JP4346629B2 - Method for measuring nitrogen concentration in silicon crystal, method for determining conversion coefficient for calculating nitrogen concentration in silicon crystal, method for manufacturing silicon wafer, and method for manufacturing semiconductor device - Google Patents

Description

  The present invention relates to a method for measuring a nitrogen concentration of a silicon crystal and a method for determining a coefficient for evaluating nitrogen concentration. Furthermore, the present invention relates to a method for manufacturing a silicon wafer including a step of obtaining a nitrogen concentration, a silicon wafer to be manufactured, and a method for manufacturing a semiconductor device using the silicon wafer.

As a semiconductor integrated circuit substrate, an epitaxial wafer is often used in which a p-type silicon layer is epitaxially grown on the surface of a p ⁺ -type silicon substrate (specific resistance of about 0.01 Ωcm) doped with boron at a high concentration. This silicon substrate is obtained by slicing a silicon ingot pulled up by the Czochralski method, and boron is doped during crystal pulling. Boron doped in a silicon substrate has a large gettering ability against d-electron heavy metal contaminating atoms such as iron (Fe). Thereby, impurity atoms such as Fe contained in the epitaxial layer are attracted and captured in the silicon substrate.

When a plurality of epitaxial wafers are loaded into a heating furnace and heat treatment is performed, boron in the highly doped substrate is vaporized and doped into the epitaxial layer of the substrate. This phenomenon is called boron autodoping. In order to prevent boron auto-doping, a shield oxide film for preventing vaporization of boron is formed on the back surface (surface on the p ⁺ -type silicon substrate side) of the epitaxial wafer.

  The shield oxide film formed on the back surface of the epitaxial wafer increases the manufacturing cost of the epitaxial wafer. Further, both sides of the 300 mm wafer are mirror-polished from the viewpoint of flatness. For this reason, it is not realistic to form a shield oxide film on the back surface.

  In order to avoid autodoping without forming a shield oxide film, it is necessary to reduce the boron concentration of the silicon substrate. For example, it is necessary to reduce the specific resistance of the silicon substrate to about 0.1 Ωcm to 1 Ωcm. However, if the boron concentration is reduced to this extent, the gettering effect by boron cannot be expected. For this reason, heavy metal contamination atoms contained in the epitaxial layer must be attracted and captured by other methods.

  A technique (intrinsic gettering technique) that attracts and captures heavy metal contamination atoms in oxide precipitates formed inside a silicon substrate is a promising candidate. However, in general, a part of oxygen precipitation latent nuclei (micro oxygen precipitates formed during crystal growth) disappears by high-temperature treatment for forming an epitaxial layer of an epitaxial wafer.

  In order to deposit a sufficient amount of oxygen in the wafer process, a method of doping nitrogen as an impurity into the silicon substrate can be considered. Impurity nitrogen in the silicon substrate has the effect of significantly promoting the precipitation of oxygen. Further, nitrogen reduces void defects (defects manifested in the 16M-DRAM generation, called crystal-originating particles (COP)) formed during the growth of silicon crystals.

In order to increase the gettering effect of heavy metal contamination atoms and reduce COP, it is important to control the impurity nitrogen concentration in the silicon substrate. Quantification of impurity nitrogen in silicon crystals produced by the floating zone method can be performed by measuring infrared absorbance. Due to the NN pair contained in the silicon crystal, an absorption peak appears at a wave number of 963 cm ⁻¹ . The impurity nitrogen concentration can be obtained by multiplying the intensity (absorbance) of this absorption peak by a known proportional conversion factor.

  However, the absorbance of silicon crystals pulled up by the Czochralski method varies greatly between samples. Even when the impurity nitrogen concentrations of the silicon crystal produced by the floating zone method and the silicon crystal pulled by the Czochralski method are equal, the silicon crystal pulled by the Czochralski method shows a smaller absorption peak intensity. Note that the nitrogen concentration in these crystals is measured in advance by a secondary ion mass spectrometry (SIMS) method or the like. Although the nitrogen concentration is the same, the silicon crystal pulled up by the Czochralski method has a low absorbance, and it is assumed that there are defects that do not absorb infrared light other than the NN pair.

  Thus, a technique for measuring the nitrogen concentration in the silicon crystal pulled up by the Czochralski method with high accuracy has not yet been established.

  An object of the present invention is to provide a nitrogen concentration measurement method capable of measuring the impurity nitrogen concentration in a silicon crystal with high accuracy. Another object of the present invention is to provide a method for determining a nitrogen concentration evaluation coefficient for evaluating the nitrogen concentration in a silicon crystal. Still another object of the present invention is to provide a method for producing a silicon wafer in which the nitrogen concentration is required. Still another object of the present invention is to provide a silicon wafer manufactured by this manufacturing method. Still another object of the present invention is to provide a method of manufacturing a semiconductor device using this silicon wafer.

According to one aspect of the invention,
A first heating step of heating the silicon crystal to a quasi-thermal equilibrium condition under a condition that an absorption peak corresponding to a defect caused by the NNO ₂ composite disappears;
In the quasi-thermal equilibrium state after the first heating step, the infrared absorption spectrum of the silicon crystal is measured, and from the absorption spectrum, the intensity of the first absorption peak corresponding to the defect caused by the NN pair and the NNO complex are obtained. Determining the intensity of the second absorption peak corresponding to the resulting defect;
After obtaining the intensity of the first and second absorption peaks, ripening at a temperature higher than the heating temperature of the first heating step, the second heating step of bringing the silicon crystal to a quasi-thermal equilibrium state;
In the quasi-thermal equilibrium state after the second heating step, the infrared absorption spectrum of the silicon crystal is measured, and from the absorption spectrum, the intensity of the third absorption peak corresponding to the defect caused by the NN pair and the NNO complex are obtained. Determining the intensity of the fourth absorption peak corresponding to the resulting defect;
The difference between the intensity of the first absorption peak and the intensity of the third absorption peak is defined as a first difference, and the difference between the intensity of the second absorption peak and the intensity of the fourth absorption peak is defined as a second difference. NNO based on the first conversion factor for calculating the density of the NN pair from the ratio of the first difference and the second difference and the intensity of the absorption peak due to the NN pair. Calculating a second conversion factor for calculating the density of the NNO complex from the intensity of the absorption peak due to the complex;
The density of the NN pair is calculated from the intensity of the absorption peak due to the NN pair and the first conversion factor, and the density of the NNO complex is calculated from the intensity of the absorption peak due to the NNO complex and the second conversion factor. Calculating the density;
And a step of determining the nitrogen concentration in the silicon crystal based on the density of the NN pair and the NNO complex.

According to yet another aspect of the invention,
The silicon crystal is heated to the first temperature, the intracrystalline defect reaction is brought to a quasi-thermal equilibrium state, the infrared absorption spectrum is measured in a state where the intracrystalline defect concentration in the quasi-thermal equilibrium state is maintained, and the first due to the NN pair. Determining the absorption peak intensity of 1 and the second absorption peak intensity resulting from the NNO complex;
The infrared absorption spectrum in a state where the silicon crystal is heated to a second temperature higher than the first temperature to bring the intracrystalline defect reaction to a quasi-thermal equilibrium state, and the intracrystalline defect concentration in the quasi-thermal equilibrium state is maintained. Measuring the third absorption peak intensity attributed to the NN pair and the fourth absorption peak intensity attributed to the NNO complex,
Determining the difference between the first absorption peak intensity and the third absorption peak intensity, and the difference between the second absorption peak intensity and the fourth absorption peak intensity, and determining the ratio between the two,
Based on the conversion coefficient for calculating the density of the NN pair from the absorption peak intensity caused by the NN pair, and the ratio obtained in the step of obtaining the ratio, the NNO complex is obtained from the absorption peak intensity caused by the NNO complex. A method for determining a coefficient for evaluating nitrogen concentration in a silicon crystal is provided which includes a step of calculating a conversion coefficient for calculating the density of the silicon crystal.

  According to still another aspect of the present invention, a conversion factor for calculating the density of the NNO complex determined by the method for determining a coefficient for evaluating the nitrogen concentration in the silicon crystal, and a density for calculating the density of the NN pair. There is provided a method for measuring a nitrogen concentration in a silicon crystal, comprising a step of calculating a nitrogen concentration in the silicon crystal using a conversion factor.

According to yet another aspect of the invention,
Using a CZ method to pull up a silicon single crystal to which nitrogen has been added;
Cutting a plurality of silicon wafers from the pulled silicon single crystal ingot;
Heating at least one silicon wafer to a quasi-thermal equilibrium condition under a condition where an absorption peak corresponding to a defect caused by the NNO ₂ composite disappears;
The infrared absorption spectrum of the silicon wafer is measured, and from the absorption spectrum, the intensity of the first absorption peak corresponding to the defect caused by the NN pair and the intensity of the second absorption peak corresponding to the defect caused by the NNO complex The process of seeking
A step of calculating the density of the NN pair based on the conversion factor for calculating the density of the NN pair from the intensity of the absorption peak caused by the NN pair, and the intensity of the first absorption peak;
The density of the NNO complex is calculated based on the conversion factor for calculating the density of the NNO complex determined by the method according to claim 7 and the intensity of the second absorption peak. And a process of
Determining the nitrogen concentration of the silicon wafer based on the density of the NN pair and the density of the NNO complex;
And a step of determining whether or not the nitrogen concentration is within a predetermined range.

According to yet another aspect of the invention,
A semiconductor device manufacturing method is provided in which a semiconductor integrated circuit is manufactured on a silicon wafer that has been accepted by the silicon wafer manufacturing method.

In the state where the absorption peak corresponding to the defect caused by the NNO ₂ complex disappears, it is considered that the NNO ₂ complex is not substantially present. For this reason, the nitrogen concentration in the silicon crystal can be quantified by adding the density of the NN pair and the density of the NNO complex.

  When the coefficient for evaluating the nitrogen concentration in the silicon crystal is determined, the nitrogen concentration can be obtained by measuring the infrared absorption spectrum of the silicon crystal in the quasi-thermal equilibrium state.

A silicon wafer evaluation method according to a first reference example of the present invention will be described. A silicon single crystal doped with nitrogen is pulled up using the Czochralski method. Doping of nitrogen is performed by a method of adding Si ₃ N ₄ into the silicon melt or a method of pulling up in a nitrogen atmosphere. A silicon wafer is cut out from the pulled silicon single crystal ingot.

  The silicon wafer is loaded into an electric furnace and heat treatment is performed. Suitable heat treatment temperature and time will be described later. After heat treatment for a predetermined time, the silicon wafer is taken out of the electric furnace and rapidly cooled. For example, it is taken out in a room temperature atmosphere in 2 to 3 seconds. An infrared absorption spectrum of a silicon wafer cooled to room temperature is measured.

FIG. 1 shows an example of an infrared absorption spectrum. The horizontal axis represents the wave number in the unit “cm ⁻¹ ”, and the vertical axis represents the absorbance. Absorption peaks A, B, and C appear at the positions of wave numbers 963 cm ⁻¹ , 996 cm ⁻¹ , and 1026 cm ⁻¹ , respectively. Peak A is absorption due to the NN pair, and peaks B and C are absorption due to the NNO complex. Here, the NN pair means a pair of one impurity nitrogen atom existing between lattices in the silicon crystal and an impurity nitrogen atom existing between lattices of silicon atoms within the fourth proximity from the nitrogen atom. To do. The NNO complex means a defect composed of an NN pair and impurity oxygen atoms existing between lattices of silicon atoms within the fourth proximity from the NN pair.

FIG. 2 shows the relationship between absorbance and heat treatment conditions. The horizontal axis represents heat treatment time in the unit “time”, and the vertical axis represents absorbance. The heat treatment temperature is 800 ° C., and the thickness of the measurement target silicon wafer is 10 mm. White circle in the figure represents an intensity of infrared light yield peak A appeared to the position of the wave number 963Cm ^-1, black circle indicates the intensity of the infrared ray yield peak B appeared to the position of the wave number 996cm ^-1. Here, the intensity (absorbance) of the absorption peak means the height of the peak when the base level is used as a reference.

  In the region where the heat treatment time is less than 2 hours, the absorbance varies depending on the heat treatment time. However, when the heat treatment time is 2 hours or more, the absorbance is hardly affected by the heat treatment time and becomes almost constant.

  The reason why the absorbance fluctuates depending on the heat treatment time is considered to be that an intracrystalline defect reaction occurs during the heat treatment and the density of the NN pair or NNO complex changes. A state where the heat treatment is performed for 2 hours or more and the absorbance hardly changes is called a quasi-thermal equilibrium state. In this state, the intracrystalline defect reaction remains in thermal equilibrium on a time scale of at least several hours. The inventor of the present application discovered for the first time that the infrared absorbance due to the NN pair or NNO complex fluctuates due to the heat treatment, and that the quasi-thermal equilibrium state is reached when the heat treatment time is 2 hours or longer. Conventionally, the reason that the infrared absorbance of the silicon single crystal pulled up by the Czochralski method has a large variation among samples is considered that the absorbance was measured in a state before reaching the quasi-thermal equilibrium state.

By heating the silicon substrate until it reaches the quasi-thermal equilibrium state and then rapidly cooling, the defect state due to the NN pair and the NNO complex in the quasi-thermal equilibrium state can be fixed (maintained). This makes it possible to measure the absorbance of infrared rays in the thermal equilibrium state of the defect reaction. In the first reference example , the silicon crystal was taken out from the heating atmosphere to the room temperature atmosphere in 2 to 3 seconds, but it may be taken out in a short time within 10 seconds.

In the first reference example , the heat treatment temperature is set to 800 ° C., but it is considered that the time required to reach the quasi-thermal equilibrium state varies when the heat treatment temperature is changed.

  FIG. 3 shows the relationship between the time required to reach the quasi-thermal equilibrium state and the heat treatment temperature. The horizontal axis represents the heat treatment temperature in the unit “° C.”, and the vertical axis represents the heat treatment time required to reach the quasi-thermal equilibrium state in the unit “time”. It can be seen that as the heat treatment temperature increases, the heat treatment time required to reach the quasi-thermal equilibrium state is shortened. In order to cause the defect reaction in the silicon crystal to reach a quasi-thermal equilibrium state, heat treatment may be performed under the conditions of temperature and time on the upper right side of the curve shown in FIG. Expressing this condition as an inequality,

(Equation 1)
t ≧ 10.56 × exp (−0.0022T) (1)
It becomes. Here, T is the heat treatment temperature expressed in the unit “° C.”, and t is the heat treatment time expressed in the unit “time”.

Next, the physical meaning of each peak shown in FIG. 1 will be described. Peak A appearing at the position of the wave number 963cm ^-1 is due to the NN pair, peak B and C appearing at the position of the wave number 996cm ^-1 and a wavenumber 1026cm ^-1 is attributed to NNO complex. However, it is not clear whether peak B and peak C are caused by the same NNO complex or different forms of NNO complexes.

FIG. 4 shows the relationship between the heat treatment temperature and the absorbance corresponding to each absorption peak. The horizontal axis represents the reciprocal of temperature in the unit “1000 / K”, and the vertical axis represents absorbance. White circles in the figure, black circles, and white squares, each wavenumber 963cm ^-1, indicating the absorbance at 996cm ^-1, and 1026cm ^-1. The absorbance slopes at wave numbers 996 cm ⁻¹ and 1026 cm ⁻¹ are almost equal. That is, the activation energies of the intracrystalline defect reactions corresponding to wave numbers 996 cm ⁻¹ and 1026 cm ⁻¹ are equal. This indicates that the two absorption peaks are absorptions by different localized modes belonging to the same defect. The formation energy of this defect is estimated to be about 0.4 eV from the slope of the graph.

From the above considerations, it can be seen that in order to determine the concentration of the NNO complex, it is only necessary to measure the intensity of one of the absorption peaks appearing at wave numbers of 996 cm ⁻¹ and 1026 cm ⁻¹ . Further, in order to determine the concentration of the NN pair, the absorbance that appears at the position of a wave number of 963 cm ⁻¹ may be measured.

Next, a method for determining the impurity nitrogen concentration will be described. The impurity nitrogen concentration of the silicon crystal pulled up by the Czochralski method is measured by SIMS or the like. An infrared absorption spectrum in a quasi-thermal equilibrium state of the same silicon crystal is measured, and an absorbance that appears at at least one position of wave numbers 963 cm ⁻¹ , 996 cm ⁻¹ , and 1026 cm ⁻¹ is determined. By performing the same measurement on a plurality of samples having different impurity nitrogen concentrations, the correspondence relationship between the impurity nitrogen concentration and the absorbance is obtained. For example, a conversion table or a graph for calculating the impurity nitrogen concentration from the absorbance is created.

An infrared absorption spectrum in a quasi-thermal equilibrium state of a silicon crystal whose impurity nitrogen concentration is unknown is measured, and an absorbance that appears at at least one position of wave numbers 963 cm ⁻¹ , 996 cm ⁻¹ , and 1026 cm ⁻¹ is determined. The impurity nitrogen concentration can be determined from the absorbance using a conversion table or the like prepared in advance.

According to the first reference example described above, the impurity nitrogen concentration in the silicon crystal can be easily obtained without using a large measuring device such as SIMS. This makes it possible to maintain a sufficient gettering effect of the silicon crystal substrate used for the epitaxial wafer.

In the first reference example , the peak intensities appearing at wave numbers of 963 cm ⁻¹ , 996 cm ⁻¹ , and 1026 cm ⁻¹ were obtained, but the position of this peak is shifted depending on the sample temperature at the time of measurement. Therefore, it is necessary to measure the absorbance at the wave number corresponding to the sample temperature at the time of measurement.

Moreover, you may obtain | require the intensity | strength of the absorption peak which appears in another position. For example, the absorption peak due to the NN pair, also appear in the position of the wave number 763Cm ^-1, absorption peaks caused by NNO complex, also appears at the position of the wave number 796cm ^-1 and 801cm ^-1. You may obtain | require the intensity | strength of these absorption peaks.

It should be noted that the S / N ratio in impurity nitrogen concentration measurement can be increased by using the largest one of the plurality of absorption peaks. As shown in FIG. 4, when the heat treatment temperature for reaching the quasi-thermal equilibrium state is 800 ° C., the intensity of the absorption peak appearing at the wave number of 963 cm ⁻¹ is maximized. Therefore, it is preferable that the heat treatment temperature for reaching the quasi-thermal equilibrium state is 700 ° C. to 900 ° C., and the intensity of the absorption peak appearing at a wave number of 963 cm ⁻¹ is obtained.

In the first reference example , the thickness of the sample for measuring the infrared absorption spectrum is 10 mm, but other thicknesses may be used. However, since it is difficult to observe the absorption peak if the sample is too thin, the thickness of the sample is preferably 0.5 mm or more. The measurement resolution is preferably 8 cm ⁻¹ or more.

Moreover, in the said 1st reference example , although the infrared absorption spectrum was measured at room temperature, you may measure in the state cooled or heated. For example, the measurement may be performed at a constant temperature within a temperature range of 1.5K to 473K.

Next, the method for determining the coefficient for evaluating the nitrogen concentration in the silicon crystal according to the first embodiment of the present invention will be described.

As shown in FIG. 1, the infrared absorption spectrum of silicon shows a peak D at a position of 1018 cm ⁻¹ . Peak D is absorption due to the NNO ₂ complex. Here, the NNO ₂ complex means a defect composed of an NN pair and two impurity oxygen atoms existing between lattices of silicon atoms within the fourth proximity from the NN pair.

FIG. 5 shows the relationship between the temperature and the absorbance that reach the quasi-thermal equilibrium state. The horizontal axis represents the temperature at which the silicon wafer is heated in order to reach a quasi-thermal equilibrium state in the unit “° C.”, and the vertical axis represents the absorbance. White circles, white squares, and white triangles in the figure indicate absorbances at wavelengths where absorption due to the NN pair, the NNO complex, and the NNO ₂ complex appears, respectively. That is, each wave number 963cm ^-1, indicating the absorbance at 996cm ^-1, and 1018 cm ^-1.

As the heating temperature is increased from 600 ° C., the absorbance due to the NNO complex and the NNO ₂ complex decreases. This is because oxygen was dissociated from each complex, and the density of these complexes decreased. In the temperature range from 600 ° C. to 720 ° C., a phenomenon in which oxygen is dissociated from the NNO ₂ complex and an NN pair or an NNO complex is generated, and a phenomenon in which oxygen is dissociated from the NNO complex and an NN pair is generated. Is happening. For this reason, the light absorbency resulting from a NN pair is increasing.

When the heating temperature is 720 ° C. or higher, the absorption peak due to the NNO ₂ complex disappears. This is presumably because oxygen in the NNO ₂ complex was dissociated and all NNO ₂ complexes were changed to NN pairs and NNO complexes. For this reason, in the temperature range of 720 ° C. to 800 ° C., the reaction related to the NNO ₂ complex does not occur, and it is considered that only the phenomenon in which oxygen of the NNO complex is dissociated and NN pairs are generated. When the temperature is 800 ° C. or higher, the absorbance due to the NN pair starts to decrease.

When the temperature of the quasi-thermal equilibrium increases from 720 ° C. to 800 ° C., the reduced number of NNO complexes is considered to be equal to the increased number of NN pairs. At this time, the amount of decrease in absorbance due to the NNO complex was about 0.72 × 10 ⁻³ , and the amount of increase in absorbance due to the NN pair was about 1 × 10 ⁻³ . This is because the ratio of the vibrator strength of the NN pair (corresponding to the absorbance of one NN pair) and the vibrator strength of the NNO complex (corresponding to the absorbance of one NNO complex) is 1: 0.72. It represents something.

When the absorption coefficient due to the NN pair in wavenumber 963cm ^-1 and I ₉₆₃ (cm ^-1), the density of the NN pair [NN] (cm ^-3) is,

(Equation 2)
[NN] = 1.83 × 10 ¹⁷ × I ₉₆₃ (2)
It is known that The proportionality coefficient multiplied by the absorption coefficient I ₉₆₃ resulting from the NN pair on the right side of Equation (2) is called a conversion coefficient.

Since the oscillator strength of the NNO composite is 0.72 times the oscillator strength of the NN pair, the density [NNO] (cm ^{− of the} NNO composite is obtained from the absorption coefficient I ₉₉₆ (cm ⁻¹ ) caused by the NNO composite. ³ ) It is considered that the conversion factor for obtaining 1) is 1 / 0.72 times the conversion factor of Equation (2). That is, the conversion factor is considered to be 2.54 × 10 ¹⁷ . Therefore, the density [NNO] (cm ⁻³ ) of the NNO complex is

(Equation 3)
[NNO] = 2.54 × 10 ¹⁷ × I ₉₉₆ (3)
It is expressed.

Hereinafter, the steps according to the first embodiment will be described. First, the silicon crystal is heated to a first temperature (for example, 720 ° C.) to bring the intracrystalline defect reaction to a quasi-thermal equilibrium state. The infrared absorption spectrum was measured while keeping the crystal in the defect concentration in the quasi-thermal equilibrium, wavenumber 996cm ^-1 due to the first absorption peak intensity and NNO complex appearing at the position of the wave number 963cm ^-1 due to the NN pair The second absorption peak intensity appearing at the position is obtained.

  The silicon crystal is heated to a second temperature higher than the first temperature (for example, 800 ° C.) to bring the intracrystalline defect reaction to a quasi-thermal equilibrium state. An infrared absorption spectrum is measured in a state where the defect concentration in the crystal in the quasi-thermal equilibrium state is maintained, and a third absorption peak intensity attributed to the NN pair and a fourth absorption peak intensity attributed to the NNO complex are obtained.

  The fourth absorption peak intensity is subtracted from the second absorption peak intensity, and the value is taken as A. The first absorption peak intensity is subtracted from the third absorption peak intensity, and the value is defined as B. The conversion coefficient for calculating the density of the NN pair is multiplied by B / A from the absorption coefficient obtained from the absorption peak intensity caused by the NN pair. This result is a conversion coefficient for calculating the density of the NNO complex from the absorption coefficient obtained from the absorption peak intensity caused by the NNO complex.

To indicate absorption silicon crystal before the heat treatment is due to NNO ₂ complex, the first temperature described above, it is necessary to more than a temperature at which the absorption peak due to NNO ₂ complex disappears. When heated to this temperature, the NNO ₂ complex is considered substantially absent. For this reason, the nitrogen concentration in the silicon crystal can be quantified by adding the density of the NN pair and the density of the NNO complex.

  In addition, the second temperature described above needs to be equal to or lower than a temperature at which the absorption peak intensity due to the NN pair starts to decrease.

Next, a second reference example of the present invention will be described. First, a plurality of silicon crystals having different concentrations of nitrogen are prepared. The nitrogen concentration in these silicon crystals is measured by a method such as SIMS.

  Each of the silicon crystals is heated to a first temperature (eg, 720 ° C.) to bring the intracrystalline defect reaction to a quasi-thermal equilibrium state. For each silicon crystal, an infrared absorption spectrum is measured in a state where the concentration of defects in the crystal in the quasi-thermal equilibrium state is maintained. From this measurement result, the first absorption peak intensity attributed to the NN pair and the second absorption peak intensity attributed to the NNO complex are determined.

  For each of the silicon crystals, the density of the NN pair is obtained by multiplying the absorption coefficient obtained from the first peak intensity by the conversion coefficient. The density of the NN pair is reduced from the nitrogen concentration measured by SIMS or the like. This result corresponds to the density of the NNO complex. Thereby, the relationship between the absorption coefficient calculated | required from the absorption peak intensity resulting from an NNO complex, and the density of an NNO complex is obtained.

  FIG. 6 shows the relationship between the density of the NN pair and NNO complex and the absorption coefficient. The horizontal axis represents the density, and the vertical axis represents the absorption coefficient. The absorption coefficient is located on a straight line passing through the origin. From the slope of the graph shown in FIG. 6, a conversion coefficient that associates the absorption coefficient with the density can be obtained.

  The absorption peak intensity resulting from the NN pair and NNO complex of silicon crystal whose nitrogen concentration is unknown is measured to determine the absorption coefficient. From this measurement result and FIG. 6, the density of the NN pair and the NNO complex can be obtained. By adding these two densities, the nitrogen concentration can be obtained.

Next, a second embodiment of the present invention will be described. A plurality of silicon wafers are cut out from a silicon ingot pulled up by the Czochralski method. At least one evaluation wafer is extracted from the plurality of silicon wafers. The infrared absorption spectrum of the evaluation wafer is measured, and the density of the NN pair and the NNO complex is obtained from the measurement result. By adding these two densities, the concentration of impurity nitrogen contained in the wafer for evaluation can be obtained.

  It is determined whether or not the quantified nitrogen concentration is within a predetermined reference range of nitrogen concentration. If the quantitative result is within the range, the silicon wafer cut out from the ingot is accepted, and if it is outside the range, the silicon wafer is rejected. A semiconductor integrated circuit is fabricated on a silicon wafer that is determined to be acceptable.

In the second embodiment, silicon wafers whose nitrogen concentration does not meet the criteria can be eliminated before being introduced into the wafer process.

Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. It will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made .

It is a graph which shows an example of the infrared absorption spectrum of a silicon crystal. It is a graph which shows the relationship between the temperature of the heat processing for making a silicon crystal reach a quasi-thermal equilibrium state, and an infrared light absorbency. It is a graph which shows the relationship between the heat processing time and heat processing temperature for making a silicon crystal reach a quasi-thermal equilibrium state. It is a graph which shows the relationship between the heat processing temperature for making a silicon crystal reach | attain a quasi-thermal equilibrium state, and infrared rays absorbance. It is a graph which shows the relationship between the light absorbency in a quasi-thermal equilibrium state, and the temperature which leads to a quasi-thermal equilibrium state. It is a graph which shows the relationship between an absorption coefficient and the density of a NN pair and a NNO complex.

Claims (15)

A first heating step of heating the silicon crystal to a quasi-thermal equilibrium condition under a condition that an absorption peak corresponding to a defect caused by the NNO ₂ composite disappears;
In the quasi-thermal equilibrium state after the first heating step, the infrared absorption spectrum of the silicon crystal is measured, and from the absorption spectrum, the intensity of the first absorption peak corresponding to the defect caused by the NN pair and the NNO complex are obtained. Determining the intensity of the second absorption peak corresponding to the resulting defect;
After obtaining the intensity of the first and second absorption peaks, ripening at a temperature higher than the heating temperature of the first heating step, the second heating step of bringing the silicon crystal to a quasi-thermal equilibrium state;
In the quasi-thermal equilibrium state after the second heating step, the infrared absorption spectrum of the silicon crystal is measured, and from the absorption spectrum, the intensity of the third absorption peak corresponding to the defect caused by the NN pair and the NNO complex are obtained. Determining the intensity of the fourth absorption peak corresponding to the resulting defect;
The difference between the intensity of the first absorption peak and the intensity of the third absorption peak is defined as a first difference, and the difference between the intensity of the second absorption peak and the intensity of the fourth absorption peak is defined as a second difference. NNO based on the first conversion factor for calculating the density of the NN pair from the ratio of the first difference and the second difference and the intensity of the absorption peak due to the NN pair. Calculating a second conversion factor for calculating the density of the NNO complex from the intensity of the absorption peak due to the complex;
The density of the NN pair is calculated from the intensity of the absorption peak due to the NN pair and the first conversion factor, and the density of the NNO complex is calculated from the intensity of the absorption peak due to the NNO complex and the second conversion factor. Calculating the density;
And a step of determining a nitrogen concentration in the silicon crystal based on the density of the NN pair and the NNO complex.   The method for measuring a nitrogen concentration in a silicon crystal according to claim 1, wherein the heating temperature in the first heating step is in a range of 720 to 800 ° C.   The method for measuring a nitrogen concentration in a silicon crystal according to claim 1 or 2, wherein the heating temperature in the second heating step is 800 ° C or higher. 4. The method for measuring a nitrogen concentration in a silicon crystal according to claim 1, wherein the absorption peak of the NN pair is one that appears at a wave number of 963 cm ⁻¹ or 763 cm ⁻¹ . ^In the absorption peak of the NNO ^{complex, 796cm -1, 801cm -1, 996cm} -1, any one of claims 1 to 4, wherein the use of what appears to any one of wave numbers of 1026cm ^-1 2. A method for measuring a nitrogen concentration in a silicon crystal according to 1. 2. The method for measuring a nitrogen concentration in a silicon crystal according to claim 1 , wherein the first conversion factor is 1.83 × 10 ¹⁷ cm ⁻² . The silicon crystal is heated to the first temperature, the intracrystalline defect reaction is brought to a quasi-thermal equilibrium state, the infrared absorption spectrum is measured in a state where the intracrystalline defect concentration in the quasi-thermal equilibrium state is maintained, and the first due to the NN pair. a step of determining the intensity of the second absorption peak attributable to the strength and NNO complex of 1 absorption peak,
The infrared absorption spectrum in a state where the silicon crystal is heated to a second temperature higher than the first temperature to bring the intracrystalline defect reaction to a quasi-thermal equilibrium state, and the intracrystalline defect concentration in the quasi-thermal equilibrium state is maintained. It was measured, and obtaining a fourth intensity of the absorption peak of that caused by the third magnitude and NNO complex of absorption peak attributable to NN pair,
Obtaining a difference between the intensity of the first absorption peak and the intensity of the third absorption peak , and a difference between the intensity of the second absorption peak and the intensity of the fourth absorption peak , and obtaining a ratio between them; ,
Based on the conversion factor for calculating the density of the NN pair from the intensity of the absorption peak due to the NN pair and the ratio obtained in the step of obtaining the ratio, the NNO from the intensity of the absorption peak due to the NNO complex A method for determining a conversion factor for calculating a nitrogen concentration in a silicon crystal, comprising a step of calculating a conversion factor for calculating the density of the composite. It said first temperature, the method of determining the conversion factor of the silicon crystal in nitrogen concentration for calculation according to claim 7, wherein the absorption peak attributed to NNO ₂ complex is extinguished temperature or more. The conversion factor for calculating the density of the NN pair is 1.83 × 10 ¹⁷ cm ⁻² . Determination of the conversion factor for calculating the nitrogen concentration in the silicon crystal according to claim 7 or 8. Method. A conversion coefficient for calculating the density of the NNO complex determined by the method for determining the coefficient for evaluating nitrogen concentration in silicon crystal according to any one of claims 7 to 9 , and the density of the NN pair. A method for measuring a nitrogen concentration in a silicon crystal, comprising a step of calculating a nitrogen concentration in the silicon crystal using a conversion coefficient for the purpose. The step of calculating the nitrogen concentration further includes adding the silicon crystal to the first temperature, bringing the intracrystalline defect reaction to the quasi-thermal equilibrium state, and maintaining the intracrystalline defect concentration in the quasithermal equilibrium state. It was measured, comprising the step of determining the intensity of the second absorption peak attributable to the strength and NNO complex of the first absorption peak due to the NN pair,
Silicon based on the intensity of the first absorption peak and the intensity of the second absorption peak , a conversion factor for calculating the density of the NNO complex, and a conversion factor for calculating the density of the NN pair The method for measuring a nitrogen concentration in a silicon crystal according to claim 10 , wherein the nitrogen concentration in the crystal is calculated. 12. The method for measuring a nitrogen concentration in a silicon crystal according to claim 10 , wherein the silicon crystal is a silicon wafer added with nitrogen cut out from a silicon ingot pulled up by a CZ method. Using a CZ method to pull up a silicon single crystal to which nitrogen has been added;
Cutting a plurality of silicon wafers from the pulled silicon single crystal ingot;
Heating at least one silicon wafer to a quasi-thermal equilibrium condition under a condition where an absorption peak corresponding to a defect caused by the NNO ₂ composite disappears;
The infrared absorption spectrum of the silicon wafer is measured, and from the absorption spectrum, the intensity of the first absorption peak corresponding to the defect caused by the NN pair and the intensity of the second absorption peak corresponding to the defect caused by the NNO complex The process of seeking
A step of calculating the density of the NN pair based on the conversion factor for calculating the density of the NN pair from the intensity of the absorption peak caused by the NN pair, and the intensity of the first absorption peak;
The density of the NNO complex is calculated based on the conversion factor for calculating the density of the NNO complex determined by the method according to claim 7 and the intensity of the second absorption peak. And a process of
Determining the nitrogen concentration of the silicon wafer based on the density of the NN pair and the density of the NNO complex;
And a step of determining whether or not the nitrogen concentration is within a predetermined range. If the nitrogen concentration is within a predetermined range, the determining step is acceptable for the plurality of silicon wafers, and if the nitrogen concentration is outside the predetermined range, the plurality of silicon wafers. The method of manufacturing a silicon wafer according to claim 13 , wherein: A method for manufacturing a semiconductor device, comprising: manufacturing a semiconductor integrated circuit on a silicon wafer that has been accepted by the method for manufacturing a silicon wafer according to claim 14 .

Images