Light detection device, film thickness measurement device, semiconductor process monitoring device, and noise detection method

Abstract

In a light detection device according to the present invention, a processing unit sets a threshold value for each of a plurality of channels and identifies, on the basis of the threshold values for each of the plurality of channels and a difference between a first signal and a second signal, a channel in which noise is included in the first signal or the second signal. The threshold value is set on the basis of the first signal and the second signal for each of the plurality of channels.

Description

Photodetector, film thickness measuring device, semiconductor process monitoring device, and noise detection method

[0001] One aspect of the present disclosure relates to a photodetector, a film thickness measurement device, a semiconductor process monitoring device, and a noise detection method.

[0002] Patent Document 1 describes an image sensor that continuously observes two lines (first and second lines) consisting of a plurality of photosensitive pixels, and a method for capturing images using the image sensor. In the method described in Patent Document 1, to detect pixels damaged by the impact of cosmic particles, the difference between the digital values ​​from the pixels of the first and second lines is calculated, the result of the subtraction is compared with a threshold value, and if the result is less than the threshold value, the digital value is deemed acceptable, while if the result is greater than the threshold value, the digital value is deemed unacceptable. The threshold value is set to a constant value greater than the read noise of the sensor.

[0003] Special Publication No. 2017-522823

[0004] The present inventors have found that, for example, when the method described in Patent Document 1 is applied to the detection of sudden noise caused by cosmic rays, false detection may occur, and the noise may not be detected accurately.

[0005] An object of one aspect of the present disclosure is to provide a photodetector, a film thickness measurement device, a semiconductor process monitoring device, and a noise detection method that are capable of detecting noise with high accuracy.

[0006] A light detection device according to one aspect of the present disclosure includes: [1] "a light detector that detects measurement light; and a processing unit that processes data output from the light detector, wherein the light detector has a plurality of first pixel rows arranged along a first direction, each of the plurality of first pixel rows including at least one pixel arranged along a second direction perpendicular to the first direction; and a first detection unit that has a plurality of second pixel rows arranged along the first direction, each of the plurality of second pixel rows including at least one pixel arranged along the second direction, wherein the first detection unit and the second detection unit are arranged adjacent to each other in the second direction; and wherein the light detector detects one-dimensional data in the first direction of the measurement light detected by the first detection unit. and a second signal which is one-dimensional data of the measurement light in the first direction detected by the second detection unit, to the processing unit, and where one channel is formed by the first pixel row and the second pixel row aligned in the second direction, the photodetector has a plurality of channels aligned along the first direction, the processing unit sets a threshold for each of the plurality of channels, and identifies the channel in which noise is included in the first signal or the second signal based on the threshold for each of the plurality of channels and the difference between the first signal and the second signal, and the threshold is set for each of the plurality of channels based on the first signal and the second signal.

[0007] In this photodetector, the processing unit sets a threshold for each of the multiple channels and identifies a channel in which noise is included in the first signal or the second signal based on the threshold for each channel and the difference between the first signal and the second signal. The threshold is set for each channel based on the first signal and the second signal. By setting the threshold for each channel based on the first signal and the second signal in this way, it is possible to detect noise with higher accuracy than, for example, when the threshold is set to a constant value regardless of the channel.

[0008] The photodetector according to one aspect of the present disclosure may be [2] "the photodetector according to [1], in which the threshold is set for each of the plurality of channels based on the magnitude of the first signal and the magnitude of the second signal in the channel." In this case, the threshold can be suitably set, and noise can be detected with higher accuracy.

[0009] A photodetector according to one aspect of the present disclosure may be [3] "the photodetector according to [1] or [2], in which the threshold is set based on a difference between the value of the first signal in the channel and the value of the first signal in a first adjacent channel adjacent to the channel, and a difference between the value of the second signal in the channel and the value of the second signal in a second adjacent channel adjacent to the channel." In this case, the threshold can be set appropriately, enabling more accurate noise detection.

[0010] A photodetector according to one aspect of the present disclosure may be the photodetector according to any one of [1] to [3], [4] "wherein the processing unit creates threshold sum data by adding the threshold to one of the first signal and the second signal, whichever has a smaller value, for each of the plurality of channels, and compares the first signal and the second signal with the threshold sum data to identify the channel in which noise is included in the first signal or the second signal." In this case, noise detection can be preferably performed.

[0011] The photodetector according to one aspect of the present disclosure may be [5] "the photodetector according to any one of [1] to [4], wherein the first signal is the one-dimensional data of the measurement light detected by the first detector over a first exposure time, the second signal is the one-dimensional data of the measurement light detected by the second detector over a second exposure time, and the second exposure time is longer than the first exposure time." In this case, the dynamic range can be increased by combining data having different exposure times.

[0012] A photodetector according to one aspect of the present disclosure may be [6] "the photodetector according to any one of [1] to [5], wherein the processing unit interpolates the value of the channel identified as including the noise in the first signal or the second signal based on the values ​​of the channels on both sides of the channel identified as including the noise." In this case, the effect of noise on photodetection can be reduced.

[0013] A photodetector according to one aspect of the present disclosure may be [7] "the photodetector according to any one of [1] to [6], wherein the photodetector is a CCD photodetector having a first horizontal shift register to which charges generated in the plurality of first pixel columns of the first detection unit are transferred, and a second horizontal shift register to which charges generated in the plurality of second pixel columns of the second detection unit are transferred." In this case, an increase in readout noise when reading out charges generated in each pixel can be avoided.

[0014] A photodetector according to one aspect of the present disclosure may be the photodetector according to any one of [1] to [6], wherein the photodetector is a CCD photodetector having: a plurality of first accumulation units in which charges generated in the plurality of first pixel columns of the first detection unit are accumulated, respectively; a plurality of second accumulation units in which charges generated in the plurality of second pixel columns of the second detection unit are accumulated, respectively; a plurality of first readout units electrically connected to the plurality of first accumulation units, respectively, and outputting electrical signals corresponding to the magnitudes of the charges accumulated in the plurality of first accumulation units; and a plurality of second readout units electrically connected to the plurality of second accumulation units, respectively, and outputting electrical signals corresponding to the magnitudes of the charges accumulated in the plurality of second accumulation units. In this case, too, it is possible to avoid an increase in readout noise when reading out charges generated in each pixel.

[0015] A photodetector according to one aspect of the present disclosure may be [9] "the photodetector according to any one of [1] to [6], wherein the photodetector is a CMOS photodetector having a first readout unit that outputs an electrical signal corresponding to the magnitude of the charge accumulated in each of the pixels of the first detection unit, and a second readout unit that outputs an electrical signal corresponding to the magnitude of the charge accumulated in each of the pixels of the second detection unit." In this case, power consumption can be reduced compared to, for example, a CCD photodetector that reads out charge on a column-by-column basis.

[0016] A spectrometer according to one aspect of the present disclosure is

[10] "a spectrometer including a spectroscopic unit that spectroscopically separates the measurement light and forms an image across a first detection unit and a second detection unit, and the photodetector according to [1] to [9] that detects the measurement light dispersed by the spectroscopic unit in the first detection unit and the second detection unit." This spectrometer enables noise to be detected with high accuracy for the reasons described above.

[0017] A film thickness measurement device according to one aspect of the present disclosure is

[11] "a film thickness measurement device including the spectroscope described in

[10] and a light source unit that generates light to be irradiated onto an object, wherein the processing unit measures the film thickness of the object based on the data from which the noise has been removed." With this film thickness measurement device, for the reasons described above, it is possible to measure the film thickness with high accuracy based on data from which the noise detected with high accuracy has been removed.

[0018] A film thickness measurement device according to one aspect of the present disclosure is

[12] "a film thickness measurement device including the spectroscope described in

[10] and a light source unit that generates light to be irradiated onto an object, wherein the processing unit measures the film thickness of the object based on the data from which the noise has been removed." With this film thickness measurement device, for the reasons described above, it is possible to measure the film thickness with high accuracy based on the data from which the noise has been removed that has been accurately detected.

[0019] A noise detection method according to one aspect of the present disclosure includes

[13] "a photodetector including a first detection unit having a plurality of first pixel rows arranged along a first direction, each of the plurality of first pixel rows including at least one pixel arranged along a second direction perpendicular to the first direction, and a second detection unit having a plurality of second pixel rows arranged along the first direction, each of the plurality of second pixel rows including at least one pixel arranged along the second direction, the first detection unit and the second detection unit being arranged adjacent to each other in the second direction, wherein if one channel is formed by the first pixel row and the second pixel row arranged in the second direction, then a photodetector having a plurality of channels arranged along the first direction a step of outputting a first signal, which is one-dimensional data in the first direction of the measurement light detected by the first detection unit, and a second signal, which is one-dimensional data in the first direction of the measurement light detected by the second detection unit, using the photodetector; a step of setting a threshold for each of the plurality of channels; and a step of identifying a channel in which noise is included in the first signal and the second signal based on the threshold for the channel and a difference between the first signal and the second signal, wherein in the step of setting the threshold, the threshold is set for each of the plurality of channels based on the first signal and the second signal. This noise detection method enables noise to be detected with high accuracy for the reasons described above.

[0020] According to one aspect of the present disclosure, it is possible to provide a photodetector, a film thickness measuring device, a semiconductor process monitoring device, and a noise detection method that are capable of detecting noise with high accuracy.

[0021] 1 is a diagram illustrating a configuration of a spectrometer according to an embodiment; FIG. 2 is a diagram illustrating a configuration of a photodetector; FIG. 3 is a diagram illustrating a detection unit of a photodetector; FIG. 4 is a graph illustrating an example of a first signal, and FIG. 5 is a graph illustrating an example of a second signal; FIG. 6 is a graph illustrating a method for detecting noise based on a difference between a first signal and a second signal; FIG. 7 is a graph illustrating an example when a threshold is set to a constant value; FIG. 8 is a graph illustrating an example when a threshold is set by a first method; FIG. 9 is a graph illustrating another example when a threshold is set by the first method; FIG. 10 is a graph illustrating examples of a first signal and a second signal; FIG. 11 is a diagram illustrating optical distortion occurring in a first signal and a second signal; FIG. 12 is a diagram illustrating the magnitude of optical distortion; FIG. 13 is a diagram illustrating signal discretization; FIG. 14 is a diagram illustrating a second method; FIG. 15 is a diagram illustrating a first example of the second method; 1A and 1B are diagrams for explaining a second example of the second method; 1A and 1B are diagrams for explaining a third example of the second method; 1A and 1B are diagrams for explaining a fourth example of the second method; 1A and 1B are diagrams for explaining a comparative example; 1B are graphs showing an example of a combination of the first method and the second method; 1B are diagrams for explaining threshold sum data; 1B are graphs showing an example of a result of application of the second method; 1B are graphs showing another example of a result of application of the second method; 1B, 1B, and 1C are tables for explaining actual measurement results; 1B are diagrams showing a first modified example of a photodetector; 1B are diagrams showing a first modified example of a photodetector; 1B are diagrams showing a second modified example of a photodetector; 1C are diagrams showing an example of a film thickness measurement device;

[0022] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the following description, the same or equivalent elements will be designated by the same reference numerals, and redundant description will be omitted. [Spectrometer]

[0023] 1 , the spectroscope 1 includes a spectroscopic unit 2, a photodetector 3, and a processing unit 4. The spectroscope 1 disperses measurement light L arriving from an object S using the spectroscopic unit 2, and detects the dispersed measurement light L using the photodetector 3 to obtain spectral data of the measurement light L.

[0024] In the example of FIG. 1 , the spectrometer 1 constitutes a semiconductor process monitoring apparatus 100. The semiconductor process monitoring apparatus 100 includes the spectrometer 1 and a semiconductor process equipment 6 that performs a predetermined process on a semiconductor device. In this case, the target object S may be a semiconductor substrate (semiconductor wafer) (semiconductor device) processed in the semiconductor process, a gas used in the semiconductor process, or plasma generated in the semiconductor process. The semiconductor process equipment 6 may be, for example, a dry etching equipment used in a dry etching process or a film deposition equipment used in a film deposition process. The spectrometer 1 acquires spectral data of the measurement light L arriving from the target object S in the semiconductor process equipment 6. FIG. 1 illustrates a case in which the semiconductor process equipment 6 is a dry etching equipment, with a semiconductor substrate W disposed in a chamber 6 a. During etching, an etching gas is introduced into the chamber 6 a to generate plasma PL, which etches the semiconductor substrate W. The measurement light L from the dry etching equipment is incident on the spectrometer 2, for example, through a window 6 b formed in the chamber 6 a. The measurement light L includes light originating from the etching gas and light originating from the material of the semiconductor substrate W. In this case, the processing unit 4 of the spectrometer 1 monitors abnormalities in the semiconductor process based on the detection results of the measurement light L. In particular, in this semiconductor process monitoring device 100, the processing unit 4 identifies channels containing noise using a noise detection method described below, removes the noise, and monitors abnormalities in the semiconductor process based on the data from which the noise has been removed.

[0025] Alternatively, as shown in FIG. 29 , the spectrometer 1 may constitute a film thickness measurement device 101. The film thickness measurement device 101 includes the spectrometer 1, a light source unit 7 that generates light to be irradiated onto the object S, and a measurement head 8 that irradiates the object S with the light. In this case, the object S is an optical film, a display panel, a semiconductor substrate (semiconductor wafer) (semiconductor device), or the like. The light source unit 7 is, for example, a white light source. The measurement head 8 is configured to include, for example, a half mirror and a condenser lens. Light from the light source unit 7 is irradiated onto the object S via the measurement head 8. Light reflected by or transmitted through the object S is input to the spectrometer 1 via the measurement head 8 as measurement light L. The film thickness measurement device 101 measures, for example, the film thickness of an optical film, which is the object S, the thickness of a thin film formed on a display panel, or the thickness of a thin film formed on a semiconductor wafer. In this case, the processing unit 4 of the spectrometer 1 measures the film thickness of the object S based on the detection result of the measurement light L. In particular, in this semiconductor process monitoring system 100, the processing unit 4 identifies channels containing noise using a noise detection method described below, removes the noise, and measures the film thickness of the object S based on the noise-removed data. Alternatively, the object S may be a measurement object for luminescence quantum yield measurement placed in an integrating sphere. In this case, the measurement light L includes excitation light for exciting the measurement object and light (e.g., fluorescence) generated by the measurement object.

[0026] The spectroscopic unit 2 disperses the measurement light L arriving from the object S and forms a spectral image on the detection area of ​​the photodetector 3. The spectroscopic unit 2 includes a spectroscopic element such as a grating or a prism, and splits the measurement light L into individual wavelength components. A spectrometer with good imaging performance can be used as the spectroscopic unit 2. The spectroscopic element constituting the spectroscopic unit 2 may be, for example, a Czerny-Turner spectroscopic element, a Dyson spectroscopic element, an Offner spectroscopic element, or the like, which is capable of astigmatism correction.

[0027] The photodetector 3 detects the measurement light L dispersed by the spectroscopic unit 2. Details of the photodetector 3 will be described later. The processing unit 4 processes data output from the photodetector 3. The processing unit 4 is configured by a computer including a processor such as a CPU, and storage media such as RAM and ROM. The processing unit 4 may be configured by a smart device such as a smartphone or tablet terminal that is integrated with a display unit and an input unit, or may be configured by a microcomputer or FPGA (Field-Programmable Gate Array). The photodetector 3 and the processing unit 4 can be considered to constitute a photodetection device 5 that detects the measurement light L dispersed by the spectroscopic unit 2. [Photodetector]

[0028] The photodetector 3 will be described in detail with reference to FIGS. 2 and 3 . As shown in FIGS. 2 and 3 , the photodetector 3 includes a first detection unit 11 and a second detection unit 21. The first detection unit 11 includes a plurality of first pixel columns 12 arranged along a first direction D1 (row direction). Each first pixel column 12 includes a plurality of pixels 10 arranged along a second direction D2 (column direction) perpendicular to the first direction D1. The second detection unit 21 includes a plurality of second pixel columns 22 arranged along the first direction D1. Each second pixel column 22 includes a plurality of pixels 10 arranged along the second direction D2. The second detection unit 21 is disposed adjacent to (facing) the first detection unit 11 in the second direction D2.

[0029] From a different perspective, the photodetector 3 has a plurality of pixels 10 arranged in a first direction D1 (row direction) and a second direction D2 (column direction), and the photodetector 3 has two detection units, a first detection unit 11 and a second detection unit 21, which are divided in the second direction D2. In this example, the number of pixels along the first direction D1 is greater than the number of pixels along the second direction D2, and the detection region formed by the first detection unit 11 and the second detection unit 21 is rectangular with long sides parallel to the first direction D1. The number of pixels 10 arranged in the first detection unit 11 and the second detection unit 21 is equal in both the first direction D1 and the second direction D2. That is, the first detection unit 11 and the second detection unit 21 are configured symmetrically with respect to a center line C that passes through the center in the second direction D2 and is parallel to the first direction D1. Note that the first detection unit 11 and the second detection unit 21 may be configured by different (mutually separated) photodetection elements or may be configured by a common (single) photodetection element.

[0030] Each second pixel column 22 is arranged to be aligned with a corresponding first pixel column 12 in the second direction D2 (i.e., to be positioned on a straight line parallel to the second direction D2). Therefore, there are corresponding pixel columns among the multiple first pixel columns 12 and the multiple second pixel columns 22, and the corresponding first pixel columns 12 and second pixel columns 22 are referred to as one channel. In other words, in the photodetector 3, multiple channels each extending in the second direction D2 are arranged along the first direction D1. Hereinafter, the description will be given assuming that one channel is constituted by the first pixel columns 12 and the second pixel columns 22 aligned in the second direction D2. In this case, the photodetector 3 has multiple channels aligned along the first direction D1.

[0031] A spectral image P formed by the spectroscopic unit 2 is incident on the detection region (first detection unit 11 and second detection unit 21) of the photodetector 3 (FIG. 3). In the example of FIGS. 2 and 3, the first direction D1 is along the wavelength resolution direction by the spectroscopic unit 2, and the second direction D2 is along the charge transfer direction in which charges generated in the pixels 10 are transferred. The spectral image P is incident on the detection region of the photodetector 3 so as to extend across the first detection unit 11 and the second detection unit 21. In other words, the spectroscopic unit 2 is arranged so that the spectral image P is formed (imaged) across the first detection unit 11 and the second detection unit 21. Each pixel 10 receives the wavelength-resolved spectral image P and generates and accumulates charges according to the intensity of the light. FIG. 3 illustrates three spectral images P with different wavelengths. The three spectral images P are imaged while spaced apart from each other in the first direction D1. In this way, different channels are incident for each wavelength. That is, the spectral images P of the measurement light L incident on each channel have different wavelengths. Each spectral image P has, for example, a slit-like (vertically elongated) shape and extends along the second direction D2. The correspondence between channels and wavelengths is the same in the first detection unit 11 and the second detection unit 21.

[0032] The first exposure time of each pixel 10 in the first detection unit 11 and the second exposure time of each pixel 10 in the second detection unit 21 can be set independently of each other. In this embodiment, the second exposure time is set to be longer than the first exposure time.

[0033] The photodetector 3 outputs to the processing unit 4 a first signal S1, which is one-dimensional data in the first direction D1 of the measurement light L detected by the first detection unit 11 during the first exposure time, and a second signal S2, which is one-dimensional data in the first direction D1 of the measurement light L detected by the second detection unit 21 during the second exposure time. This output operation will be described below. As described above, in this embodiment, the first exposure time is shorter than the second exposure time, so in FIG. 2, the first signal S1 is labeled "Short" and the second signal S2 is labeled "Long." Furthermore, in FIG. 3, "Short" is labeled next to the first detection unit 11, and "Long" is labeled next to the second detection unit 21.

[0034] 2 , the photodetector 3 is a CCD (Charge Coupled Device) photodetector having a first horizontal shift register 31 and a second horizontal shift register 32. The first horizontal shift register 31 and the second horizontal shift register 32 are signal processing units that process signals generated in the first detection unit 11 and the second detection unit 21.

[0035] In the first detection unit 11, the electric charges generated and accumulated in each pixel 10 are transferred to the first horizontal shift register 31 along the second direction D2, and the electric charges of the pixels 10 in each first pixel column 12 are added together for each column in the first horizontal shift register 31 (hereinafter, this operation will be referred to as "vertical transfer"). That is, in the photodetector 3, signals from the multiple pixels 10 included in the first pixel column 12 are added together (analog binned) to generate an output signal from the first pixel column 12. Therefore, it can be considered that the multiple pixels 10 included in one first pixel column 12 (channel) constitute one pixel.

[0036] Next, the charges added up for each column in the first horizontal shift register 31 are sequentially read out from the first horizontal shift register 31 (hereinafter, this operation will be referred to as "horizontal transfer"). Then, an electrical signal (e.g., a signal indicating a voltage value) corresponding to the amount of charge read out from the first horizontal shift register 31 is output from the first amplifier 33, and the electrical signal is AD (analog-to-digital) converted by an AD converter into a digital value. The digital value is output to the processing unit 4. In this way, a first signal S1 is output to the processing unit 4. The first signal S1 is an intensity signal of the measurement light L in each channel and is spectrum data of the measurement light L.

[0037] In the second detection unit 21, the electric charges generated and accumulated in each pixel 10 are transferred to the second horizontal shift register 32 along the second direction D2, and the electric charges of the pixels 10 in each second pixel column 22 are added together for each column in the second horizontal shift register 32 (vertical transfer). That is, in the photodetector 3, signals from the multiple pixels 10 included in the second pixel column 22 are added together (analog binned) to generate an output signal from the second pixel column 22. Therefore, it can be considered that the multiple pixels 10 included in one second pixel column 22 (channel) constitute one pixel.

[0038] Next, the charges added up for each column in the second horizontal shift register 32 are sequentially read out from the second horizontal shift register 32 (horizontal transfer). Then, an electrical signal (e.g., a signal indicating a voltage value) corresponding to the amount of charge read out from the second horizontal shift register 32 is output from the second amplifier 34, and the electrical signal is AD converted by an AD converter into a digital value. The digital value is output to the processing unit 4. In this way, the second signal S2 is output to the processing unit 4. The second signal S2 is an intensity signal of the measurement light L in each channel and is spectrum data of the measurement light L.

[0039] As shown in FIGS. 4( a) and 4(b), the first signal S1 and the second signal S2 have a constant multiplication relationship. For example, the left side of FIGS. 4(a) and 4(b) shows examples of the first signal S1 and the second signal S2. The right side of FIG. 4(a) shows a signal obtained by multiplying the intensity of the first signal S1 by 1 (i.e., the first signal S1). The right side of FIG. 4(b) shows a signal obtained by multiplying the intensity of the second signal S2 by (1 / N) (the normalized second signal S2). N corresponds to the ratio of the second exposure time to the first exposure time. As shown in FIGS. 4(a) and 4(b), the first signal S1 corresponds to a signal obtained by multiplying the second signal S2 by (1 / N). The normalized second signal S2 is used in the following noise detection method. [Noise Detection Method]

[0040] The noise detection operation (detection method) in the spectrometer 1 will be described below. The following noise detection operation is performed by the processing unit 4 based on the data (first signal S1 and second signal S2) output from the photodetector 3. The noise to be detected is noise that occurs suddenly, such as noise caused by cosmic rays. When cosmic rays enter the detection area of ​​the photodetector 3, the intensity of the first signal S1 or the second signal S2 in the channel into which the cosmic rays entered increases sharply, as shown by the dashed line in the graph in the upper left of Figure 5.

[0041] The following method can be considered as a method for detecting such noise. Because the probability of cosmic rays being incident on the photodetector 3 during a certain period is low and cosmic rays are extremely small, corresponding in size to one or several pixels 10, they are likely to be detected by only one of the first detection unit 11 and the second detection unit 21 ( FIG. 3 ). Therefore, it is considered possible to detect noise caused by cosmic rays by calculating the intensity difference between the first signal S1 and the second signal S2 in the same channel and monitoring the symmetry of the two spectral data. Since real-time detection of such noise is preferable, it is desirable to perform detection using simple processing that requires as little computational effort as possible.

[0042] For example, in the example of Figure 5, the first signal S1 contains noise due to cosmic rays, as indicated by the dashed line. In this case, as shown in Figure 5, it is considered possible to detect noise due to cosmic rays by calculating the difference between the first signal S1 and the second signal S2 for each channel (wavelength) and comparing the difference with a predetermined threshold. That is, if the difference is less than the threshold, it is determined that no noise is contained, while if the difference is equal to or greater than the threshold, it is determined that noise is contained. The threshold may be set to a constant value greater than the readout noise and dark current noise (i.e., noise originating from the sensor) of the photodetector 3, for example.

[0043] However, the inventors have found that this method may result in false positives and may not accurately detect noise. For example, in the example shown in FIG. 6 , a channel containing noise due to cosmic rays is indicated by the symbol N. In this example, the difference between the first signal S1 and the second signal S2 in the channel containing noise due to cosmic rays is greater than the threshold. On the other hand, the two channels indicated by the symbol M are not affected by cosmic rays, but the difference in these channels is also greater than the threshold. In this case, the signal in these channels would be erroneously detected as containing noise. Note that in the graph in FIG. 6 , the unit of intensity "DN" on the vertical axis represents the count value. This also applies to FIGS. 7 to 9 , which will be described later.

[0044] Therefore, in the noise detection method of this embodiment, a threshold is set for each channel based on the first signal S1 and the second signal S2. That is, instead of setting a constant threshold regardless of the channel as in the above-described method, a threshold is (dynamically) set for each channel depending on the first signal S1 and the second signal S2. In a specific noise detection operation, the processing unit 4 sets a threshold for each channel and identifies (detects) a channel in which noise is included in the first signal S1 or the second signal S2 based on the threshold for each channel and the difference between the first signal S1 and the second signal S2.

[0045] First, a first technique of the noise detection method of this embodiment will be described. In the first technique, a threshold is set for each channel based on the magnitude of the first signal S1 and the second signal S2 in that channel. Specifically, in this example, σ expressed by the following equation (1) is calculated for each channel, and the threshold is set based on δ. In this case, the threshold may be a value obtained by multiplying δ by a predetermined coefficient K, for example, 5σ (coefficient K is 5). Note that the coefficient K by which σ is multiplied is not limited to 5 and may be any positive value greater than 1.

[0046] In formula (1), σ １ａcorresponds to the read noise and dark current noise for the first signal S1 (first detection unit 11). １ｂ corresponds to the shot noise of the first signal S1 (first detection unit 11). The shot noise is noise that increases in proportion to the magnitude of the detection signal, and σ １ｂ is proportional to the magnitude of the first signal S1. ２ａ corresponds to the read noise and dark current noise for the second signal S2 (second detection unit 21). ２ｂ corresponds to the shot noise of the second signal S2 (second detection unit 21). ２ｂ is proportional to the magnitude of the second signal S2. The readout noise and dark current noise are noises originating from the sensor, and the shot noise is noise originating from the measurement light L.

[0047] FIG. 7 shows an example in which the threshold is set using the first method. In this example, the first signal S1 and the second signal S2 are set in the same manner as in the example in FIG. 6. In this example, too, the difference between the first signal S1 and the second signal S2 is greater than the threshold in the channel (N) where noise due to cosmic rays is occurring. Furthermore, the difference is smaller than the threshold in the two channels (M) where no cosmic rays are incident. In this case, it can be correctly determined that the signals of those channels do not contain noise. In this way, the first method enables more accurate noise detection than, for example, setting the threshold to a constant value regardless of the channel.

[0048] The inventors further discovered that, although the first method can detect noise more accurately than the above-described case where the threshold is set to a constant value, there is a possibility of false positives. For example, in the example shown in Figure 8, the difference between the first signal S1 and the second signal S2 is larger than the threshold in the channel (N) where noise due to cosmic rays is occurring, but the difference is also larger than the threshold in two channels (M) where cosmic rays are not incident. In this case, the signal of that channel is erroneously detected as containing noise.

[0049] Therefore, in a second technique of the noise detection method of this embodiment, a threshold is set for each channel based on the magnitude of the first signal S1 and the second signal S2 in that channel and in channels adjacent to that channel. The second technique will be described below.

[0050] First, optical distortion that may occur in the first signal S1 and the second signal S2 will be described with reference to FIGS. 9 and 10 . Ideally, the first signal S1 and the second signal S2 should match each other. However, as shown in FIG. 9 , in reality, there may be a discrepancy between the first signal S1 and the second signal S2. This is thought to be because, as shown in FIG. 10 , the spectral image P incident on the detection region (first detection unit 11 and second detection unit 21) of the photodetector 3 is tilted with respect to the second direction D2 due to optical distortion, becoming asymmetric with respect to the center line C. In other words, this is thought to be because the spectral image P of a certain wavelength is formed across multiple channels. Such optical distortion may occur, for example, due to an installation error (misalignment) between the spectroscopic unit 2 and the photodetector 3.

[0051] The deviation between the first signal S1 and the second signal S2 due to optical distortion is usually smaller than the width of several channels (e.g., two or three channels), and in many cases smaller than the width of one channel (i.e., at the sub-channel level). Due to this deviation, as in the examples of Figures 6 and 8 described above, it is thought that the difference between the first signal S1 and the second signal S2 may become larger than the threshold value even in two channels (M) where no cosmic rays are incident. Therefore, in the second technique of the noise detection method of this embodiment, a threshold value is set for each channel, taking into account the influence of optical distortion.

[0052] The second technique will be described with reference to Figures 11 to 15. The second technique is based on the idea that the difference X1 between the first signal S1 and the second signal S2 in a certain channel, as shown in Figure 11(a), should be smaller than the difference X2 between the first signal S1 in that channel and the first signal S1 in an adjacent channel that is close to that channel, as shown in Figure 11(b). If the difference X1 is larger than the difference X2, such a deviation cannot be explained by the optical distortion described above.

[0053] The adjacent channel is, for example, a channel adjacent to the channel in question, but it may be any channel that is close to the channel in question, and may be a channel within one to five channels from the channel in question. The adjacent channel may be the channel two channels away from the channel in question, or the channel three channels away from the channel in question. In other words, the adjacent channel may be a channel within one to three channels from the channel in question.

[0054] 12(a) and 12(b), in the second technique of the noise detection method of this embodiment, first, the first signal S1 and the second signal S2 are discretized from continuous values ​​to discrete values ​​for each channel. Note that although only the first signal S1 is shown in FIGS. 12(a) and 12(b), the second signal S2 is also discretized in the same manner.

[0055] 13(a) and 13(b) illustrate the first signal S1 and the second signal S2 when there is a 1 / 4-channel offset between the first signal S1 and the second signal S2, as well as the first signal S1 (the first signal S1 in the adjacent channel) when there is an offset of just one channel. FIG. 14(a) corresponds to FIG. 13(b). FIG. 14(b) illustrates the difference X1 between the first signal S1 and the second signal S2 in each channel for the example of FIG. 14(a). Also, FIG. 14(b) illustrates the difference X2 between the first signal S1 in that channel and the first signal S1 when it is offset by just one channel from that channel. As shown in FIG. 14(b), the difference X1 is smaller than the difference X2 in each channel.

[0056] In the second technique of the noise detection method of this embodiment, at least one of the differences A1, A2, B1, and B2 (difference values) shown in FIGS. 15A and 15B is used to set the threshold. The difference A1 is the difference between the value of the first signal S1 in the target channel and the value of the first signal S1 in an adjacent channel (first adjacent channel) adjacent to the target channel on one side of the first direction D1 (the right side in FIG. 15). The difference A2 is the difference between the value of the first signal S1 in the target channel and the value of the first signal S1 in an adjacent channel adjacent to the target channel on the other side of the first direction D1 (the left side in FIG. 15). The difference B1 is the difference between the value of the second signal S2 in the target channel and the value of the second signal S2 in an adjacent channel adjacent to the target channel on one side of the first direction D1. The difference B2 is the difference between the value of the second signal S2 in the target channel and the value of the second signal S2 in an adjacent channel (second adjacent channel) adjacent to the target channel on the other side of the first direction D1.

[0057] 16(a) and 16(b) are diagrams illustrating a first example of the second technique. In the first example, the threshold is set to the average value of the differences A1, A2, B1, and B2 (difference values). That is, the threshold is set to (difference A1 + difference A2 + difference B1 + difference B2) / 4. FIG. 16(a) shows an example of the first signal S1 and the second signal S2, and the differences in each case. As shown in FIG. 16(b), setting the threshold as in the first example enables accurate detection of noise caused by cosmic rays.

[0058] 17(a) and 17(b) are diagrams illustrating a second example of the second technique. In the second example, the threshold is set to the average value of the differences A2 and B1 (difference values). That is, the threshold is set to (difference A2 + difference B1) / 2. FIG. 17(a) corresponds to FIG. 16(a). As shown in FIG. 17(b), even when the threshold is set as in the second example, noise caused by cosmic rays can be detected with high accuracy.

[0059] 18(a) and 18(b) are diagrams illustrating a third example of the second technique. In the third example, the threshold is set to the average value of the differences A1 and B1 (difference values). That is, the threshold is set to (difference A1 + difference B1) / 2. FIG. 18(a) corresponds to FIG. 16(a). As shown in FIG. 18(b), even when the threshold is set as in the third example, noise caused by cosmic rays can be detected with high accuracy.

[0060] 19(a) and 19(b) are diagrams illustrating a fourth example of the second technique. In the fourth example, the threshold is set to the average value of the differences A2 and B2 (difference values). That is, the threshold is set to (difference A2 + difference B2) / 2. FIG. 19(a) corresponds to FIG. 16(a). As shown in FIG. 19(b), even when the threshold is set as in the fourth example, noise caused by cosmic rays can be detected with high accuracy.

[0061] 20(a) and 20(b) are diagrams illustrating a comparative example. In the comparative example, the threshold is set to the average value of the differences A1 and A2 (difference values). That is, the threshold is set to (difference A1 + difference A2) / 2. FIG. 20(a) corresponds to FIG. 16(a). In the comparative example, there is a possibility that noise caused by cosmic rays cannot be detected accurately. This is because, as described above, cosmic rays are likely to be detected by only one of the first detection unit 11 and the second detection unit 21. Therefore, if the threshold is set based on the differences A1 and A2 calculated from only the first signal S1, noise cannot be detected when it is contained in the second signal S2.

[0062] As described above, by setting the threshold based on the difference value (difference A1, A2, B1, or B2), the influence of noise due to cosmic rays on the threshold setting can be mitigated, and noise due to cosmic rays can be detected with high accuracy. Furthermore, as described above, by setting the threshold based on a difference value including at least one of the differences A1 and A2 and one of the differences B1 and B2, noise due to cosmic rays can be detected with high accuracy. In particular, by setting the threshold based on the four differences A1, A2, B1, and B2, as in the first example, noise due to cosmic rays can be detected with even higher accuracy.

[0063] Furthermore, when selecting two differences as difference values, the combination of difference A1 and difference B2 and the combination of difference A2 and difference B1 (second example, FIG. 17 ) are more preferable than the combination of difference A1 and difference B1 (third example, FIG. 18 ) and the combination of difference A2 and difference B2 (fourth example, FIG. 19 ). This is because, as shown in FIG. 10 , the spectral image P is asymmetric with respect to the center line C due to optical distortion, and therefore it is preferable to combine differences on both sides of the first direction D1. For this reason, while the threshold is appropriately set for each channel in the second example of FIG. 17 , the difference between the first signal S1 and the second signal S2 is considered to be close to the threshold in channel 8 in the third example of FIG. 18 and in channels 7 and 9 in the fourth example of FIG. 19 .

[0064] FIG. 21 is a graph showing an example of a combination of the first and second methods. In this example, the first and second methods are combined, and the threshold is set based on the above-described K×σ and the average value of the differences A1, A2, B1, and B2. For example, the threshold may be set to a value obtained by adding K×σ to the average value of the differences A1, A2, B1, and B2. In the example of FIG. 21, the difference between the first signal S1 and the second signal S2 is greater than the threshold in the channel (N) where noise due to cosmic rays is occurring. Furthermore, the difference is smaller than the threshold in the two channels (M) where cosmic rays are not incident. In this way, even when the first and second methods are combined, noise due to cosmic rays can be detected with high accuracy.

[0065] 22(a) and 22(b) are diagrams illustrating threshold-added data. For noise detection, as shown in FIG. 22(a), the difference between the first signal S1 and the second signal S2 for each channel may be calculated and the calculated difference may be compared with a threshold. That is, the processing unit 4 may compare the difference between the first signal S1 and the second signal S2 for each channel with a threshold to determine that the first signal S1 or the second signal S2 for that channel contains noise. More specifically, if the difference is greater than the threshold, the processing unit 4 may determine that the first signal S1 or the second signal S2 for that channel contains noise, and if the difference is equal to or less than the threshold, the processing unit 4 may determine that the first signal S1 or the second signal S2 for that channel does not contain noise.

[0066] Alternatively, as shown in FIG. 22( b), the processing unit 4 may create threshold addition data for each channel by adding a threshold to the smaller of the first and second signals S1 and S2, and compare the first and second signals S1 and S2 with the threshold addition data to identify a channel in which noise is included in the first or second signal S2. More specifically, the processing unit 4 may determine that noise is included in the first or second signal S1 or S2 in a channel if the larger of the first and second signals S1 and S2 is greater than the threshold addition data, and may determine that noise is not included in the first or second signal S1 or S2 in a channel if the larger of the first and second signals S1 and S2 is equal to or less than the threshold addition data. By using the threshold addition data, a threshold can be set for the first and second signals S1 and S2, rather than the difference between the first and second signals S1 and S2.

[0067] 23(a) and 23(b) are graphs showing an example of the results of applying the second method. In FIG. 23(b), the portion indicated by the oval in FIG. 23(a) is shown enlarged. In this example, similar to the first example described above, the threshold is set to the average value of the differences A1, A2, B1, and B2. In this example, noise caused by cosmic rays is added by simulation to the measured data of the second signal S2. The data marked "before correction" is the larger value of the first signal S1 or the second signal S2, which in this example is the second signal S2.

[0068] 23(b), the second signal S2 is larger than the threshold sum data in the channel containing noise due to cosmic rays, and the second signal S2 is smaller than the threshold sum data in the other channels not containing cosmic rays. In this way, by using the second method, noise due to cosmic rays can be detected with high accuracy.

[0069] Furthermore, as shown in FIG. 23( b), after identifying a channel containing noise, the value of that channel may be interpolated using the values ​​of the channels on either side of that channel. In FIG. 23( b), the data after the interpolation is shown as "after correction." For example, the value of that channel may be interpolated using a line segment that smoothly connects the values ​​of the channels on either side of that channel. That is, the processing unit 4 may interpolate the value of a channel identified as containing noise in the first signal S1 or the second signal S2 based on the values ​​of the channels on either side of that channel.

[0070] 24(a) and 24(b) are graphs showing another example of the application results of the second method. In FIG. 24(b), the oval portion in FIG. 24(a) is shown enlarged. As shown in FIG. 24(b), this example also shows that noise caused by cosmic rays can be detected with high accuracy by using the second method. Furthermore, although the example in FIG. 24 adds noise that is steeper than the example in FIG. 23, it can be seen that such noise can also be detected well. Therefore, the second method can be suitably used when detecting measurement light L having a broad spectrum or when the brightness of cosmic rays is high.

[0071] Figures 25(a), 25(b), and 25(c) are tables for explaining the results of actual measurements. Figure 25(a) shows the results of measurements performed without a light source (with no light incident on the photodetector 3) while detecting noise with a constant threshold value. In the experiment shown in Figure 25(a), the number of channels was 2048, and 6208 measurements were performed. The number of cosmic rays detected was 501, and the average number of channels into which cosmic rays entered was 1.13. The measurement time was 6208 seconds, and the cosmic ray incidence rate was 4.84 / min.

[0072] FIG. 25(b) shows the results of measurements performed with a light source (light incident on the photodetector 3) while detecting noise with a constant threshold value. In the experiment shown in FIG. 25(b), 2,048 channels were used, and 12,465 measurements were performed. 789 cosmic rays were detected, and the average number of channels through which cosmic rays entered was 1.23. The measurement time was 623 seconds, and the cosmic ray incidence rate was 75.99 / min. Since the cosmic ray incidence rate in FIG. 25(b) is significantly higher than in FIG. 25(a), it can be seen that detecting noise with a constant threshold value resulted in many false positives.

[0073] FIG. 25(c) shows the results of an actual measurement performed with a light source (light incident on the photodetector 3) while detecting noise by setting the threshold to the average value of the differences A1, A2, B1, and B2, as in the first example of the second method. In the experiment shown in FIG. 25(c), the number of channels was 2048, and 12,465 measurements were performed. The number of cosmic rays detected was 41, and the average number of channels into which cosmic rays entered was 1.12. The measurement time was 623 seconds, and the cosmic ray incidence rate was 3.95 / min. Since the cosmic ray incidence rate in FIG. 25(c) is similar to that in FIG. 25(a), it can be seen that noise caused by cosmic rays can be detected with high accuracy by using the second method. [Actions and Effects]

[0074] In the photodetector 5, the processing unit 4 sets a threshold for each channel and identifies a channel in which noise is contained in the first signal S1 or the second signal S2 based on the threshold for each channel and the difference between the first signal S1 and the second signal S2. In the first and second methods described above, the threshold is set for each channel based on the first signal S1 and the second signal S2. By setting the threshold for each channel based on the first signal S1 and the second signal S2 in this way, noise can be detected more accurately than, for example, when the threshold is set to a constant value regardless of the channel. Furthermore, in the first and second methods described above, the threshold is set for each channel based on the magnitude of the first signal S1 and the magnitude of the second signal S2 in that channel and / or channels adjacent to that channel. That is, in the first method, the threshold is set for each channel based on the magnitude of the first signal S1 and the second signal S2 in that channel, and in the second method, the threshold is set for each channel based on the magnitude of the first signal S1 and the second signal S2 in that channel and channels adjacent to that channel. This makes it possible to detect noise with high accuracy.

[0075] In the first method described above, the threshold value is set for each channel based on the magnitude of the first signal S1 and the magnitude of the second signal S2 (shot noise) in that channel, which allows the threshold value to be set appropriately and enables noise to be detected with greater accuracy.

[0076] In the second method described above, the threshold is set based on the difference (difference A1 or difference A2) between the value of the first signal S1 in the target channel and the value of the first signal S1 in the adjacent channel (first adjacent channel) that is close to the target channel, and the difference (difference B1 or difference B2) between the value of the second signal S2 in the target channel and the value of the second signal S2 in the adjacent channel that is close to the target channel. This allows the threshold to be set appropriately, making it possible to detect noise with even greater accuracy.

[0077] In the photodetector 5, the processing unit 4 may create threshold sum data by adding a threshold to the smaller value of the first signal S1 or the second signal S2 for each channel, and identify channels in which noise is included in the first signal S1 or the second signal S2 by comparing the first signal S1 and the second signal S2 with the threshold sum data. In this case, noise detection can be performed effectively.

[0078] In the light detection device 5, the first signal S1 is one-dimensional data of the measurement light L detected by the first detection unit 11 over a first exposure time, and the second signal S2 is one-dimensional data of the measurement light L detected by the second detection unit 21 over a second exposure time. The second exposure time is longer than the first exposure time. This allows the dynamic range to be increased by combining data (the first signal S1 and the second signal S2) having different exposure times.

[0079] The processing unit 4 interpolates the value of a channel identified as containing noise in the first signal S1 or the second signal S2 based on the values ​​of the channels on both sides of the channel, thereby reducing the effect of noise on light detection.

[0080] In the photodetection device 5, the photodetector 3 is a CCD photodetector having a first horizontal shift register 31 to which charges generated in the plurality of first pixel rows 12 of the first detection unit 11 are transferred, and a second horizontal shift register 32 to which charges generated in the plurality of second pixel rows 22 of the second detection unit 21 are transferred. This makes it possible to avoid an increase in readout noise when reading out charges generated in each pixel 10.

[0081] The spectrometer 1 includes a spectroscopic unit 2 that splits the measurement light L, and a photodetector 5 that detects the measurement light L split by the spectroscopic unit 2 in a first detection unit 11 and a second detection unit 21. For the reasons described above, the spectrometer 1 enables noise to be detected with high accuracy.

[0082] The noise detection method according to the embodiment includes the steps of: using a photodetector 3 to output a first signal S1 and a second signal S2; setting a threshold for each channel; and identifying a channel containing noise in the first signal S1 and the second signal S2 based on the threshold for the channel and the difference between the first signal S1 and the second signal S2. In the step of setting the threshold, the threshold is set based on a difference value. The difference value includes a difference (difference A1 or difference A2) between the value of the first signal S1 in a target channel and the value of the first signal S1 in an adjacent channel (first adjacent channel) adjacent to the target channel, and a difference (difference B1 or difference B2) between the value of the second signal S2 in the target channel and the value of the second signal S2 in an adjacent channel adjacent to the target channel. For the reasons described above, this noise detection method enables accurate noise detection.

[0083] The present disclosure is not limited to the above embodiment. For example, as shown in FIGS. 26 and 27 , the photodetector 3 may be a CCD photodetector having a plurality of first storage units 45A, a plurality of second storage units 45B, a plurality of first readout units 46A, and a plurality of second readout units 46B. The first storage units 45A respectively store electric charges generated in the plurality of first pixel columns 12 of the first detection unit 11 (arrow A1). The second storage units 45B respectively store electric charges generated in the plurality of second pixel columns 22 of the second detection unit 21 (arrow A2). The first readout units 46A are electrically connected to the plurality of first storage units 45A, respectively, and output electric signals corresponding to the magnitudes of the electric charges stored in the plurality of first storage units 45A. The second readout units 46B are electrically connected to the plurality of second storage units 45B, respectively, and output electric signals corresponding to the magnitudes of the electric charges stored in the plurality of second storage units 45B. Even in this case, it is possible to avoid an increase in readout noise when reading out charges generated in each pixel 10. Fig. 26 shows a transistor 51A and a bonding pad 52A that constitute the first readout unit 46A. The transistor 51A is electrically connected to a bonding pad 54A via a wiring 53A. Fig. 27 shows a transistor 51B and a bonding pad 52B that constitute the second readout unit 46B. The transistor 51B is electrically connected to a bonding pad 54B via a wiring 53B.

[0084] Alternatively, as shown in FIG. 28 , the photodetector 3 may be a CMOS (Complementary Metal Oxide Semiconductor) photodetector having a first readout unit 47A and a second readout unit 47B. In this case, the first readout unit 47A outputs an electrical signal corresponding to the magnitude of the charge accumulated in each pixel 10 of the first detection unit 11, and the second readout unit 47B outputs an electrical signal corresponding to the magnitude of the charge accumulated in each pixel 10 of the second detection unit 21. In this case, power consumption can be reduced compared to, for example, a CCD photodetector that reads out charge column by column. Furthermore, an increase in readout noise when reading out charge generated in each pixel can be avoided.

[0085] FIG. 28 shows the photodiode 61 and amplifier 62 that constitute the pixel 10. The photodiode 61 accumulates electrons (photoelectrons) generated by the input of the measurement light L as electric charges. The amplifier 62 converts the electric charges accumulated in the photodiode 61 into an electric signal (e.g., a signal indicating a voltage value) and amplifies the electric signal. The electric signal amplified by the amplifier 62 is transferred to vertical signal lines 64 connecting the pixels 10 in the row direction by switching the selection switch 63 of each pixel 10. A CDS (correlated double sampling) circuit 65 is disposed on each vertical signal line 64. The CDS circuit 65 reduces readout noise between the pixels 10 and temporarily stores the electric signal transferred to the vertical signal line 64. 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 detection unit 11 is output to the processing unit 4 via the first readout unit 47A. The digital value corresponding to the second detection unit 21 is output to the processing unit 4 via the second readout unit 47B.

[0086] In the above embodiment, the first pixel row 12 includes a plurality of pixels 10. However, the first pixel row 12 may include at least one pixel 10, and may be composed of only one pixel 10. Similarly, the second pixel row may include at least one pixel 10, and may be composed of only one pixel 10. That is, each of the first detection unit 11 and the second detection unit 21 may be composed of pixels 10 lined up in a line along the first direction D1. The photodetector 3 may be a TDI (Time Delay Integration) line sensor or a one-dimensional line sensor.

[0087] In the above example, the adjacent channel was a channel adjacent to the target channel. However, as described above, the adjacent channel may be a channel within a range of one to three adjacent channels from the target channel. In the above example, only the difference between the first signal S1 or the second signal S2 and one adjacent channel adjacent to the target channel on one side of the first direction D1 was considered. However, differences between two or more adjacent channels may also be considered. For example, instead of the difference A1, the difference A1 between the channel immediately adjacent to the target channel and the average value of the differences between the channel two adjacent to the target channel may be used to set the threshold. Similarly, instead of the difference A1, the average value (moving average value) of the differences between five channels within a range of five adjacent channels from the target channel may be used to set the threshold.

[0088] The threshold value may be set to a value corresponding to the difference value, and need not necessarily be set to the average value of the difference value. The average value is not limited to an arithmetic mean and may be a weighted mean, etc. The first exposure time of the first detection unit 11 may be equal to the second exposure time of the second detection unit 21. In the above example, the processing unit 4 interpolated the value of a channel identified as containing noise in the first signal S1 or the second signal S2 based on the values ​​of the channels on both sides of the channel. However, this interpolation is not necessary. For example, the value of the channel identified as containing noise may simply be excluded from the acquired data. The noise detection method of this embodiment is not limited to detecting noise caused by cosmic rays, and can also be applied to detecting other types of noise that occur suddenly. The photodetector 5 may be used for applications other than the spectrometer 1.

[0089] 1...spectroscope, 2...spectrometry section, 3...photodetector, 4...processing section, 5...photodetection device, 10...pixel, 11...first detection section, 12...first pixel row, 21...second detection section, 22...second pixel row, 31...first horizontal shift register, 32...second horizontal shift register, 45A...first storage section, 45B...second storage section, 46A, 47A...first readout section, 46B, 47B...second readout section, 100...semiconductor process monitoring device, 101...film thickness measurement device, A1, A2, B1, B2...difference, D1...first direction, D2...second direction, L...measurement light, S1...first signal, S2...second signal.

Claims

1. A light detector that detects measurement light; and a processing unit that processes data output from the light detector, wherein the light detector has a first detection unit having a plurality of first pixel rows arranged along a first direction, each of the plurality of first pixel rows including at least one pixel arranged along a second direction perpendicular to the first direction, and a second detection unit having a plurality of second pixel rows arranged along the first direction, each of the plurality of second pixel rows including at least one pixel arranged along the second direction, and the first detection unit and the second detection unit are arranged adjacent to each other in the second direction, the light detector outputs to the processing unit a first signal that is one-dimensional data of the measurement light in the first direction detected by the first detection unit, and a second signal that is one-dimensional data of the measurement light in the first direction detected by the second detection unit, and wherein the first pixel row and the second pixel row arranged in the second direction form one channel, the light detector has a plurality of channels arranged along the first direction, The processing unit sets a threshold for each of the multiple channels, and identifies the channel in which the first signal or the second signal contains noise based on the threshold for each of the multiple channels and a difference between the first signal and the second signal, and the threshold is set for each of the multiple channels based on the first signal and the second signal.

2. The optical detection device of claim 1, wherein the threshold is set for each of the plurality of channels based on the magnitude of the first signal and the magnitude of the second signal in the channel.

3. An optical detection device as described in claim 1 or 2, wherein the threshold is set based on the difference between the value of the first signal in the channel and the value of the first signal in a first adjacent channel adjacent to the channel, and the difference between the value of the second signal in the channel and the value of the second signal in a second adjacent channel adjacent to the channel.

4. The optical detection device according to any one of claims 1 to 3, wherein the processing unit creates threshold addition data for each of the multiple channels by adding the threshold to the smaller value of the first signal or the second signal, and identifies the channel in which the first signal or the second signal contains noise by comparing the first signal and the second signal with the threshold addition data.

5. An optical detection device as described in any one of claims 1 to 4, wherein the first signal is the one-dimensional data of the measurement light detected in the first detection unit at a first exposure time, the second signal is the one-dimensional data of the measurement light detected in the second detection unit at a second exposure time, and the second exposure time is longer than the first exposure time.

6. An optical detection device as claimed in any one of claims 1 to 5, wherein the processing unit interpolates the value of the channel identified as containing the noise in the first signal or the second signal based on the values ​​of the channels on both sides of the channel.

7. A photodetection device according to any one of claims 1 to 6, wherein the photodetector is a CCD photodetector having a first horizontal shift register to which charges generated in the plurality of first pixel rows of the first detection unit are transferred, and a second horizontal shift register to which charges generated in the plurality of second pixel rows of the second detection unit are transferred.

8. The photodetector according to any one of claims 1 to 6, wherein the photodetector is a CCD photodetector having: a plurality of first accumulation sections in which charges generated in the plurality of first pixel rows of the first detection section are respectively accumulated; a plurality of second accumulation sections in which charges generated in the plurality of second pixel rows of the second detection section are respectively accumulated; a plurality of first readout sections electrically connected respectively to the plurality of first accumulation sections and outputting an electrical signal corresponding to the magnitude of the charge accumulated in the plurality of first accumulation sections; and a plurality of second readout sections electrically connected respectively to the plurality of second accumulation sections and outputting an electrical signal corresponding to the magnitude of the charge accumulated in the plurality of second accumulation sections.

9. A photodetector device according to any one of claims 1 to 6, wherein the photodetector is a CMOS photodetector having: a first readout section that outputs an electrical signal corresponding to the magnitude of the charge accumulated in each of the pixels of the first detection section; and a second readout section that outputs an electrical signal corresponding to the magnitude of the charge accumulated in each of the pixels of the second detection section.

10. A spectrometer comprising: a spectroscopic section which disperses the measurement light and forms an image across the first detection section and the second detection section; and a light detection device according to any one of claims 1 to 9, which detects the measurement light dispersed by the spectroscopic section in the first detection section and the second detection section.

11. A film thickness measuring device comprising: the spectrometer according to claim 10; and a light source unit which generates light to be irradiated onto an object, wherein the processing unit measures the film thickness of the object based on the data from which the noise has been removed.

12. A semiconductor process monitoring device comprising: the spectrometer according to claim 10; and a semiconductor process equipment which processes a semiconductor device in a predetermined process, wherein the photodetector detects the measurement light generated within the semiconductor process equipment by the predetermined process, and the processing unit monitors anomalies in the semiconductor process based on the data from which the noise has been removed.

13. A photodetector comprising: a first detection unit having a plurality of first pixel rows arranged along a first direction, each of the plurality of first pixel rows including at least one pixel arranged along a second direction perpendicular to the first direction; and a second detection unit having a plurality of second pixel rows arranged along the first direction, each of the plurality of second pixel rows including at least one pixel arranged along the second direction, the first detection unit and the second detection unit being arranged adjacent to each other in the second direction, wherein, assuming that one channel is formed by the first pixel rows and the second pixel rows arranged in the second direction, a step of outputting, using the photodetector having a plurality of channels arranged along the first direction, a first signal which is one-dimensional data in the first direction of the measurement light detected by the first detection unit, and a second signal which is one-dimensional data in the first direction of the measurement light detected by the second detection unit; and a step of setting a threshold value for each of the plurality of channels. and identifying the channel in which noise is included in the first signal and the second signal based on the threshold for the channel and a difference between the first signal and the second signal, wherein in the threshold setting step, the threshold is set for each of the multiple channels based on the first signal and the second signal.

Images