the processus | protrusion which made the crystal defect obtained in embodiment of this invention a vertex. It is an observation figure by the SEM photograph of the processus | protrusion obtained by the composition of the predetermined etching gas concerning embodiment of this invention. It is an observation figure by the SEM photograph of the front-end | tip part of protrusion by the etching which concerns on embodiment of this invention. It is a figure explaining the formation principle of the front-end | tip shape of the processus | protrusion of FIG. It is an observation figure by the SEM photograph of the processus | protrusion obtained by etching with the various etching gas composition concerning embodiment of this invention. It is an observation figure by the SEM photograph of the crystal defect detection result at the time of changing the flow rate of oxygen gas concerning the embodiment of the present invention, and a projection. It is an observation figure by the SEM photograph of the processus | protrusion which shows the influence which the source power and cathode electrode temperature concerning embodiment of this invention have on a processus | protrusion. It is an observation figure of the Si sample surface after etching on the conditions from which the heat processing time before the etching which concerns on embodiment of this invention differs.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention (hereinafter referred to as embodiments) will be described with reference to the drawings.

  FIG. 1 shows the defect detection principle according to the present embodiment, FIG. 2 shows a crystal defect evaluation procedure, and FIG. 3 shows a method for obtaining the depth (position) from the surface of the crystal defect based on the obtained protrusion. .

  In the present embodiment, as a semiconductor material, a Si wafer (hereinafter referred to as a Si sample) is used, and crystal defects existing in the Si sample are subjected to dry etching with a high selection ratio with respect to the crystal defects. As shown in FIG. 4, the cone-shaped projections having the crystal defects as apexes are exposed on the bottom surface of the etching.

Specifically, RIE (reactive ion etching) is employed for the etching. The etching gas has a low etching rate (high selectivity) with respect to silicon oxide (SiO ₂ ) with respect to defects such as oxygen precipitation defects and cavity defects present in the Si sample, and the top of the crystal defects. The gas having the function of maintaining the protruding shape is used. In this embodiment, at least chlorine (Cl ₂ ) gas and oxygen (O ₂ ) gas are used as such etching gas. Chlorine gas has an etching function for Si samples. Oxygen gas has a high selectivity to crystal defects and has an anisotropic etching function for Si samples, and functions to maintain the shape of the protrusion by forming a silicon oxide layer on the side of the protrusion with the crystal defect at the top. Have In addition, sulfur hexafluoride (SF ₆ ) gas having a silicon oxide etching function is adopted to adjust the detection sensitivity and noise of crystal defects (formation of protrusions not caused by defects due to adhesion of etching products). Is possible.

  Hereinafter, the defect detection method according to the present embodiment will be described. First, before the crystal defect revealing etching, the Si sample is heat-treated (S1). By this heat treatment, oxygen precipitation defects are deposited from the defect nuclei as shown in FIGS. In this heat treatment, the longer the treatment time, the larger the number and area of defects grown from oxygen precipitation defect nuclei. As a result, the number of protrusions formed with the oxygen precipitation defect as a vertex by etching increases, and the tip end area of the protrusion increases. Therefore, detection sensitivity can be increased by increasing the processing time. High sensitivity, such as when the oxygen precipitation defect concentration due to the oxygen concentration assumed to be contained in the Si sample is low, or when the Si sample to be detected may be exposed to long-term heat treatment in practical use When defect detection is required, it is preferable to lengthen the heat treatment time.

  This heat treatment is preferably performed for at least 3 hours. For example, when the heat treatment is performed for 4 hours or more, extremely high detection accuracy is realized as described later. On the other hand, when the oxygen concentration in the Si sample is high, if the heat treatment time is too long, the projections having the crystal defects as apexes overlap, which affects the evaluation accuracy of the number of defects and the density. Moreover, since the whole detection processing time becomes long, it is preferable to set it as about 16 hours or less, for example. This step is omitted when evaluating a Si sample that is not subjected to heat treatment.

After the heat treatment and before the crystal defect detection etching, the natural oxide film formed on the surface of the Si sample during the heat treatment is removed (S2). For example, the removal of the natural oxide film can be performed by an RIE method using CF ₄ gas. Although it can be removed by wet etching such as HF solution, it can be removed by dry etching, especially by RIE, and etching and etching for revealing crystal defects to be described later can be performed by changing the etching gas employed. Devices can be shared.

  In the present embodiment, after the natural oxide film is removed as described above, etching for revealing crystal defects is performed (FIGS. 1C and 1D and S3 in FIG. 2).

As described above, at least Cl ₂ gas and O ₂ gas are used as the etching gas for revealing crystal defects. When the gas composition is expressed as a flow ratio to the etching chamber, for example, it is preferable to set the O ₂ flow rate within a range of about 0.6 to 1.0 with respect to the Cl ₂ flow rate 7. Further, the SF ₆ flow rate is in the range of 0 to about 0.4 with respect to the Cl ₂ flow rate 9. The SF ₆ flow rate is in the range of 2 to 0 and less than half of the O ₂ flow rate 4-5. Note that the SF ₆ flow rate is 0, that is, SF ₆ is not contained as an etching gas.

  Chlorine gas has the effect of selectively etching Si crystals.

Oxygen gas has a high selectivity to crystal defects and has an anisotropic etching function for Si samples. It reacts with silicon, which is a semiconductor material, in the etching chamber to form silicon oxide, resulting in a reaction product. objects (SiO ₂₎ is deposited on the side surface of the cone-shaped projection formed crystal defects as the vertex. In the anisotropic etching mainly composed of chlorine gas, the reaction product adhering to the side surface of the cone functions as a protective film, so that the cone shape is maintained even if etching proceeds.

Further, the sulfur hexafluoride gas has a silicon oxide etching function, and the oxide etching rate can be increased by using this gas. Therefore, the sensitivity of the crystal defect revealing etching can be adjusted by the concentration of the sulfur hexafluoride gas. Here, when NF ₃ which has been conventionally used is adopted as a gas having a silicon oxide etching function, a detoxification step different from chlorine gas is required for detoxification of etching reaction products. However, this sulfur hexafluoride gas can perform the detoxification of the etching reactant simultaneously with the detoxification of the reactant caused by the chlorine gas.

  When the pressure in the etching chamber is a low pressure of several tens of mTorr (several Pa) or less, the mean free path of reactive ions in RIE becomes long, the direction of ion movement becomes uniform, and etching anisotropy increases. Therefore, it is suitable for forming a projection having a crystal defect as a vertex. However, when the pressure is lowered, the density of reaction products in the plasma gas phase (deposited on protrusions with crystal defects as the apex to form a film that protects from etching) decreases, so that the film that protects the defects In some cases, the deposition rate of the liquid drops, leading to a decrease in detection sensitivity. Therefore, the pressure in the etching chamber is set to, for example, 20 mTorr (2.67 Pa) or more and 30 mTorr (4.0 Pa) or less, more preferably 21 mTorr (2.79 Pa) or more and 25 mTorr (3.33 Pa) or less.

  The temperature of the cathode electrode disposed in the etching chamber is preferably about 50 ° C. to less than 90 ° C. In the present embodiment, since the Si sample to be etched is disposed on the cathode electrode of the RIE apparatus, the cathode electrode temperature greatly affects the temperature of the Si sample. When the cathode electrode temperature increases, the amount of reaction product adhering to the side wall decreases, so the decomposition of the reaction gas proceeds, the side wall protective film formation rate decreases, the base angle of the protrusion increases, and the steepness increases. Protrusions are obtained. On the other hand, if the temperature is too high, it is difficult to maintain the shape of the protrusions. However, in a sample with a high crystal defect density, the base angle of the protrusions is increased, so that the crystal defect measurement resolution is improved. Therefore, the cathode electrode temperature is preferably set between about 50 ° C. and about 70 ° C.

  The power (W: Watt) of the etching source is preferably set between 750 W and 825 W. Increasing the source power increases the electron density in the plasma and promotes gas dissociation, resulting in a highly reactive plasma. Accordingly, the higher the source power, the faster the Si etching rate. On the other hand, the plasma density of reaction products that inhibit the reaction also increases, so that the protective film on the side wall becomes thicker and the base angle of the protrusion becomes smaller. When it exceeds 825 W, the etching of Si is also stopped. An increase in electron density due to an increase in source power has the effect of reducing the potential difference between the plasma sheath regions (between the wafer and the bulk plasma), and has the effect of reducing the energy of ions incident on the wafer. As a result, since physical ion sputtering can be suppressed, the detection sensitivity of crystal defects increases.

  A commercially available apparatus can be used as the etching apparatus. For example, an apparatus capable of batch etching the entire surface by introducing a silicon wafer of 300 mm or more in a chamber (etching chamber) is preferable. In this embodiment, by adopting the above conditions as the crystal defect revealing etching, for example, on the entire surface of a wafer having a diameter of 300 mm, a projection having the top of the crystal defect is formed with high sensitivity under the same conditions at the same time. It becomes possible to reveal crystal defects.

  After forming the projection with the crystal defect at the apex, the crystal defect detection process is executed based on the projection, but in order to enable accurate detection of the projection shape, in this embodiment, before this detection process, The side wall protective film formed on the side surface of the cone is removed (S4). The removal of the sidewall protective film can be performed by, for example, wet etching using dilute hydrofluoric acid. However, the protective film removal process can be omitted.

  After removing the protective film, the cone-shaped projections are captured and observed, and for example, the size of the bottom surface of the projection is evaluated (S5). An optical microscope, SEM, or the like is used for capturing and observing the protrusions. The sample can be observed from above (etching direction) with an optical microscope, and the number of cone-shaped projections on the etching exposed surface can be measured with a data processing device, a particle counter, or the like. Also, for example, the diameter d of the bottom surface of the cone is measured from image data obtained with an optical microscope, SEM, etc. (diameter in the case of a cone, side length in the case of a cone other than a cone, or the length of the hypotenuse of the bottom surface) d You can also

  The base angle θ of the cone can also be measured using an optical microscope or SEM. If the conditions of the high selectivity anisotropic etching applied to the Si sample are the same, the base angle θ of the cone-shaped projection formed with the crystal defect as the apex by this etching depends on the size of the cone. Almost constant. Therefore, when the cone base angle θ corresponding to the performed anisotropic etching condition is known, the calculation process of the base angle θ using SEM or the like is not necessarily required.

The etching depth D from the surface of the semiconductor material before etching is determined by measuring the distance from the surface to the etching exposed surface by masking so that a part of the surface of the Si sample remains at the time of etching as shown in FIG. It is calculated | required by calculation from the known etching rate according to the etching gas to be used, and the processed etching time. As described above, based on the calculation principle shown in FIG. 3 using the obtained measurement and calculation values, the following formula (1)
X _OX = D− (d / 2) tan θ (1)
Is calculated to obtain the depth (position) X _OX from the surface of the crystal defect.

  By adopting such a method, in addition to the distribution in the planar direction, the distribution of crystal defects in the depth direction can be obtained. The obtained result is output to an output device such as a printer or a monitor (S6), and the evaluator can know the distribution of crystal defects from the output result, and can perform crystal defect evaluation.

FIG. 4 is an SEM of the etching exposed surface of the Si sample when crystal defect revealing etching is performed on the Si sample that has been subjected to the heat treatment (800 ° C.) for 16 hours under the high selectivity etching condition according to the present embodiment. It is a photograph. Etching conditions are Cl ₂ (9 times the flow rate of O ₂ gas), O ₂ (reference flow rate), SF ₆ (1/10 times the flow rate of O ₂ gas), etching chamber pressure 25 mTorr, source power 750 W, cathode electrode The temperature was 50 ° C. In the sample shown in this photograph, the side wall protective film remains attached.

The photograph in FIG. 4 is an enlarged image of the etching exposed surface. As shown in FIG. 4, the tip of the cone-shaped protrusion remains well and the shape of the cone is maintained, and the base angle θ is substantially the same regardless of the height of the protrusion. . Therefore, if the height of the protrusion is obtained based on the size d of the bottom surface and the base angle θ, the crystal defect depth X _OX can be obtained accurately.

  In addition, when a silicon wafer having a diameter of 300 mm is subjected to high-selectivity anisotropic etching on the above conditions and the flatness of the etching bottom surface is measured, the error in the entire bottom surface is within ± 3%, and a large area such as 300 mm is obtained. As a method for detecting crystal defects in a semiconductor material, the above-described high selectivity anisotropic etching is effective.

  FIG. 5 shows an example of an observation view by SEM photograph of the tip portion of the cone-shaped protrusion formed by high selective anisotropic etching due to crystal defects (oxygen precipitation defects, cavity defects) in the Si sample. ing. FIG. 5A shows the tip of the protrusion due to the cavity defect existing in the Si sample. Although it does not appear clearly, a recess having a shape in which two quadrangular pyramids are joined to each other near the base angle is formed at the tip of the conical protrusion. FIG. 5B shows the tip of the protrusion due to the plate-like oxygen precipitates present in the Si sample. The tips of the protrusions shown in FIG. 5B maintain a flat plate shape without being recessed.

  When there is an oxygen precipitation defect or a cavity defect in the Si sample, when the highly selective anisotropic etching according to the present embodiment is performed, first, as shown in FIG. 40e is formed. When the defect 40 is removed by etching, when the etching further proceeds from the etching exposed surface 10e1 toward the etching exposed surface 10e2, the side wall is not eroded and the cone is further dug. At this time, the tip 42 of the protrusion 40e is directed from the uppermost surface of the protrusion 40e toward the inside of the cone, and the etching selectively proceeds into a shape corresponding to the type of the crystal defect 40. The shape of the tip 42 of the protrusion 40e differs depending on the type of defect. For example, in the case of an oxygen precipitation defect along the <110> plane of a silicon crystal, the shape of a square pyramid as shown in FIG. Become. In the case of a cavity defect along <111> of the silicon crystal, an octahedral structure in which two quadrangular pyramids are joined as shown in FIG. Further, when a defect 40 remains due to a flat oxygen precipitation defect, a protrusion is formed with this flat plate as an apex, so that the tip of the protrusion becomes a flat surface as shown in FIG.

  As described above, by adopting the crystal defect revealing etching condition in which the shape of the protrusion is maintained, the type of crystal defect is determined not only from the distribution of the crystal defect in the Si sample but also from the observation of the tip shape of the protrusion. It becomes possible. It is also possible to obtain the distribution in the depth direction for each type of crystal defect and the plane direction of the Si sample.

  Hereinafter, examples of crystal defect etching conditions and evaluation thereof will be described.

In Table 1 below, the flow rate of O ₂ gas, the flow rate of SF ₆ gas, and the pressure conditions in the etching chamber are changed for Si samples when the heat treatment time before etching is 3 hours, 4 hours, 8 hours, and 16 hours. The result of crystal defect revealing etching in the case of being made is shown.

(Evaluation 1: Flow rate of O ₂ gas and SF ₆ gas)
In Table 1 (a), the flow rate of O ₂ gas was set to −10% and −20% from the reference flow rate (± 0) and the reference flow rate. In this case, the Cl ₂ gas flow rate is 9 times the O ₂ gas reference flow rate (Cl ₂ gas reference flow rate), and the SF ₆ gas flow rate is 1/10 of the O ₂ gas reference flow rate (SF ₆ gas reference flow rate). Yes. The pressure in the etching chamber is 25 mTorr, the cathode electrode temperature is 50 ° C., and the source power is 750 W.

From Table 1 (a), protrusions (BMD: bulk micro defects) cannot be observed (BMD-a [3h]) for any Si gas sample subjected to heat treatment for 3 hours under any O ₂ gas flow rate condition. However, with respect to the Si sample (BMD-b [4h]) subjected to the heat treatment for 4 hours, in the case of O ₂ gas ± 0, the presence and shape of protrusions due to crystal defects could be confirmed (Δ). However, it could not be confirmed under other O ₂ gas flow rate conditions (×).

For the Si sample (BMD-c [8h]) heat-treated for 8 hours, the presence and shape of protrusions due to crystal defects can be well confirmed when the O ₂ gas flow rate is ± 0 and the flow rate is −10%. ○), even when the flow rate was −20%, the presence could be confirmed (Δ). For Si sample (BMD-d [16h]) heat-treated for 16 hours, the presence of protrusions due to crystal defects at any O ₂ gas reference flow rate ± 0, −10%, −20% Was well confirmed (○).

Table 1 (b) shows the SF ₆ gas flow rate with respect to the SF ₆ gas reference flow rate (± 0), −100% (0: not supplied), + 150% (2.5 times), + 300% (4 times the reference flow rate). The result of etching is shown. Cl ₂ gas flow rate to the reference flow rate (9 times the _{O 2} gas flow rate), _{O 2} gas flow rate to the reference flow rate, the pressure in the etch chamber is 25 mTorr, the cathode electrode temperature is 50 ° C., the source power is 750W.

From Table 1 (b), the flow rate of SF ₆ gas was most sensitive when not supplied, and the presence of protrusions was also confirmed in the Si sample that was heat-treated for 3 hours. As the flow rate of the SF ₆ gas increases, the etching of the side wall protective film of the protrusion caused by crystal defects proceeds, so that the protrusion becomes difficult to detect.

7A to 7D show SEM photographs of the Si sample when etching is performed by changing the flow ratio of the O ₂ gas and the SF ₆ gas. In each of FIGS. 7A to 7D, the right-side photograph is an SEM photograph observed at a higher magnification than the left-hand side. In the etching of each Si sample in FIG. 7, the heat treatment time is 16 hours, the Cl ₂ gas flow rate is the reference flow rate, the pressure in the etching chamber is 25 mTorr, the cathode electrode temperature is 50 ° C., and the source power is 750 W.

As can be understood from the comparison between FIGS. 7A and 7B, when the SF ₆ gas increases when the O ₂ gas flow rate is constant at the reference flow rate, the tip of the protrusion is removed by etching, and the protrusion cone The body shape is difficult to maintain. For this reason, it can be seen that the smaller the SF ₆ gas flow rate, the higher the crystal defect detection sensitivity.

FIG. 7 (c), the a 0 to SF ₆ gas as (d), if the reduced _{O 2} gas flow rate, towards the _{O 2} gas flow rate less FIG 7 (d) is, the etching of the tip portion of the projection Progresses. This is presumably because the O ₂ gas affects the formation amount of the side wall protective layer of the cone.

FIG. 8 shows the results of evaluating the influence of O ₂ gas on crystal defect detection etching by increasing its flow rate. In this evaluation, as shown in FIG. 8A, the O ₂ gas flow rate was set to the reference flow rate (± 0) and increased to + 10% and + 20%. FIGS. 8B and 8C are upper surface SEM photographs of the sample after etching the Si sample as + 10% when the O ₂ gas flow rate is the reference flow rate.

As can be seen from comparison with FIG. 8B, in FIG. 8C, a spot-like region is confirmed on the paper surface and on the center of the sample. This stain-like region, by the O ₂ gas is excessive, a region where the silicon oxide is deposited which is a reaction product of O ₂ gas and Si in a region free of inherent defects.

When the O ₂ gas flow rate was + 20%, more reactants were confirmed to be attached than in FIG.

As described above, the O ₂ gas is important in forming the sidewall protective film of the protrusion and maintaining the shape of the protrusion. However, when the flow rate is increased, an excessive reactive oxide is generated, which is a cause of erroneous detection. Become. For this reason, in the example of the present embodiment, the O ₂ gas flow rate is preferably in the range of 0.6 to about 1 with respect to the Cl ₂ gas flow rate 7 as described above.

In the evaluation of FIG. 8, the Cl ₂ gas flow rate is the reference flow rate, the SF ₆ gas flow rate is 0 (−100%), and the conditions of the pressure in the etching chamber are 25 mTorr, the source power is 750 W, and the cathode electrode temperature is 50 ° C. is doing.

When a semiconductor material such as a Si sample becomes larger than, for example, 300 mm, the amount of the O ₂ gas in the etching gas increases in the etching chamber of silicon that is a raw material of silicon oxide. Therefore, in this case, it is preferable to reduce the flow rate of O ₂ gas in the etching gas. The SF ₂ flow rate may be increased without decreasing.

(Evaluation 2: Pressure in the etching chamber)
Next, the evaluation result about the pressure in the etching chamber will be described. Table 1 (c) above shows the results of changing the pressure in the etching chamber during the crystal defect detection etching for each of the Si samples for 3 hours, 4 hours, 8 hours, and 16 hours of heat treatment. The pressure was set to 25 mTorr as a reference (± 0), −2 mTorr (23 mTorr), and −4 mTorr (21 mTorr). As shown in Table 1 (c), although it is possible to detect a crystal defect at any pressure, it can be seen that when it decreases, the detection sensitivity of the crystal defect tends to decrease. Note that the other conditions at the time of evaluation in Table 1 (c), (9 times the _{O 2} gas) Cl ₂ gas flow rate reference flow rate, _{O 2} gas flow rate as a reference flow rate, SF ₆ gas flow rate 0, the source The conditions of power 750 W and cathode electrode temperature 50 ° C. were adopted.

  Next, the effect of increasing the pressure will be described with reference to Table 2.

In Table 2, a wafer that has been heat-treated for 16 hours is used as the Si sample, and the etching gas is a Cl ₂ gas flow rate (reference flow rate), an O ₂ gas flow rate (reference flow rate), and an SF ₆ gas flow rate of 0. Was 750 W, and the cathode electrode temperature was 50 ° C. Under these conditions, as shown in Table 2 (d), the pressure in the etching chamber was set to +2 mTorr (27 mTorr) and +4 mTorr (29 mTorr), with 25 mTorr being the reference (± 0).

Here, the notation “X” in Table 2 indicates whether or not etching is hindered by a reaction byproduct of O ₂ and Si in a region that is originally free from defects. In other words, x means a process region that cannot be used for defect detection because etching does not proceed in a region where Si originally has no crystal defect to be etched.

  Therefore, it can be seen from Table 2 (d) that the pressure used in the defect detection process should be less than +2 mTorr based on the reference conditions.

(Evaluation 3: Evaluation of etching source power and cathode electrode temperature)
The evaluation results of the source power for supplying reactive ions into the etching chamber are shown in Table 2 (e). In Table 2 (e), the source power was set to +75 W (825 W) and +150 W (900 W) with a source power of 750 W as a reference (± 0) and other conditions being constant. As a result, it can be seen that the source power should be less than +150 W from the reference condition.

  The evaluation results of the cathode electrode temperature at which the sample to be etched is arranged in the etching chamber are shown in Table 2 (f) above. In Table 2 (f), the cathode electrode temperature is based on 50 ° C. (± 0), and other conditions are constant, and the cathode electrode temperature is + 20 ° C. (70 ° C.), +40 by operating the heater in the etching chamber. It was evaluated as ° C (90 ° C). Here, the cathode electrode temperature corresponds to the temperature of the Si sample mounted on the cathode electrode. From Table 2 (f), it can be understood that the cathode electrode temperature should be less than + 40 ° C. from the reference temperature (50 ° C.).

FIG. 9 shows the relationship between the base angle θ (taper angle) of the protrusion with the crystal defect as the apex, the source power (X), and the cathode electrode temperature (Y), and the Si samples (BMD-a, The result evaluated using b, c, d) is shown. In the figure, “Y = OFF” represents that the heater was turned off and the cathode electrode temperature was set to 50 ° C. as a reference. In FIG. 9, conditions other than the source power, cathode electrode temperature, and heat treatment time are fixed, and the Cl ₂ gas flow rate (reference flow rate), the O ₂ gas flow rate (reference flow rate), and the SF ₆ gas flow rate are 0, and the pressure in the etching chamber. Is 25 mTorr. FIG. 9A is an SEM photograph of the protrusions of each Si sample when the source power is 750 W and the cathode electrode temperature is 50 ° C., and FIG. 9B is the cathode electrode temperature of 50 ° C. and the source power is 825 W (+75 W). ) Is an SEM photograph of the protrusion of each Si sample. It can be seen that the base angle θ of the protrusions obtained by etching under the conditions of FIGS. 9A and 9B is substantially the same, and the influence of the source power on the base angle θ is relatively low.

  FIG. 9C is an SEM photograph of the protrusions of each Si sample when the source power is 750 W and the cathode electrode temperature is 70 ° C. (+ 20 ° C.). From a comparison between FIG. 9C and FIG. 9A, when the cathode electrode temperature rises, the base angle θ increases and the heat treatment time differs for the protrusions of the Si sample at any heat treatment time. It can be understood that the difference in the base angle θ of the protrusions in the Si sample becomes small.

  From the above results, it can be understood that the crystal defect revealing etching can be properly performed at a source power of 825 W or less and a cathode electrode temperature of 70 ° C. or less, and the detection resolution of crystal defects can be adjusted by the cathode electrode temperature. When the obtained base angle θ is large, the bottom area of the protrusion can be reduced. Therefore, even when the oxygen concentration in the Si sample is high and a large number of crystal defects exist, a plurality of protrusions are difficult to overlap, and it becomes easy to form protrusions having independent shapes with each of the large number of crystal defects as vertices. For this reason, detection sensitivity can be improved.

(Evaluation 4: Heat treatment time)
Next, the heat treatment time performed on the Si sample before the etching treatment will be described. FIG. 10 employs common etching conditions, and heat treatment times performed before etching are 3 hours (FIG. 10 (a)), 4 hours (same (b)), 8 hours (same (c)), 16 It is the photograph which image | photographed the upper surface after the etching of Si sample made into time (same (d)). The etching conditions were Cl ₂ gas flow rate (reference flow rate), O ₂ gas flow rate (reference flow rate), SF ₆ gas flow rate 0, pressure in the etching chamber 25 mTorr, source power 750 W, and cathode electrode temperature 50 ° C.

  In any of FIGS. 10A to 10D, it can be understood that a plurality of stripes exist concentrically from the center of the silicon wafer. A region in which a plurality of projections with crystal defects as apexes is formed is observed as the streaks, and the crystal defects are generated concentrically as shown in these photographs in the plane of the wafer. Further, the longer the heat treatment time is, the thicker the stripes are, and it can be understood that the oxygen precipitation defect nuclei spread as defects by the heat treatment. In addition, it can be understood from this result that it is effective to increase the heat treatment time before the crystal defect detection etching when it is necessary to increase the detection accuracy of the crystal defects that may occur in the semiconductor process.

  Further, it can be seen from the results of FIG. 10 that the distribution of crystal defects in the plane direction of the sample can be easily detected by observing the surface of the Si sample after etching. In addition, the surface of each Si sample of FIG. 10 is expanded as described above, and the distribution in the depth direction of the sample, the defect type, and the like are observed by observing individual protrusion shapes (height, base angle, tip shape). Can be evaluated.

  10 Semiconductor material (Si sample), 40 Crystal defect, 40e Projection caused by crystal defect, 42 Tip of projection.

Claims (7)

A crystal defect detection method comprising:
An etching gas containing at least Cl ₂ and O ₂ is used to perform high-selectivity anisotropic etching on crystal defects contained in a semiconductor material that is a crystal defect detection target. The semiconductor material projections that are formed with the crystal defects of
A crystal defect detection method, characterized by detecting a crystal defect based on the semiconductor material protrusion. The crystal defect detection method according to claim 1,
_2. The crystal defect detection method according to claim 1, wherein the flow rate of O ₂ is set in a range of about 0.6 to about 1 with respect to the flow rate 7 of Cl ₂ to the etching chamber. The crystal defect detection method according to claim 1,
The etching gas further contains SF ₆
The crystal defect detection method, wherein a flow rate of the SF ₆ to the etching chamber is set in a range of less than or equal to a half of the flow rate of the O ₂ and greater than 0. The crystal defect detection method according to claim 1,
The etching gas further contains SF ₆
The flow rate of O ₂ is in the range of about 0.6 or more and about 1.0 or less, and the flow rate of SF ₆ is more than 0 and about 0.4 or less with respect to the Cl ₂ flow rate 7 to the etching chamber. A crystal defect detection method characterized by being set in a range. In the crystal defect detection method according to any one of claims 1 to 4,
A crystal defect detection method, wherein a pressure in an etching chamber for performing the etching is set to 20 mTorr or more and 30 mTorr or less. In the crystal defect detection method according to any one of claims 1 to 5,
A crystal defect detection method, wherein a temperature of a cathode electrode disposed in an etching chamber for performing the etching is set to 50 ° C. or higher and lower than 90 ° C. In the crystal defect detection method according to any one of claims 1 to 6,
A crystal defect detection method, wherein the semiconductor material is subjected to heat treatment for 3 hours to 16 hours before the etching.