Plasma treatment method and plasma treatment device

Abstract

A plasma processing method and a plasma processing apparatus are provided that accurately estimate the amount of etching in the transverse direction and determine the endpoint based on the estimated amount of etching. A plasma treatment method according to the present invention is a plasma treatment method for plasma treating a wafer, comprising: a first step of irradiating light onto the wafer; a second step of receiving light reflected from the wafer at a predetermined plurality of times during plasma treatment of the wafer; a third step of performing signal processing on light quantity data for each of the plurality of wavelengths of the received light; a fourth step of determining the amount of etching of the wafer during plasma treatment using the data to be processed after the signal processing; and a fifth step of determining the end point of the plasma treatment based on the amount of etching.

Description

Technology Field

[0001] The present invention relates to a plasma treatment method and a plasma treatment apparatus for treating a substrate-shaped sample, such as a semiconductor wafer, placed in a treatment chamber inside a vacuum vessel using plasma formed in said treatment chamber. In particular, the invention relates to a plasma treatment method and a plasma treatment apparatus for performing treatment while detecting the amount of wafer treatment using light from inside the sample chamber. Background Technology

[0002] In the manufacture of semiconductor devices, a process is performed to form components or wiring that interconnects them as a single block of circuits that perform various functions on the surface of a wafer. The formation of these components or wiring is carried out by repeating processing steps such as forming film layers of various materials, including conductors, semiconductors, or insulators, that are pre-formed on the substrate-shaped surface of a semiconductor wafer, or removing unnecessary parts of these film layers. In the process of removing such unnecessary parts, dry etching treatment (process) using plasma is widely used.

[0003] Dry etching using such plasma (plasma etching) involves introducing a processing gas into a processing chamber located inside a vacuum vessel of a processing device, and simultaneously supplying a high-frequency electric field by high-frequency power supplied from a high-frequency power source into the processing chamber. The atoms or molecules of the introduced gas are excited, ionized, or dissociated to form a plasma, and the surface of a sample installed in the processing chamber is exposed to the plasma and brought into contact with it, thereby causing a reaction between the particles in the plasma and the film layer to be processed. At this time, anisotropic or isotropic etching of the film layer to be processed is performed through physical reactions such as sputtering caused by charged particles such as ions in the plasma, or chemical reactions caused by radicals (reaction-active particles, active species). On the wafer surface, various treatments having different characteristics are appropriately selected and applied, thereby forming components or wiring having a circuit structure that performs the various functions mentioned above.

[0004] If the shape of the processed material produced by plasma etching differs from the design, the various components formed cannot realize their functions. For this reason, many process monitor technologies have been proposed to monitor and stabilize the etching process. Process monitors that measure the thickness of a film deposited on a wafer or the depth of grooves or holes formed on the wafer by measuring reflected light from the wafer during processing are called film thickness and depth monitors and have been used for determining the end point of the etching process.

[0005] Patent Document 1 describes a method for increasing processing precision using a film thickness and depth monitor. In this method, a film thickness and depth monitor using plasma light as a light source is used to detect just before the film to be processed is completely removed, thereby terminating the etching process.

[0006] Patent Document 2 describes a technology for increasing the precision of film thickness and depth measurement in a film thickness and depth monitor. In this method, an external light source is used instead of plasma light as the light source irradiating the wafer. By doing so, the variation in the amount of light from the light source is reduced, thereby realizing high-precision measurement of film thickness and depth. Prior art literature

[0007] Japanese Public Notice No. 11-260799, Japanese Public Notice No. 2004-507070 The problem to be solved

[0008] In the case of a semiconductor device that has undergone three-dimensionalization, there is a process for selectively etching one of the films in a transverse direction for a multilayer film (stacked structure) in which two or more films are stacked together. For example, in the process of forming a gate electrode of a next-generation 3D-NAND flash memory, a process is included for etching a tungsten film in a transverse direction (horizontal direction relative to the vertical depth direction of the groove) from a fine width groove with a high aspect ratio formed in a stacked structure of a metal film and an insulating film.

[0009] Conventionally, for such transverse etching, the etching amount was controlled according to the etching time. The etching amount can be measured by the difference in wafer weight before and after etching. However, since the etching amount in the transverse direction of the metal film cannot be strictly controlled in this method, a technique is required to monitor the transverse etching amount and determine the etching endpoint.

[0010] As a method for monitoring the etching amount, a method can be conceived in which, similar to Patent Document 1, the spectrum of wafer reflected light is measured during etching, and the measured spectrum is compared with a database (DB) prepared in advance for reference in which the spectrum corresponds to the etching amount, thereby calculating the etching amount in the transverse direction during etching. In this method, the relationship between the etching amount and the spectrum must always be constant. Here, in 3D-NAND, since the stacking of metal films and insulating films consists of tens to hundreds of layers, for example, if the thickness of the insulating film changes by 1 nm due to film deposition deviation, the height of the entire stacked film changes significantly to 100 nm when the number of insulating film layers is 100. It is known that when detecting the spectrum of reflected light of a wafer formed only with insulating films, the peak position or the number of peaks in the spectrum changes significantly depending on the difference in the thickness of the insulating film. As such, when the total height of the stacked film changes, even if a spectrum attributable to transverse etching of the metal film is obtained, the variation in the spectrum attributable to the difference in the total height of the stacked film is included as noise, and thus the correspondence between the measured spectrum and the amount of etching does not match the spectrum and amount of etching of the reference data. Therefore, when the total height of the stacked film changes, it becomes difficult to determine the etching endpoint by estimating the amount of transverse etching using the spectrum of the wafer reflected light. Patent Document 1 and Patent Document 2 do not disclose an awareness of the problem regarding the change in the spectrum attributable to such a difference in the total height of the stacked film.

[0011] The present invention provides a plasma processing method and a plasma processing apparatus that solve the above-mentioned problem, accurately estimate the amount of etching in the transverse direction, and determine the endpoint based on the estimated amount of etching. means of solving the problem

[0012] To solve the above problem, a representative plasma processing method of the present invention is a plasma processing method for plasma processing a film structure disposed on the surface of a wafer, wherein a film to be processed and a film not to be processed are stacked vertically, comprising: a first step of irradiating light onto the wafer; a second step of receiving reflected light reflected from the wafer at a predetermined plurality of times during plasma processing of the wafer; a third step of performing signal processing on light quantity data for each of the plurality of wavelengths of the received reflected light; a fourth step of detecting the amount of etching in the transverse direction of the film to be processed during plasma processing using the data to be processed after signal processing; and a fifth step of determining the end point of the plasma processing based on the amount of etching. In the second step, the reflected light that has passed through the film not to be processed is received, and in the third step, the light quantity data has a vibration component whose magnitude increases or decreases with a plurality of maximum values ​​or a plurality of minimum values ​​with respect to changes in the plurality of wavelengths or wavenumbers. Signal processing includes a seventh step of removing vibration components of a frequency greater than or equal to a predetermined frequency from the light quantity data, or an eighth step of detecting a lower envelope or an upper envelope of the light quantity data. Effects of the invention

[0013] According to the present invention, it is realized to accurately estimate the amount of etching in the transverse direction and to determine the endpoint based on the estimated amount of etching. Other problems, configurations, and effects other than those mentioned above will become apparent from the description of the embodiments below. Brief explanation of the drawing

[0014] FIG. 1 is a diagram schematically showing the configuration of a plasma processing device according to a first embodiment. FIG. 2 is a diagram showing the configuration of the functional blocks of a digital signal processing unit. FIG. 3 is a cross-sectional view schematically showing the membrane structure of the subject to treatment. FIG. 4 is a diagram showing the amount of light obtained during the etching process of a target. Figure 5 is a diagram showing spectra when the SiO2 film thicknesses are different. FIG. 6 is a diagram showing an example of a spectrum with data formatted. FIG. 7 is a diagram showing the result of comparing the processed data, on which signal processing has been performed, with the pattern data. FIG. 8 is a diagram showing the case using the lower envelope. FIG. 9 is a drawing showing the case using the upper envelope. Specific details for implementing the invention

[0015] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Furthermore, the present invention is not limited by these embodiments. Also, in the description of the drawings, the same parts are indicated by the same reference numerals.

[0016] In the present disclosure, “upward” means the upward direction in a vertical direction relative to the substrate surface of a Si substrate placed on a sample stand, and “downward” means the downward direction. Additionally, “transverse direction” means the direction horizontal to the substrate surface.

[0017] Furthermore, "light quantity data" and "light quantity" include not only direct data on light quantity (light intensity), such as the reflectance of reflected light from the wafer, but also data related to the amount of change in direct data, such as the difference in reflectance.

[0018] [First embodiment]

[0019] With reference to FIGS. 1 and 2, a plasma treatment method and a plasma treatment apparatus according to a first embodiment of the present invention will be described.

[0020] FIG. 1 is a schematic diagram showing the configuration of a plasma processing device according to a first embodiment. The plasma processing device (1) shown in FIG. 1 (a) comprises a vacuum processing chamber (10), a light source unit (18), an optical system (50), a detection unit (28), an etching amount calculation unit (30), and a control unit (40).

[0021] The vacuum processing chamber (10) generates plasma (12) by exciting and decomposing etching gas introduced from a gas introduction means (not shown) inside using power or microwaves generated by a high-frequency power source (not shown). By this plasma (12), a processing target (16), such as a semiconductor wafer installed on a sample stand (14), is etched (plasma processed). The control unit (40) performs the introduction of gas into the vacuum processing chamber (10), the generation and control of plasma (12), and the application of voltage to the processing target (16) by a high-frequency power source, etc., thereby synchronizing and adjusting the timing between each device so that the desired etching process is realized. When the plasma (12) is pulsed, the control of the pulsed plasma is also performed by the control unit (40). At this time, the plasma (12) is pulsed by modulating the on / off of the voltage application and microwave irradiation by a high-frequency power source, etc., that plasmaizes the etching gas. In addition, the plasma is pulsed by time modulating the introduction of the etching gas.

[0022] The plasma processing device (1) is equipped with a mechanism for measuring the film thickness and depth of a processing target (16). Light emitted from a light source unit (18) is introduced into a vacuum processing chamber (10) through an optical system (50) and an introduction lens (20), and irradiation light (22) is irradiated onto the processing target (16). The light source unit (18) may use continuous light ranging from ultraviolet to infrared, but it is also possible to measure the film thickness and depth using a specific wavelength. Reflected light (24) from the processing target (16) is introduced into a detection unit (28) through a detection lens (26) and an optical system (50).

[0023] The detection unit (28) is configured to use, for example, a spectrometer, to spectrally analyze the introduced light and detect the amount of light for each wavelength. When measuring film thickness and depth using a specific wavelength, the detection unit is not limited to a spectrometer and may use a photodetector, etc. In this case, if the light introduced to the detection unit (28) is only the desired specific wavelength, a photodetector can be used directly, and if continuous light is introduced, a mechanism to select only the specific wavelength, such as a monochromator, can be installed in front of the photodetector.

[0024] Here, in FIG. 1(a), the introduction lens (20) for introducing light into the vacuum processing chamber (10) and the detection lens (26) for detecting reflected light are installed so that their positions do not overlap. In this configuration, in order to detect reflected light (24) most efficiently, it is preferable to install the introduction lens (20) and the detection lens (26) at an angle so that they face each other on the same light path consisting of the irradiating light (22) and the reflected light (24) with the processing target (16) as the reflective surface.

[0025] The configuration of the introduction lens (20) and the detection lens (26) is not limited to that of FIG. 1 (a), and as a fully coaxial configuration, the introduction lens (20) and the detection lens (26) may be combined into a single lens. In this case, it is preferable to have the direction of the light rays of the lens perpendicular to the processing target (16) so as to have a configuration capable of detecting the vertical reflected light obtained as a result of vertical irradiation. Additionally, FIG. 1 (a) describes a system for introducing a pair of irradiation light (22) and a system for detecting reflected light (24), but when measuring film thickness and depth at multiple locations of the processing target (16), multiple measurement systems may be installed.

[0026] Although FIG. 1 (a) describes the case where light is incident from an external light source unit (18) as a light source, the light source unit (18) does not need to be used when using light from a plasma (12) as a light source. Even when using a plasma (12) as a light source, the light emitted from the plasma (12) is reflected by the target to be processed (16), and the reflected light (24) is detected in the same way as when using a light source unit (18). The data from the detection unit (28) is introduced into the etching amount calculation unit (30) to determine the film thickness and depth.

[0027] FIG. 1(b) is a diagram showing the configuration of the etching amount calculation unit (30). FIG. 1(b) is a block diagram showing the configuration of the etching amount calculation unit (30) shown in FIG. 1(a) divided into blocks for each part that performs a function, and showing the exchange or flow of data or information between them with lines or arrows.

[0028] The time series data (D1) of the amount of light of each wavelength, which is output from the detection unit (28) and introduced into the etching amount calculation unit (30), is processed by the digital signal processing unit (100) to remove and correct various noises and fluctuations, and is supplied to the waveform comparator (102) as time series data (D2).

[0029] In the first embodiment, time series data (D2) output from the digital signal processing unit (100) is received by a waveform comparator (102), and in the waveform comparator (102), a comparison is performed using an arithmetic unit with at least one pattern data representing the correlation between the etching amount and the light intensity of each wavelength, which is previously acquired as data stored in the waveform pattern database (122). In the waveform comparator (102), the data (D2)(i) of the time series data (D2) at each sampling time (i) is compared with the pattern data having a wavelength as a parameter, in which the light intensity values ​​of multiple wavelengths correspond to multiple values ​​of the etching amount or time after the start of processing in the waveform pattern database (122), and among the pattern data of the light intensity (light intensity) of multiple wavelengths for each time after the start of processing, the one with the smallest difference from the data (D2)(i) is detected as the closest pattern data. In this way, pattern matching is performed by detecting the pattern data closest to the data.

[0030] As for the pattern data with the smallest difference, for example, one can be used where the standard deviation between multiple wavelength data is minimized. The etching amount corresponding to this closest pattern data is calculated as the etching amount for the corresponding sampling time (i). The etching amount for each sampling time (i) calculated by the waveform comparator (102) is transmitted to the etching amount memory unit (104) and output as time series data (D3)(i), and is stored in a memory device such as a hard disk, semiconductor RAM, or ROM that is connected to the etching amount memory unit (104) for data communication.

[0031] The light quantity data for each wavelength in the waveform pattern database (122) is data processed by signal processing performed in the digital signal processing unit (100), and it is preferable that the signal processing be the same as that of the time series data (D2), but they may be different. Here, if there are multiple databases of pattern data for the etching amount and the light quantity for each wavelength in the waveform pattern database (122), the film thickness and depth (D3) determined using each database may be supplied to the etching amount memory unit (104).

[0032] The etching amount memory unit (104) outputs time series data (D4) of the etching amount to the etching amount correction unit (106).

[0033] The etching amount correction unit (106) can correct the etching amount at each time point based on the time series trend of the calculated etching amount. For example, if there is a fluctuation caused by noise or the like in the time series trend of the calculated etching amount, the etching amount at each time point is corrected by linearly approximating the time trend of the etching amount. The etching amount corrected by the etching amount correction unit (106) outputs etching amount data to the outside.

[0034] FIG. 2 is a diagram showing the configuration of a functional block of a digital signal processing unit (100). In the digital signal processing unit (100), the spectrum input to the digital signal processing unit (100) from the detection unit (28) undergoes signal processing in the pre-processing unit (202) for the purpose of removing noise or offset and correcting light intensity fluctuations. For example, a low-pass filter is used to remove noise on the time axis of each wavelength. For example, a second-order Butterworth type low-pass filter may be used, and the time series data (D2) is obtained by the following equation.

[0035]

[0036] Here, Dk(i) represents the data at time i of each data Dk, and the coefficients b and a differ in value depending on the sampling frequency and cutoff frequency. Additionally, the coefficient differences of the digital filter are, for example, a2=-1.143, a3=0.4128, b1=0.067455, b2=-0.013491, b3=0.067455 (sampling frequency 10Hz, cutoff frequency 1Hz). When removing the light intensity offset of each wavelength and observing the temporal change in light intensity, signal processing that calculates the amount of change in light intensity or the derivative value between times can be used. For example, the time series data output by using the Savitzky-Golay Method (SG) along the time axis becomes the derivative value. This derivative value is calculated using the polynomial-fit smoothed derivative method and is calculated by the following equation.

[0037]

[0038] Here, regarding the weighting factor wj, for the first derivative calculation, for example, w-2=-2, w-1=-1, w0=0, w1=1, w2=2 is used. Also, for the second derivative calculation, for example, w-2=2, w-1=-1, w0=-2, w1=-1, w2=2 is used.

[0039] In addition, for data Dk(i) at a given sampling time, if the light intensity values ​​of all wavelengths of the detected target among the data are changing at the same rate over time, a process can be applied to normalize the light intensity values ​​of each wavelength by the average value or the sum of the absolute values ​​of all corresponding wavelengths.

[0040] The signal output from the preprocessing unit (202) forms data points in the wavelength direction in the data formatting unit (204). For example, when digital signal processing is performed based on the wave axis (inverse of the wavelength) in the subsequent wavelength direction signal processing unit (206), the spectrum at each time point converts the wavelength into wavenumbers and resamples the spectrum data so that they are equally spaced along the wave axis. For example, spline interpolation is used for resampling. Additionally, for example, when digital signal processing is performed based on the wavelength axis in the wavelength direction signal processing unit (206), the spectrum at each time point resamples the spectrum data so that they are equally spaced along the wavelength axis.

[0041] The signal output from the data formatting unit (204) undergoes signal processing in the wavelength direction by the wavelength direction signal processing unit (206). For example, when removing the vibration component of the light quantity in the wavelength direction, an LPF is applied along the wavelength axis. Although an LPF can be applied along the wavelength axis in this manner, if vibrations in the light quantity along the wavelength axis or wave axis occur due to optical interference, the light quantity vibrates at a frequency close to the wave axis, so applying an LPF along the wave axis may be more effective for removing vibrations.

[0042] Additionally, for example, to detect the envelope of the vibrational components of light intensity along the wavelength direction, a Hilbert transform or peak or bottom detection is performed with respect to the wavelength axis or the wave axis. In the Hilbert transform, if the light intensity vibrates at a frequency close to the wave axis, it is preferable to perform signal processing along the wave axis rather than the wavelength axis, as described above. On the other hand, for envelope detection using peak or bottom detection, since the data does not need to be equally spaced along the wavelength axis or the wave axis, either the wavelength axis or the wave axis may be used.

[0043] The signal output from the wavelength direction signal processing unit (206) is supplied to the post-processing unit (208). In the post-processing unit (208), if there is signal processing such as an LPF for removing noise in the time direction of the input signal or noise removal not performed in the pre-processing unit (202), such processing is performed. Additionally, since the spectrum of each time point is processed separately in the wavelength direction signal processing unit (206), spectrum smoothing processing between each time point is performed for the purpose of ensuring their temporal continuity. The signal (D2) that has undergone these signal processing is output from the digital signal processing unit (100) and input to the waveform comparator (102).

[0044] The plasma processing device shown in FIG. 1 (a) performs an endpoint determination using a signal indicating an etching amount output from an etching amount calculation unit (30). That is, an endpoint determination device that receives a signal from the etching amount calculation unit (30) compares the etching amount indicated by the signal with a predetermined target etching amount, and if it is determined to be within a predetermined allowable range, it is determined that the processing has reached the endpoint; if it is outside the allowable range, it is determined that it has not reached the endpoint. When it is determined that the target etching amount has been reached, the reaching is reported by a reporting device such as a monitor, lamp, or signal device not shown, and at the same time, a control unit (40) that receives the signal indicating the reaching sends a signal to the plasma processing device to stop the etching process or change the processing conditions.

[0045] In the plasma processing device (1), based on the received etching stop signal, the etching process of the target film layer of the processing target (16) that detected etching is stopped, or, after changing the processing conditions, the processing process for the next processing target (16) is performed. Through this operation, the plasma processing device (1) can perform an endpoint determination operation using an etching amount monitor.

[0046] Next, with reference to FIG. 3, a film structure to be etched while detecting the amount of etching using a plasma processing apparatus according to the first embodiment will be described. FIG. 3 is a cross-sectional view schematically showing the film structure to be processed.

[0047] In the film structure of the processing target (16), an oxide film (301) and a metal film (302) are alternately stacked in the vertical direction on a Si substrate (303). As shown in FIG. 3, this film structure has a trench (304) which is a groove or hole-shaped structure formed in the vertical direction (vertical direction) in the central part, and has a side wall surface (305) facing the inside of the trench (304).

[0048] In the plasma treatment apparatus (1) of the first embodiment, in the etching process performed on the treatment target (16), the metal film (302) of the film structure is selectively etched in the transverse direction from the surface of the side wall surface (305) facing the trench (304). FIG. 3 (a) shows the state before the start of the etching process. Since it is in an unetched state, the position of the end of the metal film (302) facing the trench (304) is the same as the position of the end of the oxide film (301) adjacent to the upper or lower side. Meanwhile, FIG. 3 (b) shows the state after the etching process has started and proceeded. The portion of the metal film (302) facing the trench (304) is removed, and a recess is formed in the trench (304) in the transverse direction relative to the end of the oxide film (301).

[0049] Thus, the etching process of the first embodiment is to remove a desired amount of the metal film (302) from the side wall surface (305) prior to the etching process, thereby making it concave (retracted), with respect to a multilayer stacked film structure in which an oxide film (301) and a metal film (302) are stacked in the vertical direction. Here, in the first embodiment, the film thickness of the oxide film (301) and the metal film (302) is 25 nm each, and the number of stacked metal films (302) is 100 layers. Because of this, the total height of the multilayer film is a very thick structure of 5 μm or more. The width of the trench (304) is 200 nm, and the trench (304) is formed on the Si substrate (303) with a pitch of 1 μm.

[0050] Next, with reference to Figures 4 and 5, the influence of film thickness on the etching amount will be explained.

[0051] Figure 4 shows an example of a spectrum detected from light obtained from the surface of a processing target (16) placed in the vacuum processing chamber (10) of the plasma processing apparatus (1) of the present embodiment. Figure 4 is a diagram showing the amount of light obtained during the etching process of the processing target.

[0052] FIG. 4(a) shows the amount of light of reflected light (24) from the surface of the target (16) to be processed when the amount of etching at the start of the processing of the metal film (302) is 25 nm and 30 nm, as a reflectance spectrum at multiple wavelengths. This corresponds to the data of D1. Here, the amount of etching of the metal film (302) is defined as the depth value of the concave portion in the transverse direction of the metal film (302) from the end edge facing the trench (304) of the oxide film (301), that is, based on the position of the end of the metal film (302) facing the trench (304) before the start of the etching process, which is set to 0, and the transverse direction in which the etching proceeds is taken as positive. As shown in FIG. 4(a), it can be confirmed that the spectrum changes as the etching proceeds.

[0053] The reflectance spectrum shown in Fig. 4(a) appears to change smoothly as the wavelength changes, but in fact, from this figure, it can be seen that the spectrum obtained during the etching of the stacked film structure vibrates in the direction of the wavelength axis. This is because, in the film structure of the semiconductor wafer, there is a region where light can be transmitted to the Si substrate (303) part, such as a repeating structure between the vacuum part where the oxide film and the metal film have been removed or a trench groove, and the height of this region is several μm and is very large.

[0054] When light reflected from the Si substrate (303) after passing through such a few micrometers interferes with light reflected from the surface of the top layer oxide film, their difference in optical path length varies depending on the wavelength. Since the change in the difference in optical path length due to this wavelength increases in proportion to the height of the stacked portion of the oxide film and vacuum or the trench, the strengthening and weakening of interference in the wavelength direction is repeatedly observed in this film structure, and as a result, the amount of light (intensity) in the wavelength direction is observed to oscillate.

[0055] As the etching progresses, the reflectance spectrum changes. FIG. 4(b) is the result of calculating the amount of light data of reflected light (24) using the change in reflectance spectrum (hereinafter also referred to as "spectral difference" or "light amount difference"). It corresponds to the output data of the preprocessing unit (202). The spectrum of the solid line is the spectrum difference when the metal film (302) changes from 20 nm to 25 nm, and the spectrum of the dashed line is the spectrum difference when the metal film (302) changes from 25 nm to 30 nm. Since the change in the spectrum difference can also be confirmed as the etching progresses, it can be seen that the amount of etching of the metal film (302) can be estimated by using the spectra and spectrum difference of FIG. 4(a) and FIG. 4(b).

[0056] Here, since the laminated film of FIG. 3 (a) has 100 layers, when manufacturing this structure, variations in the film thickness of the oxide film (301) or metal film (302), etc., occur in the height direction. For example, if the oxide film increases by 1 nm, the height of the entire laminated film changes by 100 nm.

[0057] Figure 5 is a diagram illustrating spectra when the SiO2 film thickness is different. Figure 5 shows the spectral difference observed when the SiO2 film thickness is 25 nm and when it becomes 26 nm due to film deposition deviation. The spectral difference represented by the solid line is for a SiO2 film thickness of 25 nm, and the spectral difference represented by the dashed line is for a SiO2 film thickness of 26 nm. In both cases, the etching amount of the metal film is 25 nm, and the spectral difference is represented as the difference between an etching amount of 20 nm and 25 nm. It can be seen that the spectrum changes as the thickness of the SiO2 film changes, even though the etching amount of the metal film is the same. For this reason, when estimating the etching amount from the spectrum, it is difficult to detect the etching amount by pattern matching of the spectrum with high precision, as the total height of the stacked film varies between wafers, by only registering pattern data, which is a comparison target data having a correlation between a specific spectrum and the etching amount, in the waveform pattern database (122) of FIG. 1 (b).

[0058] Next, referring to FIGS. 6 and FIGS. 7, a process for detecting the etching amount of the film structure of the target to be processed will be described.

[0059] In the following description, the etching amount is detected by the etching amount calculation unit (30). Additionally, the wafer is etched, and the etching amount of the metal film is estimated. Also, the thickness of the SiO2 film of the film structure formed on the surface of the processing target (16) is 25 nm.

[0060] The light quantity data processed in the preprocessing unit (202) is transmitted to the data formatting unit (204), and signal processing is performed as needed and the spectrum data for each time point is resampled. In this embodiment, for the spectrum of each sampling time point, the wavelength axis is converted into wavenumbers, and the spectrum data is resampled using spline interpolation to divide the wavenumber axis into 512 equal intervals of wavenumbers 1 / 300 nm to 1 / 900 nm (the denominator is the wavelength) to generate data formatted spectrum data.

[0061] FIG. 6 is a diagram showing an example of a data-formed spectrum. FIG. 6(a) shows an example of a spectrum difference in which the horizontal axis is converted into wavenumber in the data-formed section (204). The horizontal axis represents the wavenumber, and the vertical axis represents the difference in reflectance derived based on the difference in light intensity, and represents the spectrum difference between the etching amount of the metal film of 20 nm and 25 nm at a SiO2 film thickness of 25 nm. As shown in FIG. 6(a), the vibration in the wavelength direction vibrates like a sinusoidal wave at a frequency that is almost constant with respect to the wavenumber axis. In other words, the spectrum vibrates with an amplitude that includes multiple maximum or minimum values ​​with respect to changes in wavenumber or wavelength.

[0062] The signal representing the spectrum data after such data formatting is transmitted to the wavelength direction signal processing unit (206) for further signal processing. In the wavelength direction signal processing unit (206), for the spectrum data in which light intensity values ​​are arranged at equal intervals along the wave axis, removal of high-frequency components by an LPF, or envelope detection by Hilbert transform or peak / bottom detection is performed.

[0063] In the first embodiment, an LPF is performed in the wavelength direction, and vibration components in the wavelength direction are removed.

[0064] FIG. 6(b) shows an example of a spectrum difference in which wavelength-direction vibrations are removed in the wavelength-direction signal processing unit (206). In the example of the spectrum difference shown in FIG. 6(a), the vibration component is at almost a constant frequency, so it can be seen that the vibration component can be sufficiently removed by the LPF in FIG. 6(b).

[0065] Here, the cutoff frequency of the LPF was set to half the vibration frequency based on the vibration frequency of Fig. 6(a). The cutoff frequency should be less than or equal to the vibration frequency, but if it is too low, not only high-frequency components but also low-frequency components are removed, which may distort the spectrum shape. Therefore, it is desirable to set the cutoff frequency to a level that can remove the vibration frequency.

[0066] The signal-processed data in the wavelength direction is input to the post-processing unit (208), and light intensity smoothing in the time direction for each wavelength is performed. For example, in the wavelength direction signal processing unit (206), signal processing is performed independently for the spectrum of each time point, and as a result, the change in light intensity between times of each wavelength may become discontinuous. In that case, an LPF is applied to the light intensity of each wavelength in the time direction, or a moving average is calculated. In this embodiment, for the spectrum of each time point, data smoothing is performed by a moving average of the past 1 second.

[0067] The data to be processed obtained by performing these processes is transmitted to a waveform comparator (102) and compared with data stored in a waveform pattern database (122) to perform pattern matching. In the waveform pattern database (122), pattern data is registered in which the etching amount of a metal film (302) corresponds to a spectrum of a predetermined plurality of wavelengths obtained by etching a processing target (16) having a film structure shown in FIG. 3 in advance. In this embodiment, a pattern data corresponding to a stacked film with an SiO2 film thickness of 24 nm is used, and a pattern data corresponding to a film structure with a different configuration from the SiO2 film thickness of 25 nm of the actual processing target (16) on which the etching process is performed is used.

[0068] Meanwhile, the pattern data is composed of a spectrum obtained by performing the same signal processing as the signal processing performed on the spectrum obtained from the actual processing target (16) on which the etching process is being performed. In the waveform comparator (102), the spectrum data obtained at any sampling time during processing is compared with the spectrum of the pattern data stored in the database, and the etching amount corresponding to the pattern data with the smallest difference as a result of pattern matching is detected as the etching amount at that time.

[0069] Here, regarding the data of a time series spectrum obtained at any time or multiple times during processing, if data processing is not performed in the data formatting unit (204), wavelength direction signal processing unit (206), and post-processing unit (208) shown in FIG. 2, the spectrum for the same etching amount of the metal film (302) becomes different values ​​as shown in FIG. 5, and even if high-precision pattern data obtained in advance is pattern-matched with the spectrum data obtained during processing, it is hindered from detecting the etching amount with high precision. On the other hand, in this embodiment, by performing signal processing in the digital signal processing unit (100) shown in FIG. 2, it becomes possible to obtain a spectrum that accurately represents the actual amount of etching even when the overall height of the film structure having multiple stacked films of the processing target (16) changes.

[0070] FIG. 7 is a diagram showing the result of comparing the processed data, on which signal processing has been performed, with the pattern data. FIG. 7 shows the result of comparing the spectrum data detected during plasma processing of the processing target (16) when the etching amount is 25 nm with the spectrum of the pattern data stored in the waveform pattern database (122). In FIG. 7, the solid line represents the spectrum difference data with the wavelength obtained during processing of the processing target (16), which has an oxide film (301) thickness of 25 nm, as the horizontal axis, and the dashed line represents the spectrum difference data of the pattern data with an oxide film (301) thickness of 26 nm stored in the waveform pattern database (122). Both spectrum differences represent the difference from the spectrum in the case of a predetermined etching amount. As shown in FIG. 7, the distribution of the two spectral differences matches with high precision, so that the data of the spectral difference obtained using the configuration shown in the present embodiment can be matched with the pattern data stored in the waveform pattern database (122) with high precision and the amount of etching can be detected.

[0071] The etching amount detected in this way is processed in the etching amount memory unit (104) and the etching amount correction unit (106), and then input into the control unit (40) which is connected to communicate with the plasma processing device as data representing the etching amount, and the data is stored in the internal memory device. The control unit (40) determines whether the etching amount at the sampling time indicated by the data has reached the target value, and if it is determined that it has reached it, it sends a command to stop the etching process to the plasma processing device, and the etching process of the processing target (16) that is undergoing the etching process is stopped and terminated.

[0072] (Actions / Effects)

[0073] As described above, in the first embodiment, even when the total height of the film structure, in which a plurality of film layers formed in advance on the surface of the processing target (16) are stacked, varies between the plurality of processing targets (16), signal processing is performed using light quantity data of the spectrum obtained from the reflected light (24) from the surface of the processing target (16), thereby enabling the detection of the transverse etching amount of the metal film (302) of the processing target with high precision, and making it possible to accurately determine the end point of the etching process of the processing target (16).

[0074] Here, in the first embodiment, the factors of deviation in the film structure of the processing target (16), the material of the film layer, and the height of the stacking are examples, and it is possible to use this to maintain a constant correlation between the etching amount and the spectrum in cases where the vibration frequency or amplitude of the spectrum vibrating in the wavelength direction varies due to deviations in the structure or material other than the part to be etched. For example, it is also applicable to variations in the spectrum when the material of the film layer is non-uniform and there is a difference in the refractive index. In addition, although pattern matching using spectral difference was used as a method for determining the etching amount in this embodiment, the method for determining the etching amount from the spectrum is not limited to this embodiment, and the etching amount may be determined using light quantity data of a specific wavelength of the spectrum or feature quantity data extracted from the spectrum.

[0075] In addition, the amount of etching was determined using the spectral difference (amount of change in the spectrum), but the present invention is not limited thereto. For example, it is also possible to determine the amount of etching by comparing the spectrum data obtained from reflected light with the spectrum of pattern data. In addition, data obtained by converting the wavelength to the wavenumber with respect to the spectrum data obtained from reflected light was used, but the present invention is not limited thereto. For example, it is also possible to determine the amount of etching using the data as the wavelength itself.

[0076] [Second embodiment]

[0077] In the first embodiment above, the data formatting unit (204) performed conversion of the wavelength axis and resampling of the light quantity data. Even without performing these data processings, the amount of etching can be detected with high precision by using lower envelope detection in the wavelength direction signal processing unit (206). In the second embodiment, the configuration for detecting such an amount of etching is described. In the following description, configurations other than the above differences are the same as those described in the embodiments of FIGS. 1 to 7, so descriptions are omitted unless particularly necessary.

[0078] FIG. 8 is a diagram showing the case where a lower envelope is used. FIG. 8(a) shows the result of detecting the lower envelope for spectrum data obtained during the processing of a processing target (16) in which the thickness of the SiO2 film as the oxide film (301) shown in FIG. 5 is 25 nm. The dashed line, which is the original signal, represents a spectrum of multiple wavelengths detected from reflected light obtained at any sampling time during the processing of the processing target (16), and the solid line represents the spectrum data detected as the lower envelope of the spectrum data shown by the dashed line. In the second embodiment, the lower envelope shown in FIG. 8(a) is detected as a line connecting the minimum value (bottom part) of the light amount (parameter indicated by the vertical axis on the drawing) in a specific wavelength range, and is used as spectrum data to detect the amount of etching.

[0079] In the example of FIG. 8(a), the lower envelope was used as spectrum data without changing the wavelength axis or resampling the light quantity data in the data formatting unit (204) of FIG. 2. However, if the wavelength axis is treated as a spectrum, the vibration in the wavelength direction appears as if the frequency is changing, and envelope detection may not function well. In such cases, it is desirable to convert the spectrum to the wave axis and perform envelope detection processing.

[0080] The result of pattern matching in the waveform comparator (102) using the spectrum data of FIG. 8 (a) is shown in FIG. 8 (b). The two spectrum data are for the case where the etching amount is 25 nm, the solid line represents the spectral difference obtained during the processing of the film structure of the target (16) for processing, where the film thickness of the SiO2 film as the oxide film (301) is 25 nm, and the dashed line represents the spectral difference created using the pattern data stored in the waveform pattern database (122) for the case where the film thickness of the SiO2 film as the oxide film (301) is 26 nm. As shown in the figure, it can be seen that the spectrum data corresponding to the same etching amount matches with high precision.

[0081] (Actions / Effects)

[0082] From the above, as in the second embodiment, by using the spectrum obtained using the lower envelope, the amount of etching can be detected with high precision, just like in the first embodiment shown in FIGS. 1 to 7, and thus, endpoint determination can be realized with high precision.

[0083] [Third Embodiment]

[0084] In the second embodiment, an example using the lower envelope is shown, and as a result, even when the upper envelope is detected by the wavelength direction signal processing unit (206) and used as spectrum data for pattern matching, the amount of etching can be detected with high precision. In the third embodiment, the configuration for detecting such an amount of etching is described. In the following description, configurations other than the above differences are the same as those described in the first and second embodiments, so descriptions are omitted unless particularly necessary.

[0085] FIG. 9 is a diagram showing a case using an upper envelope. FIG. 9 shows an example in which an upper envelope is detected for spectrum data obtained during the processing of a film structure in which the film thickness of the SiO2 film is 25 nm as the oxide film (301) shown in FIG. 6 (a). The dashed line, which is the original signal, represents a spectrum of multiple wavelengths detected from reflected light obtained at any sampling time during the processing of the target (16), and the solid line represents the spectrum data detected as the upper envelope of the spectrum data represented by the dashed line.

[0086] In the detection of the upper envelope shown in Fig. 9, an LPF, a High Pass Filter (HPF), and a Hilbert transform are used. First, for the same spectrum data shown in Fig. 6 (a), the result of low-pass filtering using the LPF and the result of high-pass filtering using the HPF are obtained, respectively, by the LPF and HPF with the same cutoff frequency. After high-pass filtering by the HPF, the low-frequency components of the spectrum data are removed, and only vibrations in the wavenumber direction are extracted.

[0087] In addition, in the third embodiment, a Hilbert transform is performed on such high-pass filtered spectrum data, and the envelope of the vibrational component is calculated as a spectrum. The upper envelope is detected by combining the calculated envelope spectrum with the spectrum of the low-frequency component after low-pass filtering.

[0088] The upper envelope obtained in this manner is used as spectrum data for detecting the etching amount, and the result of pattern matching with the spectrum of the pattern data stored in the waveform pattern database (122) in the waveform comparator (102) is shown in FIG. 9 (b). The two spectrum data are for the case where the etching amount is 25 nm, the solid line represents the spectrum difference obtained during the processing of the film structure of the target (16) in which the film thickness of the SiO2 film as the oxide film (301) is 25 nm, and the dashed line represents the spectrum difference created using the pattern data stored in the waveform pattern database (122) in which the film thickness of the SiO2 film as the oxide film is 26 nm. As shown in the figure, it can be seen that the spectrum data corresponding to the same etching amount matches with high precision.

[0089] (Actions / Effects)

[0090] As described above, by using the spectrum obtained using the upper envelope as in the third embodiment, the amount of etching can be detected with high precision, just like in the first embodiment shown in FIGS. 1 to 7, and thus, endpoint determination can be realized with high precision.

[0091] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and various modifications are possible within the scope of not departing from the gist of the present invention.

[0092] For example, in the second embodiment, the lower envelope was used, but the upper envelope may also be used as spectral data to detect the amount of etching, and in the third embodiment, the upper envelope was used, but the lower envelope may also be used as spectral data to detect the amount of etching. Explanation of the symbols

[0093] 1… Plasma processing unit 10… Vacuum processing chamber 12… Plasma 14… Sample stand 16… Target for processing 18… Light source 20… Introduction lens 22… Illumination light 24… Reflected light 26… Detection lens 28… Detector 30… Etching amount calculation unit 40… Control unit 50… Optical system 100… Digital signal processing unit 102… Waveform comparator 104… Etching amount memory unit 106… Etching amount correction unit 122… Waveform Pattern Database 202… Preprocessing Unit 204… Data Shaping Unit 206… Wavelength Direction Signal Processing Unit 208… Post-processing unit 301… Oxide film 302… Metal film 303… Si substrate 304… Trench 305… Sidewall D1… Time series data output from the detection unit D2… Time series data output from the digital signal processing unit D3… Etching amount data output from the waveform comparator D4… ​​Etching amount data output from the etching amount memory unit

Claims

Claim 1 A plasma processing method for plasma processing a film structure disposed on the surface of a wafer, wherein a film to be processed and a film not to be processed are stacked vertically, comprising: a first step of irradiating light onto the wafer; a second step of receiving reflected light reflected from the wafer at a predetermined plurality of times during plasma processing of the wafer; a third step of performing signal processing on light quantity data for each of the plurality of wavelengths of the received reflected light; a fourth step of detecting the amount of etching in the transverse direction of the film to be processed during plasma processing using the data to be processed after signal processing; and a fifth step of determining the end point of the plasma processing based on the amount of etching. In the second step, the reflected light that has passed through the film not to be processed is received. In the third step, the light quantity data has a vibration component whose magnitude increases or decreases with a plurality of maximum values ​​or a plurality of minimum values ​​with respect to changes in the plurality of wavelengths or wavenumbers. The signal processing includes a seventh step of removing the vibration component with a frequency greater than or equal to a predetermined frequency from the light quantity data. Plasma treatment method. Claim 2 A plasma processing method for plasma processing a film structure disposed on the surface of a wafer, wherein a film to be processed and a film not to be processed are stacked vertically, comprising: a first step of irradiating light onto the wafer; a second step of receiving reflected light reflected from the wafer at a predetermined plurality of times during plasma processing of the wafer; a third step of performing signal processing on light quantity data for each of the plurality of wavelengths of the received reflected light; a fourth step of detecting the amount of etching in the transverse direction of the film to be processed during plasma processing using the data to be processed after signal processing; and a fifth step of determining the end point of the plasma processing based on the amount of etching. In the second step, the reflected light that has passed through the film not to be processed is received. In the third step, the light quantity data has a oscillating component whose magnitude increases or decreases with a plurality of maximum values ​​or a plurality of minimum values ​​with respect to changes in the plurality of wavelengths or wavenumbers. The signal processing includes an eighth step of detecting the lower envelope or the upper envelope of the light quantity data. Plasma treatment method. Claim 3 A plasma processing method according to claim 1 or 2, wherein the signal processing comprises a sixth step of converting the plurality of wavelengths into wavenumbers and interpolating and resampling the light quantity data such that the wavenumber intervals are equal on the axis of the plurality of wavenumbers. Claim 4 A plasma treatment method according to claim 1 or 2, wherein in the fourth step, the amount of etching of the wafer is determined from the data to be treated during the plasma treatment by using pattern data in which the amount of etching corresponds to the data to be treated obtained in advance before performing the plasma treatment. Claim 5 A plasma treatment method according to claim 1 or 2, wherein the membrane structure comprises a multilayer film having an insulating film formed on the surface and a film to be treated for plasma treatment including a metal, which are alternately stacked in the vertical direction. Claim 6 A plasma treatment method according to claim 5, wherein the insulating film is composed of a material including silicon oxide, and the surface of the target film is composed of a material that reflects light. Claim 7 A plasma processing apparatus for plasma processing a film structure disposed on the surface of a wafer, wherein a film to be processed and a film not to be processed are stacked vertically, the apparatus comprises: a digital signal processing unit that performs signal processing on light quantity data for each of a plurality of wavelengths obtained from reflected light from the wafer in relation to irradiation light; a waveform comparator that detects the amount of etching in the transverse direction of the film to be processed during plasma processing using the processed data that has undergone signal processing; and a control unit that determines the end point of the plasma processing based on the amount of etching. The reflected light is that which has passed through the film not to be processed in the film structure, and the light quantity data has a vibration component whose magnitude increases or decreases with a plurality of maximum values ​​or a plurality of minimum values ​​with respect to a change in the plurality of wavelengths or wavenumbers. The digital signal processing unit has a wavelength direction signal processing unit that removes the vibration component above a predetermined frequency from the light quantity data. Claim 8 A plasma processing apparatus for plasma processing a film structure disposed on the surface of a wafer, wherein a film to be processed and a film not to be processed are stacked vertically, the apparatus comprises: a digital signal processing unit that performs signal processing on light quantity data for each of a plurality of wavelengths obtained from reflected light from the wafer in relation to irradiation light; a waveform comparator that detects the amount of etching in the transverse direction of the film to be processed during plasma processing using the processed data that has undergone signal processing; and a control unit that determines the end point of the plasma processing based on the amount of etching. The reflected light is that which has passed through the film not to be processed in the film structure, and the light quantity data has a vibration component whose magnitude increases or decreases with a plurality of maximum values ​​or a plurality of minimum values ​​with respect to a change in the plurality of wavelengths or wavenumbers. The digital signal processing unit has a wavelength direction signal processing unit that detects the lower envelope or the upper envelope of the light quantity data. Claim 9 A plasma processing device according to claim 7 or 8, wherein the digital signal processing unit converts the plurality of wavelengths into wavenumbers and interpolates and resamples the light quantity data such that the wavenumber intervals are equal on the axis of the plurality of wavenumbers. Claim 10 A plasma processing apparatus according to claim 7 or 8, wherein the waveform comparator determines the etching amount of the wafer from the data to be processed during the plasma processing by using pattern data in which the etching amount corresponds to the data to be processed obtained in advance before the plasma processing is performed. Claim 11 In claim 7 or 8, the above-mentioned membrane structure is a plasma treatment apparatus having a multilayer membrane in which an insulating membrane formed on the surface and a membrane to be treated for plasma treatment comprising a metal are alternately stacked in the vertical direction. Claim 12 A plasma treatment apparatus according to claim 11, wherein the insulating film is composed of a material including silicon oxide, and the surface of the target film is composed of a light-reflecting material. Claim 13 delete Claim 14 delete

Images