JP7687310B2 - Method for evaluating silicon substrates and method for managing the manufacturing process of silicon substrates - Google Patents

Description

The present invention relates to a method for evaluating silicon substrates (silicon wafers) and a method for managing the manufacturing process of silicon substrates.

In the manufacturing process of single crystal silicon wafers for semiconductor devices, the main surface is finished in a polishing process. In addition, a cleaning process is performed to remove abrasives and metal impurities that have adhered to the silicon wafer surface in the polishing process. This cleaning process uses a cleaning method called RCA cleaning. This RCA cleaning is a cleaning method that combines SC1 (Standard Cleaning 1) cleaning, SC2 (Standard Cleaning 2) cleaning, and DHF (Diluted Hydrofluoric Acid) cleaning depending on the purpose.
The SC1 cleaning is a cleaning method using an alkaline cleaning solution in which ammonia water and hydrogen peroxide water are mixed at an arbitrary ratio, and the adhering particles are lifted off by etching the silicon wafer surface, and the particles are removed while suppressing re-adhesion to the silicon wafer by utilizing the electrostatic repulsion between the silicon wafer and the particles. The SC2 cleaning is a cleaning method in which metal impurities on the silicon wafer surface are dissolved and removed using a cleaning solution in which hydrochloric acid and hydrogen peroxide water are mixed at an arbitrary ratio. The DHF cleaning is a cleaning method in which a natural oxide film on the silicon wafer surface is removed by dilute hydrofluoric acid. In addition, ozone water cleaning, which has a strong oxidizing power, may also be used to remove organic matter adhering to the silicon wafer surface and form a natural oxide film (also called a chemical oxide film) on the silicon wafer surface after DHF cleaning. Surface quality such as particles and surface roughness of the silicon wafer after cleaning is important, and these cleaning methods are combined according to the purpose.

Semiconductor elements such as MOS (Metal Oxide Semiconductor) capacitors and transistors are formed on the surface of semiconductor silicon wafers. The insulating films formed on these semiconductor elements, such as the gate oxide film, are used under high electric field strength, and silicon oxide film, which is easy to form, is often used as this insulating film.

An ellipsometer is one method for evaluating the thickness of an oxide film on a silicon substrate (Patent Document 1). An ellipsometer is a device that irradiates a polarized light onto a substrate sample and measures the change in the polarization state of the incident light and the reflected light to determine the phase difference (Δdelta) and amplitude ratio (Ψpsi). Taking a silicon oxide film on a silicon substrate as an example, the polarization state of the incident light changes as it is reflected by the silicon oxide film on the top surface and the interface between the silicon oxide film and the silicon substrate. There are two types of ellipsometers: a single-wavelength type that uses a laser as a light source and a spectral type that uses a white light source containing multiple wavelength components. The short-wavelength type is a method for measuring delta and psi for a specific wavelength (e.g., 633 nm), while the spectral type can measure delta and psi for each wavelength, and it is known that the spectral type with a large amount of information can be used to evaluate the film thickness more accurately.

As mentioned above, the information obtained by ellipsometer measurements is the phase difference and amplitude ratio, and it is not possible to directly calculate film thickness. To calculate film thickness, a model corresponding to the substrate sample is created, and the delta and psi theoretically obtained from this model are compared with the delta and psi obtained by ellipsometer measurements. The model is created by setting conditions according to the physical properties of the sample, and the set conditions include the material of the substrate and film, the film thickness of each film layer, and the optical constants of the substrate and film. In addition, a known reference corresponding to the sample and a required dispersion equation that shows the wavelength dependence of the dielectric constant and has multiple parameters are usually used to set each item.

Furthermore, a process is carried out to change the parameters of the dispersion equation and the film thickness of each film layer of the model so that the degree of difference between the two is minimized for the above comparison (also called fitting). The difference between the two is usually found by calculation using the least squares method, and when it is determined that the result obtained by the least squares method has become small to a certain extent through fitting, the refractive index and extinction coefficient of the film are found from the parameter values of the dispersion equation at that time, and the film thickness at that time is specified as the film thickness of the film possessed by the sample, thereby making it possible to find the film thickness. Note that model creation and fitting are generally performed manually or automatically using a computer based on the required program.

Patent Document 2 describes that the thickness of a native oxide film on a silicon wafer obtained by an ellipsometer varies depending on the surface roughness. Specifically, the rougher the surface, the thicker the film thickness, and discloses a method for quantitatively evaluating surface roughness based on the correlation between roughness and the thickness of the native oxide film.
Furthermore, an index of surface roughness on a silicon substrate is the haze value obtained by a particle counter. Haze is expressed as cloudiness and is widely used as an index of roughness on a silicon surface, with a high haze level indicating a rough wafer surface.

By the way, dense silicon oxide films with high insulating properties are produced by thermally oxidizing silicon wafers, but because silicon wafers shipped with natural oxide films formed during cleaning have natural oxide films present at the time of shipment in order to prevent particle adhesion, thermal oxidation is often carried out on silicon wafers with natural oxide films formed on them. In this case, it is known that the thickness of the thermal oxide film is affected by the film quality (thickness and structure) of the natural oxide film before thermal oxidation.
In recent years, with the miniaturization and multi-layering of semiconductor integrated circuits, there is a demand for further thinning of various films, including the insulating film constituting the elements. This thinning requires the formation of extremely thin insulating films, i.e., silicon oxide films, uniform and reproducible within a surface or between substrates. However, as the thermal oxide film becomes thinner, the ratio of the native oxide film before thermal oxidation to the film thickness after thermal oxidation increases, making the film more susceptible to the influence of the native oxide film before thermal oxidation.

Patent Document 3 describes the relationship between silicon wafers cleaned under various conditions and the oxide film thickness after thermal oxidation. Specifically, it discloses that increasing the NH ₄ OH concentration in the SC1 cleaning solution increases the amount of OH groups contained in the chemical oxide film, resulting in a thicker film after thermal oxidation, and that a method of controlling the film thickness after thermal oxidation is disclosed by using the correlation between the structure of the chemical oxide film and the film thickness after thermal oxidation.

Thus, the thickness of the film after thermal oxidation varies depending on the cleaning conditions of the wafer. In view of this background, in the manufacturing process of silicon wafers, in order to make the thermal oxide film of the silicon wafers to be manufactured uniform between wafers, it is desirable to clean them under the same conditions as much as possible. The cleaning process is often a combination of SC1 cleaning, SC2 cleaning, hydrofluoric acid cleaning, and ozone water cleaning. It is desirable to make the cleaning flow the same as well as the chemical concentration, chemical temperature, and cleaning time. However, in the actual cleaning process, it is difficult to completely avoid slight changes in the chemical concentration, for example, at the beginning, middle, and end of the chemical life, or slight fluctuations in temperature due to deterioration of the temperature sensor that controls the chemical temperature.
Even such a slight change in conditions can change the etching behavior, especially in SC1 cleaning, and cause the film thickness after thermal oxidation to vary. Therefore, to check whether such film thickness variations have occurred, it is desirable to extract a wafer during the cleaning process, for example, and evaluate the thermal oxide film.

Generally, to evaluate the film thickness after thermal oxidation, it is necessary to actually perform thermal oxidation treatment using an oxidation furnace, but this is time-consuming and laborious, and the maintenance and management of the oxidation furnace is also difficult. Therefore, as in Patent Document 3, it is a very effective method to estimate the film thickness after thermal oxidation from the structure of the chemical oxide film after cleaning without performing thermal oxidation treatment. In this case, the higher the throughput and the higher the reproducibility, the more suitable it is for manufacturing process management in particular.
Furthermore, if the thermal oxide film can be quickly evaluated without performing thermal oxidation treatment, cleaning conditions for controlling the thermal oxide film thickness can be easily determined. In addition, for example, in failure analysis, there are cases where the thermal oxide film thickness of a silicon wafer with an unknown cleaning history needs to be estimated. In such cases, if the thermal oxide film thickness can be estimated from the information on the native oxide film, it will be useful analysis information.

JP 2005-283502 A Japanese Patent Application Publication No. 06-163662 Patent No. 6791453

As described above, there is a need for a method for estimating the thickness of a thermal oxide film formed on a silicon substrate with good throughput without performing thermal oxidation processing. As described above, Patent Document 3 discloses a method for estimating the thickness of a thermal oxide film from the structure of a chemical oxide film before thermal oxidation without performing thermal oxidation processing. This method is excellent in that it does not require actual thermal oxidation processing, but the analysis to evaluate the structure of this chemical oxide film and the analysis take time. In other words, there is a need for an index that is correlated with the thermal oxide film thickness and can be analyzed more easily and with better throughput than the structure of such a chemical oxide film.

The present invention has been made to solve the above problems, and aims to provide a method for estimating the thickness of a thermal oxide film when a thermal oxidation process is performed, for example, in a device manufacturing process, with better throughput and more easily than conventional methods that use the composition of a chemical oxide film as an index. In addition, the present invention aims to provide a method for controlling with high accuracy the manufacturing process of silicon substrates (particularly the cleaning process), which affects the thickness of thermal oxide films, and a method for estimating the cleaning history (particularly the presence or absence of SC1 cleaning) of silicon substrates with unknown cleaning history.

In order to achieve the above object, the present inventors have conducted extensive research into indicators on silicon substrates before thermal oxidation that are correlated with the thickness of the thermal oxide film. In particular, since the thickness of the thermal oxide film depends on the cleaning conditions, they focused on whether there is an indicator that reflects the cleaning history of the silicon substrate before thermal oxidation. Furthermore, they focused on the fact that, as the etching action progresses in SC1 cleaning, the composition of the formed chemical oxide film changes (for example, the amount of OH groups increases) and at the same time, the surface roughness of the silicon substrate also deteriorates. They also considered that such changes in the composition of the chemical oxide film and deterioration in the surface roughness can be interpreted as the cleaning history of the silicon substrate.
Here, we thought that there was a correlation between the change in the composition of this chemical oxide film and the surface roughness of the silicon substrate. Based on the above, we thought that the degree of surface roughness of the silicon substrate formed by SC1 cleaning is also the SC1 cleaning history, and if there is an index that reflects such surface roughness, this index also indirectly reflects the composition of the chemical oxide film, so a correlation with the thermal oxide film thickness can be obtained, and we conducted extensive research.
As a result, the inventors discovered that, for a silicon substrate whose thermal oxide film thickness is to be estimated, the native oxide film is removed under non-etching conditions, and then the thickness of an evaluation oxide film formed under oxidation conditions without etching action correlates with the thickness of the thermal oxide film, and that the thickness of the thermal oxide film can be estimated by using the thickness of this evaluation oxide film as an index, thereby completing the present invention.

That is, the present invention is a method for evaluating a silicon substrate, comprising the steps of:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
an evaluation step of evaluating the evaluation substrate based on the thickness of the evaluation oxide film measured in the film thickness measurement step without performing a thermal oxidation treatment on the evaluation substrate,
In the evaluation step,
(i) a film thickness estimation evaluation step of estimating and evaluating a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate;
(ii) a cleaning history estimation evaluation step of estimating and evaluating the cleaning history of the prepared evaluation substrate.

In the silicon substrate evaluation method, the thickness of the thermal oxide film can be estimated by using the thickness of the evaluation oxide film, which is correlated with the thermal oxide film, as an index. In addition, the evaluation method can be performed more simply and with a higher throughput than the conventional analysis of the composition of a chemical oxide film.
In a normal manufacturing process, the cleaning history of a manufactured silicon substrate is stored as data, but it is possible that the cleaning history is not stored for some reason, such as when a certain period of time has passed. In such a case, the evaluation method of the present invention can be performed to estimate the cleaning history of a silicon substrate whose cleaning history is unknown (in particular, whether SC1 cleaning has been performed).

At this time, in the oxide film removal process, the native oxide film can be completely removed by cleaning with hydrofluoric acid.

In the present invention, it is important to reflect the cleaning history of the silicon substrate whose thermal oxide film thickness is to be estimated in the film thickness of the oxide film to be evaluated. Hydrofluoric acid cleaning is suitable because it has no etching effect on the silicon substrate and can completely remove the native oxide film.

In addition, in the evaluation oxide film formation process, the evaluation oxide film can be formed using ozone water or hydrogen peroxide water.

As mentioned above, it is preferable to form the oxide film for evaluation using ozone water or hydrogen peroxide water, which does not have any unnecessary etching effect, because the thickness of the oxide film for evaluation can reflect the cleaning history of the silicon substrate, which is the original purpose.

In addition, when the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measuring step of forming a thermal oxide film on another of the plurality of preliminary substrates by performing a thermal oxidation process on the other preliminary substrate and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
Based on the correlation, the thickness of the thermal oxide film of the evaluation substrate can be estimated from the thickness of the evaluation oxide film of the evaluation substrate.

By obtaining a correlation in advance through such preliminary testing, it is possible to easily estimate the thickness of the thermal oxide film on the substrate being evaluated.

Further, based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
The cleaning history of the prepared evaluation target substrate can be estimated and evaluated.

In this way, the thermal oxide film thickness can be easily evaluated, and the cleaning history can be easily estimated using the estimated thermal oxide film thickness as an index.

In addition, when estimating and evaluating the cleaning history of the prepared evaluation substrate, it is possible to determine whether SC1 cleaning has been performed.

In this way, the evaluation oxide film or estimated thermal oxide film thickness can be used as an indicator to easily determine whether or not SC1 cleaning was performed in the cleaning history.

The present invention also provides a method for managing a manufacturing process of a silicon substrate, comprising the steps of:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
a management step of managing the manufacturing process based on the film thickness of the evaluation oxide film measured in the film thickness measurement step without performing a thermal oxidation process on the prepared evaluation substrate,
In the management step,
When the thickness of the evaluation oxide film deviates from a predetermined thickness control value set in advance,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one of a cleaning history investigation of the substrate to be evaluated, and a cleaning management process for modifying process conditions for the cleaning treatment in the manufacturing process.

The method of using the evaluation oxide film as an index in the silicon substrate evaluation method of the present invention described above has better throughput and is simpler than conventional methods, so it can also be used as a method for managing the silicon substrate manufacturing process. With such a management method, it is possible to quickly detect fluctuations in the thermal oxide film thickness caused by unintended and unpredictable slight fluctuations in the cleaning conditions. In addition, these results can be used to provide feedback to the manufacturing process (cleaning process) to correct (improve) the conditions.

In the manufacturing process control method of the present invention, the native oxide film can be completely removed by washing with hydrofluoric acid in the oxide film removal process, and the evaluation oxide film can be formed by using ozone water or hydrogen peroxide water in the evaluation oxide film formation process.

It is important that the cleaning history of the silicon substrate is reflected in the thickness of the oxide film to be evaluated. Hydrofluoric acid cleaning is suitable because it has no etching effect on the silicon substrate and can remove the native oxide film.
Similarly, it is preferable to form the oxide film for evaluation using ozone water or hydrogen peroxide water, which have no unnecessary etching effect, because the thickness of the oxide film for evaluation can reflect the cleaning history of the silicon substrate, which is the original purpose.

The present invention also provides a method for managing a manufacturing process of a silicon substrate, comprising the steps of:
The method further includes a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step of the silicon substrate evaluation method described above,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one of a cleaning history investigation of the substrate to be evaluated, and a cleaning management process for modifying process conditions for the cleaning treatment in the manufacturing process.

The above-mentioned manufacturing process control method uses the evaluation oxide film as an index, but instead of the evaluation oxide film, the thermal oxide film thickness estimated using the above-mentioned evaluation method can also be used as an index to control the manufacturing process. This type of control method can also quickly detect fluctuations in the thermal oxide film thickness, allowing feedback and correction of cleaning process conditions.

The silicon substrate evaluation method of the present invention allows the thermal oxide film thickness to be evaluated more simply and with a higher throughput than in the past, and also allows the cleaning history (particularly whether the silicon substrate has been subjected to SC1 cleaning) to be known.
Furthermore, the method for controlling the manufacturing process of silicon substrates of the present invention makes it possible to quickly detect unintended and unexpected slight variations in the thickness of the thermal oxide film, and to feed back the results to the manufacturing process to improve the cleaning process.

1 is a flowchart showing an example of a main test of the silicon substrate evaluation method (estimated evaluation of thermal oxide film thickness) of the present invention. 1 is a graph showing haze values before and after cleaning with hydrofluoric acid and ozone water, or SC1 cleaning, and the increase (difference) after cleaning. 13 is a graph showing the thickness of an oxide film for evaluation after additional cleaning with hydrofluoric acid and cleaning with ozone water. 1 is a graph showing the dependency of etching amount on chemical temperature when SC1 cleaning is performed for 3 minutes. 1 is a graph showing the influence of haze and natural oxide film thickness on the thermal oxide film thickness of each sample. 1 is a graph showing the influence of the amount of OH groups and suboxides on the thermal oxide film thickness of each sample. 1 is a graph showing the influence of the evaluation oxide film thickness of the present invention on the thermal oxide film thickness of each sample. 1 is a flowchart showing an example of a preliminary test of the silicon substrate evaluation method (estimated evaluation of thermal oxide film thickness) of the present invention. 1 is a flowchart showing an example of a method for evaluating a silicon substrate according to the present invention (estimated evaluation of cleaning history: judgment evaluation of SC1 cleaning). 1 is a flowchart showing an example of a method for managing a manufacturing process of a silicon substrate according to the present invention. 1 is a graph showing the correlation between the thickness of an oxide film for evaluation and the thickness of a thermal oxide film obtained in the verification of Example 1. 1 is a flowchart showing an investigation of the relationship between wafer surface roughness and oxide film thickness. 1 is a graph showing the relationship between the surface roughness (haze) and the native oxide film thickness of a silicon substrate that has been roughened by CMP and SC1 cleaning. 1 is a graph showing the relationship between the surface roughness (haze) and the native oxide film thickness of a silicon substrate that has been roughened by single-wafer cleaning.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings, but the present invention is not limited thereto.
<Preliminary investigation by the present inventors>
As described above, the present invention is an evaluation method for estimating the thickness of a thermal oxide film by finding an index that is correlated with the thickness of the thermal oxide film. First, generally, silicon wafers are cleaned using a combination of SC1, SC2, hydrofluoric acid, and ozone water. Among these cleaning methods, SC1 cleaning in particular is a cleaning method that involves an etching action, and this etching action progresses more with a higher ammonia concentration, a higher chemical temperature, and a longer cleaning time.

This etching action changes the structure of the chemical oxide film on the silicon substrate (for example, the amount of OH groups increases), and the surface roughness also worsens. The stronger the etching action, the more significant this change or deterioration becomes. For example, in SC1 cleaning, if the chemical solution concentration and cleaning time are fixed and the chemical solution temperature is changed to 70°C and 40°C, the etching action is stronger at 70°C than at 40°C, so the structure of the chemical oxide film changes significantly, the surface roughness worsens, and the thermal oxide film becomes thicker.
Therefore, it can be said that there is a certain correlation between the structure of the chemical oxide film and the wafer surface roughness. And since this wafer surface roughness is caused by the etching action of SC1 cleaning, the rougher the surface roughness, the stronger the etching action. In other words, if an index that reflects the wafer surface roughness formed by the etching action of SC1 cleaning applied to a silicon substrate can be found, this index will be an index that reflects the structure of the chemical oxide film, and a correlation will be obtained with the thermal oxide film thickness.

Next, we will explain the index reflecting the wafer surface roughness formed by the etching action of SC1 cleaning. First, we will explain the relationship between the thickness of the oxide film and various wafer surface roughnesses formed in the silicon wafer manufacturing process. Figure 12 is the investigation flow chart.
The prepared silicon substrate was subjected to roughening treatment to form roughness by changing the CMP processing conditions and SC1 cleaning conditions. Next, the oxide film was completely peeled off by hydrofluoric acid cleaning in a batch cleaning machine, and then an oxide film was formed by cleaning with ozone water. In both levels (CMP level, SC1 level), the oxide film of the roughening treatment was completely peeled off by hydrofluoric acid cleaning, and the oxide film (reformed natural oxide film) formed by the subsequent ozone water was formed, so it can be interpreted that the oxide film was formed by the same method. After that, haze measurement was performed using a particle counter, and the thickness of the natural oxide film was evaluated by spectroscopic ellipsometry.

Figure 13 shows the results of the natural oxide film (reformed as described above) for each level of haze. At the CMP level (■), the thickness of the natural oxide film was the same even when the haze value exceeded 10 ppm, but at the SC1 cleaning level (●), there was a tendency for the natural oxide film to become thicker as the haze increased. Since the oxide film was formed using the same method, it was assumed that the film thickness would be the same, as with the CMP level, but this was not the case with the SC1 cleaning level.

Furthermore, Fig. 14 shows the results of the roughening process performed by single-wafer cleaning combining hydrofluoric acid and ozone water cleaning. Hydrofluoric acid cleaning and ozone water cleaning after roughening were also performed by single-wafer cleaning instead of batch-type. Since the oxide film formation method is different from the batch-type ozone water cleaning, it is not possible to compare the film thickness at the above-mentioned SC1 level and CMP level with the film thickness at this single-wafer cleaning level, but it is possible to discuss the influence of haze within the single-wafer cleaning level. As a result, the single-wafer cleaning level had the same natural oxide film thickness even if the haze changed, just like the CMP level.
To summarize the above results, it has been newly discovered that the roughness formed by CMP and single-wafer cleaning that combines hydrofluoric acid and ozone water cleaning does not affect the thickness of the oxide film, while the roughness formed by SC1 cleaning increases the thickness of the oxide film.

These results show that at the SC1 level, the stronger the SC1 etching action, the worse the surface roughness becomes, and as a result, the thickness of the oxide film (the above-mentioned re-formed native oxide film) also increases due to the effect of surface roughness. In other words, if the oxide film (the above-mentioned re-formed native oxide film) evaluated using this method is thick, it can be determined that the silicon substrate has been subjected to SC1 cleaning, which has a strong etching action.

<Silicon Substrate Evaluation Method of the Present Invention>
[Evaluation of Thermal Oxidation Film Thickness on Silicon Substrate]
Based on the results of the above preliminary investigation, the evaluation method of the present invention will be described. Figure 1 shows an example of the evaluation method of the thermal oxide film thickness of a silicon substrate according to the present invention. This is a flow for evaluating the evaluation substrate, which is the actual evaluation target (main test flow). A preliminary test is performed to obtain a correlation (relationship between the evaluation oxide film thickness and the thermal oxide film thickness) using a spare substrate prepared separately from the evaluation substrate in the main test, and this correlation can be used to evaluate the thermal oxide film thickness of the evaluation substrate in the main test. The flow of this preliminary test will be described later.
The flow of this test is described in detail below.
(S1: Substrate preparation process)
First, in step S1, a silicon substrate for which the thermal oxide film thickness is to be estimated is prepared as a substrate to be evaluated.
In this case, there is no restriction on the conductivity type, aperture (diameter), or shape of the silicon substrate. However, it is preferable to prepare the surface roughness taking the following points into consideration. Specifically, it is preferable that the arithmetic mean height Sa is 0.5 nm or less. The reason for this will be explained using Patent Document 2 and Patent Document 3.
Patent Document 3 describes the surface roughness Ra after SC1 cleaning or ozone water cleaning, and the value is about 0.06 to 0.12 nm. Such Ra values are the roughness values of silicon wafers that have been used in recent years.
Patent Document 2 discloses that the roughness of the wafer surface affects the thickness of the native oxide film measured by an ellipsometer. In this case, the surface roughness value is 0.22 to 2.05 nm in terms of the Ra value measured by AFM, which is very high compared to the above-mentioned surface roughness value of 0.06 to 0.12 nm.

In addition, while the natural oxide film is generally known to be about 1 nm thick, in Patent Document 2, the natural oxide film thickness is 0.097 nm when the Ra value is 0.22 nm, 1.586 nm when the Ra value is 1.23 nm, and 3.313 nm when the Ra value is 2.05 nm, all of which are far from about 1 nm. In this way, the surface roughness and thickness of the natural oxide film in Patent Document 2 are significantly different from the surface roughness and thickness of silicon wafers used in recent years. The reason for this is thought to be that Patent Document 2 intentionally roughens the surface using a mixture of hydrofluoric acid and nitric acid, which is not used in normal silicon wafer cleaning solutions. In other words, it is difficult to discuss the roughness and natural oxide film thickness in the range of Ra of 0.06 to 0.12 nm, for example, using the correlation disclosed in Patent Document 2, and it is presumed that it can be applied to very rough surfaces, such as those with an Ra value of more than 1 nm.

From these considerations, it is preferable that Sa is 0.5 nm or less for the evaluation of the present invention. If the Sa value is 0.5 nm or less, it is possible to prevent the thickness of the evaluation oxide film described below from deviating significantly from 1 nm due to the influence of roughness. However, since the Sa value of the front surface of a silicon wafer after CMP is generally 0.1 nm or less and the Sa value of the back surface (DSP surface) is about 0.2 to 0.4 nm, the cleaning method of the present invention is suitable for wafers at least after DSP processing or CMP processing. In fact, wafers that undergo thermal oxidation processing for device fabrication are silicon wafers that have undergone the DSP process, CMP process, and cleaning process, so the surface roughness Sa of silicon substrates for which the thermal oxide film thickness is to be evaluated is usually 0.5 nm or less in most cases.

(S2: Oxide film removal step)
Next, in step S2, the native oxide film on the evaluation substrate is peeled off (completely removed) under conditions that do not involve any etching action.
Generally, a natural oxide film is formed on a silicon wafer after a cleaning process by SC1 cleaning or _O3 cleaning in order to reduce particle adhesion. However, the thickness of these natural oxide films varies depending on the cleaning conditions, and as described in Patent Document 3, the thickness of the thermal oxide film does not depend on the thickness of the natural oxide film, so it cannot be an indicator of the thermal oxide film. Therefore, it is necessary to completely peel off the natural oxide film that was once formed and form a new natural oxide film (oxide film for evaluation) in the next S3 process.

Here, the etching action will be explained. The etching action means that, for example, silicon is removed by applying a chemical solution or plasma to silicon. In the cleaning process, SC1 cleaning exhibits this etching action, and hydroxide ions ionized from ammonia cause the etching action. It is generally known that as etching progresses, the surface roughness after etching worsens. As described above, in the present invention, the film thickness reflecting the cleaning history applied to the silicon substrate is used as an index, so if an unnecessary etching action is applied to the substrate to be evaluated in the S2 process, the influence caused by the S2 process will be taken into account in addition to the cleaning history that is originally desired to be reflected, so it is necessary to peel off the oxide film under conditions without an etching action in the S2 process. However, since the oxide film is peeled off with a chemical solution or plasma, it is difficult to completely eliminate the etching action and to completely prevent the surface from becoming rough when focusing on the silicon substrate surface from a macroscopic or microscopic perspective. Therefore, in the evaluation method of the silicon substrate of the present invention (and the management method of the manufacturing process of the silicon substrate of the present invention described later), "no etching action" is defined as no etching action that creates a surface roughness that affects the evaluation method (management method) of the present invention.
In other words, a very slight, negligible etching effect is acceptable.

An example of this etching effect is shown in Figure 2. Figure 2 shows the haze before and after cleaning, and the difference before and after cleaning, for a silicon substrate that was cleaned with hydrofluoric acid and ozone water and that was cleaned with SC1. The SC1 cleaning was performed at 40°C to 70°C for 3 minutes.
Overall, it can be seen that the higher the temperature conditions, the stronger the etching effect of SC1, the greater the deterioration of haze after cleaning. In contrast, when only hydrofluoric acid and ozone water cleaning was used, the haze value deteriorated slightly, but it was much smaller than the SC1 level. Therefore, although the surface became slightly rough after hydrofluoric acid cleaning and ozone water cleaning, the degree of the deterioration was very small, and it can be said that there was no etching effect in the present invention.

Next, in relation to the level shown in FIG. 2, the thickness of the native oxide film after additional washing with hydrofluoric acid and ozone water is shown in FIG.
Due to the influence of the wafer surface roughness formed by the etching action of the SC1 cleaning described above, the thickness of the native oxide film tends to become thicker as the cleaning temperature increases. Compared to the case without SC1 cleaning, the film thickness was the same when the SC1 cleaning temperature was up to 40°C, but became thicker when the temperature was 45°C or higher. That is, in this case, the etching action was weak up to the chemical temperature of 40°C, and no wafer surface roughness that would affect the film thickness was created, but conversely, it is thought that at temperatures of 45°C or higher, wafer surface roughness was created, causing the thickness of the native oxide film to fluctuate.

Therefore, in this case, if the cleaning temperature is at a level of 40° C. or less, a slight etching effect occurs, but it is thought that this does not affect the evaluation method of the present invention. The reason for this lies in the etching amount of SC1 cleaning. That is, in the S2 step, as described above, it is necessary to completely remove the native oxide film formed on the evaluation substrate. And in the S3 step, it is necessary to form an evaluation oxide film. Also, since the thickness of the native oxide film is generally about 1 nm or more, when using SC1 cleaning in the S2 step, the etching amount must be 1 nm or more in order to completely remove the native oxide film.
Here, Fig. 4 shows the etching amount when etching was performed for 3 minutes under the same cleaning conditions as those in Fig. 3. From this graph, it can be seen that an etching amount exceeding 1 nm cannot be ensured unless the cleaning temperature is at least 70°C or higher. However, at a cleaning temperature of 70°C, as shown in the graph in Fig. 3, the film thickness is significantly thicker, which is considered to be due to the influence of the above-mentioned S2 step.
It is easily conceivable to increase the cleaning time at a cleaning temperature of 40° C., but since the roughness formed by SC1 cleaning is due to the amount of etching, it is difficult to avoid the effect on the evaluation method of the present invention even if etching is performed for a long time at a low etching rate.

Although the method of removing the oxide film in the S2 step is not particularly limited except for "conditions without etching action," the above results suggest that it is difficult to avoid the creation of surface roughness that affects the evaluation method of the present invention while completely removing the native oxide film with SC1 cleaning, and therefore the inventors have investigated other methods. As a result, it was found that the most suitable method of removing the oxide film is hydrofluoric acid cleaning. Note that, as long as the native oxide film can be completely removed, the hydrofluoric acid cleaning conditions are not particularly limited. One example of the conditions is a hydrofluoric acid concentration of 0.3 to 5.0 wt%, a temperature of 10 to 30°C, and a cleaning time of 60 to 360 seconds.
Other oxide film removal methods include dry etching methods such as ALE (Atomic Layer Etching), in which conditions can be applied for selectively etching only SiO ₂ and not etching Si.

(S3: Evaluation oxide film formation process)
Next, in step S3, an oxide film for evaluation is formed on the evaluation substrate under oxidation conditions that have no etching effect.
Here too, as in the S2 step, it is necessary to avoid applying unnecessary etching to the evaluation substrate. Note that here too, "no etching effect" is defined as no etching effect that creates surface roughness that would affect the evaluation method (control method) of the present invention. Both ozone water and hydrogen peroxide water caused very little haze deterioration before and after cleaning.
Therefore, it is preferable to oxidize the wafer surface with ozone water or hydrogen peroxide water, which has no etching effect, and ozone water, which has a stronger oxidizing power, is more preferable. Ozone water or hydrogen peroxide water can easily form an oxide film (reformed natural oxide film). For example, the concentration of ozone water used is in the range of 3 to 25 ppm, the temperature is 10 to 30° C., and the cleaning time is 60 to 360 seconds. The concentration of hydrogen peroxide water used is 0.2 to 5.0 wt%, the temperature is 30 to 90° C., and the cleaning time is 60 to 360 seconds. In this manner, an evaluation oxide film can be formed. In addition, since the thickness of this evaluation oxide film is used as an index in the step S5 described later, when there are multiple substrates to be evaluated, it is particularly preferable to use the same method for forming the oxide film for each of the substrates. For example, a more accurate evaluation can be achieved by comparing the thicknesses of evaluation oxide films formed by the same method, rather than comparing the thicknesses of evaluation oxide films formed by ozone water and hydrogen peroxide water. Furthermore, even when the ozone water cleaning is performed, it is preferable that the concentration and cleaning time are similar, since this results in a more accurate evaluation, rather than when the concentration and cleaning time are significantly different.
As another oxidation method, an oxide film can be formed by using plasma or the like in a dry method such as ALD (Atomic Layer Deposition). In particular, the CVD method is effective because it does not have an etching effect on the silicon substrate.

(S4: Film thickness measurement process)
Next, in step S4, the thickness of the evaluation oxide film is measured.
This oxide film for evaluation can be measured, for example, by a spectroscopic ellipsometer, and a known method may be used. In this S4 step, a film thickness reflecting the cleaning history of the silicon substrate can be obtained. In this case, it is preferable to make the elapsed time from the end of the S3 step to the S4 step as uniform as possible. This is because the thickness of the oxide film for evaluation formed in the S3 step is oxidized by moisture and oxygen in the atmosphere over time, and the oxide film becomes slightly thicker. In order to suppress this variation, evaluation can be performed with higher accuracy by evaluating the film after a preset elapsed time, such as about 1 hour or 24 hours after the S3 step.

(S5: Evaluation step [film thickness estimation evaluation step])
Next, in step S5, the substrate is evaluated based on the evaluation oxide film thickness obtained in step S4 without performing a thermal oxidation process on the substrate. Here, the thickness of the thermal oxide film formed when the substrate is subjected to a thermal oxidation process is estimated from the evaluation oxide film thickness.
The conditions of the thermal oxidation treatment are not particularly limited. For example, dry oxidation using oxygen gas or wet oxidation using water vapor can be used. The target film thickness varies depending on the thermal oxidation conditions in the device fabrication process, so the maximum temperature and time can be adjusted as appropriate.
A more detailed explanation is given below. Here, 11 levels of samples were prepared by changing the cleaning and polishing conditions, with three samples for each level. Of the 11 levels, two were cleaned with hydrofluoric acid and ozone water, and the remaining nine were cleaned with SC1.
One of the three pieces was used to obtain haze using a particle counter, and the thickness of the native oxide film was evaluated using a spectroscopic ellipsometer. An oxide film for evaluation was then formed using hydrofluoric acid and ozone water cleaning, and the thickness of the oxide film was evaluated using a spectroscopic ellipsometer.
The other piece was subjected to a thermal oxidation treatment as it was, and the thermal oxide film was evaluated by spectroscopic ellipsometry.
For the last piece, the amount of OH groups and the ratio of suboxides were evaluated by ATR-FT-IR and XPS following Patent Document 3.

Figure 5 shows the influence of haze and native oxide film thickness on the thermal oxide film thickness at each level. The correlation coefficient ^R2 value is small in both cases, and it can be seen that there is simply no correlation between the haze of the wafer surface roughness or the thickness of the native oxide film and the thermal oxide film thickness. This is because the haze also takes into account other roughness components such as those from the CMP process in addition to the SC1 cleaning roughness. It can also be seen that the thickness of the native oxide film alone cannot be used as an indicator.
Fig. 6 shows the composition of the chemical oxide film, with the amount of OH groups and the ratio of suboxides as indicators. The ^R2 value is higher than that in Fig. 5, and it is clear that there is a correlation with the thermal oxide film. Therefore, as described in Patent Document 3, it is possible to estimate the thickness of the thermal oxide film from the composition of the chemical oxide film.
FIG. 7 shows the case where the thickness of the oxide film for evaluation of the present invention is used as an index. ^{The R2} value is very high at 0.9 or more, and a good correlation was observed between the oxide film for evaluation and the thermal oxide film. Therefore, it can be estimated that the thicker the oxide film for evaluation is, the thicker the thermal oxide film will be. In this way, the thickness of the oxide film for evaluation of the present invention has a very good correlation with the thickness of the thermal oxide film, so that the thermal oxide film can be estimated with high accuracy. Note that the two on the left side of the graph (thin oxide film for evaluation) are those cleaned with hydrofluoric acid and ozone water, and the nine on the right side are those cleaned with SC1.
For example, by comparing the thickness of each of the evaluation oxide films on a plurality of evaluation substrates, it is possible to perform a comparative evaluation of the size of each of the thermal oxide films on those evaluation substrates.

Here, a case where a preliminary test is performed will be described.
In the evaluation method of the present invention, the thermal oxide film thickness can be more specifically estimated by obtaining a correlation between the evaluation oxide film thickness and the thermal oxide film thickness as shown in Fig. 7 in advance. A flow chart of a typical preliminary test in such a case is shown in Fig. 8. The preliminary test may be performed before the evaluation step [thickness estimation evaluation step] of the main test.
(S11: Spare-substrate preparation process)
First, in step S11, a plurality of spare substrates (two substrates in this example) are prepared as a reference, which are silicon substrates and have the same cleaning conditions. For each set of two substrates, a plurality of sets are prepared, each set having different cleaning conditions, for example. The number of sets is preferably as large as possible, since a highly accurate correlation can be obtained, but the number of sets can be appropriately determined in consideration of the effort and time required. The above FIG. 7 can be said to have 11 sets (11 levels).
(S12: preliminary oxide film removal process, preliminary oxide film formation process for evaluation, preliminary film thickness measurement process, and preliminary thermal oxide film thickness measurement process)
Next, in step S12, steps similar to steps S2-S4 in FIG. 1 are carried out on one of the spare substrates, and the thickness of the evaluation oxide film on the spare substrate is measured.
The other side is subjected to a thermal oxidation process to evaluate the thickness of the thermal oxide film. The conditions for this thermal oxidation process should be similar to those for the thermal oxidation process assumed in the above-mentioned main test.
(S13: correlation acquisition step)
Next, in step S13, the correlation between the evaluation oxide film thickness (the evaluation oxide film thickness of the preliminary substrate) and the thermal oxide film thickness (the thermal oxide film thickness of the preliminary substrate) is obtained.
This completes the flow of the preliminary test.

The above-mentioned preliminary test may be carried out before the main test, and then steps S1 to S4 of the main test may be carried out.
Then, in step S5, the specific value of the thermal oxide film thickness of the evaluation substrate, whose thermal oxide film thickness is unknown, can be estimated very easily based on the correlation in S13 from the evaluation oxide film thickness of the silicon substrate (substrate to be evaluated in this test).
A specific example of the correlation is shown in Fig. 7. If the thickness of the evaluation oxide film on a silicon substrate (substrate to be evaluated in this test) whose thermal oxide film thickness is unknown is 1.265 nm, the thermal oxide film thickness can be estimated to be 5.147 nm by substituting the thickness into the correlation formula (correlation) in Fig. 7.

This method is extremely useful when there are only a few thermal treatment conditions that are expected to be applied to a silicon substrate, as it allows the actual thermal oxide film thickness to be estimated. For example, when there are two thermal oxidation conditions, one level for the dry method and one level for the wet method, the thermal oxide film thickness can be easily estimated by obtaining a correlation in advance. When there are a large number of thermal oxidation treatments that are applied to a silicon substrate, it is particularly preferable to obtain a correlation for each thermal oxidation condition.

The evaluation method of the present invention (evaluation of thermal oxide film thickness) described above can estimate thermal oxide films more easily and with better throughput than conventional methods of analyzing the composition of chemical oxide films, and can be said to be an extremely effective evaluation method.

[Estimation of cleaning history of silicon substrate (judgment of SC1 cleaning)]
In a normal manufacturing process, the cleaning history of a manufactured silicon wafer is stored as data, but it is possible that the cleaning history is not stored for some reason, such as when a certain period of time has passed. In such a case, the cleaning history can be estimated by carrying out another form of the evaluation method of the present invention using the evaluation oxide film thickness as an index. The content of this cleaning history is not particularly limited, but here, a method for determining whether the silicon wafer has been subjected to SC1 cleaning will be described. Figure 9 shows a flow of the method for determining and evaluating SC1 cleaning.
As for the procedure, first, steps similar to steps S1 to S4 in FIG. 1 are carried out, and the thickness of the evaluation oxide film on the evaluation substrate is measured.
(S5': Evaluation step [SC1 judgment evaluation step])
Thereafter, in step S5', without carrying out a thermal oxidation process on the prepared evaluation substrate, it is determined whether or not SC1 cleaning has been performed in the cleaning history of the prepared evaluation substrate based on the evaluation oxide film thickness.

The criteria for determining whether or not SC1 cleaning is performed are not particularly limited, but for example, the correlation between the cleaning conditions (with SC1 cleaning and without SC1 cleaning) and the evaluation oxide film thickness (and, if necessary, the thermal oxide film thickness) can be obtained and the criteria for determination can be determined based on the correlation. For example, in the sample shown in FIG. 7, there are two levels of cleaning with hydrofluoric acid and ozone water as described above, and the thicknesses of the evaluation oxide films are 1.236 nm and 1.237 nm, respectively, which are very thin within these levels. The other levels are levels where SC1 cleaning is performed, and the thickness of the evaluation oxide film at the level with the smallest amount of etching by SC1 is 1.257 nm, which is thicker than the above-mentioned 1.236 nm and 1.237 nm.
Therefore, if the thickness of the evaluation oxide film on a silicon wafer (substrate to be evaluated) with an unknown cleaning history is at least 1.257 nm, it can be assumed that the wafer has been subjected to SC1 cleaning.

The judgment evaluation of the presence or absence of SC1 cleaning in this cleaning history may be performed using the thermal oxide film thickness estimated from a correlation prepared in advance as an index. That is, after performing a film thickness estimation evaluation process to estimate and evaluate the thermal oxide film thickness, the thermal oxide film thickness can be used to judge the presence or absence of SC1 cleaning. For example, in FIG. 7, the second sample from the left (cleaned with hydrofluoric acid and ozone water) has a thermal oxide film thickness of 5.125 nm, and the third sample from the left (SC1 cleaning) has a thermal oxide film thickness of 5.127 nm, and a judgment criterion for judging that SC1 cleaning has been performed is 5.127 nm or more. In addition, in the case of such an evaluation, it is preferable to align the elapsed time between the S3 process and the S4 process as described above.
By the way, when there is no need to estimate the thermal oxide film thickness, the judgment evaluation based on the evaluation oxide film thickness requires fewer steps and is much simpler. However, even when the thermal oxide film thickness for the estimated evaluation is used as an index, it is still sufficiently simple because there is no need to actually form a thermal oxide film on the evaluation substrate.

The evaluation method of the present invention (determining and evaluating SC1 cleaning) described above can be useful in determining whether SC1 cleaning has been performed on a silicon substrate, even if the cleaning history of the substrate is unknown.

In the evaluation method of the present invention, the evaluation step may include at least one of the film thickness estimation evaluation step using the evaluation oxide film thickness described above and the cleaning history estimation evaluation step, and of course, both may be performed as necessary.

<Method of controlling the manufacturing process of silicon substrate according to the present invention>
The above-mentioned silicon substrate evaluation method of the present invention is a technique that utilizes the evaluation oxide film thickness as an index, but this can also be used for managing the manufacturing process of silicon substrates.
A typical flow chart is shown in FIG.
(S21: Substrate preparation step)
In step S21, for example, a silicon wafer during the manufacturing process (in the manufacturing line) is extracted and prepared as a substrate to be evaluated. Specifically, it is desirable to use a silicon wafer that has undergone the same manufacturing process as the silicon wafer actually used in device fabrication. In general, silicon wafers after CMP processing and cleaning can be used. In this case, the possibility cannot be ruled out that there is a slight unintended and unexpected change in the cleaning conditions (change in chemical solution concentration or temperature) in the cleaning tank. In particular, when SC1 cleaning is used, such changes can also change the etching action, which can result in a change in the thermal oxide film thickness when a thermal oxidation process is performed (the thermal oxide film thickness changes compared to normal times before the change in cleaning conditions), so special care is required. Therefore, when using a chemical solution, the thermal oxide film thickness can be managed with higher accuracy by extracting the solution at the beginning, middle, and end of the chemical solution life.

(S22: Oxide film removal process to film thickness measurement process)
Next, in step S22, steps similar to steps S2 to S4 (and steps S11 to S13 as necessary) in the silicon substrate evaluation method of the present invention shown in FIG. 1 are performed, and the thickness of the evaluation oxide film is measured.

(S23 to S27: Management process [film thickness management process] [cleaning management process])
In step S23, the thickness of the evaluation oxide film is compared with a predetermined film thickness control value (also simply referred to as a control value) that has been set in advance.
Here, the control value can be calculated by performing the steps S21 and S22 multiple times on silicon wafers manufactured at different times to obtain the oxide film thickness for evaluation in advance, and then setting the center value, upper control limit value, and lower control limit value from these oxide film thicknesses for evaluation by a known method. For example, the average value of the film thickness can be set as the center value, and the standard deviation σ of the center value can be calculated, and +3σ from the center value can be set as the upper control limit value, and -3σ can be set as the lower control limit value. However, there is no particular limit to the method of calculating the control value. The above control value can be determined while also taking into consideration whether the thickness of the thermal oxide film formed by the thermal oxidation process in the device manufacturing process downstream of the manufacturing line of a silicon wafer manufactured at the same time as the silicon wafer extracted to obtain the oxide film thickness for evaluation as described above falls within the allowable range.

If the evaluation oxide film thickness is within the control value in step S24, it is determined that there is no problem, and the silicon wafer is sent to the next step, such as step S25.
On the other hand, if the control value is exceeded in step S24, it can be determined in step S26 that the thermal oxide film thickness is fluctuating in the thermal oxidation process of the downstream device process or the like during the silicon substrate manufacturing process. Also, in step S27, the cleaning history is investigated and fed back to the cleaning process, so that the process conditions of the cleaning process can be corrected and improved. Abnormalities can be detected early and a prompt response can be made.

For example, in a cleaning line using SC1 cleaning, if the thickness of the evaluation oxide film exceeds the upper control limit, it can be easily assumed that the etching action of SC1 is stronger than usual, and that the thermal oxide film thickness is thicker. In this case, it is effective to investigate the history of the chemical concentration, chemical temperature, and cleaning time. Also, if the control values are often exceeded at the end of the chemical life, for example, slight fluctuations in the chemical concentration are suspected. One countermeasure is to make a new chemical solution.
Conversely, if the thickness of the evaluation oxide film falls below the lower control limit, it is determined that the thermal oxide film is thin. In this case, the etching effect of the SC1 cleaning is expected to be small, and similarly, by investigating the cleaning history and taking measures, the thermal oxide film thickness can be stabilized, enabling the production of good silicon wafers.

In the control method of the present invention, at least one of the film thickness control process using the evaluation oxide film and the cleaning control process described above may be performed. Figure 10 shows an example in which both processes are performed.

By the way, the above-mentioned manufacturing process control method uses the evaluation oxide film thickness as an index, but the manufacturing process can also be controlled using the thermal oxide film thickness as an index. That is, after estimating and evaluating the thermal oxide film thickness by performing the film thickness estimation evaluation step of the above-mentioned silicon substrate evaluation method of the present invention, the estimated and evaluated thermal oxide film thickness can be used instead of the evaluation oxide film thickness, and the silicon substrate manufacturing process can be controlled by performing steps similar to the above-mentioned film thickness control step and cleaning control step using a separately set predetermined control value (thermal oxide film thickness control value) as a reference.
In addition, when there is no need to estimate the thermal oxide film thickness, the management method based on the evaluation oxide film thickness requires fewer steps and is much simpler. However, even when the estimated thermal oxide film thickness is used as an index, it is still sufficiently simple because there is no need to actually form a thermal oxide film on the evaluation substrate.

The present invention will be further described below with reference to examples. However, these examples are presented for illustrative purposes and should not be construed as limiting.
First, the method for estimating the thermal oxide film thickness will be described. After preparing seven levels of silicon wafers (samples A to G) for which the thermal oxide film thickness was to be estimated, which were different in CMP processing conditions and cleaning conditions, the following evaluation was performed.
Specifically, the CMP processing conditions and cleaning conditions for each sample were varied as shown in Table 1 below.

Example 1
One silicon wafer for each level was washed with hydrofluoric acid to completely remove the native oxide film, and then washed with ozone water to form an oxide film for evaluation. The hydrofluoric acid washing was performed with a concentration of 0.5 wt%, at a temperature of 25°C, and for a washing time of 3 minutes, while the ozone water washing was performed with a concentration of 20 ppm, at a temperature of 25°C, and for a washing time of 3 minutes. Next, the thickness of the oxide film for evaluation was evaluated using a spectroscopic ellipsometer M-2000V manufactured by J. A. Woollam Co. The oxide film thickness for evaluation is shown in Table 2.

As shown in Table 2, differences in film thickness were observed between the levels. The order of thickness was (thick) G>B>E>F>C>A>D (thin). The thicker the oxide film for evaluation, the thicker the film after thermal oxidation. Therefore, the film thickness after thermal oxidation was estimated to be (thick) G>B>E>F>C>A>D (thin), and the thermal oxide film thickness could be evaluated. It took about one hour from preparing the sample to estimating the thermal oxide film.

Here, we verified whether the above evaluation was valid. A separate piece from each level was subjected to dry thermal oxidation treatment (maximum temperature reached: 900°C, time: 60 minutes), and the thermal oxide film thickness formed by the thermal oxidation treatment was evaluated at M-2000V. It was confirmed that the thermal oxide film thickness between samples was G>B>E>F>C>A>D (thin), which was the same as the results of the above evaluation method. In other words, it was confirmed that the above evaluation method was valid.

Example 2
Next, the correlation between the thickness of the evaluation oxide film measured in the verification in Example 1 and the thickness of the thermal oxide film was obtained, which resulted in the correlation shown in Fig. 11. The correlation coefficient ^R2 value was very good, and it was confirmed that there is a good correlation between the evaluation oxide film and the thermal oxide film.
Obtaining this correlation was a preliminary test, and then as a main test, two silicon wafers for which the thermal oxide film thickness was to be estimated were prepared, and the thicknesses of the evaluation oxide films were measured in the same manner as in Example 1. The thicknesses of the evaluation oxide films were found to be 1.271 nm and 1.314 nm, respectively. Next, by substituting these values into the correlation (correlation formula) in Figure 11 to determine the thermal oxide film thicknesses, the thicknesses were estimated to be 5.153 nm and 5.223 nm.
When a thermal oxidation process (under the same conditions as in the verification of Example 1) was performed on silicon wafers similar to each of the two silicon wafers prepared above, the thicknesses of the thermal oxide films were almost the same as the estimated thermal oxide film thicknesses evaluated above, and it was confirmed that the evaluation method described above is appropriate and excellent.

Comparative Example
According to the evaluation method of Patent Document 3, one silicon wafer for each level was subjected to ATR-FT-IR measurement, and the amount of OH groups was calculated as a component of the chemical oxide film. The results are shown in Table 3.

As shown in Table 3, differences were observed in the amount of OH groups. Here, the amount of OH groups was (large) B>G>E>F>D>A>C (small). The thermal oxide film thickness estimated from this result was (thick) B>G>E>F>D>A>C (thin). However, this did not completely match the order of film thickness after the thermal oxidation process verified in Example 1. Therefore, it was shown that the evaluation method of the present invention in Examples 1 and 2, which matched the verification results, had a higher accuracy in estimating the thermal oxide film.
The time required from preparation of the samples to estimation of the order of thermal oxide film thickness was about 6 hours.
Therefore, it was also shown that Example 1 had a better throughput and was capable of more reliably estimating the thickness of the thermal oxide film.

Example 3
The control method of the present invention was carried out as a method for controlling the manufacturing process of silicon wafers. First, in order to determine the range of the control value, two silicon wafers were taken out after the cleaning process, and the thickness of the evaluation oxide film was measured in the same manner as in Example 1. This measurement was carried out five times on different manufacturing dates (note that, in silicon wafers manufactured at the same time as the taken out silicon wafers, the thermal oxide film thickness formed during the thermal oxidation treatment in the device manufacturing process downstream of the manufacturing line was all within the allowable range). The average value of these evaluation oxide film thickness values was calculated to be 1.241 nm. Next, the standard deviation σ was 0.005 nm. From these results, the center value was set to the average value of 1.241 nm. The upper control limit was set to 1.256 nm at the center line +3σ, and the lower control limit was set to 1.226 nm at the center line -3σ.

Next, a silicon wafer (substrate to be evaluated) was sampled during the manufacturing process, and the thickness of the oxide film for evaluation was measured in the same manner as in Example 1. As a result, the oxide film for evaluation was found to have a thickness of 1.238 nm, which was within the range of the control value, so it was determined that there was no abnormality in the manufacturing process, and manufacturing was continued.
In fact, when a thermal oxidation process was performed on a silicon wafer manufactured at the same time as the extracted silicon wafer, the thickness of the thermal oxide film was within the allowable range.

Furthermore, when a similar sampling evaluation was performed at a different timing from the above, the film thickness was 1.259 nm, which was a value exceeding the upper control limit. From this result, it was determined that there was a fluctuation in the thermal oxide film thickness in the thermal oxidation process (thermal oxidation process during device fabrication) (the film thickness was larger than the film thickness value that is usually formed), and a manufacturing history investigation (cleaning history investigation) was performed. As a result, it was found that there was a malfunction in the thermocouple that controls the chemical solution temperature in the cleaning process, and the chemical solution temperature was higher than the set temperature. Therefore, it was estimated that the thickness of the evaluation oxide film became thicker because the chemical solution temperature became too high and etching proceeded too far.
After that, by replacing the thermocouple, it was confirmed that the chemical temperature was equal to the set temperature, and then the silicon wafer was cleaned, and the thickness of the evaluation oxide film was evaluated in the same manner as in Example 1, and it was found to be 1.244 nm, which was within the range of the control value. After this confirmation, the production of silicon wafers was resumed.
In fact, when the thermal oxide film thickness of a silicon wafer manufactured at the same time as the sampled silicon wafer was checked, it was found to be outside the allowable range.
The above-described method for controlling the manufacturing process of silicon wafers makes it possible to detect slight variations in the thermal oxide film and provide feedback to the manufacturing process, thereby enabling the conditions in the cleaning process to be improved as quickly as possible.

The present specification includes the following aspects.
[1]: A method for evaluating a silicon substrate, comprising:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
an evaluation step of evaluating the evaluation substrate based on the film thickness of the evaluation oxide film measured in the film thickness measurement step, without performing a thermal oxidation treatment on the evaluation substrate;
In the evaluation step,
(i) a film thickness estimation evaluation step of estimating and evaluating a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate;
(ii) a cleaning history estimation and evaluation step of estimating and evaluating the cleaning history of the prepared evaluation target substrate.
[2]: The method for evaluating a silicon substrate according to the above [1], wherein in the oxide film removing step, the native oxide film is completely removed by washing with hydrofluoric acid.
[3]: The method for evaluating a silicon substrate according to the above [1] or [2], wherein in the evaluation oxide film forming step, the evaluation oxide film is formed using ozone water or hydrogen peroxide water.
[4]: when the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measurement step of performing a thermal oxidation process on another of the plurality of prepared preliminary substrates to form a thermal oxide film and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
The method for evaluating a silicon substrate according to any one of the above [1] to [3], further comprising estimating and evaluating a thickness of a thermal oxide film of the evaluation substrate from a thickness of an evaluation oxide film of the evaluation substrate based on the correlation.
[5]: based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
The method for evaluating a silicon substrate according to any one of the above [1] to [4], further comprising estimating and evaluating a cleaning history of the prepared evaluation substrate.
[6]: The method for evaluating a silicon substrate according to any one of [1] to [5] above, which determines whether SC1 cleaning has been performed when estimating the cleaning history of the prepared evaluation substrate.
[7]: A method for managing a manufacturing process of a silicon substrate, comprising:
The method includes a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step of the silicon substrate evaluation method according to any one of [1] to [6] above,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one of a cleaning history investigation of the substrate to be evaluated, and a cleaning management process for modifying processing conditions for a cleaning treatment in the manufacturing process.
[8]: A method for managing a manufacturing process of a silicon substrate, comprising:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
a management step of managing the manufacturing process based on the film thickness of the evaluation oxide film measured in the film thickness measurement step without performing a thermal oxidation process on the prepared evaluation substrate,
In the management step,
When the thickness of the evaluation oxide film deviates from a predetermined thickness control value set in advance,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one of a cleaning history investigation of the substrate to be evaluated, and a cleaning management process for modifying processing conditions for a cleaning treatment in the manufacturing process.
[9]: The method for managing a silicon substrate manufacturing process according to the above [8], wherein in the oxide film removing step, the native oxide film is completely removed by washing with hydrofluoric acid.
[10]: The method for managing a silicon substrate manufacturing process according to the above [8] or [9], wherein in the evaluation oxide film forming step, the evaluation oxide film is formed using ozone water or hydrogen peroxide water.

The present invention is not limited to the above-described embodiments. The above-described embodiments are merely examples, and anything that has substantially the same configuration as the technical idea described in the claims of the present invention and exhibits similar effects is included within the technical scope of the present invention.

Claims (18)

A method for evaluating a silicon substrate, comprising the steps of:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
an evaluation step of evaluating the evaluation substrate based on the thickness of the evaluation oxide film measured in the film thickness measurement step without performing a thermal oxidation treatment on the evaluation substrate,
In the evaluation step,
(i) a film thickness estimation evaluation step of estimating and evaluating a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate;
(ii) a cleaning history estimation and evaluation step of estimating and evaluating the cleaning history of the prepared evaluation substrate. The silicon substrate evaluation method according to claim 1, characterized in that in the oxide film removal process, the native oxide film is completely removed by washing with hydrofluoric acid. The silicon substrate evaluation method according to claim 1, characterized in that in the evaluation oxide film formation process, the evaluation oxide film is formed using ozone water or hydrogen peroxide water. The silicon substrate evaluation method according to claim 2, characterized in that in the evaluation oxide film formation process, the evaluation oxide film is formed using ozone water or hydrogen peroxide water. When the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measurement step of performing a thermal oxidation process on another of the plurality of prepared preliminary substrates to form a thermal oxide film and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
2. The method for evaluating a silicon substrate according to claim 1, further comprising estimating and evaluating a thickness of a thermal oxide film of the evaluation substrate from a thickness of an evaluation oxide film of the evaluation substrate based on the correlation. When the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measurement step of performing a thermal oxidation process on another of the plurality of prepared preliminary substrates to form a thermal oxide film and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
3. The method for evaluating a silicon substrate according to claim 2, further comprising estimating and evaluating a thickness of a thermal oxide film of the evaluation substrate from a thickness of an evaluation oxide film of the evaluation substrate based on the correlation. When the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measurement step of performing a thermal oxidation process on another of the plurality of prepared preliminary substrates to form a thermal oxide film and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
4. The method for evaluating a silicon substrate according to claim 3, further comprising estimating and evaluating a thickness of a thermal oxide film of the evaluation substrate from a thickness of an evaluation oxide film of the evaluation substrate based on the correlation. When the film thickness estimation evaluation step is performed in the evaluation step,
a spare-substrate preparation process for preparing a plurality of spare silicon substrates that serve as references and that have been cleaned under the same conditions;
a preliminary oxide film removal step of completely removing the native oxide film of one of the prepared preliminary substrates under conditions without etching;
a preliminary-evaluation oxide film forming step of forming an evaluation oxide film on the preliminary substrate after the native oxide film has been completely removed under oxidation conditions that have no etching effect;
a preliminary film thickness measuring step of measuring a film thickness of the evaluation oxide film formed in the preliminary evaluation oxide film forming step;
a preliminary thermal oxide film thickness measurement step of performing a thermal oxidation process on another of the plurality of prepared preliminary substrates to form a thermal oxide film and measuring the thickness of the thermal oxide film on the preliminary substrate;
a correlation acquisition step of acquiring a correlation between a thickness of the evaluation oxide film on the preliminary substrate and a thickness of the thermal oxide film on the preliminary substrate;
5. The method for evaluating a silicon substrate according to claim 4, further comprising estimating and evaluating a thickness of a thermal oxide film of the evaluation substrate from a thickness of an evaluation oxide film of the evaluation substrate based on the correlation. Based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
9. The method for evaluating a silicon substrate according to claim 1, further comprising the step of estimating and evaluating a cleaning history of the prepared evaluation target substrate. The method for evaluating a silicon substrate according to any one of claims 1 to 8, characterized in that when estimating and evaluating the cleaning history of the prepared evaluation substrate, it is determined whether SC1 cleaning has been performed. The silicon substrate evaluation method according to claim 9, characterized in that when estimating and evaluating the cleaning history of the prepared evaluation substrate, it is determined whether SC1 cleaning has been performed. A method for managing a manufacturing process of a silicon substrate, comprising the steps of:
The method for evaluating a silicon substrate according to any one of claims 1 to 8 further comprises a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one step of investigating a cleaning history of the substrate to be evaluated, and a cleaning management step of modifying processing conditions for a cleaning treatment in the manufacturing process. A method for managing a manufacturing process of a silicon substrate, comprising the steps of:
The evaluation method of the silicon substrate according to claim 9 further comprises a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one step of investigating a cleaning history of the substrate to be evaluated, and a cleaning management step of modifying processing conditions for a cleaning treatment in the manufacturing process. A method for managing a manufacturing process of a silicon substrate, comprising the steps of:
The evaluation method of the silicon substrate according to claim 10 further comprises a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one step of investigating a cleaning history of the substrate to be evaluated, and a cleaning management step of modifying processing conditions for a cleaning treatment in the manufacturing process. A method for managing a manufacturing process of a silicon substrate, comprising the steps of:
The evaluation method of the silicon substrate according to claim 11 further comprises a management step of managing the manufacturing process based on the thickness of the thermal oxide film estimated and evaluated in the thickness estimation evaluation step of the evaluation step,
In the management step,
When the estimated thickness of the thermal oxide film is out of a predetermined thickness control value,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one step of investigating a cleaning history of the substrate to be evaluated, and a cleaning management step of modifying processing conditions for a cleaning treatment in the manufacturing process. A method for managing a manufacturing process of a silicon substrate, comprising the steps of:
a substrate preparation step of preparing the silicon substrate as a substrate to be evaluated;
an oxide film removing step of completely removing the native oxide film of the evaluation substrate under conditions that do not cause an etching effect;
a step of forming an oxide film for evaluation on the substrate after the native oxide film has been completely removed, under oxidation conditions that have no etching effect;
a thickness measuring step of measuring a thickness of the evaluation oxide film formed in the evaluation oxide film forming step;
a management step of managing the manufacturing process based on the film thickness of the evaluation oxide film measured in the film thickness measurement step without performing a thermal oxidation process on the prepared evaluation substrate,
In the management step,
When the thickness of the evaluation oxide film deviates from a predetermined thickness control value set in advance,
(i) a film thickness control step of evaluating whether a film thickness of a thermal oxide film formed when the thermal oxidation treatment is performed on the prepared evaluation substrate varies;
(ii) A method for managing a manufacturing process of a silicon substrate, comprising at least one step of investigating a cleaning history of the substrate to be evaluated, and a cleaning management step of modifying processing conditions for a cleaning treatment in the manufacturing process. The method for managing the manufacturing process of silicon substrates according to claim 16, characterized in that in the oxide film removal process, the native oxide film is completely removed by washing with hydrofluoric acid. The method for managing the manufacturing process of silicon substrates according to claim 16 or 17, characterized in that in the evaluation oxide film formation process, the evaluation oxide film is formed using ozone water or hydrogen peroxide water.

Images