Semiconductor process monitoring device and semiconductor process monitoring method

Abstract

In a semiconductor process monitoring apparatus, a spectroscopic unit performs wavelength decomposition of measurement light from a chamber in a first direction and forms a spectroscopic image of each wavelength in a second direction intersecting the first direction, and a detection unit has a first pixel region and a second pixel region divided in the second direction. The determination unit determines the presence or absence of an abnormality in the semiconductor process on the basis of first spectral data obtained at a first exposure time in the first pixel region, and determines the end point of the semiconductor process on the basis of second spectral data obtained at a second exposure time longer than the first exposure time in the second pixel region.

Description

Technical Field

[0001] The present disclosure relates to a semiconductor process monitoring device and a semiconductor process monitoring method. Background Art

[0002] Spectrometry is a technique for detecting a spectral image of measurement light generated by an object and analyzing the object based on the spectral data of the spectral image. In spectrometry, it is sometimes necessary to obtain a spectrum with a high dynamic range, depending on the type of object, etc. (see Patent Document 1). For example, in the process of dry etching an object using a plasma process, light is generated by the etching gas and also by the material of the object being etched. In such semiconductor processes, the wavelength band of the light originating from the gas tends to differ from the wavelength band of the light originating from the material. In addition, the intensity of the light originating from the material tends to be weaker than the intensity of the light originating from the gas. [Prior art literature] [Patent Document]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2020-118477 Summary of the Invention [Problems to be solved by the invention]

[0004] With the recent miniaturization and increasing stacking of semiconductor devices, the intensity of the measurement light generated by the target during semiconductor manufacturing processes has tended to become weaker. Therefore, in order to accurately monitor semiconductor manufacturing processes, it is important to measure the intensity of the measurement light with a high dynamic range and high accuracy.

[0005] An object of the present disclosure is to provide a semiconductor process monitoring device and a semiconductor process monitoring method that can monitor a semiconductor process with high precision by measuring the intensity of measurement light with a high dynamic range and high precision. [Technical means to solve the problem]

[0006] A semiconductor process monitoring device according to one aspect of the present disclosure includes: a spectrometer that spectrometers measurement light from a chamber; a detection unit that detects a spectral image of the measurement light from the spectrometer; a judgment unit that judges the progress of the semiconductor process in the chamber based on data obtained from the detection result of the spectral image of the measurement light; and the spectrometer decomposes the measurement light by wavelength in a first direction and forms a spectral image of each wavelength in a second direction intersecting the first direction; the detection unit has a first pixel area and a second pixel area divided in the second direction; the judgment unit judges whether there is any abnormality in the semiconductor process based on first spectral data obtained in the first pixel area with a first exposure time, and judges the end point of the semiconductor process based on second spectral data obtained in the second pixel area with a second exposure time longer than the first exposure time.

[0007] In this semiconductor process monitoring device, the presence of anomalies in the semiconductor process is determined based on first spectral data obtained in a first pixel region with a first exposure time, and the endpoint of the semiconductor process is determined based on second spectral data obtained in a second pixel region with a second exposure time longer than the first. By combining the detection of spectral images of measurement light in the first and second pixel regions with different exposure times, the intensity of the spectral images of measurement light can be measured with high dynamic range and high precision. Therefore, even in situations where the semiconductor process contains a mixture of high-intensity measurement light originating from gas and low-intensity measurement light originating from materials, the presence of anomalies in the semiconductor process and the endpoint of the process can be monitored with high precision.

[0008] It is feasible to set the first exposure time in the first pixel region to be shorter than the first frame time in the first pixel region, and to set the second exposure time in the second pixel region to be the same as the second frame time in the second pixel region. With this configuration, by combining the detection of the spectral images of the measurement light in the first and second pixel regions with different exposure times, the intensity of the spectral image of the measurement light can be measured with high dynamic range and high precision. Therefore, even in a situation where high-intensity measurement light originating from gas and low-intensity measurement light originating from materials coexist in a semiconductor process, it is possible to monitor the presence of abnormalities in the semiconductor process and the process endpoint with high precision.

[0009] It is feasible to set the first exposure time in the first pixel region so that the measurement light is not saturated in at least the long-wavelength region of the measurement light, and to set the second exposure time in the second pixel region so that the measurement light is not saturated in at least the short-wavelength region of the measurement light. With this configuration, first spectral data can be acquired in the long-wavelength region without saturating the higher-intensity measurement light originating from gas. Furthermore, second spectral data related to the lower-intensity measurement light originating from materials can be acquired in the short-wavelength region with a good signal-to-noise ratio. This can further improve the accuracy of monitoring the presence of abnormalities and process endpoints in semiconductor manufacturing processes.

[0010] Alternatively, the generation unit can combine the long-wavelength data of the first spectral data with the short-wavelength data of the second spectral data and output the combined data to the determination unit. This configuration allows for the generation of spectral data of the measurement light over a wide dynamic range, encompassing the entire wavelength band of the monitored light. This allows for simple and highly accurate determination of abnormalities and process endpoints.

[0011] The determination unit may determine the presence of an abnormality in the semiconductor process based on the presence of a first peak at a predetermined wavelength in the long-wavelength region of the first spectral data, and determine the endpoint of the semiconductor process when a second peak appears at a predetermined wavelength in the short-wavelength region of the second spectral data. By performing determinations based on the appearance trends of peaks of material-derived and gas-derived measurement light in the semiconductor process, the accuracy of monitoring the presence of abnormalities and the endpoint of the process can be further improved.

[0012] Alternatively, the detection unit may be a CCD photodetector having a first horizontal shift register for transferring charges generated by each column of the first pixel region and a second horizontal shift register for transferring charges generated by each column of the second pixel region. With this configuration, using the CCD photodetector can avoid an increase in readout noise when reading out charges generated by pixels in each column.

[0013] Alternatively, the detection unit may be a CCD photodetector having a first storage unit that stores charges generated by each column of the first pixel region; a second storage unit that stores charges generated by each column of the second pixel region; a first readout unit that outputs an electrical signal for each column corresponding to the magnitude of the charges stored in the first storage unit; and a second readout unit that outputs an electrical signal for each column corresponding to the magnitude of the charges stored in the second storage unit. In this configuration, the use of a CCD photodetector can also avoid an increase in readout noise when reading out the charges generated by pixels in each column.

[0014] The detection unit may be a CMOS photodetector having a first readout unit that outputs an electrical signal corresponding to the magnitude of charge accumulated by each pixel in the first pixel region, and a second readout unit that outputs an electrical signal corresponding to the magnitude of charge accumulated by each pixel in the second pixel region. With this configuration, for example, power consumption can be suppressed to a lower level compared to a CCD photodetector that reads charge for each column.

[0015] The first exposure time of the first pixel area can be controlled by an electronic shutter. This configuration allows for highly accurate adjustment of the first exposure time even when the second exposure time is set to a shorter value and the first exposure time is set sufficiently shorter than the second exposure time. By making the first exposure time sufficiently shorter than the second exposure time, the dynamic range of measurement can be further significantly increased.

[0016] One aspect of the present disclosure provides a semiconductor process monitoring method comprising: a spectroscopic step of spectroscopies measurement light from a chamber; a detection step of detecting a spectral image of the measurement light spectroscopies by the spectroscopic step; and a judgment step of judging the progress of the semiconductor process in the chamber based on data obtained from the detection result of the spectral image of the measurement light; and in the spectroscopic step, the measurement light is wavelength-decomposed in a first direction, and the spectral image of each wavelength is formed in a second direction intersecting the first direction; in the detection step, the spectral image is detected in a light detector having a first pixel area and a second pixel area divided in a second direction; in the judgment step, the presence or absence of an abnormality in the semiconductor process is judged based on first spectral data obtained in the first pixel area with a first exposure time, and the end point of the semiconductor process is judged based on second spectral data obtained in the second pixel area with a second exposure time longer than the first exposure time.

[0017] In this semiconductor process monitoring method, the presence of anomalies in the semiconductor process is determined based on first spectral data obtained in a first pixel region with a first exposure time, and the endpoint of the semiconductor process is determined based on second spectral data obtained in a second pixel region with a second exposure time longer than the first exposure time. This semiconductor process monitoring device combines detection of spectral images of measurement light in the first and second pixel regions with different exposure times, enabling high-dynamic-range and high-precision measurement of the spectral images of measurement light. Therefore, even in situations where high-intensity measurement light originating from gases coexists with lower-intensity measurement light originating from materials, it is possible to accurately monitor the presence of anomalies in the semiconductor process and the endpoint of the process. [Effects of the Invention]

[0018] According to the present disclosure, a semiconductor process monitoring device and a semiconductor process monitoring method can be provided that can monitor a semiconductor process with high precision by measuring the intensity of measurement light with a high dynamic range and high precision. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a block diagram showing the structure of a semiconductor process monitoring device according to one embodiment of the present disclosure. Figure 2 To display Figure 1 The diagram shows the structure of the detection unit. Figure 3 A diagram showing an example of a CCD photodetector constituting a detection unit. Figure 4 (a) is displayed Figure 3 1 is a timing chart showing an example of the operation of the first pixel region of the detection unit. Figure 4 (b) shows Figure 3 1 is a timing chart showing an example of the operation of the second pixel region of the detection unit shown. Figure 5 This is a diagram showing another example of a CCD photodetector constituting the detection section. Figure 6 This is a diagram showing another example of a CCD photodetector constituting the detection section. Figure 7 (a) is displayed Figure 5 and Figure 6 1 is a timing chart showing an example of the operation of the first pixel region of the detection unit. Figure 7 (b) shows Figure 5 and Figure 6 1 is a timing chart showing an example of the operation of the second pixel region of the detection unit shown. Figure 8 A diagram showing an example of a CMOS photodetector constituting a detection unit. Figure 9 (a) is displayed Figure 8 1 is a timing chart showing an example of the operation of the first pixel region of the detection unit. Figure 9 (b) shows Figure 8 1 is a timing chart showing an example of the operation of the second pixel region of the detection unit shown. Figure 10 FIG. 1 is a flow chart showing a semiconductor process monitoring method according to an embodiment of the present disclosure. Figure 11 It is a detailed flowchart showing a semiconductor process monitoring method according to one embodiment of the present disclosure. Figure 12 This is a diagram showing an example of the first spectrum data. Figure 13 This is a diagram showing an example of the second spectrum data. Figure 14 A diagram for explaining generation of spectral data regarding measurement light. Figure 15 It is a detailed flowchart showing a semiconductor process monitoring method according to one embodiment of the present disclosure. Figure 16 (a) is a graph showing spectral data in a case where there is no abnormality. Figure 16 (b) is a graph showing spectral data in a case where an abnormality occurs. Figure 17 (a) is a graph showing spectrum data for determining whether a semiconductor process is continued. Figure 17 (b) is a graph showing spectral data for determining the end point of a semiconductor process. DETAILED DESCRIPTION

[0020] Hereinafter, preferred embodiments of a semiconductor process monitoring device and a semiconductor process monitoring method according to one embodiment of the present disclosure will be described in detail with reference to the drawings.

[0021] Figure 1 This is a block diagram showing the structure of a semiconductor process monitoring device 1 according to one embodiment of the present disclosure. The semiconductor process monitoring device 1 includes: a light guiding unit 2, a spectroscopic unit 3, a detection unit 4, a generating unit 5, and a judging unit 6. The generating unit 5 and the judging unit 6 are computers 7 that physically include a processor such as a CPU, a storage medium such as a RAM, and a ROM. The computer 7 can be an intelligent device such as a smartphone or a tablet terminal that integrally includes a display unit and an input unit. The computer 7 can be composed of a microcomputer or an FPGA (Field-Programmable Gate Array). The semiconductor process monitoring device 1 generates spectral data of the measurement light L1 arriving from the dry etching device 8 that is the measurement object, and performs control based on the spectral data.

[0022] The dry etching device 8 is a device used in the dry etching process. The dry etching device 8 has a chamber 81 and a monitoring window 82. The monitoring window 82 is, for example, colorless transparent glass and is embedded in the side wall of the chamber 81. The operator of the dry etching device 8 moves the substrate W into the chamber 81, allows the etching gas to flow into the chamber 81, and generates plasma PL. The substrate W is etched by the plasma PL. The measurement light L1 from the dry etching device 8 is output to the outside through the monitoring window 82. The measurement light L1 includes light caused by the gas used for etching and light caused by the material of the substrate W. The details will be described later. The light caused by the gas appears in a wavelength region with a longer wavelength and tends to have a higher intensity. The light caused by the material appears in a wavelength region with a shorter wavelength and tends to have a lower intensity.

[0023] The light guide unit 2 guides the measurement light L1 incident from the dry etching device 8 to the spectrometer 3 in the subsequent stage. The light guide unit 2 is, for example, an optical fiber. The light guide unit 2 can be a single-core optical fiber or a bundled fiber formed by bundling multiple optical fibers. The light guide unit 2 has an incident end 2a and an output end 2b. The measurement light L1 is incident on the incident end 2a and is output from the output end 2b. The incident end 2a is opposite to the monitoring window 82 of the dry etching device 8 and is arranged at a predetermined distance from the monitoring window 82. The output end 2b is opposite to the spectrometer 3 and is optically connected to the spectrometer 3 via, for example, a connector.

[0024] The spectrometer 3 splits the measuring light L1 guided by the light-guiding unit 2, and forms a spectroscopic image on the pixel area of the detection unit 4. The spectrometer 3 splits the measuring light L1 into wavelength components by a spectroscopic component such as a grating or a prism. As the spectrometer 3, a spectrometer with good imaging performance is used. As the spectrometer constituting the spectrometer 3, for example, a Czerny-Turner spectrometer, a Dyson spectrometer, an Offner spectrometer, etc. capable of performing aberration correction are cited. The spectrometer 3 decomposes the measuring light L1 by wavelength in a first direction parallel to the wavelength axis of the spectroscopic image, and images the spectroscopic image of each wavelength on the pixel area 41 of the detection unit 4 in a second direction intersecting the first direction.

[0025] The detection unit 4 detects the measurement light L1 dispersed by the spectroscopic unit 3 . Figure 2 To display Figure 1 The structure of the detection unit 4 is shown in FIG. Figure 2 As shown in FIG. 4 , a pixel region 41 for capturing a spectroscopic image P is provided. In the pixel region 41 , a plurality of pixels 42 are arranged in the row direction and the column direction. Figure 2 In the example, the row direction corresponding to the first direction is along the wavelength separation direction of the spectroscopic image P by the spectroscopic unit 3, and the column direction corresponding to the second direction is along the charge transfer direction of each pixel 42. Each pixel 42 receives the wavelength-separated spectroscopic image P and generates and accumulates charges corresponding to the intensity of the light. Figure 2 In the example of , five spectral images P are formed in a state where they are separated from each other along the row direction. Each of the spectral images P extends linearly along the column direction of the pixels 42 .

[0026] exist Figure 2 In the example, the pixel area 41 forms a horizontally long rectangular area in which the number of pixels in the row direction is greater than the number of pixels in the column direction. The pixel area 41 includes a first pixel area 41A and a second pixel area 41B divided in the column direction. Here, the first pixel area 41A and the second pixel area 41B are divided in the center of the column direction. That is, in the pixel area 41, the pixel 42 on one side of the center in the column direction belongs to the first pixel area 41A, and the pixel 42 on the other side of the center in the column direction belongs to the second pixel area 41B. Here, the direction along the column direction from the center in the column direction to the first pixel area 41A is defined as column direction A1, and the direction along the column direction from the center in the column direction to the second pixel area 41B is defined as column direction A2. The column direction A1 and the column direction A2 are opposite directions to each other.

[0027] The first exposure time of each pixel 42 in the first pixel region 41A and the second exposure time of each pixel 42 in the second pixel region 41B can be set independently of each other. In this embodiment, the second exposure time is set longer than the first exposure time.

[0028] Data obtained from the detection results of the spectral image P of the measurement light L1 is input to the generator 5 from the detector 4. The generator 5 generates spectral data S of the measurement light L1 based on the detection results of the spectral image P. The generator 5 outputs the spectral data S of the measurement light L1 to the determination unit 6. The determination unit 6 determines the progress of the semiconductor process in the chamber 81 based on the spectral data S input from the generator 5. The generator 5 or the determination unit 6 can control the detector 4.

[0029] use Figures 3 to 9 A more detailed description will be given of a specific example of the detection unit 4. The detection unit 4 is composed of, for example, a CCD photodetector or a CMOS photodetector. Figure 3 This diagram shows an example of a CCD photodetector that constitutes the detection unit 4A. The detection unit 4A includes a pixel region 41 including the first pixel region 41A and the second pixel region 41B described above; a first horizontal shift register 43A corresponding to the first pixel region 41A; a second horizontal shift register 43B corresponding to the second pixel region 41B; and dummy pixels.

[0030] In the first pixel area 41A, the charges generated and accumulated by each pixel 42 are transferred to the first horizontal shift register 43A along the column direction A1, and the charges of the pixels 42 located in each column are added for each column in the first horizontal shift register 43A (hereinafter referred to as "vertical transfer"). Thereafter, the charges added for each column in the first horizontal shift register 43A are sequentially read out from the first horizontal buffer 43A (hereinafter referred to as "horizontal transfer"). Then, an electrical signal (e.g., a signal representing a voltage value) corresponding to the amount of charge read out from the first horizontal buffer 43A is output from the first amplifier 44A, and this electrical signal is converted into a digital value by an A / D converter. The digital value is output to the generating unit 5.

[0031] In the second pixel area 41B, the charge generated and accumulated by each pixel 42 is transferred to the second horizontal shift register 43B along the column direction A2. The charges of the pixels 42 located in each column are summed for each column in the second horizontal shift register 43B (vertical transfer). The charges summed for each column in the second horizontal shift register 43B are then sequentially read out from the second horizontal buffer 43B (horizontal transfer). An electrical signal (e.g., a signal representing a voltage value) corresponding to the amount of charge read out from the second horizontal buffer 43B is then output from the second amplifier 44B. This electrical signal is then converted to a digital value by an A / D converter. The digital value is then output to the generator 5.

[0032] Figure 4 To display Figure 3The timing diagram of the detection unit 4A is shown in FIG. This diagram is a timing diagram for a case where a full-frame transfer type CCD image sensor is used. In this case, half of the charge accumulated during a vertical transfer period is transferred to the horizontal shift register during that vertical transfer, and the remaining half is transferred to the horizontal shift register during the next vertical transfer. Figure 4 (a) is a timing chart showing an example of the operation of the first pixel region 41A. The first exposure time T1 can be set by an electronic shutter. The electronic shutter can be implemented by using, for example, an anti-blooming gate (ABG).

[0033] The electronic shutter switches between the accumulation of electric charges generated in the first pixel region 41A and the discharge of the accumulated electric charges. Figure 4 In the example (a), the time from the start of the horizontal transfer of the first horizontal shift register 43A to the end of the next vertical transfer is set as the first frame time FT1. The first frame time FT1 is equivalent to the inverse of the frame rate of the first pixel area 41A. In this case, the detection unit 4A discharges the charge at the prescribed time DT from the start of the first frame time FT1 by switching the electronic shutter during the horizontal transfer. Thus, the detection unit 4A sets the time excluding the prescribed time DT in the first frame time FT1 as the first exposure time T1. Therefore, in Figure 4 In the example of (a), the first exposure time T1 is shorter than the first frame time FT1.

[0034] Figure 4 (b) is a timing diagram showing an operation example of the second pixel region 41B. In the operation example of the second pixel region 41B, the electronic shutter is always in the OFF state, and the charge is always continuously accumulated. Figure 4 In the example (b), the time from the start of the horizontal transfer of the second horizontal shift register 43B to the end of the next vertical transfer is set as the second frame time FT2. The second frame time FT2 is equivalent to the inverse of the frame rate of the second pixel area 41B. The second exposure time T2 is consistent with the second frame time FT2. Figure 4 In the example, the first frame time FT1 and the second frame time FT2 are the same time, but the first frame time FT1 and the second frame time FT2 may be different.

[0035] Even without using an electronic shutter, the second exposure time T2 can be extended relative to the first exposure time T1. For example, the detection unit 4A can set the period of digital data output from the second pixel area 41B to the generator 5 to be five times the period of digital data output from the first pixel area 41A to the generator 5. In this case, the second exposure time T2 is approximately five times the first exposure time T1. Alternatively, the second frame time FT2 can be set to five times the first frame time FT1, or the second frame rate can be set to 1 / 5 of the first frame rate. In these cases, the second exposure time T2 is approximately five times the first exposure time T1.

[0036] Figure 5 and Figure 6 This diagram shows another example of a CCD photodetector that constitutes the detection unit 4B. The detection unit 4B includes a first accumulation unit 45A that accumulates the charge generated by each column of the first pixel region 41A; a second accumulation unit 45B that accumulates the charge generated by each column of the second pixel region 41B; a first readout unit 46A that outputs an electrical signal for each column corresponding to the magnitude of the charge accumulated by the first accumulation unit 45A; and a second readout unit 46B that outputs an electrical signal for each column corresponding to the magnitude of the charge accumulated by the second accumulation unit 45B.

[0037] The first accumulation unit 45A is arranged at the end of the first pixel region 41A in the column direction A1, for each column. Charges generated by the pixels 42 in each column of the first pixel region 41A are vertically transferred along the column direction A1 and accumulated in the first accumulation unit 45A. The second accumulation unit 45B is arranged at the end of the second pixel region 41B in the column direction A2, for each column. Charges generated by the pixels 42 in each column of the second pixel region 41B are vertically transferred along the column direction A2 and accumulated in the second accumulation unit 45B.

[0038] The first readout section 46A is positioned downstream of the first accumulator section 45A at the end portion on the first pixel region 41A side, and the second readout section 46B is positioned downstream of the second accumulator section 45B at the end portion on the second pixel region 41B side. The first readout section 46A outputs a first electrical signal for each column corresponding to the magnitude of the charge accumulated in the first accumulator section 45A. The second readout section 46B outputs a second electrical signal for each column corresponding to the magnitude of the charge accumulated in the second accumulator section 45B. The first and second electrical signals are, for example, signals representing voltage values.

[0039] The first reading unit 46A is as follows Figure 5As shown, the device includes a transistor 51A and a bonding pad 52A for signal output. The control terminal (gate) of transistor 51A is electrically connected to the first storage unit 45A. One current terminal (drain) of transistor 51A is electrically connected to a bonding pad 54A via a common wiring 53A across all columns of the first pixel region 41A. A predetermined voltage is constantly applied to bonding pad 54A.

[0040] The other current terminal (source) of transistor 51A is electrically connected to a signal output bonding pad 52A. A voltage corresponding to the first electrical signal output from the first storage unit 45A is applied to the control terminal of transistor 51A. A current corresponding to the applied voltage is output from the other current terminal of transistor 51A and is extracted via the signal output bonding pad 52A. The first electrical signal output from the signal output bonding pad 52A is amplified by the first amplifier and then output to the AD converter. The first electrical signal is converted by the AD converter into a digital value. The digital value is output to the generator 5.

[0041] The second reading unit 46B is as follows Figure 6 As shown, the second pixel region 41B includes a transistor 51B and a bonding pad 52B for signal output. The control terminal (gate) of transistor 51B is electrically connected to the second storage unit 45B. One current terminal (drain) of transistor 51B is electrically connected to a bonding pad 54B via a common wiring 53B across all columns of the second pixel region 41B. A predetermined voltage is constantly applied to bonding pad 54B.

[0042] The other current terminal (source) of transistor 51B is electrically connected to a signal output bonding pad 52B. A voltage corresponding to the second electrical signal output from the second storage unit 45B is applied to the control terminal of transistor 51B. A current corresponding to the applied voltage is output from the other current terminal of transistor 51B and is taken out via the signal output bonding pad 52B. The second electrical signal output from the signal output bonding pad 52B is amplified by the second amplifier and then output to the AD converter. The second electrical signal is converted by the AD converter into a digital value. The digital value is output to the generator 5.

[0043] Figure 7 To display Figure 5 and Figure 6 The timing chart of the operation example of the detection unit 4B shown in FIG. Figure 7 (a) is a timing chart showing an example of the operation of the first pixel region 41A. Figure 7 (b) is a timing diagram showing an example of the operation of the second pixel region 41B. Figure 4 The point is different from the example of the operation of the detection unit 4A described in . Figure 7In the example of (a), the first readout section 46A defines a readout period from the start of reading out the charge accumulated in the first accumulation section 45A to the output of the first electrical signal of each column corresponding to the magnitude of the charge. Figure 7 In the example (a), the time from the start of the readout period to the end of the next vertical transfer is defined as the first frame time FT1. The detection unit 4B switches the electronic shutter during the readout period, and discharges the charge at the predetermined time DT from the start of the first frame time FT1. Thus, the detection unit 4B defines the time excluding the predetermined time DT in the first frame time FT1 as the first exposure time T1. Figure 7 In the example (a), the first exposure time T1 is shorter than the first frame time FT1. Figure 7 In the example of (b), the period from the start of the readout period to the end of the next vertical transfer is referred to as the second frame time FT2. The second exposure time T2 coincides with the second frame time FT2.

[0044] Figure 8 4 is a diagram showing an example of a CMOS photodetector constituting the detection unit 4C. In the detection unit 4C, there are: a first pixel area 41A, a second pixel area 41B, a first readout unit 47A, and a second readout unit 47B. Each pixel 42 of the first pixel area 41A and the second pixel area 41B has a photodiode 61 and an amplifier 62. The photodiode 61 accumulates the electrons (photoelectrons) generated by the input of the measurement light L1 as electric charge. The amplifier 62 converts the charge accumulated in the photodiode 61 into an electrical signal (for example, a signal representing a voltage value) and amplifies it. In addition, the amplifier 62 may include: a capacitor portion (not shown) connected between the input terminal and the output terminal, and a reset switch (not shown) connected in parallel to the capacitor portion. In this case, when the detection unit 4C receives a reset signal from an external control circuit (for example, the generating unit 5), the charge accumulated in the capacitor portion is reset (discharged) by turning on the reset switch.

[0045] The electrical signal amplified by the amplifier 62 is transmitted to the vertical signal line 64 that connects the pixels 42 in the row direction by switching the selector switch 63 of each pixel 42. A CDS (correlated double sampling) circuit 65 is arranged on each vertical signal line 64. The CDS circuit 65 reduces readout noise between the pixels 42 and temporarily stores the electrical signal transmitted to the vertical signal line 64.

[0046] The AD converter 66 converts the voltage value stored in the CDS circuit 65 into a digital value. The digital value corresponding to the first pixel region 41A is output to the generator 5 via the first readout section 47A. In other words, the first readout section 47A outputs an electrical signal corresponding to the magnitude of the charge accumulated by each pixel 42 in the first pixel region 41A. The digital value corresponding to the second pixel region 41B is output to the generator 5 via the second readout section 47B. In other words, the second readout section 47B outputs an electrical signal corresponding to the magnitude of the charge accumulated by each pixel 42 in the second pixel region 41B.

[0047] Figure 9 To display Figure 8 The timing chart of the operation example of the detection unit 4C shown in FIG. Figure 9 (a) is a timing chart showing an example of the operation of the first pixel region 41A. Figure 9 (b) is a timing diagram showing an example of the operation of the second pixel area 41B. The detection unit 4C discharges the accumulated charge while receiving a reset signal from an external control circuit (e.g., the generating unit 5). The reset signal is, for example, a pulse. The detection unit 4C discharges the accumulated charge while the pulse is at a high level. When the detection unit 4C receives a readout signal from an external control circuit (e.g., the generating unit 5), the detection unit 4C outputs a digital value to the generating unit 5 via the first readout unit 47A and outputs a digital value to the generating unit 5 via the second readout unit 47B. The readout signal is, for example, a pulse. Figure 9 In the example of , the period from the falling edge of the reset signal to the rising edge of the readout signal is set as the exposure time. Figure 9 In the example of , the period from the falling edge of the reset signal to the rising edge of the readout signal can be adjusted by the electronic shutter. In this case, the first exposure time T1 and the second exposure time T2 can be set individually by the electronic shutter. Figure 9 In the example (a), the first exposure time T1 corresponds to the first frame time FT1. Figure 9 In the example (b), the second exposure time T2 corresponds to the second frame time FT2. Alternatively, the time from the falling edge of the reset signal to the falling edge of the next reset signal can be set as the first frame time FT1 and the second frame time FT2. In this case, it is feasible that the first exposure time T1 is shorter than the first frame time FT1, and the second exposure time T2 is shorter than the second frame time FT2.

[0048] Here, the relationship between the first exposure time T1 and the second exposure time T2 and the dynamic range is explained. The dynamic range is calculated by multiplying the ratio of the maximum detectable level to the minimum detectable level by the ratio of the second exposure time T2 to the first exposure time T1. The maximum detectable level is, for example, the maximum number of output bits of the AD converter. Specifically, in the case of 16 bits, this is 65535 in decimal notation. The minimum detectable level is the noise level detectable by the detection unit 4. Since the ratio of the maximum detectable level to the minimum detectable level is determined by the specifications of the detection unit 4, the dynamic range depends on the ratio of the second exposure time T2 to the first exposure time T1.

[0049] In the detection units 4A, 4B, and 4C, the dynamic range can be set within a wide range by using an electronic shutter to control the first exposure time T1. In the detection units 4A and 4B, the frame time of the image sensor is as short as about 10 ms. When the electronic shutter is not used, if the dynamic range (the ratio of the second exposure time T2 to the first exposure time T1) is set to 100, the second exposure time T2 is 1000 ms. In this way, when monitoring the semiconductor process, the sampling interval is too large. For this reason, when the electronic shutter is used to set the first exposure time T1 to 1 ms shorter than the first frame time FT1, the second exposure time T2 is 100 ms. Compared to the case where the electronic shutter is not used, the sampling interval can be shortened to 1 / 10. In the detection unit 4C, the dynamic range can also be expanded by using an electronic shutter to set the first exposure time T1 and the second exposure time T2 individually.

[0050] Figure 10 The flowchart of the semiconductor process monitoring method according to one embodiment of the present disclosure includes a spectroscopic step S11 , a detection step S12 , a generation step S13 , and a determination step S14 . Figure 11 It is a detailed flowchart showing a semiconductor process monitoring method according to one embodiment of the present disclosure. Figure 11 Corresponding to Figure 10 The spectroscopic step S11, the detection step S12, and the generation step S13 are shown. The spectroscopic unit 3 spectroscopically separates the measurement light L1 from the chamber 81 (step S21). The detection unit 4 detects the spectroscopic measurement light L1 (step S22). The detection unit 4 outputs the data obtained from the result of detecting the measurement light L1 to the generation unit 5. The generation unit 5 generates first spectrum data based on the detection result of the spectroscopic image P in the first pixel area 41A (step S23). The generation unit 5 generates second spectrum data based on the detection result of the spectroscopic image P in the second pixel area 41B (step S24). Steps S23 and S24 can be performed simultaneously, or one of them can be performed first.

[0051] Figure 12: is a diagram showing an example of the first spectrum data S1. Figure 12 As shown, in the first spectrum data S1, measurement light L1 is not saturated in any wavelength band, and generation unit 5 obtains an intensity that is not saturated in any wavelength band. Here, the wavelength band where measurement light L1 has high intensity tends to have longer wavelengths than the wavelength band where it has low intensity. There is a tendency for long-wavelength, high-intensity measurement light L1 to be caused by etching gas, while short-wavelength, low-intensity measurement light L1 is caused by the material being etched. Therefore, the wavelength region containing the wavelength band where measurement light L1 has high intensity is defined as long-wavelength region Δλ1, and the wavelength region containing the wavelength band where measurement light L1 has low intensity is defined as short-wavelength region Δλ2. The wavelength at the boundary between long-wavelength region Δλ1 and short-wavelength region Δλ2 is, for example, 400 nm, the boundary between the wavelengths of ultraviolet light and visible light.

[0052] Figure 13 This figure shows an example of second spectral data S2. In the second spectral data S2, the long-wavelength region Δλ1 includes a wavelength band in which measurement light L1 is saturated (saturated wavelength band Δλ3). On the other hand, the short-wavelength region Δλ2 includes a wavelength band in which measurement light L1 is unsaturated (unsaturated wavelength band Δλ4). To address this, in the first spectral data S1, the intensity can be accurately acquired in the wavelength band in the long-wavelength region Δλ1 corresponding to the saturated wavelength band Δλ3 of the second spectral data S2. That is, in the first pixel area 41A, the first exposure time T1 is set so that measurement light L1 is unsaturated in at least the long-wavelength region Δλ1 of measurement light L1.

[0053] In the first spectral data S1, noise is superimposed in the wavelength band of the short-wavelength region Δλ2 corresponding to the unsaturated wavelength band Δλ4 of the second spectral data, resulting in a poor signal-to-noise ratio. On the other hand, in the second spectral data S2, noise is not superimposed even in the short-wavelength region Δλ2 (unsaturated wavelength band Δλ4), allowing the generator 5 to obtain highly accurate data. Specifically, in the second pixel region 41B, the second exposure time T2 is set so that the measurement light L1 is not saturated in at least the short-wavelength region Δλ2 (unsaturated wavelength band Δλ4) of the measurement light L1.

[0054] Next, the generator 5 combines (joins) a portion of the first spectral data S1 with a portion of the second spectral data S2 to generate spectral data S for the measuring light L1 (step S25). For example, the generator 5 may combine (join) data for a wavelength band corresponding to the saturated wavelength band Δλ3 of the second spectral data S2 with data for the unsaturated wavelength band Δλ4 of the second spectral data S2 to generate spectral data S for the measuring light L1. Alternatively, the generator 5 may combine (join) data for the long-wavelength region Δλ1 of the first spectral data S1 with data for the short-wavelength region Δλ2 of the second spectral data S2 to generate spectral data S for the measuring light L1. Figure 14 This figure illustrates the generation of spectral data S for measuring light L1. For example, the generator 5 generates spectral data S for measuring light L1 as follows. For the first spectral data S1, the generator 5 divides the first spectral data S1 by the first exposure time T1 and multiplies the result by the reference exposure time. Conversely, for the second spectral data S2, the generator 5 divides the second spectral data S2 by the second exposure time T2 and multiplies the result by the reference exposure time. Thus, the generator 5 equalizes the ratios of the first spectral data S1 and the second spectral data S2 and generates spectral data S for measuring light L1. In the spectral data S, the measuring light L1 is not saturated in any wavelength band, and noise is not superimposed, achieving an intensity that is free of saturation in all wavelength bands.

[0055] Figure 15 It is a detailed flowchart showing a semiconductor process monitoring method according to one embodiment of the present disclosure. Figure 15 Corresponding to Figure 10 Judgment step S14. The judgment unit 6 judges whether there is an abnormality in the semiconductor process based on the spectral data S of the measurement light L1. First, the judgment unit 6 judges whether the first peak appears in the wavelength specified in the long wavelength region Δλ1 of the spectral data S (for example, a wavelength of 600nm to 800nm) (step S31). When the first peak appears (step S31: yes), the judgment unit 6 judges that there is an abnormality in the semiconductor process (step S32). In this case, the judgment unit 6 can issue an alarm to the operator of the device to end the etching process. Thereafter, a series of judgment processes end. In addition, step S34 includes: the judgment unit 6 judges that there is an abnormality in the semiconductor process based on the spectral data of both the long wavelength region Δλ1 and the short wavelength region Δλ2 of the spectral data S.

[0056] Figure 16 A diagram for explaining a process for determining whether or not there is an abnormality in a semiconductor process. Figure 16 The data of the long wavelength region Δλ1 of the spectral data S is enlarged and shown in the figure. Figure 16 (a) shows the spectrum data S when there is no abnormality. Figure 16(b) shows the spectral data S of an abnormal situation. Figure 16 In (b), Figure 16 A first peak P1 is generated in the wavelength band where no peak is generated in (a). The first peak P1 is generated, for example, by nitrogen gas mixed in from the outside. For example, during the repeated opening and closing of the door of the chamber 81, the sealing member between the door and the chamber 81 body is worn, and when the chamber is set to a vacuum, there is a possibility that nitrogen gas will be mixed in from the outside. At this time, by detecting the first peak P1 generated when the nitrogen gas mixed in from the outside is greater than a specified value, the state of excessive gas mixing into the etching chamber from the outside can be detected as an abnormal state with high precision.

[0057] If the first peak P1 is not detected (step S31: No), the determination unit 6 determines that there is no abnormality in the semiconductor process (step S33). Next, the determination unit 6 determines the end point of the semiconductor process. First, the determination unit 6 determines whether a second peak is detected within a wavelength specified in the short-wavelength region Δλ2 of the spectral data S (e.g., a wavelength of 300 nm to 400 nm) (step S34). If the second peak is not detected (step S34: No), the determination unit 6 determines that the semiconductor process continues, and the determination process returns to step 31. On the other hand, if the second peak is detected (step S34: Yes), the determination unit 6 determines that the semiconductor process has ended (step S35), and the determination process ends. Furthermore, step S34 includes the determination unit 6 determining that there is no abnormality in the semiconductor process based on the spectral data of both the long-wavelength region Δλ1 and the short-wavelength region Δλ2 of the spectral data S.

[0058] Figure 17 A diagram for explaining a process for determining an endpoint of a semiconductor process. Figure 17 The data of the short wavelength region Δλ2 of the spectral data S is enlarged and shown in the figure. Figure 17 (a) shows the spectrum data S for judging whether the semiconductor process is continued. Figure 17 (b) shows the spectrum data S for determining the end point of the semiconductor process. Figure 17 In (b), Figure 17 A second peak P2 is generated in the wavelength band where no peak is generated in (a). The second peak P2 is generated by the material to be etched. For example, assume that a base layer is formed on a substrate W, and an etching target layer is formed on the base layer. In this case, the second peak P2 is generated by the base layer. When the etching process progresses and the etching target layer is etched, the base layer may be exposed at the bottom of the etching pattern. In this case, by detecting the second peak P2, it is possible to detect with high precision that the etching process has ended and the base layer has been exposed.

[0059] As described above, in a semiconductor process monitoring device 1 according to one aspect of the present disclosure, the presence or absence of an abnormality in the semiconductor process is determined based on first spectral data S1 obtained in first pixel region 41A with first exposure time T1, and the endpoint of the semiconductor process is determined based on second spectral data S2 obtained in second pixel region 41B with second exposure time T2, which is longer than first exposure time T1. Furthermore, the presence or absence of an abnormality in the semiconductor process can be determined based on the spectral data S of both first spectral data S1 and second spectral data S2. In this semiconductor process monitoring device, by combining the detection of spectral images P of measurement light L1 in first pixel region 41A and second pixel region 41B with different exposure times, the intensity of spectral image P of measurement light L1 can be measured with high dynamic range and high accuracy. Therefore, even in a situation where high-intensity gas-derived measurement light L1 and low-intensity material-derived measurement light L1 coexist in the semiconductor process, the presence or absence of abnormalities in the semiconductor process and the endpoint of the process can be monitored with high accuracy.

[0060] In first pixel region 41A, first exposure time T1 is set shorter than first frame time FT1 in first pixel region 41A, while in second pixel region 41B, second exposure time T2 is set equal to second frame time FT2 in second pixel region 41B. With this configuration, by combining the detection of spectral image P of measurement light L1 in first pixel region 41A and second pixel region 41B with different exposure times, the intensity of spectral image P of measurement light L1 can be measured with high dynamic range and high precision. Therefore, even in a situation where high-intensity gas-derived measurement light L1 and lower-intensity material-derived measurement light L1 coexist in a semiconductor process, it is possible to accurately monitor the presence of abnormalities in the semiconductor process and the process endpoint.

[0061] In the first pixel region 41A, the first exposure time T1 is set so that the measurement light L1 is not saturated in at least the long-wavelength region Δλ1 of the measurement light L1 (the wavelength band of the long-wavelength region Δλ1 corresponding to the saturation wavelength band Δλ3 of the second spectral data S2). In the second pixel region 41B, the second exposure time T2 is set so that the measurement light L1 is not saturated in at least the short-wavelength region Δλ2 of the measurement light L1. With this configuration, the first spectral data S1 can be acquired in the long-wavelength region Δλ1 without saturating the high-intensity gas-derived measurement light L1. Furthermore, in the short-wavelength region Δλ2, the second spectral data S2 associated with the low-intensity material-derived measurement light L1 can be acquired with a good signal-to-noise ratio. This further improves the accuracy of monitoring the presence of abnormalities and the process endpoint in the semiconductor process.

[0062] Generator 5 combines (concatenates) the data in the long-wavelength region Δλ1 of first spectral data S1 with the data in the short-wavelength region Δλ2 of second spectral data S2, and outputs the combined data to determination unit 6. This configuration allows generation of spectral data S for measurement light L1 across a wide dynamic range, encompassing the entire wavelength band to be monitored. This enables simple and highly accurate determination of the presence of abnormalities and the endpoint of a process.

[0063] Determination unit 6 determines the presence of an abnormality in the semiconductor process based on the presence of a first peak P1 at a predetermined wavelength in the long-wavelength region Δλ1 of first spectral data S1. Determination unit 6 determines the endpoint of the semiconductor process when a second peak P2 appears at a predetermined wavelength in the short-wavelength region Δλ2 of second spectral data S2. By performing determinations based on the appearance trends of the second peak P2 of measurement light L1 (material-derived) and the first peak P1 of measurement light L1 (gas-derived) in the semiconductor process, the accuracy of monitoring the presence of abnormalities and the endpoint of the semiconductor process can be further improved.

[0064] The detection unit 4 is a CCD photodetector comprising a first horizontal shift register 43A to which charges generated by each column of the first pixel region 41A are transferred, and a second horizontal shift register 43B to which charges generated by each column of the second pixel region 41B are transferred. This configuration, using a CCD photodetector, avoids an increase in readout noise when reading out charges generated by pixels 42 in each column.

[0065] The detection unit 4 is a CCD photodetector comprising a first accumulation unit 45A that accumulates the charge generated by each column of the first pixel region 41A; a second accumulation unit 45B that accumulates the charge generated by each column of the second pixel region 41B; a first readout unit 46A that outputs an electrical signal for each column corresponding to the magnitude of the charge accumulated by the first accumulation unit 45A; and a second readout unit 46B that outputs an electrical signal for each column corresponding to the magnitude of the charge accumulated by the second accumulation unit 45B. In this configuration, the use of a CCD photodetector also avoids an increase in readout noise when reading out the charge generated by each column of pixels 42.

[0066] The detection unit 4 is a CMOS photodetector comprising a first readout unit 47A that outputs an electrical signal corresponding to the magnitude of the charge accumulated by each pixel 42 in the first pixel region 41A, and a second readout unit 47B that outputs an electrical signal corresponding to the magnitude of the charge accumulated by each pixel 42 in the second pixel region 41B. This configuration allows for lower power consumption compared to, for example, a CCD photodetector that reads charge for each column.

[0067] The first exposure time T1 of the first pixel area 41A is controlled by an electronic shutter. This configuration allows for highly accurate adjustment of the first exposure time T1 even when the second exposure time T2 is set relatively short, and the first exposure time T1 is set sufficiently shorter than the second exposure time T2. By making the first exposure time T1 sufficiently shorter than the second exposure time T2, the dynamic range of measurement can be further significantly improved.

[0068] In the semiconductor process monitoring method according to one aspect of the present disclosure, for the same reasons as the semiconductor process monitoring apparatus 1 described above, the intensity of the measurement light is measured with high accuracy over a high dynamic range, thereby enabling high-precision monitoring of the semiconductor process.

[0069] As described above, in a semiconductor process monitoring method according to one aspect of the present disclosure, the presence or absence of an abnormality in the semiconductor process is determined based on first spectral data S1 obtained in first pixel region 41A with first exposure time T1, and the endpoint of the semiconductor process is determined based on second spectral data S2 obtained in second pixel region 41B with second exposure time T2, which is longer than first exposure time T1. In this semiconductor process monitoring method, by combining the detection of spectral images P of measurement light L1 in first pixel region 41A and second pixel region 41B with different exposure times, the intensity of spectral image P of measurement light L1 can be measured with high dynamic range and high precision. Therefore, even in a situation where measurement light L1 of relatively high intensity originating from gas and measurement light L1 of relatively low intensity originating from materials coexist in the semiconductor process, the presence or absence of an abnormality in the semiconductor process and the endpoint of the process can be monitored with high precision.

[0070] As mentioned above, although embodiment of this disclosure was demonstrated, this disclosure is not necessarily limited to the said embodiment, Various changes are possible within the range which does not deviate from the summary.

[0071] The generating unit 5 may not combine the data of the long wavelength region Δλ1 of the first spectrum data S1 with the data of the short wavelength region Δλ2 of the second spectrum data S2 to generate the spectrum data S of the measurement light L1. The generating unit 5 may omit Figure 11 Step S25 is performed. In this case, the determination unit 6 can determine whether there is an abnormality in the semiconductor process based on the first spectral data S1. Furthermore, the determination unit 6 can determine the end point of the semiconductor process based on the second spectral data S2. By not generating the spectral data S for the measurement light L1, the computational load on the generation unit 5 can be reduced.

[0072] The gist of the present disclosure is as follows [1] to

[18] . [1] A semiconductor process monitoring device, comprising: a spectrometer that splits measurement light from a chamber; a detection unit that detects a spectral image of the measurement light after being split by the spectrometer; and a judgment unit that judges the progress of the semiconductor process in the chamber based on data obtained from the detection result of the spectral image of the measurement light; the spectrometer decomposes the measurement light into wavelengths in a first direction and forms a spectral image of each wavelength in a second direction intersecting the first direction; the detection unit has a first pixel area and a second pixel area divided in the second direction; the judgment unit judges whether there is an abnormality in the semiconductor process based on first spectral data obtained in the first pixel area with a first exposure time, and judges the end point of the semiconductor process based on second spectral data obtained in the second pixel area with a second exposure time longer than the first exposure time. [2] A semiconductor process monitoring device as described in [1], wherein in the above-mentioned first pixel area, the above-mentioned first exposure time is set to be shorter than the first frame time in the above-mentioned first pixel area, and in the above-mentioned second pixel area, the above-mentioned second exposure time is set to be the same as the second frame time in the above-mentioned second pixel area. [3] A semiconductor process monitoring device as described in [1] or [2], wherein in the above-mentioned first pixel area, the above-mentioned first exposure time is set in a manner that the above-mentioned measuring light is not saturated in at least the long wavelength area of the above-mentioned measuring light, and in the above-mentioned second pixel area, the above-mentioned second exposure time is set in a manner that the above-mentioned measuring light is not saturated in at least the short wavelength area of the above-mentioned measuring light. [4] The semiconductor process monitoring device as described in [1] or [2] further includes a generating unit that combines the data of the long wavelength region of the above-mentioned first spectral data and the data of the short wavelength region of the above-mentioned second spectral data and outputs them to the above-mentioned judgment unit. [5] A semiconductor process monitoring device as described in any one of [1] to [4], wherein the judgment unit judges whether there is an abnormality in the semiconductor process based on whether a first peak appears at a specified wavelength in the long wavelength region of the first spectral data, and judges the end point of the semiconductor process when a second peak appears at a specified wavelength in the short wavelength region of the second spectral data. [6] A semiconductor process monitoring device as described in any one of [1] to [5], wherein the detection unit is a CCD photodetector, and the CCD photodetector has: a first horizontal shift register that transfers the charges generated by each column of the first pixel area; and a second horizontal shift register that transfers the charges generated by each column of the second pixel area. [7] A semiconductor process monitoring device as described in any one of [1] to [5], wherein the detection unit is a CCD photodetector, and the CCD photodetector has: a first storage unit that stores charges generated by each column of the first pixel area; a second storage unit that stores charges generated by each column of the second pixel area; a first readout unit that outputs an electrical signal of each column corresponding to the size of the charge accumulated by the first storage unit; and a second readout unit that outputs an electrical signal of each column corresponding to the size of the charge accumulated by the second storage unit. [8] A semiconductor process monitoring device as described in any one of [1] to [5], wherein the detection unit is a CMOS photodetector, and the CMOS photodetector has: a first readout unit that outputs an electrical signal corresponding to the size of the charge accumulated by each pixel in the first pixel area; and a second readout unit that outputs an electrical signal corresponding to the size of the charge accumulated by each pixel in the second pixel area. [9] A semiconductor process monitoring device as described in any one of [1] to [8], wherein the first exposure time of the first pixel area is controlled by an electronic shutter.

[10] A semiconductor process monitoring method, comprising: a spectroscopic step of spectroscopically ...

[11] A semiconductor process monitoring method as described in

[10] , wherein in the above-mentioned detection step, in the above-mentioned first pixel area, the above-mentioned first exposure time is set to be shorter than the first frame time in the above-mentioned first pixel area, and in the above-mentioned second pixel area, the above-mentioned second exposure time is set to be the same as the second frame time in the above-mentioned second pixel area.

[12] A semiconductor process monitoring method as described in

[10] or

[11] , wherein in the above-mentioned detection step, in the above-mentioned first pixel area, the above-mentioned first exposure time is set in a manner that the above-mentioned measuring light is not saturated in at least the long wavelength area of the above-mentioned measuring light, and in the above-mentioned second pixel area, the above-mentioned second exposure time is set in a manner that the above-mentioned measuring light is not saturated in at least the short wavelength area of the above-mentioned measuring light.

[13] The semiconductor process monitoring method as described in

[10] or

[11] further includes a generation step, which combines the data of the long wavelength region of the above-mentioned first spectral data and the data of the short wavelength region of the above-mentioned second spectral data and outputs them to the above-mentioned judgment unit.

[14] A semiconductor process monitoring method as described in any one of

[10] to

[13] , wherein in the above-mentioned judgment step, whether there is an abnormality in the above-mentioned semiconductor process is judged based on whether the first peak appears at a specified wavelength in the long-wavelength region of the above-mentioned first spectral data, and when the second peak appears at a specified wavelength in the short-wavelength region of the above-mentioned second spectral data, the end point of the above-mentioned semiconductor process is judged.

[15] A semiconductor process monitoring method as described in any one of

[10] to

[14] , wherein a CCD photodetector is used in the above-mentioned detection step, and the CCD photodetector has: a first horizontal shift register, which transfers the charges generated by each column of the above-mentioned first pixel area; and a second horizontal shift register, which transfers the charges generated by each column of the above-mentioned second pixel area.

[16] A semiconductor process monitoring method as described in any one of

[10] to

[15] , wherein a CCD photodetector is used in the above-mentioned detection step, and the CCD photodetector has: a first storage unit, which stores the charges generated by each column of the above-mentioned first pixel area; a second storage unit, which stores the charges generated by each column of the above-mentioned second pixel area; a first readout unit, which outputs an electrical signal of each column corresponding to the size of the charge accumulated by the above-mentioned first storage unit; and a second readout unit, which outputs an electrical signal of each column corresponding to the size of the charge accumulated by the above-mentioned second storage unit.

[17] A semiconductor process monitoring method as described in any one of

[10] to

[15] , wherein a CMOS photodetector is used in the above-mentioned detection unit step, and the CMOS photodetector has: a first readout unit, which outputs an electrical signal corresponding to the size of the charge accumulated by each pixel in the above-mentioned first pixel area; and a second readout unit, which outputs an electrical signal corresponding to the size of the charge accumulated by each pixel in the above-mentioned second pixel area.

[18] A semiconductor process monitoring method as described in any one of

[10] to

[17] , wherein in the above-mentioned detection step, the above-mentioned first exposure time of the above-mentioned first pixel area is controlled by an electronic shutter.

Explanation of symbols

[0073] 1…semiconductor process monitoring device, 3…spectrometer, 4, 4A, 4B, 4C…detection unit, 5…generator, 6…determination unit, 41A…first pixel region, 41B…second pixel region, 42…pixel, 43A…first horizontal shift register, 43B…second horizontal shift register, 45A…first storage unit, 45B…second storage unit, 46A, 47A…first readout unit, 46B, 47B…second readout unit, 81…chamber, L1…measurement light, P1…first peak value, P2…second peak value, S…spectral data, S1…first spectral data, S2…second spectral data, S11…spectrometric step, S12…detection step, S14…determination step, T1…first exposure time, T2…second exposure time, Δλ1…long wavelength region, Δλ2…short wavelength region.

Claims

1. A semiconductor process monitoring device comprising: a spectrometer for splitting the measurement light from the chamber; a detection unit configured to detect a spectroscopic image of the measurement light after being split by the spectroscopic unit; and a determination unit that determines a progress status of the semiconductor process in the chamber based on data obtained from a detection result of the spectral image of the measurement light, The spectroscopic unit decomposes the measuring light by wavelength in a first direction and forms a spectroscopic image for each wavelength in a second direction intersecting the first direction. The detection unit includes a first pixel region and a second pixel region divided in the second direction. The judgment unit judges whether there is any abnormality in the semiconductor process based on the first spectral data obtained in the first pixel area with the first exposure time, and judges the end point of the semiconductor process based on the second spectral data obtained in the second pixel area with the second exposure time longer than the first exposure time.

2. The semiconductor process monitoring device according to claim 1, wherein: In the first pixel region, the first exposure time is set to be shorter than the first frame time in the first pixel region, In the second pixel region, the second exposure time is set to be the same as the second frame time in the second pixel region.

3. The semiconductor process monitoring device according to claim 1, wherein: In the first pixel region, the first exposure time is set so that the measurement light is not saturated in at least a long wavelength region of the measurement light. In the second pixel region, the second exposure time is set so that the measurement light is not saturated in at least a short wavelength region of the measurement light.

4. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The system further includes a generating unit that combines the data of the long wavelength region of the first spectral data and the data of the short wavelength region of the second spectral data and outputs the combined data to the determining unit.

5. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The judgment unit judges whether there is an abnormality in the semiconductor process based on whether a first peak appears at a specified wavelength in the long wavelength region of the first spectral data, and judges the end point of the semiconductor process when a second peak appears at a specified wavelength in the short wavelength region of the second spectral data.

6. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The detection unit is a CCD light detector, and the CCD light detector includes: a first horizontal shift register for transferring charges generated by each column of the first pixel region; and The second horizontal shift register transfers the charges generated by the respective columns of the second pixel region.

7. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The detection unit is a CCD light detector, and the CCD light detector includes: a first storage unit for storing charges generated by each column of the first pixel region; a second storage unit for storing charges generated by each column of the second pixel region; a first readout section that outputs an electrical signal for each column according to the magnitude of the charge accumulated in the first accumulation section; and The second readout unit outputs an electrical signal for each column according to the magnitude of the charge accumulated in the second accumulation unit.

8. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The detection unit is a CMOS light detector, which includes: a first readout section that outputs an electric signal corresponding to the magnitude of the charge accumulated in each pixel of the first pixel region; and The second readout unit outputs an electric signal according to the magnitude of the charge accumulated in each pixel of the second pixel region.

9. The semiconductor process monitoring device according to any one of claims 1 to 3, wherein: The first exposure time of the first pixel area is controlled by an electronic shutter.

10. A semiconductor process monitoring method, comprising: a spectroscopic step of splitting the measurement light from the chamber; a detection step of detecting a spectroscopic image of the measurement light after being spectroscopically separated by the spectroscopic step; and a determination step of determining a progress status of the semiconductor process in the chamber based on data obtained from a detection result of the spectral image of the measurement light, In the spectroscopic step, the measuring light is wavelength-decomposed in a first direction, and a spectroscopic image of each wavelength is formed in a second direction intersecting the first direction. In the detecting step, the spectral image is detected by a photodetector having a first pixel region and a second pixel region divided in the second direction. In the judgment step, whether there is any abnormality in the semiconductor process is judged based on the first spectral data obtained in the first pixel area with the first exposure time, and the end point of the semiconductor process is judged based on the second spectral data obtained in the second pixel area with the second exposure time longer than the first exposure time.

Images