Method for evaluating semiconductor samples, apparatus for evaluating semiconductor samples, and method for manufacturing semiconductor wafers.

Abstract

Provided is a semiconductor sample evaluation method which includes: acquiring a decay curve by subjecting a semiconductor sample being evaluated to measurement by means of a photoconductive decay method for multiple times while changing a surface charge density; applying signal data processing using a model formula including an exponential decay term and a constant term to at least one decay curve among the decay curves obtained by the multiple measurements; obtaining a recombination lifetime τeff of the semiconductor sample from an exponential decay formula obtained by the signal data processing; obtaining a quadratic function, in which a surface charge density-related value is represented by a variable x and a value related to the constant term is represented by a variable y, from measurement results obtained by the multiple measurements; and obtaining a surface recombination lifetime τs of the semiconductor sample from the quadratic function.

Description

[Technical Field]

[0001] The present invention relates to a method for evaluating semiconductor samples, an apparatus for evaluating semiconductor samples, and a method for manufacturing semiconductor wafers. [Background technology]

[0002] Photoconductivity decay (PCD) is a method commonly known as the PCD method and is widely used for evaluating semiconductor samples. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2019-012740 [Patent Document 2] Japanese Patent Publication No. 58-181549 [Non-patent literature]

[0004] [Non-Patent Document 1] Akira Usami, Fukuyasu Sone, Koji Murai, Kenji Sano, Laser Research, 1984, Vol. 12, No. 10, pp. 585-594. [Non-Patent Document 2] Akira Usami, Shinichi Kandachi, and Takeshi Kudo, Applied Physics, 1980, Vol. 49, No. 12, pp. 1192-1197. [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] For example, for silicon wafers, which are a type of semiconductor sample, a method for measuring recombination lifetime using the PCD method has been standardized (SEMI MF1535. Test Methods for Carrier Recombination Lifetime in Silicon Wafers by Noncontact Measurement of Photoconductivity Decay by Microwave Reflectance. 2007; hereinafter referred to as "SEMI standard").

[0006] The above SEMI standard describes the first-mode method and the 1 / e lifetime method as methods for determining the recombination lifetime using the PCD method. In the first-mode method, the time constant in the range where the decay curve obtained by measurement using the PCD method can be considered exponential decay is defined as the first-mode lifetime τ1. The first-mode lifetime τ1 is equal to the SRH (Shockley-Read-Hall) recombination lifetime (i.e., bulk lifetime) τ b and surface recombination lifetime τ s Therefore, equation: 1 / τ1 = 1 / τ b +1 / τ s It is expressed as ,. On the other hand, in the 1 / e lifetime method, when the time t0 is the time when excess carriers are excited in the semiconductor sample by the light pulse, t1 is the time until the signal intensity becomes 1 / e times (V1 = V0 / e) relative to the peak value V0 of the signal intensity V, and the 1 / e lifetime τ e is, equation: τ e It can be expressed as = t1 - t0.

[0007] In the above method, the recombination lifetime is calculated under the assumption that the decay of excess carrier concentration is exponential due to the contribution of SRH recombination (i.e., bulk recombination) alone. However, as described in Patent Document 2, for example, the recombination lifetime (Patent Document 2 is referred to as "effective lifetime τ") is calculated. eff It states that the SRH recombination lifetime (bulk lifetime) τ is determined by the purity of the semiconductor crystal and crystal defects of the semiconductor sample being evaluated. bIn addition, the surface recombination lifetime τ s is also involved. For example, in a highly clean silicon wafer, the contribution of SRH recombination becomes relatively weak, so that the contribution of surface recombination and the like cannot be ignored. Therefore, for example, when performing PCD measurement on a highly clean silicon wafer, for example, the influence of surface recombination in the final stage of attenuation causes distortion in the attenuation curve, resulting in non-exponential attenuation. For example, in such a case, it is difficult to accurately measure the recombination lifetime by the above method in which the attenuation of the excess carrier concentration is assumed to be in the form of exponential decay only by the contribution of SRH recombination.

[0008] Regarding surface recombination, Patent Document 1, Patent Document 2, Non-Patent Document 1, and Non-Patent Document 2 have proposed evaluation methods that take surface recombination into account. However, the method proposed in Patent Document 1 is not suitable for samples in which the recombination lifetime has a dependence on the elapsed time from the surface treatment because two measurements are required per sample, and it is indispensable to create a database for analysis, so it has poor versatility. Also, the methods proposed in Patent Document 2, Non-Patent Document 1, and Non-Patent Document 2 are all methods of removing the deviation from exponential decay that occurs in the initial stage of attenuation due to the influence of higher-order modes rather than the first-order mode as the influence of surface recombination. Therefore, with these methods, it is not possible to reduce or remove the influence of the deviation from exponential decay in the final stage of attenuation.

[0009] In view of the above, one aspect of the present invention aims to provide a new evaluation method for accurately evaluating the recombination lifetime of a semiconductor sample.

Means for Solving the Problems

[0010] As a result of intensive studies, the present inventor has newly found that by performing signal data processing on the attenuation curve as follows, it becomes possible to accurately obtain the recombination lifetime τ eff Furthermore, the present inventor has, by using a model formula as follows, the recombination lifetime τ effIn addition, surface recombination lifetime τ s We also discovered that it is possible to obtain this.

[0011] One aspect of the present invention is as follows: [1] The semiconductor sample to be evaluated is subjected to measurement by the photoconductivity decay method to obtain a decay curve, and this process is repeated multiple times while changing the surface charge density. The signal data processing is performed on at least one of the decay curves obtained from the multiple measurements described above, using a model equation that includes an exponential decay term and a constant term. The recombination lifetime τ of the semiconductor sample can be obtained from the exponential decay equation obtained by the above signal data processing. eff To seek, From the measurement results obtained through the multiple measurements described above, we will derive a quadratic function where x is the value related to the surface charge density and y is the value related to the constant term, and From the above quadratic function, the surface recombination lifetime τ of the above semiconductor sample is obtained. s To seek, A method for evaluating semiconductor samples, including the following. [2] The following formula (11)': C min =τ eff / τ s ×n0···(11)' (In equation (11)', n0: equilibrium carrier concentration) In C min Let τ be the minimum value of y in the above quadratic function, and the above surface recombination lifetime τ s The method for evaluating a semiconductor sample described in [1], which calculates the following. [3] The above recombination lifetime τ eff and the above surface recombination lifetime τ s Therefore, the SRH recombination lifetime τ of the above semiconductor sample b To seek, The method for evaluating a semiconductor sample according to [1] or [2], further comprising: [4] The above SRH recombination lifetime τ b The following equation (1)': 1 / τ eff = 1 / τb +1 / τ s ...(1) The method for evaluating semiconductor samples described in [3], obtained from the above. [5] Thickness d of the semiconductor sample and the surface recombination lifetime τ S From this, equation (12): S r =d / (2τ S ) ···(12) The surface recombination rate S of the above semiconductor sample r To seek, A method for evaluating a semiconductor sample, further comprising any of [1] to [4]. [6] By performing the above signal data processing, the constant term in the above model equation is canceled out to obtain the above exponential decay equation, and From the above exponential decay equation, the time constant τ b -1 +τ S -1 To seek, It further includes, τ b τ is the SRH recombination lifetime, and S The method for evaluating a semiconductor sample described in any of [1] to [5] is the surface recombination lifetime. [7] The method for evaluating a semiconductor sample according to [6], wherein the signal data processing described above includes repeatedly sampling and taking the difference of a time-series signal modeled by the above model equation. [8] The method for evaluating a semiconductor sample according to [7], further comprising performing autoscaling to determine the sampling area for which the above sampling is performed. [9] The semiconductor sample evaluation method described in [8], wherein the above auto-scaling determines a region in which Auger recombination is less affected and noise is less affected as the sampling region.

[10] The above model equation is given by the following equation (10)': xi(ti) = A × exp[-(τ) b -1 +τ S -1 )ti]-C ···(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ b :SRH recombination lifetime, τ S : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample described in any of [1] to [9].

[11] The following formula (11)': C min =τ eff / τ s ×n0···(11)' (In equation (11)', n0: equilibrium carrier concentration) In C min Let τ be the minimum value of y in the above quadratic function, and the above surface recombination lifetime τ s Calculate, The above recombination lifetime τ eff and the above surface recombination lifetime τ s Therefore, the SRH recombination lifetime τ of the above semiconductor sample b This further includes seeking, The above SRH recombination lifetime τ b The following equation (1)': 1 / τ eff = 1 / τ b +1 / τ s ...(1) From, By performing the above signal data processing, the constant term in the above model equation is canceled out to obtain the above exponential decay equation, and From the above exponential decay equation, the time constant τ b -1 +τ S -1 To seek, It further includes, τ b τ is the SRH recombination lifetime, and S This is the surface recombination lifetime, The above signal data processing includes repeatedly sampling the time-series signal modeled by the above model equation and taking the difference, This further includes performing autoscaling to determine the sampling area for which the above sampling will be performed, The above auto-scaling determines the sampling region as one in which the effects of Auger recombination are minimal and the effects of noise are minimal. The above model equation is the following equation (10)': xi(ti) = A × exp[-(τ) b -1 +τ S -1 )ti]-C ···(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ b :SRH recombination lifetime, τ S : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample described in any of [1] to

[10] .

[12] The following formula (11)': C min =τ eff / τ s ×n0···(11)' (In equation (11)', n0: equilibrium carrier concentration) In C min Let τ be the minimum value of y in the above quadratic function, and the above surface recombination lifetime τ s Calculate, The thickness d of the above semiconductor sample and the surface recombination lifetime τ S From this, equation (12): S r =d / (2τ S ) ···(12) The surface recombination rate S of the above semiconductor sample r This further includes seeking, By performing the above signal data processing, the constant term in the above model equation is canceled out to obtain the above exponential decay equation, and From the above exponential decay equation, the time constant τ b -1 +τ S -1 To seek, It further includes, τ b τ is the SRH recombination lifetime, and S This is the surface recombination lifetime, The above signal data processing includes repeatedly sampling the time-series signal modeled by the above model equation and taking the difference, This further includes performing autoscaling to determine the sampling area for which the above sampling will be performed, The above auto-scaling determines the sampling region as one in which the effects of Auger recombination are minimal and the effects of noise are minimal. The above model equation is given by the following equation (10): xi(ti) = A × exp[-(τ) b -1 +τ S -1 )ti]-C ···(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ b :SRH recombination lifetime, τ S : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample described in any of [1] to

[11] . A semiconductor sample evaluation apparatus for performing the semiconductor sample evaluation method described in any of

[13] [1] to

[12] , A measurement unit that subjects the semiconductor sample to be evaluated to measurement using the photoconductivity attenuation method, A processing unit that performs signal data processing on the above-mentioned decay curve using a model equation that includes an exponential decay term and a constant term, and calculates the above-mentioned quadratic function, The evaluation apparatus, including the above-mentioned apparatus.

[14] The evaluation apparatus according to

[13] , wherein the light pulse intensity is variable in the measurement unit described above.

[15] The evaluation apparatus according to

[13] or

[14] , wherein the measurement unit further includes a corona charge processing unit.

[16] The evaluation apparatus according to any one of

[13] to

[15] , further comprising a surface charge density related value measuring unit for measuring the surface charge density related value of the semiconductor sample subjected to the above measurement.

[17] In the above measurement unit, the light pulse intensity is variable, The evaluation apparatus according to any one of

[13] to

[16] further includes a corona charge processing unit and a surface charge density related value measuring unit for measuring the surface charge density related value of the semiconductor sample subjected to the above measurement.

[18] Manufacturing a semiconductor wafer lot containing multiple semiconductor wafers, To extract at least one semiconductor wafer from the above semiconductor wafer lot, The extracted semiconductor wafers are evaluated according to one of the evaluation methods described in [1] to

[12] , and Based on the above evaluation, semiconductor wafers from the same semiconductor wafer lot as those judged to be good products will be prepared for shipment as product semiconductor wafers. A method for manufacturing semiconductor wafers containing [the specified material].

[19] Manufacturing evaluation semiconductor wafers under test manufacturing conditions, The manufactured evaluation semiconductor wafer is evaluated by any of the semiconductor sample evaluation methods described in [1] to

[12] . Based on the results of the above evaluation, either the manufacturing conditions modified from the above test manufacturing conditions will be determined as the actual manufacturing conditions, or the above test manufacturing conditions will be determined as the actual manufacturing conditions, and To manufacture semiconductor wafers under the actual manufacturing conditions determined above, A method for manufacturing semiconductor wafers containing [the specified material]. [Effects of the Invention]

[0012] According to one aspect of the present invention, the recombination lifetime τ eff This makes it possible to determine the surface recombination lifetime τ with high accuracy, and furthermore, the surface recombination lifetime τ s It also becomes possible to calculate this. [Brief explanation of the drawing]

[0013] [Figure 1] This is an explanatory diagram illustrating a specific example of signal data processing. [Figure 2] This is an explanatory diagram illustrating a specific example of signal data processing. [Figure 3]It is an explanatory diagram of a specific example of signal data processing. [Figure 4] It is an explanatory diagram of an example of a method for evaluating a semiconductor sample according to one aspect of the present invention. [Figure 5] It shows a comparison result between an example (new method) of a method for evaluating a semiconductor sample according to one aspect of the present invention and a conventional method. [Figure 6] It shows a comparison result between an example (new method) of a method for evaluating a semiconductor sample according to one aspect of the present invention and a conventional method. [Figure 7] It shows a comparison result between an example (new method) of a method for evaluating a semiconductor sample according to one aspect of the present invention and a conventional method. [Figure 8] It is a schematic diagram of a plot for deriving the surface recombination lifetime τs. [Figure 9] It shows the quadratic curve obtained in Example 1. [Figure 10] It shows the quadratic curve obtained in Example 2. [Figure 11] It shows the attenuation curve obtained by subjecting an n-type silicon wafer (single crystal silicon wafer) to measurement by the μ-PCD method after performing chemical passivation treatment as a surface treatment. [Figure 12] It shows the fitting curve obtained by fitting the attenuation curve shown in FIG. 11 to the first-order mode lifetime and the fitting curve obtained by fitting the 1 / e lifetime. [Embodiments for Carrying Out the Invention]

[0014] [Method for Evaluating Semiconductor Sample] A method for evaluating a semiconductor sample according to one aspect of the present invention includes obtaining an attenuation curve by subjecting a semiconductor sample to be evaluated to measurement by the photoconductive decay method, performing this a plurality of times while changing the surface charge density, performing signal data processing on at least one of the attenuation curves obtained by the plurality of measurements using a model equation including an exponential decay term and a constant term, and obtaining the recombination lifetime τ of the semiconductor sample from the exponential decay equation obtained by the signal data processing. effTo determine the surface charge density related value and the value related to the constant term, and to obtain a quadratic function from the measurement results obtained from the above multiple measurements, where x is the variable and y is the variable, and the surface recombination lifetime τ of the semiconductor sample is obtained from the above quadratic function. s This includes seeking. The evaluation method described above will be explained in more detail below.

[0015] <Semiconductor sample to be evaluated> The evaluation target for the above evaluation method can be any semiconductor sample. Examples of semiconductor samples include various types of semiconductor samples such as single-crystal silicon, polycrystalline silicon, and SiC. The shape and dimensions of the semiconductor sample to be evaluated are not particularly limited. For example, the semiconductor sample to be evaluated can be a wafer-shaped semiconductor sample, i.e., a semiconductor wafer, such as a single-crystal silicon wafer. However, the semiconductor sample to be evaluated can also be in a shape other than a wafer. Furthermore, the conductivity type of the semiconductor sample to be evaluated is not particularly limited and may be n-type or p-type.

[0016] <Measurement by photoconductivity attenuation method> For the measurement using the photoconductive attenuation method in the above evaluation method, known techniques can be applied. A specific example of the photoconductive attenuation method is the microwave photoconductive attenuation (μ-PCD) method. However, the measurement using the photoconductive attenuation method in the above evaluation method is not limited to the μ-PCD method. For example, considering the maximum carrier injection amount of a typical μ-PCD apparatus, if the semiconductor sample to be evaluated is p-type silicon, its resistivity is preferably around 1 to 100 Ωcm, and if it is n-type silicon, its resistivity is preferably around 0.5 to 100 Ωcm.

[0017] In the measurement, excess carriers are first excited in the semiconductor sample to be evaluated using an optical pulse. Therefore, a PCD (Pulse Controlled Diode) device is used in which the optical pulse intensity can be set to a desired value (i.e., variable) in the measurement unit, and the measurement is performed after setting the optical pulse intensity to an appropriate value. Furthermore, as will be described later, in order to perform multiple measurements with varying surface charge density, a PCD apparatus equipped with a corona charge processing unit in the measurement section can be used, and by changing the number of corona charges, for example, multiple measurements can be performed with varying surface charge density. For example, a PCD apparatus equipped with a corona charge processing unit that can apply positive charge, negative charge, or both positive and negative charge to the surface of one semiconductor sample using the corona charge method, or apply positive charge or negative charge to the surfaces of multiple semiconductor samples using the corona charge method can be used. Alternatively, in order to perform multiple measurements with varying surface charge density, the surface charge density of the semiconductor sample to be measured can also be changed by changing the concentration of the chemical used in the chemical passivation treatment (e.g., iodine, quinhydrone, etc.) and / or changing the treatment time. In this case, it is preferable to use a PCD apparatus with an apparatus configuration that is compatible with chemical passivation. Examples of such PCD devices include devices equipped with a mechanism for applying a drug solution (e.g., iodine-ethanol solution) to both the front and back surfaces of a semiconductor sample and measuring it while it is in a transparent bag made of polyethylene or the like, and devices equipped with a mechanism for applying a drug solution (e.g., iodine-ethanol solution) to both the front and back surfaces of a semiconductor sample outside the device and measuring it while it is in a transparent bag made of polyethylene or the like. Furthermore, as a PCD apparatus, an apparatus equipped with a surface charge density-related value measuring device such as surface charge amount and surface potential can be used. For example, the surface charge density in each measurement can be determined by actually measuring the surface charge amount, surface potential, etc. using a surface charge density-related value measuring device such as a Kelvin probe or a non-contact CV measuring device. An example of a surface charge density-related value measuring device is a non-contact CV measuring device that measures surface charge density-related values ​​by creating an air gap between the surface of a semiconductor sample and an electrode facing that surface and applying a voltage between them.

[0018] An attenuation curve is obtained by measurement using the photoconductive attenuation method described above. Specifically, the attenuation curve is a curve that shows the change in signal intensity over time with respect to the elapsed time after excitation light irradiation. "Elapsed time after excitation light irradiation" is, more precisely, the elapsed time from the end of excitation light irradiation. Also, for example, in the μ-PCD method, the signal intensity is the intensity of the reflected microwaves. Figure 11 shows the attenuation curve obtained by measuring an n-type silicon wafer (single-crystal silicon wafer) using the μ-PCD method after chemical passivation treatment as a surface treatment. The attenuation curve shown in Figure 11 is distorted from the mid-term region to the late-term region due to the effect of surface recombination. Figure 12 shows the fitting curve obtained by first-order mode lifetime fitting and the fitting curve obtained by 1 / e lifetime fitting as described in the SEMI standard on the attenuation curve shown in Figure 11. Neither of the two fitting curves shown in Figure 12 fits the attenuation curve well, especially from the mid-term region to the late-term region. This is because the first-order mode method and the 1 / e lifetime method are fitted under the assumption that the decay of excess carrier concentration takes the form of exponential decay due solely to the contribution of SRH recombination. In contrast, the above evaluation method applies signal data processing to the decay curve, as detailed below, thereby determining the recombination lifetime τ eff This makes it possible to determine this with high accuracy.

[0019] <Recombination lifetime τ eff Derivation of > (Signal data processing for attenuation curves) In the above evaluation method, signal data processing is performed on at least one of the attenuation curves obtained from measurements of the semiconductor sample under evaluation using the photoconductivity attenuation method, using a model equation that includes an exponential attenuation term and a constant term. The attenuation curves for which signal data processing is performed may be one or more. In the case of two or more, for example, the average of the recombination lifetimes obtained from the signal data processing for each attenuation curve is used to obtain the recombination lifetime τ of the semiconductor sample under evaluation. effIt can be adopted as. As the attenuation curve for performing signal data processing, among the plurality of attenuation curves obtained by the above-mentioned plurality of measurements, it is preferable to use the attenuation curve obtained by measurement under measurement conditions with a small surface charge density, and it is more preferable to use the attenuation curve obtained by measurement under measurement conditions with the smallest surface charge density among the plurality of measurements. For example, when changing the surface charge density by changing the number of corona charges, it is preferable to use the attenuation curve obtained by measurement with zero corona charge. It is most preferable to use the attenuation curve obtained by measurement under measurement conditions with a surface charge density of zero, but in practice, the surface charge density can generally be actually measured when it is a value exceeding 0.

[0020] In the above model formula, the constant term is included outside the exponential decay term. The present inventor believes that it is appropriate to represent the attenuation curve when surface recombination occurs by an equation including an exponential decay term and a constant term, preferably an equation in which the constant term is subtracted from the exponential decay term. By performing signal data processing to cancel out the constant term for such an equation, only the exponential decay term remains. That is, an equation of exponential decay is obtained. By using this equation of exponential decay, the value of the recombination lifetime can be obtained as a value including the effects of SRH recombination and surface recombination. Thereby, it becomes possible to accurately obtain the recombination lifetime of the semiconductor sample to be evaluated. Hereinafter, such signal data processing will be described in more detail.

[0021] If the excess carrier concentration is taken as x [1 / cm 3 and the function of time is x(t), the function x(t) can be approximately expressed by the following formula (1). x(t)=A×exp[-(τ b -1 +τ S -1 )t]-C ···(1)

[0022] In formula (1), τ b represents the SRH recombination lifetime, and the unit is, for example, μsec, and τ Srepresents the surface recombination lifetime, and its unit is, for example, μsec. A and C are each independently constants [1 / cm²]. 3 ] represents a positive constant, preferably a positive constant. A and C are τ b -1 +τ S -1 It is a constant determined by . Equation (1) above is a suitable function when both SRH recombination and surface recombination due to surface levels contribute to the recombination lifetime, and the excess carrier concentration is a function of time x(t), under the condition that the excess carrier concentration is greater than the carrier concentration at equilibrium.

[0023] Equation (10)' below can be given as a model equation for performing signal data processing based on equation (1). xi(ti) = A × exp[-(τ) b -1 +τ S -1 )ti]-C ···(10)'

[0024] In equation (10)', ti is the elapsed time after excitation light irradiation, xi(ti) is the signal intensity at elapsed time ti, and its unit is, for example, mV, τ b is the SRH recombination lifetime, and τ S is the surface recombination lifetime, and A and C are independent constants, each with units, for example, mV. The above model equation (10)' includes an exponential decay term and a constant term, specifically the exponential decay term "A × exp[-(τ b -1 +τ S -1 This is the equation obtained by subtracting the constant term "C" from "(ti)". Equation (10)' is an example of the above model equation. For example, if signal data processing is performed to cancel out the constant term in a model equation that includes a constant term like equation (10)', only the exponential decay term remains, and the time constant τ is then determined by the exponential decay approximation method. b -1 +τ S -1 (The sum of the reciprocal of the surface recombination lifetime and the reciprocal of the SRH recombination lifetime) can be calculated. The τ obtained in this way isb -1 +τ S -1 The reciprocal of the recombination lifetime τ of the semiconductor sample being evaluated. eff This value can be adopted. As for exponential decay approximation methods, general exponential decay approximation methods such as the first-order lifetime method and the 1 / e lifetime method can be used.

[0025] The following describes a specific example of signal data processing, using the case where the above model equation is equation (10)' as an example. However, the signal data processing described below is illustrative, and the present invention is not limited to such examples.

[0026] Signal data processing may include determining a sampling region in the attenuation curve obtained by measurement using the photoconductive attenuation method, and repeatedly sampling the time-series signal modeled by the above model equation (specifically, measurement points on the attenuation curve) and taking the difference within this sampling region. The sampling region can be determined, for example, by autoscaling, and specific examples include the following method. For example, the following method allows for determining a region where the effects of Auger recombination are small and the effects of noise are small as the sampling region. That is, by excluding regions with high signal intensity where the effects of Auger recombination are large and regions with low signal intensity where the effects of noise are large in the attenuation curve obtained by measurement using the photoconductive attenuation method, it becomes possible to use regions with strong effects of SRH recombination and surface recombination as the sampling region.

[0027] First, set an arbitrary signal intensity (e.g., 60% of the peak value) as the starting point of the sampling area, and perform signal data processing once. For the calculation results obtained from the signal data processing, R 2 Values ​​that can serve as indicators of the degree of fit to exponential decay, such as the sum of squared residuals, are calculated. The degree of fit to exponential decay is evaluated by determining whether the calculated values ​​of these indicators satisfy a predetermined threshold. The above calculated value is a pre-set threshold (for example, R 2If the condition (≥0.99) is satisfied, the starting point of the sampling region set above can be determined as the starting point of the sampling region when performing signal data processing. The above calculated value is a pre-set threshold (for example, R 2 If the condition (≥0.99) is not satisfied, the starting point of the sampling area is shifted to the side with lower signal intensity and recalculation is performed. Recalculation can be performed once or more times, and if the evaluation result of the recalculation satisfies a preset threshold, the starting point in that recalculation can be determined as the starting point of the sampling area. The end point of the sampling area can be set to a position where the signal-to-noise ratio (SNR) is below a predetermined threshold. The SNR can be calculated, for example, by the following formula. The threshold for the SNR can be, for example, 5 dB or less, and it is preferable to determine the end point as a position where the signal-to-noise ratio of the signal is 0 dB, i.e., a position where the noise and signal are roughly equal. SN ratio [dB]=20log 10 [(Signal dispersion at arbitrary time) / (Noise dispersion in equilibrium state)]

[0028] Figures 1 to 3 illustrate specific examples of signal data processing. Signal data processing can be performed, for example, with 3N sampling points, as follows. Here, N is any integer, for example, it can be 2 or greater. Also, if the total number of points in the signal data is 1000, then N can be 333 or less. That is, if the total number of points in the signal data is T, then N can be an integer less than or equal to "T × 1 / 3". First, using the starting point (point 1) as the reference, we calculate the average A1 from point 1 to point N and the average B1 from point N+1 to point 2N (see Figure 1). A1 = {x(t1) + x(t2) + ... + x(t N )} / N B1={x(t N+1 ) + x(t N+2 )+…+x(t 2N )} / N Let Y(t1) be the value obtained by subtracting B1 from A1. Y(t1) = A1 - B1 Next, using the second point as the baseline, we calculate the average A2 from the second point to the N+1 point and the average B2 from the N+2 point to the 2N+1 point (see Figure 2). A2 = {x(t2) + x(t3) + ... + x(t N+1 )} / N B2={x(t N+2 ) + x(t N+3 )+…+x(t 2N+1 )} / N Let Y(t2) be the value obtained by subtracting B2 from A2. Y(t2) = A2 - B2 Continuing the same calculation, we finally arrive at the average A of the N+1th to 2Nth points. N+1 and the average B of points 2N+1 to 3N. N+1 Calculate the result (see Figure 3). A N+1 ={x(t N+1 ) + x(t N+2 )+…+x(t 2N )} / N B N+1 ={x(t 2N+1 ) + x(t 2N+2 )+…+x(t 3N )} / N A N+1 From B N+1 The value obtained by subtracting Y(t) N+1 ) Y(t N+1 )=A N+1 -B N+1 The time-series signal data sequence obtained by continuing the above calculations has the constant term in equation (10)' canceled out, resulting in the following exponential decay equation (2). By applying a general exponential decay approximation method to this equation (2), the time constant τ b -1 +τ S -1 We can find the value of "τ" obtained in this way. b -1 +τ S -1 "The recombination lifetime τ of the semiconductor sample being evaluated effThis value can be adopted. Examples of the above exponential decay approximation methods include the first-order mode method and the 1 / e lifetime method. Y(t) = A' × exp[-(τ) b -1 +τ S -1 )t]···(2) (A': arbitrary constant)

[0029] (recombination lifetime τ) eff (Specific form of the calculation method) Figure 4 is an explanatory diagram illustrating an example of a semiconductor sample evaluation method according to one aspect of the present invention. The decay curve shown in the left panel of Figure 4 is the same as the decay curve shown in Figure 11, and is the decay curve obtained by measuring an n-type silicon wafer (single-crystal silicon wafer, resistivity: 10 Ωcm) using the μ-PCD method after chemical passivation as a surface treatment. The maximum carrier injection rate of the μ-PCD here is approximately 1E17 / cm². 3 That was the case. Note that "E17" is "×10 17 This indicates ". For the attenuation curve shown in the left figure of Figure 4, equation (10)' was used as the model equation, and signal data processing was performed with 3N sampling points, as explained earlier with reference to Figures 1 to 3. The starting point of the sampling region was defined as the threshold "R 2 The value was determined using the method described above, with the threshold set to "≥0.99". The end point of the sampling region was set to the position where the signal-to-noise ratio (SNR) was 0 dB, as described above. By performing the above signal data processing, the constant term in equation (10)' is canceled out, and a linear equation of equation (2) consisting only of the exponential decay term is obtained (solid line in the right figure of Figure 4). By applying the first-order mode method to this linear equation, the time constant τ b -1 +τ S -1 We find this time constant τ b -1 +τ S -1 The recombination lifetime (Example A in Table 1), calculated as the reciprocal of the above, was the value shown in Table 1. Table 1 also shows the recombination lifetimes obtained by applying the first-mode method and the 1 / e method described in the SEMI standard to the attenuation curve shown in the left figure of Figure 4.

[0030] [Table 1]

[0031] The results shown in Table 1 confirm that the conventional first-order mode method and the 1 / e lifetime method underestimate the recombination lifetime by approximately 60% compared to Example A.

[0032] Next, we confirmed that the function in equation (1) is a function of time close to the measured decay by the following method. First, in order to determine the remaining undetermined parameter constants A and C in equation (1), the following fitting was performed. The time constant τ obtained in the above example b -1 +τ S -1 Using this, equation (1) becomes exp[-(τ b -1 +τ S -1 The equation takes the form of a straight line for )t]. That is, in the sampling region determined by autoscaling, the decay curve is in the form of exp[-(τ) as shown in Figure 5 (left). b -1 +τ S -1 For t], a linear approximation can be obtained, and the constants A and C can be obtained as its slope and intercept. The time constant τ obtained in this way b -1 +τ S -1And by applying equation (1) with constants A and C, as shown in the right figure of Figure 5, a fitting curve (new method fitting in the right figure of Figure 5) that fits a wider range of the damping curve is obtained compared to fitting by the first-order mode method or the 1 / e lifetime method (first-order mode lifetime fitting and 1 / e lifetime fitting in the right figure of Figure 5). From these results, it can be confirmed that the function in equation (1) is a function of time that is close to the measured decay.

[0033] Figures 6 and 7 show a comparison between an example of a semiconductor sample evaluation method (new method) according to one aspect of the present invention and a conventional method. Figure 6 (left) shows the recombination lifetime values ​​obtained using the same method as in Example A (new method in Figure 6) and the recombination lifetime values ​​obtained using the conventional first-order mode method ("conventional method" in Figure 6) for several n-type silicon wafers that yielded a recombination lifetime of approximately 7000 μsec using the same method as in Example A above. Figure 6 (right) shows the CV values ​​(standard deviation / arithmetic mean × 100, unit: %) of the measurement results shown in Figure 6 (left). From the results shown in Figure 6, it can be confirmed that the variation in recombination lifetime values ​​is suppressed in the new method compared to the conventional method.

[0034] Figure 7 shows the fitting curves obtained by fitting the decay curves obtained by measurement using the μ-PCD method for sample 1 (largest positive difference) and sample 2 (largest negative difference) using the conventional method in Figure 6, using equation (1) as described earlier, and the fitting curve obtained by fitting using the conventional first-order mode method. Figure 7 also shows the recombination lifetime values ​​for each sample obtained using the new method and the conventional method. The results shown in Figure 7 confirm that the function of equation (1) is a function of time close to the measured decay.

[0035] <Surface recombination lifetime τ s SRH recombination lifetime τ b , surface recombination rate S rDerivation of > Surface recombination lifetime τ s Regarding the derivation of the formula, as a result of the inventors' diligent research, it was found that, using Poisson's equation for the charge distribution near the surface of a semiconductor sample (relationship between carrier concentration and surface charge density), the constant term C in equation (10)' can be expressed by equation (11). From equation (11), it became clear that the constant term is a quadratic expression of the surface charge density Qs, and thus it was newly discovered that it can be analyzed using a quadratic function. C=(τ eff / τ s ) / (2kTε s ε0)×Q s 2 +τ eff / τ s ×n0···(11)

[0036] In equation (11), k is the Boltzmann constant [eV / K], T is the absolute temperature [K], and ε s ε0 is the permittivity of the semiconductor sample [F / m], ε0 is the permittivity of vacuum [F / m], Q s The surface charge density [F / m 2 ], n0 is the equilibrium carrier concentration [cm³ -3 ]

[0037] τ in equation (11) eff As mentioned earlier, this can be derived by performing a fitting calculation using equation (10)' on the decay curve obtained by subjecting the semiconductor sample to evaluation to measurement using the photoconductivity decay method.

[0038] Next, Figure 8 shows the surface recombination lifetime τ. s This is a schematic diagram of the plot used for the derivation. τ s To derive this, the semiconductor sample under evaluation is subjected to measurement using the photoconductive decay method to obtain decay curves, which are performed multiple times with varying surface charge densities. When the constant term C calculated using equation (10)' is plotted against the surface charge density Qs for the decay signals obtained for each different surface charge density, it becomes a quadratic function as shown in equation (11) (see Figure 8). The minimum value C of the quadratic function shown in Figure 8 min The second term of the equilibrium equation (11) is "τ eff / τs The expression "×n0" is represented by the following formula (11)'.

[0039] C min =τ eff / τ s ×n0···(11)'

[0040] Therefore, the minimum value C of the quadratic function min and equilibrium carrier concentration n0 and τ eff By substituting and into equation (11)', the surface recombination lifetime τ can be obtained. S This can be calculated. However, in the μ-PCD method, the microwave reflectance signal intensity is measured in voltage, and the unit of C is voltage. Therefore, the relationship between the reflectance signal intensity and the carrier concentration is converted using the amount of carriers injected by the light pulse and the initial microwave reflectance signal intensity. The equilibrium carrier concentration n0 is the equilibrium carrier concentration of the semiconductor sample being measured.

[0041] The second term of equation (11) "τ eff / τ s The minimum value C of ×n0 min For example, as shown in Figure 8, measurement points are plotted on a graph with the vertical axis (y-axis) as C and the horizontal axis (x-axis) as surface charge density. The value of the vertical axis y is then obtained as the minimum value of the vertical axis y in a quadratic function (quadratic curve) obtained by fitting these plots using a known method such as the least squares method.

[0042] In the example shown in Figure 8, the vertical axis (y-axis) represents the absolute value of the constant term C, and the horizontal axis (x-axis) represents the surface charge density. However, the variable x on the horizontal axis is not limited to the value of the surface charge density itself. For example, the surface charge density can be changed by the number of corona charge treatments (i.e., the number of corona charges). If the surface charge density is changed by the number of corona charges, the number of corona charges can also be used as the variable x on the horizontal axis. For example, the number of corona charges multiplied by a positive charge can be taken as a positive value (for example, x=1 if it is 1), and the number of corona charges multiplied by a negative charge can be taken as a negative value (for example, x=-1 if it is 1). Furthermore, the variable y on the horizontal axis is not limited to the absolute value of the constant term C; for example, the value of the constant term C itself can be taken as the variable y on the vertical axis.

[0043] That is, by adopting surface charge density-related values ​​as the variable x and values ​​related to the constant term as the variable y, the obtained plot is fitted using a known method to the quadratic function: y = ax 2 We obtain the quadratic function y + bx + c, where a is a positive constant, and b and c are positive or negative constants or zero. The minimum value of y in this quadratic function is given by C in equation (11). min From equation (11)', the surface recombination lifetime τ s It is possible to find this.

[0044] Furthermore, equation (1)': 1 / τ eff = 1 / τ b +1 / τ s ...(1) t τ eff and τ S By substituting this, the SRH recombination lifetime (bulk lifetime) τ b It is possible to find this.

[0045] Also, the surface recombination rate S r This refers to the thickness d of the semiconductor sample and the surface recombination lifetime τ. S Therefore, it can be calculated using the following formula (12). S r =d / (2τS ) ···(12)

[0046] [Evaluation equipment for semiconductor samples] One aspect of the present invention relates to an evaluation apparatus for semiconductor samples that implements the above evaluation method. The evaluation apparatus includes a measurement unit that subjects the semiconductor sample to be evaluated to measurement by the photoconductivity decay method, and a processing unit that performs signal data processing on the decay curve using a model equation that includes an exponential decay term and a constant term, and calculates the quadratic function.

[0047] The above-described measurement unit performs measurements on the semiconductor sample to be measured using the photoconductivity decay method. The measurement unit can change the light pulse intensity, as previously described. In one embodiment, the measurement unit may include a corona charge processing unit to change the surface charge density. Details of the measurement unit are as previously described. Furthermore, as previously described for the PCD apparatus, the measurement unit may be equipped with a surface charge density-related value measuring device.

[0048] To carry out a semiconductor sample evaluation method according to one aspect of the present invention, the processing unit of the evaluation apparatus performs signal data processing on the decay curve using a model equation that includes an exponential decay term and a constant term, and calculates the quadratic function. The processing unit can be configured using a known analysis program.

[0049] Furthermore, the processing unit performs signal data processing on the decay curve using a model equation that includes an exponential decay term and a constant term, thereby canceling out the constant term in the model equation to obtain the exponential decay equation, and from the exponential decay equation the time constant τ b -1 +τ S -1 It is possible to find this.

[0050] The above signal data processing may include repeatedly sampling and taking the difference of the time-series signal modeled by the above model equation.

[0051] Furthermore, the processing unit can perform autoscaling to determine the sampling area for which the sampling will be performed. Through this autoscaling, the processing unit can determine a sampling area in which the effects of recombination are minimal and the effects of noise are minimal. This autoscaling is as described above.

[0052] The evaluation apparatus described above can derive the recombination lifetime τeff of the semiconductor sample from the exponential decay formula obtained by the signal data processing, either through a measurement unit or a calculation unit provided separately from the measurement unit. The surface recombination lifetime τs and SRH recombination lifetime τ after the quadratic function has been determined. b and surface recombination rate S r The derivation of these values ​​can also be performed by the measurement unit or a calculation unit provided separately from the measurement unit. These derivations can be performed using a known analysis program provided in the measurement unit or calculation unit.

[0053] In one embodiment, the evaluation device may further include the processing unit and / or the calculation unit, which is a PCD measuring device having a measurement unit that performs measurements using the PCD method. In another embodiment, one or more computers separate from the PCD measuring device having a measurement unit that performs measurements using the PCD method may include the processing unit and / or the calculation unit, and various information such as measurement results, processing results, and calculation results can be transmitted and received between such computers and the PCD measuring device, or between even more computers, by wired or wireless communication. One computer may include the processing unit and the calculation unit, or different computers may each include the processing unit and the calculation unit.

[0054] [Method for manufacturing semiconductor wafers] One aspect of the present invention is, Manufacturing a semiconductor wafer lot containing multiple semiconductor wafers, To extract at least one semiconductor wafer from the above semiconductor wafer lot, The extracted semiconductor wafers are evaluated using the evaluation method described above, and Based on the above evaluation, semiconductor wafers from the same semiconductor wafer lot as those judged to be good products will be prepared for shipment as product semiconductor wafers. A method for manufacturing semiconductor wafers containing (hereinafter also referred to as "Manufacturing Method 1"), Regarding.

[0055] Furthermore, one embodiment of the present invention is: To manufacture evaluation semiconductor wafers under test manufacturing conditions, The manufactured evaluation semiconductor wafer is evaluated using the above-described semiconductor sample evaluation method. Based on the results of the above evaluation, either the manufacturing conditions modified from the above test manufacturing conditions will be determined as the actual manufacturing conditions, or the above test manufacturing conditions will be determined as the actual manufacturing conditions, and To manufacture semiconductor wafers under the actual manufacturing conditions determined above, A method for manufacturing semiconductor wafers containing (hereinafter also referred to as "manufacturing method 2"), Regarding.

[0056] In manufacturing method 1, semiconductor wafers from the same lot as those determined to be good products as a result of so-called sampling inspection are prepared for shipment as product semiconductor wafers. On the other hand, in manufacturing method 2, semiconductor wafers manufactured under test manufacturing conditions are evaluated, and actual manufacturing conditions are determined based on the evaluation results. In both manufacturing method 1 and manufacturing method 2, the evaluation of the semiconductor wafers is performed by the evaluation method according to one embodiment of the present invention described above.

[0057] In manufacturing method 1, the semiconductor wafer lot can be manufactured in the same way as a general semiconductor wafer manufacturing method. For example, a polished wafer can be cited as an example of a silicon wafer, which is a form of semiconductor wafer. Polished wafers can be manufactured by a manufacturing process that includes cutting (slicing), chamfering, rough polishing (e.g., lapping), etching, mirror polishing (finish polishing), and cleaning performed between or after the above processing steps, from a silicon single crystal ingot grown by the Czochralski method (CZ method), etc. Annealed wafers can be manufactured by subjecting a polished wafer manufactured as described above to heat treatment, specifically annealing. Epitaxial wafers can be manufactured by vapor-phase growth (epitaxial growth) of an epitaxial layer on the surface of a polished wafer manufactured as described above.

[0058] The total number of semiconductor wafers included in a semiconductor wafer lot is not particularly limited. The number of semiconductor wafers taken from a manufactured semiconductor wafer lot and subjected to so-called sampling inspection is at least one, but may be two or more, and this number is not particularly limited.

[0059] A semiconductor wafer extracted from a semiconductor wafer lot is evaluated by an evaluation method according to one aspect of the present invention. For example, the quality of the evaluated semiconductor wafer can be determined using the recombination lifetime value obtained by such evaluation as an indicator. For example, the higher the amount of metal contamination, the higher the recombination lifetime τ measured by the PCD method. eff The recombination lifetime becomes shorter. Therefore, based on the recombination lifetime value measured by the PCD method, it is possible to evaluate the presence and / or degree of metal contamination of the semiconductor wafer. For example, the recombination lifetime effA value equal to or exceeding a predetermined threshold can be used as a criterion for determining whether a product is good. Such a threshold can be set according to the quality required for the product wafer. Semiconductor wafers from the same semiconductor wafer lot as those determined to be good can be prepared (e.g., packaged) for shipment as product semiconductor wafers.

[0060] Regarding manufacturing method 2, the test manufacturing conditions and actual manufacturing conditions include various conditions in the various processes for manufacturing semiconductor wafers. The various processes for manufacturing semiconductor wafers are as described above for manufacturing method 1. Note that "actual manufacturing conditions" refers to the manufacturing conditions for the finished semiconductor wafer.

[0061] In manufacturing method 2, as a preliminary step to determining actual manufacturing conditions, test manufacturing conditions are set, and evaluation semiconductor wafers are manufactured under these test manufacturing conditions. The manufactured semiconductor wafers are evaluated by an evaluation method according to one aspect of the present invention. The recombination lifetime determined by such evaluation is eff Using this value as an indicator, it is possible to determine whether the test manufacturing conditions are suitable for adoption as actual manufacturing conditions, or whether the manufacturing conditions modified from the test manufacturing conditions should be adopted as actual manufacturing conditions. For example, the recombination lifetime obtained from the same perspective as described earlier regarding manufacturing method 1. eff The value of a certain parameter being equal to or exceeding a predetermined threshold can be used as a criterion for determining whether the test manufacturing conditions are suitable for adoption as actual manufacturing conditions. Examples of manufacturing conditions that may be modified as a result of this determination include those that may cause metal contamination. One example is the modification of the heat treatment furnace used (e.g., replacement of parts, cleaning of parts, cleaning of the furnace interior, etc.).

[0062] For further details of manufacturing methods 1 and 2, known technologies relating to semiconductor wafer manufacturing methods can be applied. According to manufacturing methods 1 and 2, for example, product semiconductor wafers with low metal contamination can be stably supplied to the market. [Examples]

[0063] The present invention will be further described below based on examples. However, the present invention is not limited to the embodiments shown in the examples.

[0064] [Example 1] μPCD measurements were performed on n-type silicon wafers (single-crystal silicon wafers) with varying numbers of corona charges. The number of corona charges with positive charge was set to 1, 3, or 5, and the number of corona charges with negative charge was set to 1, 3, or 5, with the charge amount (absolute value) being the same for each charge. The number of corona charges was adopted as the surface charge density-related value, variable x, with positive values ​​for corona charges with positive charge and negative values ​​for corona charges with negative charge. The absolute value of the constant term C, obtained from fitting calculations of the decay curves obtained from each measurement, was adopted as the variable y. From among the multiple damping curves obtained from multiple measurements, the damping curve obtained from the measurement with zero corona charges was selected, and a fitting calculation was performed using equation (10)' in the same way as described earlier for Example A, and τ eff The duration was calculated to be 5300 μs. For each of the multiple damping curves obtained from multiple measurements, the constant term C was calculated from the fitting calculation results. The above values ​​were used as the values ​​for the horizontal axis (x-axis) and the vertical axis (y-axis) and plotted. The minimum value on the vertical axis (the minimum value calculated by fitting a quadratic function to the average value obtained by measuring at 50 points in each orbit) and equation (11)' (the second term of equation (11) "τ eff / τ s From ×n0"), surface recombination lifetime τ S The value was calculated to be 115,000 μs (in this embodiment, n0 is 5 mV). Furthermore, from equation (1)', the SRH recombination lifetime (bulk lifetime) τ b The calculated duration was 5560 μs.

[0065] The calculation results are shown in Table 2, and the quadratic curve obtained in Example 1 is shown in Figure 9.

[0066] [Table 2]

[0067] Furthermore, the thickness d of the silicon wafer and the surface recombination lifetime τ are also relevant. S Therefore, the surface recombination rate S of the silicon wafer is given by equation (12). r The calculated value was 0.34 cm / s.

[0068] [Example 2] An example is shown using a μPCD device equipped with a corona charge processing unit and a non-contact CV device. First, a corona charge treatment was performed on an n-type silicon wafer (single-crystal silicon wafer). Next, the surface charge amount of the wafer surface was measured using a non-contact CV device, and after calculating the surface charge density, μPCD measurement was performed. These processes, measurements, and calculations were performed by varying the number of corona charges and the sign of the charge. From among the multiple damping curves obtained from multiple measurements, the damping curve obtained from the measurement with zero corona charges was selected, and a fitting calculation was performed using equation (10)' in the same way as described earlier for Example A, and τ eff The duration was calculated to be 5300 μs. For the measurement results obtained from multiple measurements, the surface charge density was plotted on the x-axis and the absolute value of the constant term on the y-axis. The minimum value on the y-axis (the minimum value calculated by fitting a quadratic function to the average value obtained from measurements at 50 points within each surface) and equation (11)' (the second term of equation (11) "τ eff / τ s From ×n0"), surface recombination lifetime τ S The value was calculated to be 121,000 μs (n0 in this example is 5 mV). Furthermore, from equation (1)', the SRH recombination lifetime (bulk lifetime) τ b The duration was calculated to be 5540 μs.

[0069] The calculation results are shown in Table 3, and the quadratic curve obtained in Example 2 is shown in Figure 10.

[0070] [Table 3]

[0071] Furthermore, the thickness d of the silicon wafer and the surface recombination lifetime τ are also relevant. S Therefore, the surface recombination rate S of the silicon wafer is given by equation (12). r The calculated value was 0.32 cm / s. [Industrial applicability]

[0072] One aspect of the present invention is useful in the technical field of various semiconductor wafers.

Claims

1. The process involves obtaining an attenuation curve by subjecting the semiconductor sample under evaluation to measurement using the photoconductivity attenuation method, and performing this process multiple times while varying the surface charge density. The signal data processing is performed on at least one of the decay curves obtained from the multiple measurements using a model equation that includes an exponential decay term and a constant term. From the exponential decay formula obtained by the signal data processing described above, the recombination lifetime τ of the semiconductor sample ｅｆｆ To seek, From the measurement results obtained through the aforementioned multiple measurements, a quadratic function is obtained in which the surface charge density-related value is the variable x and the value related to the constant term is the variable y, and The surface recombination lifetime τ of the semiconductor sample is obtained from the above quadratic function. ｓ To seek, A method for evaluating semiconductor samples, including the following.

2. The following formula (11)': C ｍｉｎ =t ｅｆｆ / t ｓ ×n ０ ・・・(11)' (In formula (11)', n ０ (Equilibrium carrier concentration) where C ｍｉｎ is taken as the minimum value of y in the quadratic function, and the surface recombination lifetime τ ｓ is calculated. The method for evaluating a semiconductor sample according to claim 1.

3. The aforementioned recombination lifetime τ ｅｆｆ and the surface recombination lifetime τ ｓ Therefore, the SRH recombination lifetime τ of the semiconductor sample ｂ To seek, The method for evaluating a semiconductor sample according to claim 1, further comprising:

4. The SRH recombination lifetime τ ｂ The following formula (1)': 1 / t ｅｆｆ =1 / t ｂ +1 / t ｓ ・・・(1)' A method for evaluating a semiconductor sample according to claim 3, obtained from the above.

5. The thickness d of the semiconductor sample and the surface recombination lifetime τ Ｓ Therefore, equation (12): S ｒ =d / (2τ Ｓ ) ・・・(12) The surface recombination rate S of the semiconductor sample is determined by ｒ To seek, The method for evaluating a semiconductor sample according to claim 1, further comprising:

6. By performing the aforementioned signal data processing, the constant term in the model equation is canceled out to obtain the exponential decay equation, and From the above exponential decay equation, the time constant τ ｂ －１ +τ Ｓ －１ To seek, It further includes, τ ｂ τ is the SRH recombination lifetime. Ｓ The method for evaluating a semiconductor sample according to claim 1, wherein is the surface recombination lifetime.

7. The method for evaluating a semiconductor sample according to claim 6, wherein the signal data processing includes repeatedly sampling and taking the difference of a time-series signal modeled by the model equation.

8. The method for evaluating a semiconductor sample according to claim 7, further comprising performing autoscaling to determine the sampling area for which the sampling is to be performed.

9. The semiconductor sample evaluation method according to claim 8, wherein the auto-scaling determines a region in which Auger recombination is less affected and noise is less affected as the sampling region.

10. The aforementioned model equation is the following equation (10)': xi(ti)=A×exp[-(τ ｂ －１ +τ Ｓ －１ )ti]-C ・・・(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ ｂ : SRH recombination lifetime, τ Ｓ : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample according to claim 1.

11. The following formula (11)': C ｍｉｎ =t ｅｆｆ / t ｓ ×n ０ ・・・(11)' (In formula (11)', n ０ (Equilibrium carrier concentration) In C ｍｉｎ Let τ be the minimum value of y in the quadratic function, and the surface recombination lifetime τ ｓ Calculate, The aforementioned recombination lifetime τ ｅｆｆ and the surface recombination lifetime τ ｓ Therefore, the SRH recombination lifetime τ of the semiconductor sample ｂ This further includes seeking, The SRH recombination lifetime τ ｂ The following formula (1)': 1 / t ｅｆｆ =1 / t ｂ +1 / t ｓ ・・・(1)' From, By performing the aforementioned signal data processing, the constant term in the model equation is canceled out to obtain the exponential decay equation, and From the above exponential decay equation, the time constant τ ｂ －１ +τ Ｓ －１ To seek, It further includes, τ ｂ τ is the SRH recombination lifetime. Ｓ This is the surface recombination lifetime, The signal data processing includes repeatedly sampling and taking the difference of the time-series signal modeled by the model equation, The process further includes performing autoscaling to determine the sampling area for which the sampling is performed, The auto-scaling described above determines the sampling region as one in which the effects of Auger recombination are minimal and the effects of noise are minimal. The aforementioned model equation is the following equation (10)': xi(ti)=A×exp[-(τ ｂ －１ +τ Ｓ －１ )ti]-C ・・・(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ ｂ : SRH recombination lifetime, τ Ｓ : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample according to claim 1.

12. The following formula (11)': C ｍｉｎ =t ｅｆｆ / t ｓ ×n ０ ・・・(11)' (In formula (11)', n ０ (Equilibrium carrier concentration) In C ｍｉｎ Let τ be the minimum value of y in the quadratic function, and the surface recombination lifetime τ ｓ Calculate, The thickness d of the semiconductor sample and the surface recombination lifetime τ Ｓ Therefore, equation (12): S ｒ =d / (2τ Ｓ ) ・・・(12) The surface recombination rate S of the semiconductor sample is determined by ｒ This further includes seeking, By performing the aforementioned signal data processing, the constant term in the model equation is canceled out to obtain the exponential decay equation, and From the above exponential decay equation, the time constant τ ｂ －１ +τ Ｓ －１ To seek, It further includes, τ ｂ τ is the SRH recombination lifetime. Ｓ This is the surface recombination lifetime, The signal data processing includes repeatedly sampling and taking the difference of the time-series signal modeled by the model equation, The process further includes performing autoscaling to determine the sampling area for which the sampling is performed, The auto-scaling described above determines the sampling region as one in which the effects of Auger recombination are minimal and the effects of noise are minimal. The aforementioned model equation is the following equation (10)': xi(ti)=A×exp[-(τ ｂ －１ +τ Ｓ －１ )ti]-C ・・・(10)' (In equation (10)', ti: elapsed time after excitation light irradiation, xi(ti): signal intensity at elapsed time ti, τ ｂ : SRH recombination lifetime, τ Ｓ : Surface recombination lifetime, A, C: constants) The method for evaluating a semiconductor sample according to claim 1.

13. A semiconductor sample evaluation apparatus for carrying out the semiconductor sample evaluation method described in any one of claims 1 to 12, A measurement unit that subjects the semiconductor sample to be evaluated to measurement using the photoconductivity attenuation method, A processing unit that performs signal data processing on the aforementioned decay curve using a model equation that includes an exponential decay term and a constant term, and calculates the aforementioned quadratic function, The evaluation apparatus, including the evaluation apparatus.

14. The evaluation apparatus according to claim 13, wherein the light pulse intensity is variable in the measurement unit.

15. The evaluation apparatus according to claim 13, wherein the measurement unit further includes a corona charge processing unit.

16. The evaluation apparatus according to claim 13, wherein the measurement unit further includes a surface charge density related value measurement unit for measuring surface charge density related values ​​of a semiconductor sample subjected to the measurement.

17. In the measurement unit, the light pulse intensity is variable. The evaluation apparatus according to claim 13, wherein the measurement unit further includes a corona charge processing unit and a surface charge density related value measurement unit for measuring surface charge density related values ​​of a semiconductor sample subjected to the measurement.

18. Manufacturing a semiconductor wafer lot containing multiple semiconductor wafers, Extracting at least one semiconductor wafer from the aforementioned semiconductor wafer lot, The extracted semiconductor wafer is evaluated by the evaluation method described in any one of claims 1 to 12, and Based on the results of the above evaluation, semiconductor wafers from the same semiconductor wafer lot as those determined to be good products will be prepared for shipment as product semiconductor wafers. A method for manufacturing semiconductor wafers containing [the specified material].

19. To manufacture evaluation semiconductor wafers under test manufacturing conditions, The manufactured evaluation semiconductor wafer is evaluated by the semiconductor sample evaluation method described in any one of claims 1 to 12. Based on the results of the evaluation, either the manufacturing conditions modified from the test manufacturing conditions are determined as the actual manufacturing conditions, or the test manufacturing conditions are determined as the actual manufacturing conditions, and To manufacture semiconductor wafers under the actual manufacturing conditions determined above, A method for manufacturing semiconductor wafers containing [the specified material].

Images