JP5411215B2 - Plasma processing equipment - Google Patents

Description

  The present invention relates to a film thickness and etching depth measurement method for detecting an etching amount of a material to be processed in the manufacture of a semiconductor integrated circuit or the like by an emission spectroscopy method, and a method for processing a material to be processed using the same. Depth and film thickness measuring method and apparatus suitable for accurately measuring the etching amount of various layers provided on the substrate by the etching process used to obtain a desired film thickness and etching depth, and using the same The present invention relates to a processing method and apparatus for a processed material.

  In the manufacture of semiconductor wafers, dry etching is widely used for the removal or patterning of various material layers and especially dielectric material layers formed on the surface of the wafer. The most important for process parameter control is to accurately determine the etch endpoint to stop etching at the desired film thickness and depth during processing of such a layer.

  During the dry etching process of a semiconductor wafer, the emission intensity of a specific wavelength in plasma light changes as the etching of a specific film progresses. Therefore, as one method for detecting the etching end point of a semiconductor wafer, conventionally, a change in emission intensity of a specific wavelength from plasma is detected during the dry etching process, and the etching end point of a specific film is detected based on the detection result. There is a way to do it. At that time, it is necessary to prevent erroneous detection based on the fluctuation of the detected waveform due to noise. As a method for accurately detecting a change in emission intensity, a detection method using a moving average method (see, for example, Patent Document 1), a method for reducing noise by an approximation process using a first-order least square method (for example, a patent) Document 2) is known.

  With the recent miniaturization and high integration of semiconductors, the aperture ratio (area to be etched of a semiconductor wafer) has decreased, and the emission intensity of a specific wavelength taken into the photodetector from the optical sensor has become weak. As a result, the level of the sampling signal from the photodetector is reduced, and it is difficult for the end point determination unit to reliably detect the etching end point based on the sampling signal from the photodetector.

  In addition, when detecting the end point of etching and stopping the process, it is actually important that the remaining thickness of the dielectric layer is equal to a predetermined value. Conventional processes monitor the entire process using time thickness control techniques based on the assumption that the etch rate of each layer is constant. The value of the etching rate is obtained, for example, by processing a sample wafer in advance. In this method, by the time monitoring method, the etching process is stopped at the same time as the time corresponding to the predetermined etching film thickness elapses.

However, it is known that an SiO ₂ layer formed by an actual film, for example, a low pressure chemical vapor deposition (LPCVD) technique has low thickness reproducibility. Thickness tolerance due to process variations during LPCVD corresponds to about 10% of the initial thickness of the SiO ₂ layer. Therefore, the time monitoring method cannot accurately measure the actual final thickness of the SiO ₂ layer remaining on the silicon substrate. The actual thickness of the remaining layer is then measured by a standard spectroscopic interferometer technique, and if it is found to be over-etched, the wafer is discarded as rejected. become.

  In addition, in an insulating film etching apparatus, a change over time is known such that the etching rate decreases as etching is repeated. In some cases, etching may stop in the middle, and the solution is essential. In addition, it is important for the stable operation of the process to monitor the variation of the etching rate over time. However, the conventional method only monitors the time for determining the end point. There was no appropriate way to deal with major changes and fluctuations. In addition, when the etching time is as short as about 10 seconds, the end point determination must be an end point determination method that shortens the determination preparation time, and the step of the determination time needs to be sufficiently shortened, but is not necessarily sufficient. Furthermore, in the insulating film, the area to be etched is often 1% or less, and the plasma emission intensity change from the reaction product generated by etching is small. Therefore, an end point determination system that can detect even a slight change is required, but no practical and inexpensive system is found.

  On the other hand, various methods using an interferometer are known as other methods for detecting the etching end point of a semiconductor wafer. That is, the first method is a method of detecting interference light (plasma light) using three types of color filters of red, green, and blue, and detecting an etching end point (see, for example, Patent Document 3). . The second method is a method of counting the extreme value of the interference waveform (maximum of waveform, minimum: zero pass point of differential waveform) using time variation of the interference waveform of two wavelengths and its differential waveform. A method of calculating an etching rate by measuring a time until the value reaches a predetermined value, obtaining a remaining etching time until reaching a predetermined film thickness based on the calculated etching rate, and stopping an etching process based on the etching time ( For example, see Patent Document 4). The third method obtains the waveform (with the wavelength as a parameter) of the difference between the light intensity pattern of the interference light before processing (having the wavelength as a parameter) and the light intensity pattern of the interference light after processing or during processing, This is a method of measuring a step (film thickness) by comparing the difference waveform with a difference waveform stored in a database (see, for example, Patent Document 5). The fourth method relates to a spin coater, and is a method for measuring the time change of interference light over multiple wavelengths to obtain a film thickness (see, for example, Patent Document 6). In the fifth method, the characteristic behavior of the interfering light with respect to time is obtained from the measurement, and is made into a database. The completion of etching is determined by comparing the database with the measured interference waveform. This is a method for prompting a change (see, for example, Patent Document 7).

In this interferometer usage, monochromatic radiation emitted from a laser is applied at a normal incidence angle to a wafer including a laminated structure of different materials. For example, the wafer SiO ₂ layer laminated are laminated on the _Si 3 _{N 4} layer, and the emitted light reflected by the upper surface of the SiO ₂ layer, which is formed between the SiO ₂ layer and _Si 3 _{N 4} layers Interference fringes are formed by the radiated light reflected from the boundary surface. The reflected radiation is applied to a suitable detector, which generates a signal whose intensity varies with the thickness of the SiO ₂ layer being etched. As soon as the upper surface of the SiO ₂ layer is exposed during the etching process, the etching rate and the current etching thickness can be monitored continuously and accurately. There is also known a method of measuring a predetermined radiation emitted by plasma instead of a laser by a spectrometer.

The above known techniques cause the following problems.
A. In film thickness determination in a thick film processing process (such as resist etchback with a film thickness of several μ), the time change of interference light is complicated and becomes several tens of cycles or more, and a slight disturbance affects the determination. In film thickness determination in a thin film processing process (gate oxide film, oxide film etchback, etc.), it is necessary to measure a slight change in interference light, and a slight disturbance affects the determination. That is, the time change of the interference light during thin film processing is 1/2 to ¼ period or less, the change of the interference fringes is slight, and it is necessary to exclude the influence of noise in the film thickness determination.
C. In processing wafers for mass production, peripheral circuits are mixed, and various materials (mask material, material to be etched, other materials in peripheral circuits) are etched simultaneously, so interference light from different materials is generated. In addition to being complicatedly superimposed, the processing wafers vary in the film thickness of various materials within a lot or between lots, and the temporal change of interference light during the etching process varies within a lot or between lots. .
D. In a mass production process of a small quantity and a variety of products, various etching processes are mixed, so that the etching apparatus is likely to change with time, abnormal discharge occurs, and the plasma fluctuates. Therefore, the plasma emission changes and disturbance is superimposed on the interference light, which affects the determination.

  In view of the above, it has been difficult to accurately measure and control the remaining layer amount and etching depth of the layer to be processed, particularly the plasma etching process, with the required measurement accuracy.

JP-A-61-53728 Japanese Patent Laid-Open No. 63-200533 JP-A-5-179467 JP-A-8-274082 (US Pat. No. 5,658,418) JP 2000-97648 A JP 2000-106356 A US Pat. No. 6,081,334

  It is an object of the present invention to measure the film thickness or etching depth of a material to be processed that can accurately measure the actual remaining film amount and etching depth of the layer to be processed on-line in the plasma etching process in the semiconductor element manufacturing process. An etching end point determination method using the method, a plasma processing apparatus for executing the end point determination method, and a plasma processing method and a plasma processing apparatus for a material to be processed using the plasma processing apparatus are provided.

  Another object of the present invention is to provide an etching method capable of accurately controlling each layer of a semiconductor element online to a predetermined film thickness and etching depth in a semiconductor element manufacturing process.

  Another object of the present invention is to provide an apparatus for measuring the film thickness and etching depth of a material to be processed that can accurately measure the actual film thickness and etching depth of the layer to be processed online in a semiconductor element manufacturing process. There is to do.

In order to solve the above-mentioned problems of the prior art and to achieve the object of the present invention, the present invention etches the film on the surface of the sample in the vacuum vessel using plasma formed in the vacuum vessel. In the plasma processing apparatus, a detector for detecting interference light of a plurality of wavelengths from the sample surface during the processing, and the intensity of the interference light at any time during processing of the sample obtained from the output of the detector The parameter of the pattern data with the wavelength of the time differential value of the parameter and the wavelength of the time differential value of the intensity of the interference light obtained from the output of the detector in the processing of another sample obtained before processing of this sample Pattern comparison means for comparing the data of a plurality of patterns corresponding to each of a plurality of remaining film thicknesses of the film on the surface of the other sample and calculating the deviation thereof Film of the remaining film thickness corresponding to the pattern to be the minimum value when the minimum value than that value by comparing the minimum value of the difference between these and the predetermined value is smaller at the arbitrary time a deviation comparing means for outputting a first value of the thickness, the residual film thickness at the time series data recording means for recording data relating to the first value of the film thickness output from the deviation comparing means as time-series data And a determination unit for determining a second value of the film thickness calculated using the time-series data at a plurality of times recorded in the time-series data recording unit as a remaining film thickness at the arbitrary time. And the first value of the film thickness at the arbitrary time output from the deviation comparing means is the second value of the film thickness or the arbitrary time determined by the determiner It said membrane when there from the remaining film thickness to within a predetermined tolerance And a plasma processing apparatus which records as time-series data to the time-series data recording unit data related to the first value of the.

Further, the present invention provides the plasma processing apparatus according to claim 1, wherein the determination unit records the time series recorded in the remaining film thickness time series data recording unit at the plurality of times before the arbitrary time. A plasma processing apparatus for determining the second value of the film thickness obtained by regression analysis using data as the remaining film thickness at the arbitrary time.

Further, the present invention provides the plasma processing apparatus according to claim 1, wherein the determination unit records the time series recorded in the remaining film thickness time series data recording unit at the plurality of times before the arbitrary time. The plasma processing apparatus is configured to determine the second value of the film thickness obtained by interpolation using data as the remaining film thickness at the arbitrary time.

Further, the present invention provides the plasma processing apparatus according to claim 1, wherein the determination unit records the time series recorded in the remaining film thickness time series data recording unit at the plurality of times before the arbitrary time. Using the data value, the plasma processing apparatus determines the second value of the film thickness interpolated by extrapolation as the film thickness at the arbitrary time.

The present invention also provides a plasma processing apparatus for etching a film on a surface of a sample in a vacuum vessel using plasma formed in the vacuum vessel, and detecting interference light having a plurality of wavelengths from the sample surface during the processing. The detector, the pattern data using the wavelength of the time differential value of the intensity of the interference light at an arbitrary time during the processing of the sample obtained from the output of the detector, and the pattern data obtained before the processing of the sample Data of a pattern using the wavelength of the time derivative of the intensity of the interference light obtained from the output of the detector in the processing of another sample as a parameter, and a plurality of remaining film thicknesses of the film on the surface of the other sample Pattern comparison means for comparing a plurality of patterns of data corresponding to each of them and calculating the deviation; and comparing the minimum value of the deviation between them with a predetermined value, and the minimum value is determined from the value A deviation comparing means for outputting the remaining film thickness corresponding to the pattern to be the minimum value when again as the first value of the film thickness of the arbitrary time, the film output from the deviation comparing means and time-series data recording means residual thickness to be recorded as time-series data data relating to the first value of the thickness, calculated using the time-series data at a plurality of time recorded in the time-series data recording means A determination unit that determines the second value of the film thickness as the remaining film thickness at the arbitrary time, and the etching processing speed obtained from the calculation result in the determination unit is determined in advance. The plasma processing apparatus records the first value of the film thickness output from the deviation comparison unit as time series data in the time series data recording unit when it is within the allowable range.

  According to the present invention, in the plasma processing, particularly in the plasma etching processing, a method for measuring the remaining film thickness or etching depth of the processing material capable of accurately measuring the actual etching amount of the processing layer online. The processing method of the sample of the to-be-processed material using can be provided.

  In addition, it is possible to provide an etching process capable of controlling each layer of a semiconductor element (semiconductor device) on-line with high accuracy so as to have a predetermined etching amount. Furthermore, it is possible to provide an apparatus for measuring a remaining film thickness of an object to be processed or an etching depth measuring apparatus capable of accurately measuring an actual etching amount of a layer to be processed online.

1 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with an etching amount measuring apparatus according to a first embodiment of the present invention. It is a flowchart which shows the procedure which calculates | requires the residual film thickness of a to-be-processed material, when performing an etching process using the etching amount measuring apparatus of FIG. It is a figure which shows the result of the time change and film thickness transition of the interference light and reference light in the normal etching process of 1st Example of this invention. It is a figure which shows the result of the time change and film thickness transition of interference light in the case of the discharge fluctuation of 1st Example of this invention, or reference light. It is a figure which shows the result of the time change and film thickness transition of interference light at the time of performing the minimum film thickness setting process of 1st Example of this invention, or reference light. It is a figure which shows the result of the time change and film thickness transition of interference light at the time of performing the pattern matching deviation process of 2nd Example of this invention, or reference light. It is a figure which shows the result of the time change and film thickness transition of interference light at the time of performing the film thickness tolerance | permissible_range process of 3rd Example of this invention, or reference light. It is a block diagram which shows the whole structure of the etching apparatus of the semiconductor wafer provided with the etching amount measuring apparatus which performs the film thickness comparison and etching rate comparison by 4th Example of this invention. It is a flowchart which shows the procedure which calculates | requires the residual film thickness of a to-be-processed material, when performing an etching process using the etching amount measuring apparatus of FIG. It is a figure which shows the result of the time change and film thickness transition of interference light and a reference light at the time of performing the etching rate tolerance | permissible_range process of the 5th Example of this invention. It is a block diagram which shows the whole structure of the etching apparatus of the semiconductor wafer provided with the etching amount measuring apparatus by the 5th Example of this invention. It is a block diagram which shows the whole structure of the etching apparatus of the semiconductor wafer provided with the reference light measuring apparatus by the modification of the 5th Example of this invention. It is a block diagram which shows the whole structure of the etching apparatus of the semiconductor wafer provided with the reference light measuring apparatus by the modification of the 5th Example of this invention. It is a flowchart which shows the procedure which calculates | requires the remaining film thickness of a to-be-processed material, when performing an etching process using the etching amount measuring apparatus of FIG. It is a figure which shows the result of the time change and film thickness transition of the interference light of 5th Example of this invention, or reference light.

  Examples of the present invention will be described below. In the following embodiments, elements having functions similar to those of the first embodiment are denoted by the same reference numerals as those of the first embodiment, and detailed description thereof is omitted.

The first embodiment of the present invention will be described below with reference to FIGS. In the first embodiment, when the material to be processed such as a semiconductor wafer is subjected to plasma etching, the wavelength dependence of the differential value of the interference light with respect to the etching amount of each layer of the sample material to be processed (first material to be processed). indicating the sex (as a parameter a wavelength) to set the standard differential pattern P _S, then plasma etching in the actual processing of the material to be treated (second material to be treated) samples for the material to be treated the same configuration The intensity of a plurality of wavelengths of interference light at an arbitrary time from the start of processing is measured, and an actual differential pattern (Pr) indicating the wavelength dependence of the differential value of the measured interference light intensity (with wavelength as a parameter) Then, the standard differential pattern (P _s ) of the differential value is compared with the actual differential pattern (Pr) to determine the etching amount of the material to be processed.

  First, referring to FIG. 1, a plasma processing apparatus which is an etching apparatus for a semiconductor wafer on which a semiconductor element provided with an apparatus for measuring an etching amount (residual film thickness of a mask material or silicon etching depth) according to the present invention is formed. The overall configuration will be described. An etching apparatus (plasma processing apparatus) 1 includes a vacuum vessel 2, and an etching gas introduced from a gas introduction unit (not shown) is decomposed by microwave power or the like into plasma 3. The material 4 to be processed such as a semiconductor wafer on the sample stage 5 is etched. Etching amount (residual film thickness of mask material or etching depth of silicon) Multi-wavelength radiated light from a measurement light source (for example, halogen light source) included in the spectroscope 11 of the measuring device 10 is introduced into the vacuum vessel 2 by the optical fiber 8 Guided and applied to the workpiece 4 at a normal incidence angle. The interference light from the material to be processed 4 is guided to the spectroscope 11 of the etching amount measuring device 10 through the optical fiber 8, and based on the state, the silicon etching depth or the remaining film thickness of the mask material is measured and the end point of the etching process is determined. Perform the process.

  The etching amount measuring apparatus 10 includes a spectroscope 11, a first digital filter circuit 12, a differentiator 13, a second digital filter circuit 14, a differential waveform comparator 15, a differential waveform pattern database 16, and a pattern matching deviation. The comparator 115, the deviation value setting unit 116, the remaining film thickness time series data recorder 18, the regression analyzer 19, the end point determination unit 230, and the display unit 17 for displaying the end point determination unit result are provided. FIG. 1 shows a functional configuration of the etching amount measuring apparatus 10, and the actual configuration of the etching amount measuring apparatus 10 excluding the display unit 17 and the spectroscope 11 includes a CPU and a mask material remaining. Storage device consisting of a ROM that stores various data such as a film thickness measurement processing program, silicon etching depth measurement processing program, interference waveform differential waveform pattern database, etc., an external storage device, etc. An output device and a communication control device can be used.

  In the spectrometer 1, multi-wavelength radiated light from a measurement light source (for example, a halogen light source) is guided into the vacuum container 2 by an optical fiber 8 and applied to the material 4 to be processed at a vertical incident angle. The interference light from the material to be processed 4 is guided to the spectroscope 11 of the etching amount measuring device 10 through the optical fiber 8, and based on the state, the etching depth measurement of the silicon or the remaining film thickness of the mask material and the end point of the etching process are performed. Perform the determination process.

  The emission intensities of the multi-wavelength interference light captured by the spectroscope 11 become current detection signals corresponding to the emission intensity for each specific wavelength, and are converted into voltage signals. Signals of a plurality of specific wavelengths (j) output as sampling signals by the spectrometer 11 are stored in a storage device such as a RAM (not shown) as time series data yi, j. Next, the time series data yi, j at time i is smoothed by the first digital filter circuit 12 and stored in a storage device such as a RAM (not shown) as smoothed time series data Yi, j. On the basis of the smoothed time series data Yi, j, the differentiator 13 calculates the time series data di, j of the derivative value (primary differential value or secondary differential value) and stores it in a RAM or the like not shown. It is stored in the device. The time series data di, j of the derivative values is smoothed by the second digital filter circuit 14 and stored in a storage device such as a RAM (not shown) as smoothed derivative coefficient time series data Di, j. Then, a real differential pattern (Prj) = Σj (Di, j) indicating the wavelength dependence of the differential value of the interference light intensity (with wavelength j as a parameter) is obtained from the smoothed differential coefficient time series data Di, j. .

On the other hand, the differential waveform pattern database 16 corresponds to the etching depth represented by the remaining film thickness s of the material to be processed, which is the target of the etching amount measurement obtained in advance using the first (sample) material to be processed. A differential waveform pattern data value P _S j of the interference light intensity for each wavelength is set. The differential waveform comparator 15 compares the actual differential pattern Prj = Σj (Di, j) and the differential waveform pattern data value P _S j of the film thickness s. In the pattern matching deviation comparator 115, the pattern matching deviation value (minimum) is the pattern matching deviation (σs = √ (Σj (Di, j−P _S j) × (Di, j−P _S j) / j)). ) σs is obtained and compared with the pattern matching deviation value (set) σ ₀ preset in the deviation value setter 116. If the pattern matching deviation value (minimum) σs is equal to or less than the pattern matching deviation value (set) σ ₀ , The film thickness value s is stored in the film thickness time series data recorder 18 as the instantaneous film thickness Zi at time i. When the pattern matching deviation value (minimum) σs is greater than or equal to the pattern matching deviation value (setting) σ ₀ , the instantaneous film thickness Zi at time i is not stored.

  In the regression analyzer 19, the calculated film thickness F at time i is obtained by regression line approximation using instantaneous film thickness data before time i. The end point determination unit 230 determines whether or not the calculated film thickness F is equal to or less than a preset target film thickness. The result of the etching amount of the material to be processed obtained by the above processing is displayed on the result display 17.

  In the first embodiment, only one spectrometer 11 is shown. However, when it is desired to measure and control the surface of the material to be processed widely, a plurality of spectrometers 11 may be provided. .

  Next, a procedure for obtaining the etching amount of the material to be processed when performing the etching process with the etching amount measuring apparatus 10 of FIG. 1 will be described with reference to the flowchart of FIG.

First, the target etching amount (target remaining film thickness value) is set, and a differential pattern (residual film thickness standard differential pattern) P extracted from the wavelength range (at least three wavelength ranges) from the differential waveform pattern database 16. _S j and pattern matching deviation value (setting) σ ₀ are set (step 600). That is, the standard differential pattern P _S j corresponding to the etching amount (remaining film thickness) s required according to the processing conditions of the material to be processed is set in the differential waveform pattern database 16 in advance.

  In the next step, sampling of interference light from the object to be processed (for example, every 0.25 to 0.5 seconds) is started (step 601). That is, a sampling start command is issued with the start of the etching process. Multi-wavelength emission intensity that changes as the etching progresses is detected by a photodetector (spectrometer 11) as a photodetection signal having a voltage corresponding to the emission intensity. The photodetection signal for each wavelength j of the spectrometer 11 is digitally converted to obtain sampling signals yi, j.

  Next, the multi-wavelength output signal yi, j from the spectroscope 11 is smoothed by the first stage digital filter 12, and the smoothed time series data Yi, j is calculated (step 602). That is, noise is reduced by the first-stage digital filter to obtain smoothed time series data Yi, j.

Next, the differentiator 13 differentiates the smoothed time series data Yi, j and calculates the differential coefficient di, j for each wavelength by the SG method (step 603). That is, the differential coefficient (first order or second order) di, j of the signal waveform for each wavelength is obtained by differential processing (SG method). Further, the smoothed differential coefficient time series data Di, j is calculated by the second stage digital filter 14 (step 604). Then, the differential waveform comparator 15 calculates a matching pattern deviation value (minimum) σs = √ (Σ (Di, j−P _S j) ² / j) value, and matches the matching pattern deviation value ( The smallest minimum value σ of (minimum) σs is obtained (step 605).

Next, the pattern matching deviation comparator 115 determines σ ≦ σ ₀ for comparing the calculated matching pattern deviation value (minimum value) σ with the matching pattern deviation value (setting) σ ₀ (step 606), and σ If ≦ σ ₀ , the instantaneous film thickness Zi at time i is stored in the remaining film thickness time series data recorder 18 assuming that the film thickness of the material to be processed becomes the remaining film thickness s (step 607). If σ ≦ σ ₀ is not satisfied, the instantaneous film thickness Zi at time i is not obtained from the standard differential pattern database, and the instantaneous film thickness is not stored in the remaining film thickness time-series data recorder 18 (step 608). The smoothed differential coefficient time series data Di, j is compared with the differential pattern P _S j previously set in the differential waveform comparator 15 to calculate the remaining film thickness value Zi at that time (step 615).

  Next, using the stored past time series data Zi, the regression analyzer 19 uses the linear regression line Y = Xa × t + Xb (Y: amount of remaining film, t: etching time, Xa: absolute value is etching rate, Xb: initial film thickness) is obtained, and a calculated remaining film amount F at time i (current time) is calculated from this regression line (step 609). Next, the end point determination unit 230 compares the calculated remaining film amount F with the target remaining film thickness value to determine the etching amount (residual film thickness value). As a result, the result is displayed on the display unit 17 (step 609). If it is equal to or greater than the target remaining film thickness value, the process returns to step 602, and these steps are repeated. Finally, if the calculated calculation amount F is equal to or less than the target remaining film thickness value in step 610, the end of sampling is set (step 611).

  Here, calculation of the smoothed differential coefficient time series data Di at time i for a certain wavelength j will be described. As the first digital filter circuit 12, for example, a second-order Butterworth low-pass filter is used. The smoothed time series data Yi is obtained by the equation (1) using a secondary Butterworth low-pass filter.

Yi = b1yi + b2yi-1 + b3yi-2
-[A2Yi-1 + a3Yi-2] (1)

  Here, the coefficients b and a have different numerical values depending on the sampling frequency and the cutoff frequency. For example, a2 = −1.143, a3 = 0.4128, b1 = 0.067455, b2 = 0.13491, b3 = 0.067455 (sampling frequency 10 Hz, cutoff frequency 1 Hz), a2 = −0.00073612, a3 = 0.17157, b1 = 0.287271, b2 = 0.85442, b3 = 0.29271 (cut-off frequency 2.5 Hz), and the like are used.

The time-series data di of the second derivative value is obtained from the following equation (2) using the polynomial-matching smoothing differentiation method of the five-point time-series data Yi by the differentiator (derivative coefficient arithmetic circuit) 13 as follows: Calculated.
j = 2
di = ΣwjYi + j (2)
j = -2
Here, w-2 = 2, w-1 = -1, w0 = -2, w1 = -1, and w2 = 2.

  Using the time-series data di of the differential coefficient value, the smoothed differential coefficient time-series data Di is a second digital filter circuit (second-order Butterworth low-pass filter, but different from the a and b coefficients of the digital filter circuit. May be determined by the following equation (3).

Di = b1di + b2di-1 + b 3di-2
-[A 2Di-1 + a 3Di-2] (3)

  FIG. 3 shows the relationship between the interference intensity and the etching time when the polysilicon is etched and the determination is made at a polysilicon film thickness of 45 nm. The initial film thickness of the polysilicon to be etched is about 170 nm. In the figure, the interference light waveform with a wavelength of 500 nm observed from the wafer surface, its first derivative waveform, and the wafer surface are observed. 6 shows the time change (instantaneous film thickness transition) of the polysilicon film thickness during the etching process obtained by matching comparison with the plasma light (reference light) having a wavelength of 500 nm and the standard differential pattern. Here, as for the instantaneous film thickness transition, the first derivative pattern at each time is compared with the standard differential pattern corresponding to each film thickness, the one with the smallest pattern matching deviation is selected, and the film thickness change is plotted. It is a thing.

  FIG. 4 shows the variation of the instantaneous film thickness transition that occurs when the polysilicon etching is continuously performed. In the figure, the instantaneous film thickness rapidly decreases and the film thickness reaches about 10 nm during the etching process time of about 25 to about 31 seconds. This phenomenon is considered to occur when the etching plasma slightly changes due to reaction products accumulated in a part of the chamber or when the power for generating the plasma fluctuates slightly, but the etching processing time is between about 25 and about 30 seconds. Only when the instantaneous film thickness decreases rapidly, the transition of the instantaneous film thickness returns to its original state, and the etching process is normally completed. When there is such a change in the instantaneous film thickness, for example, the film thickness determination at 45 nm is less than the determination film thickness at time 25 seconds, so the etching process is finished at a film thickness of about 100 nm to manufacture a defective product. Will do. Therefore, the film thickness determination system needs to accurately determine the film thickness in response to such fluctuations.

  Here, the behavior of the interference waveform was analyzed in order to prevent a sudden drop in the instantaneous film thickness. In general, the interference waveform changes when the material is thinned, so that the interference light does not change in many wavelength regions, so the first-order differential changes at those wavelengths simultaneously approach zero. In addition, even when there is plasma fluctuation, the change occurs simultaneously in many wavelength ranges, the first derivatives of those wavelengths change simultaneously, the fluctuations ease, and the behavior of this derivative becomes thin. It is similar to the change in interference light over time. Therefore, in order to prevent such a rapid fluctuation, it is necessary not to use data having a thin film thickness as much as possible among standard differential patterns used for film thickness measurement. That is, the standard differential pattern having a film thickness smaller than the target determination film thickness is to perform pattern matching processing with the standard differential pattern so as not to be used for obtaining the instantaneous film thickness during the etching process.

  FIG. 5 shows the results when the minimum film thickness of the standard differential pattern used for the pattern matching process is 20 nm. From the figure, it can be seen that an abrupt film thickness reduction of about 25 to about 30 seconds in etching processing time with plasma fluctuation can be avoided. Also, as a result of the pattern matching process at a time with fluctuation, the pattern matching deviation is 0.05 or more, and there is no film thickness that matches the standard differential pattern and the real differential pattern. It is the initial film thickness. The standard differential pattern is set for the film thickness used for film thickness measurement. When the target remaining film thickness value is set in step 600 in the flowchart of FIG. 2, the minimum film thickness of the standard differential pattern is set based on this value. And adopt a standard differential pattern greater than this film thickness.

Next, another embodiment for avoiding plasma fluctuation will be described. Here, the fact that the pattern matching deviation becomes large when there is plasma fluctuation is utilized. In general, after the etching process is started, the differential process for determining the film thickness is started for a few seconds. The interference waveform is slightly disturbed by the influence of plasma ignition, and the pattern matching deviation value σ is deteriorated. If the pattern matching deviation value σ after that time is calculated based on the pattern matching deviation value (setting) σ ₀ at this time, and the pattern matching deviation value σ is larger than the pattern matching deviation value (setting) σ ₀ The pattern matching with the standard differential pattern is not sufficient, and the instantaneous film thickness Zi is not obtained from the standard differential pattern, but is set, for example, as the initial film thickness of a database (standard differential pattern). The instantaneous film thickness data that is the initial film thickness at this time is not used for regression linear approximation analysis for calculating the calculated film thickness F.

In FIG. 6, the instantaneous film thickness Zi at time i during the etching process is obtained using the pattern matching deviation value σ ₀ = 0.04 for 2 seconds from the start of the differentiation process, which is the second embodiment. The result of calculating the calculated film thickness F at time i from the regression line approximation based on the time series data of the instantaneous film thickness Zi is shown. From the figure, it can be seen that the obtained film thickness transition is not affected by plasma fluctuations and is stable and sufficient film thickness determination can be made. Here, the pattern matching deviation value σ ₀ is obtained for several seconds after the start of the differentiation process for each wafer process, and pattern matching determination is performed. However, a plurality of wafer processes are performed, and an average of the pattern matching deviation values σ ₀ is obtained. The value may be set in step 600 in the flowchart of FIG.

  In a mass production process in which a semiconductor wafer is processed using plasma to manufacture a semiconductor device from the semiconductor wafer, the plasma processing apparatus as in the present invention is continuously operated, and the number of processed materials is reduced. Along with the increase, conditions in the processing chamber such as product depositing and depositing on the surface of the member inside the processing chamber fluctuate, and this changes the state of the plasma generated in the processing chamber and the surface of the surface obtained as a result of the processing The shape will be changed. Therefore, it is necessary to manage the process so as to control the variation in the result of processing the material to be processed in the mass production. In this embodiment, when such mass production management is performed, the number of times that the pattern matching deviation value becomes larger than a predetermined value is monitored by processing for each wafer that is a material to be processed, and the number of times is illustrated. Count with the omitted recorder or counter. Such counting may be performed by the pattern matching deviation comparator 115.

  Furthermore, by comparing the transition of the number of times with a predetermined value (for example, a value of the number of times or a predetermined value related to the increase rate), it is possible to grasp the apparatus state and the wafer etching state. That is, when the number of times is gradually increased, the predetermined value of the number of times is used as a guideline for starting maintenance in the plasma processing apparatus such as wet cleaning, or the number of times is rapidly increased and the increase rate is larger than the predetermined value. In such a case, the user is informed or warned of the necessity of treatment such as transporting the wafer to the wafer inspection process. Such notifications and warnings are displayed on the display 17 of FIG. 1, for example, in response to a command from the pattern matching deviation comparator 115.

  Next, a third embodiment for avoiding plasma fluctuation will be described. Here, the instantaneous film thickness transition Zi is not stabilized by the above-described pattern matching deviation comparison, but the instantaneous film thickness Zi at the time i during the etching process is obtained, and based on the time series data of the instantaneous film thickness Zi before the time i. When calculating the calculated film thickness F at time i from the regression line approximation, if the difference (absolute value) between the calculated film thickness F and the instantaneous film thickness Zi is greater than or equal to a predetermined film thickness tolerance value, FIG. 7 shows the result obtained by determining that the instantaneous film thickness Zi at time i is not an accurate film thickness and using this instantaneous film pressure Zi in a method not calculating the calculated film thickness by regression line approximation after time i. . From FIG. 7, it can be seen that this method can also avoid a rapid change in the etching processing time of about 25 to about 30 seconds with the plasma fluctuation shown in FIG. Here, 20 nm was used as the allowable film thickness. The setting of the allowable film thickness can be determined from the change in the interference waveform that appears during the etching process. For example, when an initial polysilicon film thickness of 200 nm of the material to be etched is etched, the interference waveform at a wavelength of 500 nm has about 7/2 cycles, and the interference waveform is accurate within 1/4 cycle (differential sign is different). The film thickness is about 20 nm.

  Next, a description will be given of a fourth embodiment relating to avoiding erroneous determination of film thickness measurement using the fact that the etching rate during mass production processing is substantially constant and the fluctuation of the etching rate is about ± 10% at most. From the transition of the instantaneous film thickness in FIG. 3, the slope of the change in the instantaneous film thickness Zi in the normal etching process (32 to 60 seconds) is constant, and the etching rate obtained from the slope is about 123 nm / min. On the other hand, it can be seen that the slope of the instantaneous film thickness change during the time when the plasma in FIG. 4 fluctuates (25-31 seconds) is smaller than that during normal etching. In the mass production process, when the etching rate is doubled or halved, the etching process is abnormal, and it is necessary to perform a process such as wet cleaning of the etching apparatus to return the etching apparatus to normal. FIG. 8 shows a configuration of a plasma processing apparatus provided with a film pressure determiner according to the fourth embodiment, and FIG. 9 shows a flowchart of film thickness determination.

  As shown in FIG. 8, the fourth embodiment has a remaining film thickness comparator 20 between the regression analyzer 19 and the end point determination unit 230 of the etching amount measuring apparatus 10 of the plasma processing apparatus shown in FIG. This is characterized in that an etching rate comparator 21 is added.

As shown in FIG. 9, first, a target etching amount (target residual film thickness value) is set, and a differential pattern (residual film thickness) extracted from the wavelength range (at least three wavelength ranges) from the differential waveform pattern database. is performing standard differential pattern) _{P S} j and pattern matching deviation (set) sigma ₀ and film thickness tolerances _{Z 0} and setting the etching rate tolerance _{R 0} (step 1600).

  In the next step, sampling of interference light (for example, every 0.25 to 0.5 seconds) is started (step 1601). That is, a sampling start command is issued with the start of the etching process. The multi-wavelength emission intensity that changes with the progress of etching is detected by the photodetector as a photodetection signal having a voltage corresponding to the emission intensity. The light detection signal of the spectroscope 11 is digitally converted to obtain sampling signals yi, j.

  Next, the multi-wavelength output signal yi, j from the spectroscope 11 is smoothed by the first stage digital filter 12, and the smoothed time series data Yi, j is calculated (step 1602). That is, noise is reduced by the first-stage digital filter to obtain smoothed time series data Yi, j.

Next, the differentiator 13 calculates the differential coefficient di, j by the SG method (step 1603). That is, the differential coefficient (primary or secondary) di of the signal waveform is obtained by differential processing (SG method). Further, the smoothed differential coefficient time series data Di, j is calculated by the second stage digital filter 14 (step 1604). Then, the differential waveform comparator 15 calculates the matching pattern deviation value (minimum) σs = √ (Σ (Di, j−P _S j) ² / j) value, and the matching pattern deviation value (minimum) with respect to the film thickness s. ) Find the smallest minimum value σ of σs (step 1605).

Next, the pattern matching deviation comparator 115 determines σ ≦ σ ₀ comparing the calculated matching pattern deviation value (minimum value) σ with the matching pattern deviation value (setting) σ ₀ (step 1606), and σ If ≦ σ ₀ , the instantaneous film thickness Zi at the time i is stored in the remaining film thickness time-series data recorder 18 assuming that the film thickness of the material to be processed becomes the film thickness s (step 1607). If σ ≦ σ ₀ is not true, the instantaneous film thickness Zi at time i is not obtained from the standard differential pattern database, and the instantaneous film thickness is not stored in the remaining film thickness time-series data recorder 18 (step 1608).

As for the etching rate during processing, the regression analyzer 19 calculates the calculated film thickness F and its slope Xa from the linear regression line approximation based on the data of the remaining film thickness time series data recorder 18 (1609). Then, the remaining film thickness of the comparator 20, whether a film thickness of the instantaneous thickness Zi is limited than the calculated thickness F and the film thickness tolerance _{z 0 (F-z 0 ≦} Zi ≦ F + z 0) Or whether the slope Xa of the straight line obtained by regression approximation in the etching rate comparator 21 is an etching rate limited by the etching rate R and the etching rate allowable value _R0 at the time of creating the standard differential pattern. When (R−R ₀ ≦ Xa ≦ R + R ₀ ) is determined and (F−z ₀ ≦ Zi ≦ F + z ₀ ) or (R−R ₀ ≦ Xa ≦ R + R ₀ ), the instantaneous film thickness Zi remains. The film thickness is stored in the time-series data recorder 18 (step 1612), and is not stored in other cases (step 1611).

  Next, the film thickness is determined based on the calculated film thickness F, and if it is equal to or less than the target remaining film thickness value, it is assumed that the etching amount of the material to be processed has reached a predetermined value, and the result is displayed on the display 17 (step 1613). ). The state of the film thickness change during the etching is possible by displaying the calculated film thickness F. If it is equal to or greater than the target remaining film thickness value, the process returns to step 1602 to repeat these steps. Finally, the end of sampling is set (step 1614).

  In FIG. 10, when the film thickness allowable range value 20 nm, the etching rate allowable 50% (etching rate 117 nm / min) setting and the minimum film thickness 1 nm (target remaining film thickness value 50 nm) are set in the fourth embodiment. The calculated film thickness transition results are shown. From the figure, it can be seen that the calculated film thickness F is not affected by the plasma fluctuation and is stable and the target film thickness of 50 nm can be determined. Here, the target film thickness is the target film thickness to be obtained by the etching process, and the minimum film thickness is the minimum value of the film pressure that can be determined when determining the minimum value. In this embodiment, since the allowable film thickness value is 20 nm, the remaining film thickness may be the target film thickness of 50 nm ± 20 nm. If the minimum film thickness value of 30 nm can be detected, the film pressure suddenly becomes lower than that. Even if it is detected, this value can be ignored.

  The number of data of the instantaneous film thickness that is not stored in the remaining film thickness time series data recorder 18 is almost zero when the etching process is normally performed. If it changes, the matching of the interference differential pattern deteriorates and the number of data points that are not stored increases. Further, if there is a change in the specifications of the wafer to be processed, pattern matching becomes worse and the number of data points increases. Therefore, in the mass production process, the number of data of the instantaneous film thickness that is not stored in the remaining film thickness time series data recorder 18 is displayed on the display unit 17 so that the apparatus management of the etching apparatus and the production management of the processed wafer are performed. Is possible.

  Next, a description will be given of a fifth embodiment in which the interference light and reference light observed when there is plasma fluctuation are corrected and the film thickness is determined. The interference light changes due to a steep change in plasma emission accompanying the fluctuation (abnormality) of the plasma. For example, as shown in FIGS. 4 and 5, it may be difficult to obtain an accurate film thickness. In addition, since digital filter processing and polynomial-adapted smoothing differentiation processing are used to improve the S / N ratio of the observed optical signal, steep changes in light emission are mitigated by these processing and affect over a long period of time. Appears. To avoid this effect, there is a method of temporarily stopping digital filter processing and polynomial-adaptive smoothing differentiation processing when there is a sharp change in plasma emission. However, when these processing is stopped, the instantaneous film thickness is obtained. Therefore, it becomes impossible to determine the film thickness.

  Therefore, a sharp change in plasma is detected, the amount of change at each wavelength used in the measurement is obtained, the optical signal of each wavelength is corrected according to the amount of change, and the digital signal is applied to these corrected optical signals. A method for obtaining the film thickness by performing processing such as filter processing and polynomial fitting smoothing differentiation processing will be described.

  When collecting standard pattern data, which is a film thickness determination database, the amount of change at the sampling point at time i (difference from time i-1) with respect to the emission data for which the temporal change in interference light and reference light is measured Is determined for each wavelength, and the maximum amount of change during the etching process is obtained for interference light and reference light. A noise threshold is set based on the maximum amount of change per one sampling, and a sharp change in plasma is detected using the noise threshold.

  Next, when there is an amount of change equal to or greater than the noise threshold, respective correction coefficients (intensity ratios: Si, j = yi−1, j / yi, j) for each wavelength are obtained, and the optical signals yi, j is corrected by y′i, j = Si, j * yi, j. The corrected y′i, j is subjected to processing such as digital filter processing and polynomial fitting smoothing differentiation processing to determine and determine the instantaneous film thickness Zi.

  The configuration of the plasma processing apparatus provided with the film pressure determination device according to the sixth embodiment in which the fluctuation of the plasma is avoided and the film thickness is determined in this way will be described with reference to FIG. The etching apparatus 1 includes a vacuum container 2. The etching gas introduced into the inside of the etching apparatus 1 is decomposed by microwave power or the like to form plasma, and the plasma 3 etches the material to be processed 4 such as a semiconductor wafer on the sample stage 5. Is done. Multi-wavelength radiated light from a measurement light source (for example, a halogen light source) included in the spectroscope 11 of the etching amount (residual film thickness or etching depth) measuring device 10 is guided into the vacuum vessel 2 by the optical fiber 8 and processed. It is applied to the material 4 at a normal incidence angle. The interference light from the material to be processed is guided to the spectroscope 11 of the etching amount measuring device 10 through the optical fiber 8, and the silicon etching film thickness measurement and end point determination processing are performed based on the state.

  The etching amount measuring apparatus 10 includes a spectroscope 11, a sampling data comparator 110, a noise value setting unit 111 for setting a noise threshold, a correction coefficient recorder / display unit 113, a sampling data correction unit 112, a first digital filter circuit 12, Differentiator 13, second digital filter circuit 14, differential waveform comparator 15, differential waveform pattern database 16, pattern matching deviation comparator 115, deviation value setter 116, remaining film thickness time series data recorder 18, regression analyzer 19, an end point determination unit 230 and a display unit 17 for displaying the result of the determination unit are provided.

  The multi-wavelength emission intensities captured by the spectroscope 11 are converted into voltage signals as current detection signals corresponding to the emission intensities. Signals of a plurality of specific wavelengths (j) output as sampling signals by the spectroscope 11 are compared with a value preset in the noise value setting unit 111 in the sampling data comparator 110, and the change value of the signal becomes noise. If the value is greater than or equal to the value, the sampling data corrector 112 corrects the time series data yi, j so that there is no change in the signal. The correction coefficient at that time is stored in the correction coefficient recorder / display unit 113. As described above, the corrected time-series data y′i, j of the instantaneously changed signal is stored in a storage device such as a RAM. Next, the time series data y'i, j at time i is smoothed by the first digital filter circuit 12 and stored as smoothed time series data Yi, j in a storage device such as a RAM. Based on the smoothed time series data Yi, j, the differentiator 13 calculates the differential series (first or second derivative) time series data di, j and stores it in a storage device such as a RAM. The The time series data di, j of the derivative values is smoothed by the second digital filter circuit 14 and stored in a storage device such as a RAM as smoothed derivative coefficient time series data Di, j. Then, an actual pattern indicating the wavelength dependency of the differential value of the interference light intensity (with the wavelength as a parameter) is obtained from the smoothed differential coefficient time series data Di, j.

On the other hand, in the differential waveform pattern database 16, differential waveform pattern data values P _S j of the interference light intensity corresponding to the respective wavelengths corresponding to the film thickness s of the material to be processed for which the etching amount is to be measured are preset. In the differential waveform comparator 15, the actual pattern and the differential waveform pattern data value P _S j of the film thickness s are compared. In the pattern matching deviation comparator 115, pattern matching deviation (σs = √ (Σj (Di , j-P S j) × (Di, j-P S j) / j)) pattern matching deviation value is minimum (Min ) S is obtained and compared with the deviation value σ ₀ preset in the deviation value setter 116. If the pattern matching deviation value (minimum) σ s is equal to or less than the pattern matching deviation value (setting) σ ₀ , the instantaneous film at time i The film thickness value s is stored in the film thickness time series data recorder 18 as the thickness. When the pattern matching deviation value (minimum) σs is greater than or equal to the pattern matching deviation value (setting) σ ₀ , it is not stored. In the regression analyzer 19, the calculated film thickness F at the time i is obtained by regression line approximation using the instantaneous film thickness data Zi before the time i. The end point determination unit 230 determines whether or not the calculated film thickness F is equal to or less than a preset target film thickness. The result of the etching amount of the material to be processed obtained by the above processing is displayed on the result display 17.

(Modification)
Although the interference light processing means is described in the configuration diagram of FIG. 11, in the modified example of the fifth embodiment, interference light measurement using plasma light is not performed, but interference light measurement using light from an external light source. As for the processing means, as shown in FIG. 12, a plasma light measuring means 1001, a spectroscope 1003, a sampling data comparator 1110, and a noise value setting device 1111 provided on the side wall of the etching processing container 2 are provided. . As plasma light measurement means, as shown in FIG. 13, a plasma light measurement means 1002, a spectroscope 1003, a sampling data comparator 1110, and a noise value setter 1111 provided at the bottom of the etching processing container 2 are provided. You may prepare. These apparatuses operate in the same manner as the spectroscope 103, the sampling data comparator 110, and the noise value setting unit 111 in FIG. The output of the sampling data comparator 1110 is output to the first digital filter 12 via the sampling data corrector 112 in the same manner as the output of the sampling data comparator 110 in FIG.

  Next, a procedure for obtaining the etching amount of the material to be processed when performing the etching process with the etching amount measuring apparatus 10 of FIG. 11 will be described with reference to the flowchart of FIG.

First, the target etching amount (target remaining film thickness value) is set, and the differential pattern P _S j, the deviation value σ ₀ and the noise value N extracted from the wavelength range (at least three wavelength ranges) from the standard differential pattern database. Settings are made (step 2600). That is, a standard differential pattern corresponding to the etching amount s required according to the processing conditions of the material to be processed is set in the differential waveform pattern databases 15 and 25 in advance.

  In the next step, sampling of interference light (for example, every 0.25 to 0.5 seconds) is started (step 2601). That is, a sampling start command is issued with the start of the etching process. The multi-wavelength emission intensity that changes with the progress of etching is detected by the photodetector as a photodetection signal having a voltage corresponding to the emission intensity. The photodetection signal of the spectroscope 11 is digitally converted to obtain sampling signals yi, j at time i.

  Next, the difference between the multiwavelength output signal yi, j from the spectroscope 11 and the signal yi-1, j at time i-1 is obtained (step 2604). It is determined using the sampling data comparator 110 whether or not the difference yi, j−yi−1, j is larger than a value N preset in the noise value setter 111 (step 2620). In the case of the examples of FIGS. 12 and 13, the sampling data comparator 1110 is used to determine whether or not the output signal related to plasma emission at times i−1 and i is larger than the value preset in the noise value setter 1111. judge. If larger, the rate of change, that is, the correction coefficient is obtained by Si, j = yi-1, j / yi, j (step 2621). If it is smaller, the correction coefficient is Si, j = 1 (step 2622). The multi-wavelength output signal yi, j from the spectroscope is corrected to y′i, j = Si, j × yi, j by this correction coefficient (step 2623). The correction coefficient value is stored or displayed in the correction coefficient recorder / display unit 113 and used for mass production management of the etching process. The signal y′i, j corrected in this way is transmitted to the first-stage digital filter 12 and smoothed to calculate time series data Yi, j (step 2602). That is, noise is reduced by the first-stage digital filter to obtain smoothed time series data Yi, j.

Next, the differentiator 13 calculates the differential coefficient di, j by the SG method (step 2603). That is, the coefficient (primary or secondary) di of the signal waveform is obtained by differential processing (SG method). Further, the smoothed differential coefficient time series data Di, j is calculated by the second stage digital filter 14 (step 2604). Then, the differential waveform comparator 15 calculates σs = √ (Σ (Di, j−P _S j) ² / j) value, and the smallest minimum value of the matching pattern deviation value (minimum) σs with respect to the film thickness s. σ is obtained (step 2605). Next, the pattern matching deviation comparator 115, a determination is made ([sigma] s:: matching pattern deviation (minimum), sigma ₀ matching pattern deviation (set)) (step 2606), when the [sigma] s ≦ sigma _0, the The instantaneous film thickness at time i is stored in the remaining film thickness time-series data recorder 18 assuming that the film thickness of the treatment material becomes the film thickness s (step 2607). If σs ≦ σ ₀ is not satisfied, the instantaneous film thickness at time i is not obtained from the standard differential pattern database, and the instantaneous film thickness is not stored in the remaining film thickness time series data recorder 18 (step 2608). The smoothed differential coefficient time series data Di, j is compared with the differential pattern P _Z j previously set in the differential waveform comparator 15 to calculate the remaining film value Zi at that time (step 2615). Next, using the stored past time series data Zi, the regression analyzer 19 uses the linear regression line Y = Xa × t + Xb (Y: amount of remaining film, t: etching time, Xa: absolute value is etching rate, Xb: initial film thickness) is obtained, and a calculated remaining film amount F at time i (current time) is calculated from this regression line (step 2609). Next, the end point determination unit 230 compares the remaining film amount F with the target remaining film thickness value, and if the target film thickness value is equal to or less than the target remaining film thickness value, the etching amount of the material to be processed becomes a predetermined value. Is displayed on the display 17 (step 2609). If it is equal to or greater than the target remaining film thickness value, the process returns to step 2604 and these steps are repeated. Finally, the end of sampling is set (step 2611).

  Next, a specific example of the present invention will be described with respect to interference light measurement with discharge fluctuation shown in FIG. The maximum amount of change per sampling of interference light and reference light in the standard pattern data for film thickness determination used here was determined as follows. In the above-described polysilicon etching process (between 5 seconds and 55 seconds from the start of etching), the maximum change amount of the interference light of the standard pattern data is 50 counts, and the maximum change amount of the reference light is 20 counts. Therefore, the noise threshold value for detecting a steep change in plasma is set to 100 counts and 50 counts, which are predetermined values determined according to specifications between 2 to 3 times the maximum change amount.

  FIG. 15 shows a result of correcting the steep change of plasma by performing the above processing. As shown in FIG. 15, it can be seen that the fluctuation of the light emission that occurred about 25 seconds after the start of the etching shown in FIG. 4 is corrected, and abnormal changes in the interference light waveform and the reference light waveform are reduced. In this example, since the noise (abnormal light emission) of the interference waveform is small, there was no change larger than the noise threshold. However, the noise (abnormal light emission) of the reference light is large, and a change in plasma can be detected sufficiently, and correction at that time is performed on the reference light and interference light. As a result, slight fluctuations in the interference light can be corrected. It can be seen that the transition of the instantaneous film thickness required during etching can be obtained stably by these treatments.

  In this example, reference light is used as a means to detect abrupt changes in plasma, but the reflected power and matching point of the plasma generation power, or the reflected power and matching point of the bias applied to the wafer are monitored. The change in the value may be used.

Here, the correction coefficient is Si, j = yi-1, j /
Although it is obtained from yi, j, an average value of a plurality of waveform data before time i-1 may be used, and an interpolation method such as Lagrange interpolation or spline interpolation for past time series data is used. An approximate value of time i-1 by a smooth curve may be used. Further, the light emission data at time i may be further corrected by applying Lagrange interpolation or spline interpolation to the light emission data at time i corrected by the correction coefficient. Further, in the examples of FIGS. 12 and 13, as in the case of using the interference light from the sample surface, the interference light is corrected using the correction coefficient obtained based on the comparison result of the sampling data comparator 1110 related to the plasma emission. Sampling data may be corrected. Further, when it is detected that the noise threshold is exceeded at an arbitrary time, the film thickness is detected by calculation using the above-described regression analysis at a predetermined number of times at least one time, When it is detected that the noise threshold is exceeded at a later time, the film pressure may be detected using a correction coefficient.

  The number of times the noise threshold is exceeded is zero when the etching process is normally performed, but increases when the etching characteristics of the apparatus change and the plasma state deteriorates due to the time-dependent change of the etching apparatus. Therefore, in the mass production process, it is possible to manage the etching apparatus by displaying the number of times that the noise threshold is exceeded on the display unit 17.

  According to the present invention, it is possible to provide a film thickness measurement method and a process end point determination method using the film thickness measurement method by accurately measuring the etching amount of a material to be processed online in plasma processing, particularly in plasma etching processing. it can.

  In addition, it is possible to provide an etching process capable of controlling a layer to be etched of a semiconductor device on-line with high accuracy so as to have a predetermined etching amount. Furthermore, it is possible to provide an apparatus for measuring the etching amount of a material to be processed, which can accurately measure the actual etching amount of the layer to be processed online.

DESCRIPTION OF SYMBOLS 1 ... Etching device, 2 ... Vacuum container, 3 ... Plasma, 4 ... Material to be processed, 5 ... Sample stand, 8 ... Optical fiber, 9 ... Radiation, 10 ... Measuring device, 11 ... Spectrometer, 12 ... 1st digital filter circuit, 13 ... Differentiator, 14 ... 2nd digital filter circuit, 15 ... Differential waveform comparator, 16 ... Differential waveform pattern database, 115 ... Pattern matching deviation comparison 116, Deviation value setting unit, 18 ... Remaining film thickness time series data recorder, 19 ... Regression analyzer, 230 ... End point determination unit, 17 ... Display unit 17, 110 ... Comparison of sampling data 111... Noise value setter 112 112 Sampling data corrector 113 Correction coefficient recorder / display 1001 1002 Plasma light measuring means 1003 Spectrometer 1110 ‥ sampled data comparator, 1111 ‥‥ noise value setter.

Claims (5)

In the plasma processing apparatus for etching the film on the surface of the sample in the vacuum vessel using the plasma formed in the vacuum vessel,
A detector for detecting interference light of a plurality of wavelengths from the sample surface during the processing;
Pattern data using the wavelength of the time differential value of the intensity of the interference light at any time during processing of the sample obtained from the output of the detector as a parameter and another sample obtained before processing of the sample Data of a pattern using as a parameter the wavelength of the time differential value of the intensity of the interference light obtained from the output of the detector in the processing, and each of a plurality of remaining film thicknesses of the film on the surface of the other sample Pattern comparing means for comparing the data of a plurality of corresponding patterns and calculating the deviation;
When the minimum value of the deviation between them is compared with a predetermined value and the minimum value is smaller than the predetermined value, the remaining film thickness corresponding to the pattern that becomes the minimum value is determined as the film at the arbitrary time. Deviation comparison means for outputting as a first value of thickness ;
Residual film thickness time series data recording means for recording data relating to the first value of the film thickness output from the deviation comparison means as time series data;
A determiner that determines a second value of the film thickness calculated using the time-series data at a plurality of times recorded in the time-series data recording unit as the remaining film thickness at the arbitrary time; Prepared,
The first value of the film thickness at the arbitrary time output from the deviation comparing means is the second value of the film thickness or the remaining film thickness at the arbitrary time determined by the determiner. A plasma processing apparatus for recording data relating to the first value of the film thickness as time-series data in the time-series data recording means when it is within a predetermined allowable range. The plasma processing apparatus according to claim 1, wherein the determination unit uses the time series data recorded in the remaining film thickness time series data recording unit at the plurality of times before the arbitrary time. A plasma processing apparatus for determining a second value of the film thickness obtained by the regression analysis as a remaining film thickness at the arbitrary time. 2. The plasma processing apparatus according to claim 1, wherein the determination unit uses the time series data recorded in the remaining film thickness time series data recording unit at the plurality of times before the arbitrary time. A plasma processing apparatus that determines the second value of the film thickness obtained by interpolation as the remaining film thickness at the arbitrary time. 2. The plasma processing apparatus according to claim 1, wherein the determination unit includes a value of the time-series data recorded in the remaining film thickness time-series data recording unit at the plurality of times before the arbitrary time. The plasma processing apparatus determines the second value of the film thickness interpolated by extrapolation as the film thickness at the arbitrary time. In the plasma processing apparatus for etching the film on the surface of the sample in the vacuum vessel using the plasma formed in the vacuum vessel,
A detector for detecting interference light of a plurality of wavelengths from the sample surface during the processing;
Pattern data using the wavelength of the time differential value of the intensity of the interference light at any time during processing of the sample obtained from the output of the detector as a parameter and another sample obtained before processing of the sample Data of a pattern using as a parameter the wavelength of the time differential value of the intensity of the interference light obtained from the output of the detector in the processing, and each of a plurality of remaining film thicknesses of the film on the surface of the other sample Pattern comparing means for comparing the data of a plurality of corresponding patterns and calculating the deviation;
The minimum value of the deviation between them is compared with a predetermined value, and when the minimum value is smaller than the value, the remaining film thickness corresponding to the pattern that becomes the minimum value is determined as the film thickness at the arbitrary time. A deviation comparison means for outputting as a first value ;
Residual film thickness time series data recording means for recording data relating to the first value of the film thickness output from the deviation comparison means as time series data;
A determiner that determines a second value of the film thickness calculated using the time-series data at a plurality of times recorded in the time-series data recording unit as the remaining film thickness at the arbitrary time; Prepared,
The first value of the film thickness output from the deviation comparing means when the speed of the etching process obtained from the result of the calculation in the determiner is within a predetermined allowable range is the time series. A plasma processing apparatus for recording as time series data in a data recording means.

Images