Plasma processing / control device and plasma processing / control method

Abstract

The purpose of the present invention is to estimate, with high accuracy, the thickness of a processing object to be subjected to plasma etching processing, even when a spectrum pattern measured during actual processing deviates from a spectrum pattern prepared in advance with respect to the processing object.　To this end, a plasma processing / control device according to the present invention compares a measurement spectrum obtained by receiving plasma reflected light from the interior of a processing chamber at a discretionary time during etching with an estimation spectrum acquired in advance in association with the thickness of the same processing object, calculates spectrum errors between the measurement spectrum and a plurality of estimation spectra, and estimates the thickness of the processing object at the discretionary time using at least an estimation spectrum having a first minimum value and an estimation spectrum having a second minimum value among the plurality of spectrum errors.

Description

Plasma processing control device and plasma processing control method

[0001] The present invention relates to a plasma processing control device and a plasma processing control method.

[0002] In semiconductor device manufacturing, plasma-based dry etching is widely used as a process for removing unwanted portions. If the processed shape during this etching process deviates from the design, the resulting semiconductor will not achieve the desired performance, necessitating process monitoring technology to monitor and stabilize the etching process. Process monitors that measure the remaining film thickness and pattern depth on a wafer by measuring the light reflected from the wafer during processing are called film thickness / depth monitors and have been used to determine the endpoint of the etching process. In Patent Document 1, to monitor the film thickness and depth during the etching process, a spectrometer detects the spectrum of plasma light reflected from the surface of the wafer being etched. Using pattern recognition techniques such as principal component analysis, the detected spectral data is compared with principal component analysis data of spectral data obtained from previous etching processes, and the endpoint of the plasma etching process is determined based on the obtained current film thickness information. In Patent Document 2, standard deviation patterns corresponding to multiple thicknesses are prepared using pre-prepared wafers, and the deviation is calculated by comparing the interference light obtained during etching of the target wafer with the standard deviation pattern. The film thickness during the etching process is recorded from this deviation, and the endpoint is determined when a predetermined etching depth is reached. Furthermore, in order to deal with fluctuations in the interference light during etching, if the deviation does not satisfy a predetermined condition, the film thickness is corrected using linear regression approximation or the like.

[0003] JP 2001-244254 A JP 2007-234666 A

[0004] As semiconductor devices continue to become increasingly miniaturized in recent years, new miniaturization processes are becoming increasingly common. Furthermore, new materials and structures are enabling increasingly higher performance. Therefore, the development of mass-production stabilization technologies for these new device manufacturing processes is essential, and there is a growing need for highly accurate monitoring of film thickness and depth. However, both Patent Documents 1 and 2 monitor the etching depth by comparing a pre-prepared spectral pattern with a measured spectral pattern of the wafer being processed. However, differences in the timing of spectral detection can result in incomplete matching of the spectra. This can lead to discrepancies in the calculated etching depth. Furthermore, even if the etching conditions are the same, some factors can cause changes in the environment within the processing chamber, which can alter the etching rate. In this case, the spectral shape of the pre-prepared spectral pattern and the measured spectral pattern of the wafer being processed differs over time, resulting in a change in the calculated etching depth. Therefore, the present invention aims to provide a technology that can accurately estimate the thickness of a processing object and determine the end point of the processing, even when the pre-prepared spectral pattern and the measured spectral pattern during the actual processing differ due to differences in detection timing, etc.

[0005] In order to solve the above-mentioned problems, one representative plasma processing control apparatus of the present invention is a plasma processing control apparatus that etches a processing object placed in a processing chamber inside a vacuum vessel to a predetermined thickness using plasma formed in the processing chamber, and compares a measurement spectrum obtained by receiving plasma reflected light from within the processing chamber at an arbitrary time during etching with an estimation spectrum that has been previously obtained for the same processing object as the processing object and associated with its thickness, calculates a spectral error between the measurement spectrum and multiple estimation spectra, and estimates the thickness of the processing object at the arbitrary time using an estimation spectrum (referred to as a "first minimum spectrum") that has at least the first smallest value (first minimum) and an estimation spectrum (referred to as a "second minimum spectrum") that has the second smallest value (second minimum) among the multiple spectral errors.

[0006] According to the present invention, even if a spectral pattern prepared in advance for a processing object to be subjected to plasma etching processing and a spectral pattern measured during the actual processing deviate from each other due to differences in detection timing, etc., it is possible to estimate the thickness of the processing object with high accuracy and determine the end point of the processing. Problems, configurations, and effects other than those described above will become clear from the description of the following embodiments.

[0007] FIG. 1 is a schematic diagram showing an example of the configuration of a plasma processing and control device according to this embodiment. FIG. 2 is a graph showing an example of a spectrum obtained by detecting plasma reflected light. FIG. 3 is a schematic diagram showing an example of the configuration of a thickness estimation and determination unit according to this embodiment. FIG. 4 is an example of a data list used to determine the thickness of a processing object using conventional spectrum matching. FIG. 5 is a graph showing the time dependence of film thickness error that occurs in conventional spectrum matching. FIG. 6 is an example of a data list of spectral error for each estimation spectrum relative to a predetermined measurement spectrum. FIG. 7 is a graph showing the distribution relationship between film thickness and spectral error. FIG. 8 is a graph explaining a method for estimating thickness from the mixture ratio of a composite spectrum according to Example 1. FIG. 9 is a flowchart of a processing procedure for performing thickness estimation according to Example 1. FIG. 10 is a graph showing the time dependence of film thickness error that occurs when estimating thickness according to Example 1. FIG. 11 is a graph explaining a method for estimating thickness from an approximation curve according to Example 2-1. FIG. 12 is a flowchart of a processing procedure for performing thickness estimation according to Example 2-1. FIG. 13 is a graph illustrating a method for estimating thickness from resampling processing in Example 2-2. FIG. 14 is a flowchart of a processing procedure for thickness estimation in Example 2-2. FIG. 15 is a graph showing outliers (white circles) and an approximation curve generated by excluding the outliers in Modification 1 of Example 2-1. FIG. 16 is a graph showing outliers (white circles) and an approximation curve (solid line) generated by excluding the outliers in Modification 2 of Example 2-1. FIG. 17 is an example of a data list in Example 3 in which film thickness and estimation spectrum data are supplemented by interpolation.

[0008] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiment. In addition, in the description of the drawings, the same parts are designated by the same reference numerals.

[0009] <Configuration of Plasma Processing and Control Device> FIG. 1 is a schematic diagram showing an example configuration of a plasma processing and control device according to this embodiment. The plasma processing device 10 includes a processing chamber 11 capable of achieving a vacuum. An etching gas introduced into the processing chamber 11 through a gas introduction means (not shown) is excited and decomposed by power or microwaves generated by a high-frequency power source (not shown) or the like to form plasma 12. This plasma 12 etches a processing object 14, such as a semiconductor wafer, placed on a sample stage 13 inside the processing chamber 11. The processing object 14 is not particularly limited as long as it is an object to be etched, but an example of such an object is an etching film (including a laminate) formed on a semiconductor wafer. In this disclosure, the thickness of the processing object may also be referred to as the film thickness of the processing object. The introduction of gas into the processing chamber 11, the generation and control of plasma 12, and the application of voltage to the processing object 14 are controlled by a control unit 23 (described below), which synchronizes and adjusts the timing between the various components to achieve the desired etching process.

[0010] In a plasma processing apparatus 10, plasma light emitted by a plasma 12 during etching is irradiated onto a processing object 14. The irradiated light interferes and is reflected depending on the surface structure of the processing object. Reflected plasma light 15 emitted from the processing chamber is collected by a detection lens 16 and detected by a detection unit 17 via an optical fiber or the like. The detection unit disperses the reflected plasma light 15 into a predetermined wavelength band using a spectroscope, and the amount of light (light intensity) at each wavelength is detected to obtain a spectrum with wavelength as a parameter (also referred to as an "optical spectrum" or "spectral pattern"). The detected spectrum has different spectral patterns depending on the film thickness. Note that the spectrum may use wavenumber instead of wavelength as a parameter. FIG. 2 is a graph schematically illustrating an example of a spectrum obtained by detecting reflected plasma light. FIG. 2(a) is a contour plot in which the horizontal axis represents the film thickness (nm) of the processing object film and the vertical axis represents the wavelength (nm) of the processing object film, with the spectral intensity (light intensity) divided into patterns. FIG. 2(b) is a graph showing the wavelength-dependent spectral pattern when FIG. 2(a) is cut at a certain film thickness. The spectral intensity is, for example, the amount of light received per unit area by the detection unit 17 (W / cm ２ ), but this is not limiting, and other measurement methods can be used, such as converting the total amount of detected light into the number of electrons. In this way, the correlation between film thickness and spectral pattern can be confirmed.

[0011] The optical spectrum detected by the detection unit 17 is processed by the signal processing unit 20 into a form that makes it easy to monitor the spectrum. Specifically, processing such as light intensity offset and high-frequency noise removal is performed. The signal-processed optical spectrum is used as an index for determining the film thickness and depth of the film to be processed by the thickness estimation / determination unit 21, and values ​​of the film thickness, depth, and etching amount corresponding to the optical spectrum are estimated. The values ​​estimated by the thickness estimation / determination unit 21 are not particularly limited as long as they directly or indirectly indicate the dimensions of the film to be processed 14 in the thickness direction (direction perpendicular to the horizontal plane of the upper surface of the sample stage 13), and may include the thickness (i.e., remaining thickness), depth, and etching amount of the film to be processed. Unless otherwise specified in this disclosure, the thickness (remaining thickness), depth, and etching amount of the film to be processed are collectively referred to as "thickness." The estimated thickness value is displayed on a display unit 22, such as a display or printer. Furthermore, if the value estimated by the thickness estimation / determination unit 21 is determined to be the target etching film thickness, the control unit 23 terminates the etching process of the film to be processed 14.

[0012] The plasma processing / control device 30 includes a signal processing unit 20, a thickness estimation / determination unit 21, and a control unit 23 in addition to the plasma processing device 10.

[0013] The thickness estimation / determination unit 21 will be further described. FIG. 3 is a schematic diagram showing an example of the configuration of the thickness estimation / determination unit in this embodiment. Plasma reflected light 15 from the processing object whose thickness is to be measured is transmitted to the thickness estimation / determination unit 21 via the detection unit 17 and the signal processing unit 20 as a signal of a wavelength-dependent spectrum (referred to as the "measured spectrum"). The comparison / calculation unit 25 compares the optical spectra as described below, performs calculations to estimate the thickness of the processing object, and determines whether the target etching film thickness has been reached. The optical spectra to be compared are stored in the database unit 24. The database unit 24 stores multiple optical spectra (collectively referred to as "estimation spectra") by preparing, for example, a test wafer identical to the semiconductor wafer serving as the processing object 14 and etching the wafer under the same conditions in advance. Furthermore, by measuring the film thickness of the test wafer before and after etching, the optical spectra stored in the database unit 24 are associated with the film thickness. It is also possible to prepare multiple test wafers to be measured in advance, create a data library of optical spectra obtained under different etching conditions (recipes), and add this to the database unit 24. The thickness data of the object to be processed estimated by the comparison / calculation unit 25 or the judgment data based on the thickness is sent to the display unit 22 and the control unit 23 .

[0014] In the comparison / calculation unit 25, program processing is executed by a processor such as a CPU. The database unit 24 may be a random access semiconductor memory, a storage device, or a storage medium (either volatile or nonvolatile) for storing data or programs (applications). The processor and memory of the thickness estimation / determination unit 21 may be installed as standalone units, or may be integrated with the processors and memories of other functional units included in the plasma processing / control device, such as the signal processing unit 20 and the control unit 23, and designed to execute each function by software. These functional units are connected by wire or wirelessly.

[0015] <Conventional Spectral Matching> First, a conventional method for determining thickness by spectral matching will be described. Fig. 4 shows an example of a data list used when determining the thickness of an object to be processed by conventional spectral matching. Fig. 4(a) shows a part of a data list library of estimation spectra stored in the database unit 24, and shows time T marked at detection timing ΔT intervals. ｎ FIG. 4B shows a data list of measurement spectra during actual measurement, showing the measurement spectrum (I') and actual film thickness data around time T'j, spaced at detection timing ΔT' intervals. FIG. 4C shows the measurement spectrum and estimation spectrum for which spectral matching is established, the film thickness determined by matching, and the film thickness error at that time. However, data for time T or T' is not required. The spectral matching method is not particularly limited. For example, when searching for an estimation spectrum that matches the measurement spectrum I'j (see FIG. 4B) at actual measurement time T'j, the absolute value of the difference or the sum of the squared differences by wavelength between all estimation spectra (see FIG. 4A) listed in the data list library is compared. If the smallest value is obtained, a match is established (In in the example of FIG. 4C), and the film thickness Dn of the matching estimation spectrum In is determined to be the film thickness at time T'j.

[0016] However, the certified film thickness in Figure 4(c) will have an error compared to the actual film thickness. When the actual film thickness at the time of actual measurement is expressed as in Figure 4(b), the film thickness error can be expressed schematically as ΔD as in Figure 4(c). Figure 5 is a graph that shows a schematic representation of the time dependency of film thickness error that occurs in conventional spectrum matching. It can be seen that the film thickness error changes in a step-like manner as the detection timing differs.

[0017] <Correlation Estimation in an Embodiment> Next, a method for estimating thickness in an embodiment of the present invention will be described. In this embodiment, when estimating thickness when the measured spectrum is I'j(λ), the thickness estimation / determination unit 21 calculates a spectral error ΔSn between I'j(λ) and the estimation spectrum In(λ) stored in the database unit 24. ΔSn is calculated, for example, using the following equation (1): However, the definition of the spectral error is not limited to this as long as it is useful as an index. For example, it may be the sum of the absolute values ​​of the differences by wavelength. It may also be the square root or logarithm of these.

[0018] FIG. 6 shows an example of a data list of spectral errors for each estimation spectrum relative to a given measured spectrum. The data list shown in FIG. 6 is temporarily stored in the database unit 24, from which the distribution relationship between the film thickness and the spectral error corresponding to each estimation spectrum can be obtained. FIG. 7 is a graph showing the distribution relationship between film thickness and spectral error. In the example shown in FIG. 7, the spectral error ΔSn with the estimation spectrum In(λ) is minimum when the film thickness is Dn, and the spectral error increases as the film thickness increases (or decreases). In this embodiment, the correlation between the film thickness and the spectral error is estimated from the discrete distribution relationship between the film thickness and the spectral error, and ΔSmin at which the spectral error is minimum under this correlation is determined. The thickness Dmin at this point is estimated as the thickness for the measured spectrum I'j(λ). A specific estimation method will be described using the following example.

[0019] Example 1 In Example 1, thickness is estimated from the spectral error between a measurement spectrum and a synthetic spectrum obtained by synthesizing an estimation spectrum while taking into account the mixing ratio. Fig. 8 is a graph illustrating a method for estimating thickness from the mixing ratio of a synthetic spectrum in Example 1. Under the distribution relationship shown in Fig. 7, an estimation spectrum In(λ) in which the spectral error between the measurement spectrum and the estimation spectrum is the first minimum value (first minimum value) ΔSn and an estimation spectrum In+1(λ) in which the spectral error between the measurement spectrum and the estimation spectrum is the second minimum value (second minimum value) ΔSn+1 are synthesized at a mixing ratio α, and the synthetic spectrum is calculated using the following equation (2). Then, the spectral error between the synthetic spectrum and the measured spectrum I'j(λ) is expressed by the following equation (3), and it can be expected to have a spectral error ΔSmin smaller than the first minimum value ΔSn and the second minimum value ΔSn+1. Furthermore, ΔSmin is differentiated by α using the following equation (4) to obtain the minimum value α ０ Ask for. Therefore, the minimum value α ０ Then, the thickness Dmin is calculated by dividing the thickness Dn of the estimation spectrum In(λ) and the thickness Dn+1 of the estimation spectrum In+1(λ) internally, and this is estimated as the thickness of the measurement spectrum I'j(λ). This allows for highly accurate thickness estimation corresponding to a composite spectral pattern with a smaller spectral error than the first minimum value.

[0020] A processing procedure for estimating the thickness of a processing object in Example 1 will be described. FIG. 9 is a flowchart of the processing procedure for thickness estimation in Example 1. First, etching of a wafer to be evaluated is initiated, and the thickness estimation process also starts (S101). Next, the detector 17 measures the spectrum of the plasma reflected light 15 from the processing chamber 11 (S102). Next, the signal of the measured spectrum is processed for monitoring by the signal processor 20 (S103). Next, the thickness estimation / determination unit 21 calculates the spectral error ΔS between the measured spectrum and the estimation spectrum to obtain data on the distribution relationship between the spectral error and thickness, as shown in FIG. 7 (S104). Next, the thickness estimation / determination unit 21 synthesizes a composite spectrum with a mixture ratio α from the estimation spectra with the first and second minimum spectral errors, and calculates an estimated film thickness value Dmin from the minimum value of α (S105). Next, the thickness estimation / determination unit 21 determines whether the estimated film thickness has reached the target film thickness (S106). If not, the process returns to S102 and repeats the thickness estimation process using a new measured spectrum. If it is determined in S106 that the target film thickness has been reached, the control unit 23 stops the etching process and also ends the thickness estimation process (S107).

[0021] 10 is a graph showing the time dependence of the film thickness error that occurs when the thickness is estimated in Example 1. The solid line shows the time dependence of Example 1, and the dotted line shows the time dependence of the conventional example (see FIG. 5). As can be seen from FIG. 10, compared to when the film thickness is determined by conventional spectrum matching, the film thickness error is generally reduced when the minimum thickness value is estimated by adjusting the mixture ratio from the composite spectrum of the estimation spectra that have the first and second minimum values.

[0022] [Example 2-1] In Example 2-1, an approximation curve is calculated based on the distribution relationship between the spectral error and film thickness for at least the first to third minimum values ​​(three minimum points), and the thickness is estimated. Figure 11 is a graph illustrating a method for estimating thickness from the approximation curve in Example 2-1. Figure 11 shows an example in which an approximation curve F(D) is generated based on the distribution relationship between the spectral error and film thickness for six minimum values ​​(six minimum points), namely the first to sixth minimum values. The approximation curve F(D) is not particularly limited, and a quadratic or higher N-th order function is used, which is automatically generated based on parameters preset based on conditions such as the plasma etching process conditions (recipe), the thickness (film thickness) of the workpiece, and the required estimation accuracy. Another possible method is to generate and update the approximation curve using machine learning, using past data as learning data. From the approximation curve F(D), Dmin corresponding to ΔSmin, where the spectral error is at its minimum, is calculated, and Dmin is estimated as the thickness of the measured spectrum I'j(λ). This achieves highly accurate thickness estimation corresponding to a spectral pattern with even smaller spectral error than the first minimum value.

[0023] A processing procedure for estimating the thickness of a processing object in Example 2-1 will be described. FIG. 12 is a flowchart of the processing procedure for thickness estimation in Example 2-1. First, etching of the wafer to be evaluated is initiated, and the thickness estimation process also begins (S201). Next, the detector 17 measures the spectrum of the plasma reflected light 15 from the processing chamber 11 (S202). Next, the signal of the measured spectrum is processed for monitoring by the signal processor 20 (S203). Next, the thickness estimation / determination unit 21 calculates the spectral error ΔS between the measured spectrum and the estimation spectrum to obtain data on the distribution relationship between the spectral error and thickness, as shown in FIG. 7 (S204). Next, the thickness estimation / determination unit 21 performs a curve approximation of the spectral error at the minimum N points, and calculates the minimum spectral error ΔSmin and the associated estimated film thickness value Dmin (S205). Next, the thickness estimation / determination unit 21 determines whether the estimated film thickness has reached the target film thickness (S206). If not, the process returns to S202 and repeats the thickness estimation process using a new measured spectrum. If it is determined in S206 that the target film thickness has been reached, the control unit 23 stops the etching process and also ends the thickness estimation process (S207).

[0024] [Example 2-2] In Example 2-2, thickness is estimated by resampling based on the distribution relationship between the spectral error and film thickness for at least the first to third minimum values ​​(the three minimum points). Figure 13 is a graph illustrating a method for estimating thickness from the resampling process in Example 2-2. In Figure 13, new data points (white circles) are added between the black circle data points by resampling in the distribution relationship between the spectral error and film thickness for six data points (the first to sixth minimum values ​​(the six minimum points)). In the example of Figure 13, five times as many data points as the black circle data points are added by resampling. The resampling method is not particularly limited, but the number and positions of resampling points are automatically determined and generated according to parameters preset based on conditions such as the plasma etching process conditions (recipe), the thickness of the workpiece (film thickness), and the required estimation accuracy. The number of data points added by resampling may be determined from the perspective of computational cost. A method of generating and updating resampling data using machine learning using past data as learning data is also conceivable. From the resampled data points, ΔSmin and the corresponding Dmin are found where the spectral error is at its minimum, and Dmin is estimated as the thickness of the measured spectrum I'j(λ). This allows for a highly accurate thickness estimation corresponding to a spectral pattern with a spectral error even smaller than the first minimum value. Note that while interpolation using a curve (Example 2-1) is more detailed than interpolation using points (Example 2-2), it is characterized by the fact that the interpolation is constrained by the curve shape. Once the distribution relationship between the spectral error and the film thickness is obtained, it is possible to appropriately estimate ΔSmin and ΔDmin using curve approximation (Example 2-1) or resampling processing (Example 2-2) depending on the characteristics and distribution conditions of each.

[0025] A processing procedure for estimating the thickness of a processing object in Example 2-2 will be described. FIG. 14 is a flowchart of the processing procedure for thickness estimation in Example 2-2. First, etching of the wafer to be evaluated is initiated, and the thickness estimation process also begins (S301). Next, the detector 17 measures the spectrum of the plasma reflected light 15 from the processing chamber 11 (S302). Next, the signal of the measured spectrum is processed for monitoring by the signal processor 20 (S303). Next, the thickness estimation / determination unit 21 calculates the spectral error ΔS between the measured spectrum and the estimation spectrum to obtain data on the distribution relationship between the spectral error and thickness shown in FIG. 7 (S304). Next, the thickness estimation / determination unit 21 performs a resampling process from the spectral error at the minimum N points to calculate the minimum spectral error ΔSmin and the associated estimated film thickness value Dmin (S305). Next, the thickness estimation / determination unit 21 determines whether the estimated film thickness has reached the target film thickness (S306). If not, the process returns to S302 and repeats the thickness estimation process using a new measured spectrum. If it is determined in S306 that the target film thickness has been reached, the control unit 23 stops the etching process and also ends the thickness estimation process (S307).

[0026] [Variation 1 of the Example] In a plasma processing apparatus, the environment inside the processing chamber may become temporarily unstable due to some factor, causing the spectral pattern to deform. In such cases, outliers may occur in the spectral error. FIG. 15 is a graph showing outliers (white circles) and an approximation curve generated by excluding the outliers in Variation 1 of Example 2-1. FIG. 15 shows an example in which the spectral error ΔSn is small due to coincidental similarity of the spectral patterns. In Variation 1, outliers are determined, for example, by sequentially calculating the rate of change (differential) of the spectral error and determining that the value is an outlier when it reverses positive and negative or when it jumps discontinuously. In each example, spectral errors determined to be outliers in Variation 1 are excluded before performing the synthesis of the composite spectrum, generation of the approximation curve, or resampling process.

[0027] [Variation 2 of the Example] As in FIG. 15 , FIG. 16 is a graph showing outliers (white circles) and an approximation curve (solid line) generated by excluding the outliers in Variation 2 of Example 2-1. FIG. 16 illustrates an example in which a significant spectral error ΔSn-1 occurs due to a coincidental deformation of the spectral pattern. In Variation 2, the outliers are determined by calculating the fitting error between an approximation curve F(D) (solid line) generated by excluding the spectral error ΔS from ΔSn-4 in order and an approximation curve F'(D) (dotted line) generated by including the spectral error ΔS, and determining the ΔS (ΔSn-1 in FIG. 16 ) with a large fitting error as an outlier. The fitting error can be obtained, for example, by summing the absolute value of the difference between F(D) and F'(D) or the square of the difference by the film thickness D. Furthermore, significantly outliers may be excluded by the determination in Variation 1, and for those requiring more precise determination, Variation 2 may be applied to determine whether the fitting error exceeds a threshold value. In each embodiment, the spectral errors determined to be outliers in the second modification are excluded before synthesis of the synthetic spectrum, generation of the approximation curve, or resampling processing.

[0028] [Example 3] The more film thickness and estimation spectrum data included in the data list for estimation spectra stored in the database unit 24, the more accurate the thickness estimation (the smaller the film thickness error). However, preparing the same test object in advance and measuring the film thickness and spectrum is labor-intensive and cost-intensive. Therefore, in Example 3, data is supplemented by interpolation from existing data. Figure 17 shows an example of a data list in Example 3 in which film thickness and estimation spectrum data are supplemented by interpolation. Figure 17(a) shows a list of film thickness and estimation spectrum data (measured estimation data) previously measured and stored in the database unit 24 (see Figure 4(a)). Figure 17(b) shows a list supplemented with data (interpolated estimation data) calculated by interpolating film thickness and estimation spectrum from the measured estimation data. In the example of Figure 17, three points of data are added by interpolation between each measured estimation data. 17(b) shows an example of linearly interpolating the film thickness, but the estimation spectrum can also be obtained by linearly interpolating the light quantity (light intensity) of the spectrum for each wavelength. The estimation data list created in FIG. 17(b) is used to estimate the thickness of the object to be processed according to the first and second embodiments. This increases the amount of data, which is expected to improve the estimation accuracy of each embodiment.

[0029] <Functions and Effects> As described above, in this embodiment, even when the measured spectrum deviates from a previously prepared estimation spectrum, the film thickness of the measured spectrum is not simply determined based on the film thickness of the estimation spectrum that matches, but rather a correlation is found from the distribution relationship of the spectral errors, and a film thickness that minimizes the spectral error is estimated, thereby enabling accurate thickness estimation and further enabling accurate determination of the end of the etching process. Furthermore, since correlation estimation does not require changes to the hardware of the plasma processing apparatus, flexible design changes and adjustments can be made, since changes to the estimation method and additions to the estimation data list by interpolation can be realized by software.

[0030] Although the embodiments of the present invention have been described above, it goes without saying that the present invention is not limited to the above-described embodiments and can be modified in various ways without departing from the spirit of the present invention. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0031] The present invention is not limited to the following embodiments: (Aspect 1) A plasma processing and control device for etching a processing object placed in a processing chamber inside a vacuum vessel to a predetermined thickness using plasma formed in the processing chamber, the plasma processing and control device comprising: a thickness estimation and determination unit that compares a measurement spectrum obtained by receiving plasma reflected light from within the processing chamber at a given time during etching with estimation spectra previously obtained for the same processing object and associated with the thickness of the object, calculates a spectral error between the measurement spectrum and a plurality of the estimation spectra, and estimates the thickness of the processing object at the given time using an estimation spectrum (referred to as a "first minimum spectrum") having at least a first minimum value (first minimum value) and an estimation spectrum (referred to as a "second minimum spectrum") having a second minimum value (second minimum value) among the plurality of spectral errors. (Aspect 2) In the plasma processing control device according to Aspect 1, the thickness estimation and determination unit estimates the thickness of the object at the given time using a spectral error calculated between the measured spectrum and a composite spectrum obtained by combining the first and second minimum spectra while taking into account a mixing ratio. (Aspect 3) In the plasma processing control device according to Aspect 1, the thickness of the object at the given time is estimated using an estimation spectrum (referred to as a "third minimum spectrum") that has at least a third minimum value (third minimum value) among the plurality of spectral errors, and using an approximation curve automatically generated from at least the first to third minimum values ​​and the thicknesses of the object associated with the first to third minimum spectra. (Aspect 4) The plasma processing / control device according to Aspect 1, wherein an estimation spectrum (referred to as the "third minimum spectrum") that is at least the third smallest value (third minimum value) among the plurality of spectral errors is used, and the thickness of the processing object at the arbitrary time is estimated using a resampling process from at least the first to third smallest values ​​and thicknesses of the processing object associated with the first to third smallest spectra.(Aspect 5) The plasma processing and control device according to any one of Aspects 1 to 4, wherein the spectral error is the sum of squares of the difference in light intensity between the compared spectra over wavelength. (Aspect 6) A plasma processing and control method for etching a processing object placed in a processing chamber inside a vacuum vessel to a predetermined thickness using plasma formed in the processing chamber, wherein a thickness estimation and determination unit compares a measurement spectrum obtained by receiving plasma reflected light from within the processing chamber at a given time during etching with an estimation spectrum of the same processing object as the processing object that has been previously acquired and associated with a thickness of the processing object, to calculate a spectral error between the measurement spectrum and the multiple estimation spectra, and estimates the thickness of the processing object at the given time using an estimation spectrum (referred to as the "first minimum spectrum") that has at least a first minimum value (first minimum value) and an estimation spectrum (referred to as the "second minimum spectrum") that has a second minimum value (second minimum value) among the multiple spectral errors. (Aspect 7) The plasma processing and control method according to Aspect 6, wherein the thickness estimation and determination unit estimates the thickness of the object at the arbitrary time using a spectral error calculated between the measured spectrum and a composite spectrum obtained by combining the first and second minimum spectra while taking into account a mixing ratio. (Aspect 8) The plasma processing and control method according to Aspect 6, wherein the thickness of the object at the arbitrary time is estimated using an estimation spectrum (referred to as the "third minimum spectrum") that has at least a third minimum value (third minimum value) among the plurality of spectral errors, and using an approximation curve automatically generated from at least the first to third minimum values ​​and the thicknesses of the object associated with the first to third minimum spectra.(Aspect 9) The plasma processing and control method according to Aspect 6, wherein an estimation spectrum (referred to as the "third minimum spectrum") that has at least a third minimum value (third minimum value) among the plurality of spectral errors is used to estimate the thickness of the object to be processed at the arbitrary time by performing a resampling process from at least the first to third minimum values ​​and thicknesses of the object to be processed associated with the first to third minimum spectra. (Aspect 10) The plasma processing and control method according to any one of Aspects 6 to 9, wherein the spectral error is the sum of squares of the difference in light intensity between the spectra to be compared, by wavelength.

[0032] 10: Plasma processing apparatus 11: Processing chamber 12: Plasma 13: Sample stage 14: Processing object 15: Plasma reflected light 16: Detection lens 17: Detection unit 20: Signal processing unit 21: Thickness estimation and determination unit 22: Display unit 23: Control unit 24: Database unit 25: Comparison and calculation unit 30: Plasma processing and control device

Claims

1. A plasma processing and control device that etches an object placed in a processing chamber inside a vacuum vessel to a predetermined thickness using plasma formed in the processing chamber, the plasma processing and control device comprising: a measurement spectrum obtained by receiving plasma reflected light from inside the processing chamber at any time during etching, which is compared with an estimation spectrum previously obtained for the same object as the object to be processed and associated with its thickness; a thickness estimation and determination unit that calculates a spectral error between the measurement spectrum and multiple estimation spectra; and a thickness estimation and determination unit that estimates the thickness of the object to be processed at the any time using an estimation spectrum (referred to as the "first minimum spectrum") that has at least the first smallest value (first minimum value) and an estimation spectrum (referred to as the "second minimum spectrum") that has the second smallest value (second minimum value) among the multiple spectral errors.

2. A plasma processing and control device according to claim 1, wherein the thickness estimation and determination unit estimates the thickness of the object to be processed at the given time using a spectral error calculated between the measured spectrum and a composite spectrum obtained by combining the first minimum spectrum and the second minimum spectrum while taking into account the mixing ratio.

3. A plasma processing and control device according to claim 1, wherein an estimation spectrum (referred to as the "third minimum spectrum") that is at least the third smallest value (third minimum value) among the plurality of spectral errors is used to estimate the thickness of the object to be processed at the arbitrary time using an approximation curve automatically generated from at least the first to third smallest values ​​and the thicknesses of the object to be processed that correspond to the first to third smallest spectra.

4. A plasma processing and control device according to claim 1, wherein an estimation spectrum (referred to as the "third minimum spectrum") that is at least the third smallest value (third minimum value) among the plurality of spectral errors is used, and the thickness of the object to be processed at the arbitrary time is estimated using a resampling process from at least the first to third smallest values ​​and the thicknesses of the object to be processed associated with the first to third smallest spectra.

5. A plasma processing and control device according to any one of claims 1 to 4, wherein the spectral error is the sum of the squares of the differences in light intensity between the spectra being compared, by wavelength.

6. A plasma processing and control method for etching an object placed in a processing chamber inside a vacuum vessel to a predetermined thickness using plasma formed in the processing chamber, wherein a thickness estimation and determination unit compares a measurement spectrum obtained by receiving plasma reflected light from inside the processing chamber at any time during etching with an estimation spectrum of the same object as the object being processed that has been previously obtained and associated with its thickness, calculates a spectral error between the measurement spectrum and multiple estimation spectra, and estimates the thickness of the object at the any time using an estimation spectrum (referred to as the "first minimum spectrum") that has at least the first minimum value (first minimum value) and an estimation spectrum (referred to as the "second minimum spectrum") that has the second minimum value (second minimum value) among the multiple spectral errors.

7. A plasma processing and control method according to claim 6, wherein the thickness estimation and determination unit estimates the thickness of the object to be processed at the arbitrary time using a spectral error calculated between the measured spectrum and a synthesized spectrum obtained by synthesizing the first minimum spectrum and the second minimum spectrum while taking into account the mixing ratio.

8. A plasma processing and control method according to claim 6, wherein an estimation spectrum (referred to as the "third minimum spectrum") that is at least the third smallest value (third minimum value) among the plurality of spectral errors is used, and the thickness of the object to be processed at the arbitrary time is estimated using an approximation curve automatically generated from at least the first to third smallest values ​​and the thicknesses of the object to be processed that correspond to the first to third smallest spectra.

9. A plasma processing and control method according to claim 6, wherein an estimation spectrum (referred to as the "third minimum spectrum") that is at least the third smallest value (third minimum value) among the plurality of spectral errors is used, and the thickness of the object to be processed at the arbitrary time is estimated by resampling from at least the first to third smallest values ​​and the thicknesses of the object to be processed associated with the first to third smallest spectra.

10. A plasma processing and control method according to any one of claims 6 to 9, wherein the spectral error is the sum of the squares of the differences in light intensity between the spectra being compared, by wavelength.

Images