WO2018105462A1 - Signal processing method and program - Google Patents

Signal processing method and program Download PDF

Info

Publication number
WO2018105462A1
WO2018105462A1 PCT/JP2017/042854 JP2017042854W WO2018105462A1 WO 2018105462 A1 WO2018105462 A1 WO 2018105462A1 JP 2017042854 W JP2017042854 W JP 2017042854W WO 2018105462 A1 WO2018105462 A1 WO 2018105462A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
waveform
fitted
processing method
peak
Prior art date
Application number
PCT/JP2017/042854
Other languages
French (fr)
Japanese (ja)
Inventor
康敏 梅原
Original Assignee
東京エレクトロン株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 東京エレクトロン株式会社 filed Critical 東京エレクトロン株式会社
Priority to JP2018554949A priority Critical patent/JP6742435B2/en
Publication of WO2018105462A1 publication Critical patent/WO2018105462A1/en

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
    • G01N15/10Investigating individual particles
    • G01N15/14Electro-optical investigation, e.g. flow cytometers

Definitions

  • the present invention relates to a signal processing method and a program.
  • Patent Document 1 It has been proposed to detect fine particles in a body to be inspected such as a liquid by using the inspection light irradiated to the liquid or gas being shielded by the fine particles contained in the liquid or gas. See).
  • a target signal and noise in the signal to be inspected are classified according to the result of matching between the signal to be inspected and a waveform template specified in advance.
  • the frequency band of the target signal is different from the frequency band of noise. Further, the intensity (spectrum) of the target signal is larger than the noise intensity.
  • the signal-to-noise ratio (noise-to-signal ratio) is a signal with a low signal-to-noise ratio such as 2 or less, and the frequency band of the target signal is close to the frequency band of noise, It is difficult to distinguish from noise.
  • the signal to be inspected by the above-mentioned patent document is a signal having a relatively simple waveform that can be represented by a Gaussian waveform, a Lorentz shape or the like.
  • the signal processing method disclosed in the above-mentioned patent document further makes it difficult to separate the target signal and noise.
  • an object of one aspect of the present invention is to accurately determine a signal of a target particle from a detection signal regardless of the S / N ratio in detecting a particle in a liquid or a gas. To do.
  • a signal of light irradiated on a fluid inspection object is detected, a peak of the waveform of the detected signal is searched, and a waveform width of the searched peak And calculating the feature quantity of the fitted waveform, and based on whether the calculated feature quantity of the waveform is within a predetermined threshold range, the fitted waveform signal is a fine particle in the inspection object.
  • a signal processing method in which a computer executes a process for determining whether or not the signal indicates a signal.
  • route is detected is calculated using the Hermite-Gauss mode formula based on a Gaussian function.
  • the figure for demonstrating the fitting of the waveform which concerns on one Embodiment. The figure which shows an example of the threshold value of the rule of the feature-value which concerns on one Embodiment.
  • the figure which shows an example of the feature-value and determination result of the detection signal and waveform which concern on one Embodiment. The flowchart which shows an example of the signal processing according to S / N ratio which concerns on one Embodiment.
  • the cleaning apparatus 100 can be incorporated in a resist pattern forming apparatus or the like.
  • the cleaning apparatus 100 includes a substrate support portion 31 that supports the wafer W substantially horizontally, a rotation mechanism 32, and a cup 41.
  • the substrate support portion 31 is a disc that supports the center of the back surface of the wafer W from below.
  • the rotation mechanism 32 rotates the wafer W together with the substrate support unit 31.
  • the rotating mechanism 32 can be moved up and down.
  • An opening 41a having a larger diameter than the wafer W is provided on the upper surface of the cup 41, and the wafer W is transferred to and from the transfer arm via the opening 41a.
  • the cleaning apparatus 100 includes a surface cleaning nozzle 6 for cleaning the surface of the wafer W.
  • the surface cleaning nozzle 6 includes a cleaning liquid nozzle 61 and a gas nozzle 62.
  • the cleaning liquid nozzle 61 supplies a cleaning liquid (for example, DIW: De-Ionized Water) for cleaning particles adhering to the surface of the wafer W toward the surface of the wafer W.
  • the gas nozzle 62 supplies, for example, a gas such as nitrogen (N 2 ) toward the surface of the wafer W in order to promote drying of the cleaning liquid on the wafer surface.
  • the cleaning liquid nozzle 61 and the gas nozzle 62 are supported by, for example, a common support portion 63, and are configured to be movable in the radial direction of the wafer W and vertically movable up and down by a driving mechanism.
  • the cleaning liquid nozzle 61 and the gas nozzle 62 of the surface cleaning nozzle 6 are connected to a cleaning liquid (DIW) source 65 and a nitrogen gas source 66 through supply paths 61a and 62a, respectively.
  • DIW cleaning liquid
  • the cleaning liquid source 65 and the nitrogen gas source 66 are controlled by a control device 200 that controls the operation of the entire coating and developing device.
  • the control device 200 is composed of a computer having a memory.
  • the memory stores a program for controlling the operation of the entire coating and developing apparatus.
  • the control device 200 controls the operation of the cleaning device 100 according to the procedure set in the program.
  • the memory is realized by storage means such as a hard disk, a compact disk, a magnetic optical disk, a memory card, and the like.
  • the cleaning apparatus 100 described above is provided with a measurement mechanism 300 for measuring particles contained in the cleaning liquid supplied from the cleaning liquid source 65.
  • the measurement mechanism 300 is provided in the supply path 61a that is separated from the cleaning liquid nozzle 61 upstream by about 10 cm to 20 cm.
  • the measurement mechanism 300 measures the state of particles in the cleaning liquid flowing through the supply path 61a.
  • the object to be inspected in the present embodiment may be a liquid such as a chemical liquid, an organic liquid, or water for cleaning the wafer, or may be a gas such as decompressed air. Details of the measurement mechanism 300 will be described below.
  • the cleaning liquid is irradiated with laser light in order to detect particles in the cleaning liquid flowing in the supply path 61a.
  • laser light is emitted from the laser light source 301 toward the measuring unit 305 provided in the supply path 61a upstream by about 10 cm to 20 cm near the discharge port of the cleaning liquid nozzle 61.
  • the laser light is scattered by particles contained in the cleaning liquid flowing inside the measuring unit 305.
  • the detector 302 receives laser light, converts it into an electrical signal (hereinafter referred to as “detection signal”) by photoelectric conversion, analyzes the detection signal, and detects the state of the particles in the cleaning liquid. Thereby, the state of nano-level particles can be detected.
  • a first detector 303 and a second detector 304 are used as the detector 302.
  • the first detector 303 and the second detector 304 are photodetectors that convert received laser light into electricity.
  • the state of particles in the cleaning liquid is detected using a difference signal (hereinafter also referred to as “detection signal”) of detection signals detected by the first detector 303 and the second detector 304.
  • detection signal a difference signal
  • the noise components included in the detection signals detected by the first detector 303 and the second detector 304 are canceled out, and the analysis can be performed based on the signal having a small noise component.
  • the detector may be either the first detector 303 or the second detector 304.
  • the analysis is performed based on the detection signal detected by the detector, not the difference signal.
  • noise generated by the laser light source 301 and the detector 302 is included in the signal for detecting the particles. Therefore, noise is also included in the differential signal of the detection signals detected by the first detector 303 and the second detector 304, respectively.
  • the noise component in the difference signal or detection signal is in the same or relatively close frequency band as the signal, it is not easy to distinguish the signal (hereinafter referred to as “particle signal”) and noise.
  • particle signal the signal
  • the signal waveform is extracted by waveform modeling and is distinguished from the noise waveform to increase the nominal S / N ratio, and about 99% even for a signal with a S / N level lower than 2.
  • a signal processing method capable of detecting a particle signal with a probability of.
  • a waveform a is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is set to zero.
  • a waveform b is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is set to 12 (mW / ch).
  • a waveform c is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is 9 (mW / ch).
  • the noise of the laser light source 301 surrounded by a broken line frame exists in a frequency band close to the particle signal on the high frequency side of 3 ⁇ 10 5 (Hz) and 1 ⁇ 10 6 (Hz). Indicates that there is a case.
  • the light output from the laser light source 301 shown in FIG. 3 is applied to the liquid (an example of the object to be inspected) flowing through the supply path 61a and scattered by particles (fine particles) in the object to be inspected. Therefore, the detection signals detected by the first detector 303 and the second detector 304 are scattered light signals generated by particles in the object to be inspected.
  • the detection signal includes noise mainly generated by the laser light source 301, the first detector 303, and the second detector 304.
  • a signal having a low S / N ratio that is, a large noise component with respect to the signal
  • a target signal in this embodiment, a particle signal
  • a detection signal having a complicated waveform compared to a signal having a relatively simple waveform that can be represented by a Gaussian waveform or the like, it becomes difficult to further separate the particle signal and noise.
  • the horizontal axis represents time
  • the vertical axis represents signal intensity.
  • modeling with a simple Gaussian waveform is performed.
  • the waveform included in the detection signal is arbitrary, and includes, for example, an asymmetric waveform that does not necessarily generate wave peaks having different positive and negative. Therefore, in this embodiment, it is possible to define a waveform model according to the characteristics of the waveform.
  • the waveform model according to this embodiment is defined by the difference between intersecting Gaussian waveforms.
  • the waveform is modeled by the positive peak height A2 of the waveform, the negative peak height A3, and the time offset between those peaks.
  • the waveform is first modeled by aligning the peak position of the waveform and minimizing the width of the peak wave.
  • the width of the wave is indicated by “ ⁇ ” in Expression (1) indicating the Gaussian function. More specifically, when defining the waveform model of the detection signal, the width of the waveform having the positive peak height A2 and the width of the waveform having the negative peak height A3 are represented by ⁇ of the Gaussian function. By minimizing ⁇ , the width of these waveforms can be fitted.
  • I (x, y) indicates the intensity profile of the laser beam spot in the x direction and the y direction of the scattered light corresponding to the detection signal.
  • Equation (1) is not a simple Gaussian function, but is physically a “Hermite-Gauss mode laser beam pattern, and is expressed as a function combining Hermitian coefficients Hk. Combining Hermitian coefficients Hk The functions are described in, for example, https://en.wikipedia.org/wiki/Transverse_mode#Laser_modes.
  • the physical quantity represented by this function indicates a phenomenon in which a beam generated in a direction perpendicular to the traveling direction of the laser generated at the time of laser resonance is split.
  • a similar two-divided pattern is generated by transmitting the laser beam through the phase shift plate.
  • the two-divided beam spot generated at this time can be approximated by 10 patterns in the Hermite-Gauss mode.
  • this function is used.
  • the optical system is actually installed so that the beam pattern of the 00 mode in FIG. 5 crosses the divided first detector 303 and second detector 304 shown in FIG. The result is almost the same between 00 mode and 10 mode.
  • the same function as the signal waveform processing of this embodiment can be used.
  • the function represented by the above formula (1) can be used. Therefore, this embodiment can be applied to both modes.
  • the Gaussian function is a particle path, and the probability of detection when the part of the particle represented by the measurement unit 305 in FIG. 2 passes can be calculated.
  • the laser light pattern through which the particle passes is in the 10 mode of FIG. 5, instead of the probability of the particle path, the scattered light or shadow of the particle passing through the region is induced, and the change in the light Indicates the intensity of detection sensitivity when entering the photodetector.
  • this black portion is a portion where the light is canceled by interference due to the phase shift.
  • the passage of particles through this part contributes to improving the sensitivity of the entire system.
  • the system sets the positional relationship of the optical system, flow cell, and photodetector so that many particles pass through this region. As a result of the setting, the signal intensity of the particles passing through this portion increases the difference signal between the two photodetectors (the first detector 303 and the second detector 304), and thus the sensitivity is increased.
  • This black part indicates that the particles are transmitted. That is, the black portion corresponds to a region where light is blocked by particles.
  • FIG. 6 is a flowchart showing an example of signal processing by waveform modeling according to the present embodiment.
  • the signal processing by waveform modeling according to the present embodiment is performed mainly by the control device 200 by searching for a waveform ⁇ fitting a waveform ⁇ extracting a feature value ⁇ evaluating a feature value based on a rule ⁇ generating a residual signal after subtracting the waveform. ⁇ Executed in the order of signal judgment.
  • control device 200 acquires the signals detected by the first detector 303 and the second detector 304, and searches for the peak of the waveform of the detection signal using the difference as a detection signal (step S10). ).
  • the control device 200 filters the detection signal waveform (original waveform) in advance using an S-Golay Filter, A peak search may be performed on the waveform after filtering.
  • a peak search may be performed on the waveform after filtering.
  • a maximum peak (Max), a minimum peak (Min), and a peak in the vicinity thereof in 1024 data may be searched.
  • the waveform of the detection signal is not necessarily a vertically symmetric waveform, but includes an asymmetrical waveform.
  • the control device 200 sets the maximum peak height A ⁇ b> 2 and the minimum peak height A ⁇ b> 3, and uses the formula (1) indicating the Gaussian function to determine the maximum peak and minimum peak.
  • the waveform is fitted by optimizing (minimizing) the width (step S12).
  • the control device 200 first sets the maximum peak height A2 and the minimum peak height A3 from the search result.
  • the peaks of the waves that set A2 and A3 may be peaks other than the maximum and minimum peaks.
  • the control device 200 fits the width of the waveform using a Gaussian function.
  • the control device 200 aligns the peak positions and the heights A2 and A3, and fits the width of each peak. That is, as shown on the right side of the lower diagram of FIG.
  • the waveform of the maximum peak height A2 and the waveform of the minimum peak height A3 are normalized, and the sum of squares of the difference of ⁇ of the Gaussian function indicating those waveforms is obtained.
  • the minimum value is the maximum peak and the minimum peak width.
  • the control device 200 calculates the feature amount of the fitted waveform based on a preset rule (step S14).
  • FIG. 9 shows an example of the set rule and the threshold value of the feature amount indicated in the rule.
  • Each threshold in FIG. 9 is an example of a first threshold set for each feature amount.
  • the waveform having the maximum peak height A2 and the waveform having the minimum peak height A3 are normalized, and the minimum value Diff of the sum of squares of the difference between these normalized waveforms is normalized. Is defined as one of the feature quantities of the waveform.
  • the threshold value of the minimum value Diff of the sum of squares of the normalized waveform difference is set to 0 to 0.08.
  • the waveform R having the maximum peak height A2 and the waveform having the minimum peak height A3 is defined as one of the feature quantities of the waveform.
  • the waveform deviation R is the ratio (A2 / A3) of the maximum peak height A2 to the minimum peak height A3.
  • the threshold value of the waveform deviation R is set to 1.1 to 2.0.
  • the peak width Offset is defined as one of the feature quantities of the waveform.
  • the threshold value of the peak width Offset is set to 0-60.
  • the width of the waveform having the maximum peak height A2 and the width ⁇ of the waveform having the minimum peak height A3 are determined as one of the feature quantities of the waveform.
  • the width ⁇ of the waveform is obtained by the Gaussian function of Equation (1).
  • the threshold value of the waveform width ⁇ is set to 0 to 5.0.
  • the control device 200 calculates each feature amount of the fitted waveform based on each rule.
  • the control device 200 extracts the fitted waveform component and subtracts the waveform component from the detection signal (step S16). For example, as shown in FIG. 10A, the control device 200 extracts the component of the fitted waveform S1. Then, as illustrated in FIG. 10B, the control device 200 removes the component of the waveform S1 from the detection signal. The remaining signal shown in (b) of FIG. 10 obtained by removing the component of the waveform S1 from the detection signal becomes the next analysis target.
  • the control device 200 determines whether the maximum peak of the remaining signal is smaller than a predetermined threshold (second threshold) (step S18).
  • second threshold a predetermined threshold
  • the control device 200 returns to step S10, and when the maximum peak of the residual signal is smaller than the predetermined second threshold value in step S18. Until it is determined, the processes in steps S10 to S18 are repeated.
  • FIG. 10 (c) shows an example in which the signal components of the fitted waveforms S1 to S3 are sequentially subtracted from the detection signal after the feature amount is calculated.
  • step S ⁇ b> 18 when the control device 200 determines that the maximum peak of the residual signal is smaller than the second threshold value, each of the feature values of the fitted waveforms is included in the first threshold value. It is determined whether it is within a threshold range corresponding to each feature amount (step S20). Here, for example, when all of the four feature amounts shown in the rules 1 to 3 are within the respective threshold ranges, the control device 200 determines that all of the feature amounts are within the first threshold range. Then, it is determined that the signal of the waveform to be determined is a particle signal (step S22), and this process ends.
  • FIG. 11 shows an example of the determination result of the signal of the fitted waveform based on the rule of FIG. 11A is determined as noise, and FIG. 11B and FIG. 11C are determined as particle signals.
  • the fitted waveform signal when the waveform feature amount of the extracted signal is within the first threshold range, the fitted waveform signal is a particle signal. It is determined that there is. Further, when the extracted waveform feature amount is outside the first threshold range, the fitted waveform signal is determined to be noise.
  • a matching waveform is generated by comparing a predetermined template waveform and a detection signal, and a waveform having a high matching score, which is similar to the template, is generated. Judge as a signal.
  • the normalized cross-correlation method there is an original signal component in the waveform that is excluded from the signal and determined to be noise, and there is a possibility that an erroneous determination is made.
  • the signal processing according to the present embodiment instead of matching by a template, a waveform is modeled and extracted, and it is determined whether or not the feature amount of the extracted waveform is within the first threshold range.
  • the extracted waveform signal is a particle signal or noise.
  • a signal determined to be noise because the waveform is not similar to the template in the conventional normalized cross-correlation method can be determined to be a particle signal in the present embodiment.
  • FIG. 12 illustrates an example of a detection signal according to the present embodiment, a waveform feature amount extracted from the detection signal, and a determination result for the extracted waveform signal.
  • the upper part of FIG. 12 shows an example of the distribution of detection signals when the diameter of the laser light spot output from the laser light source 301 is 1.2 ⁇ m, the power of the laser light beam is 20 mW, and the number of detectors 302 is one.
  • the SN ratio of the detection signal shown in the upper left graph of FIG. 12 is 1.2
  • the SN ratio of the detection signal shown in the center graph is 1.5
  • the SN ratio of the detection signal shown in the right graph is 2. .0.
  • the lower part of FIG. 12 shows the distribution of the feature amount of the waveform extracted from each detection signal by fitting, the result of determining whether the extracted waveform is a signal or noise, and the probability of miscounting.
  • the waveform component having the feature value indicated by “ ⁇ ” in the lower part of FIG. 12 is determined as a particle signal, and the waveform component having the feature value indicated by “X” in the lower part of FIG. Is determined to be noise.
  • the signal-to-noise ratio which is difficult to recognize the signal waveform in the conventional signal processing, is 1.
  • the miscount was 0.2% even for signals of about .2.
  • the miscount was 0.05% for a signal with an SN ratio of 1.5
  • the miscount was 0% for a signal with an SN ratio of 2.0.
  • FIG. 13 is a flowchart illustrating an example of signal processing according to the SN ratio according to the present embodiment.
  • the control device 200 acquires a detection signal from the detector 302 (step S30).
  • the detection signal may be a signal detected by one detector or a difference between signals detected by two detectors.
  • control device 200 matches the detection signal with the waveform of the template using the normalized cross correlation method (XCOR), obtains a matching score (cross correlation value of the detection signal), and calculates the SN ratio from the matching score. (Step S32).
  • XCOR normalized cross correlation method
  • FIG. 14 shows an example of waveform extraction by matching with a waveform template.
  • the control device 200 matches the feature waveform template with the feature waveform portion in the detection signal.
  • the control device 200 determines the waveform as a detection signal (particle signal).
  • step S34 determines whether or not the SN ratio is 1.5 or more.
  • the control device 200 determines whether or not the detection signal is a particle signal based on a matching score by a normalized cross correlation method (XCOR) shown in FIG. Step S36). Thereafter, the control device 200 returns to step S30 to acquire the next detection signal, and repeats the processing after step S30.
  • XCOR normalized cross correlation method
  • step S34 when it is determined in step S34 that the SN ratio is less than 1.5, the control device 200 determines whether the SN ratio is 1.2 or more (step S38). When it is determined that the SN ratio is 1.2 or more, the control device 200 executes signal processing by waveform modeling according to the present embodiment shown in FIG. 6 (step S40), and returns to step S30. On the other hand, when determining in step S38 that the SN ratio is less than 1.2, the control device 200 immediately returns to step S30.
  • the signal processing using the normalized cross-correlation method (XCOR) can obtain a predetermined accuracy or more. Therefore, when the SN ratio is 1.5 or more, for example, a particle signal is detected by signal processing using the normalized cross correlation method (XCOR) shown in FIG.
  • the S / N ratio is less than 1.5-2, it is difficult to distinguish between the particle signal and noise in the detection signal.
  • the SNR is 1.5 to 2 or more, it is easy to distinguish between the particle signal in the detection signal and the noise. Therefore, by changing the signal processing method of the detection signal according to the SN ratio, it is possible to accurately determine whether the signal is a particle signal or noise regardless of the level of the SN ratio. Further, the processing addition can be reduced by changing the signal processing method of the detection signal in accordance with the SN ratio.
  • step S38 in FIG. 13 the lower limit value of the SN ratio is set to 1.2.
  • the lower limit value of the SN ratio may be a numerical value other than 1.2 (for example, 1.0), and the lower limit value of the SN ratio may not be provided.
  • step S34 in FIG. 13 the value of the S / N ratio that changes the signal processing method of the detection signal is set to 1.5.
  • the present invention is not limited to this.
  • step S34 any value in the range of 1.5 to 2, such as an SN ratio of 2 or more, may be used as the SN ratio value that changes the signal processing method of the detection signal.
  • the control device 200 is an information processing device such as a personal computer or a tablet-type terminal.
  • the control device 200 includes an input device 101, a display device 102, an external I / F 103, a RAM (Random Access Memory) 104, a ROM (Read Only Memory) 105, a CPU (Central Processing Unit) 106, a communication I / F 107, and an HDD ( Hard Disk Drive) 108 and the like are connected to each other via a bus B.
  • the input device 101 includes a keyboard and a mouse, and is used for inputting each operation signal to the control device 200.
  • the display device 102 includes a display and displays various processing results.
  • the communication I / F 107 is an interface that connects the control device 200 to a network. Thereby, the control apparatus 200 can perform data communication with other apparatuses (detector 302 etc.) via communication I / F107. Thereby, the control apparatus 200 acquires the detection signal of a laser beam from the detector 302.
  • the HDD 108 is a non-volatile storage device that stores programs and data.
  • the stored programs and data include basic software and application software that control the entire control device 200.
  • the HDD 108 may store various databases and programs.
  • External I / F 103 is an interface with an external device.
  • the external device includes a recording medium 103a.
  • the control device 200 can read and / or write the recording medium 103a via the external I / F 103.
  • the recording medium 103a includes a CD (Compact Disk), a DVD (Digital Versatile Disk), an SD memory card (SD Memory Card), a USB memory (Universal Serial Bus memory), and the like.
  • the ROM 105 is a nonvolatile semiconductor memory (storage device) that can retain internal data even when the power is turned off.
  • the ROM 105 stores programs and data such as network settings.
  • the RAM 104 is a volatile semiconductor memory (storage device) that temporarily stores programs and data.
  • the CPU 106 is an arithmetic unit that realizes control of the entire apparatus and mounting functions by reading programs and data from the storage device (for example, “HDD 108”, “ROM 105”, etc.) onto the RAM 104 and executing processing.
  • the storage device stores a signal processing program by waveform modeling according to the present embodiment, and the CPU 106 reads the program from the storage device and executes the processing according to the procedure indicated by the program.
  • the particle signal and noise included in the detection signal can be determined.
  • the target particle signal is accurately determined from the detection signal regardless of the SN ratio. be able to.
  • the signal processing target is, for example, 2 This is a differential signal of the above photodetector (for example, the first and second detectors in FIG. 2), and the signal waveform is complicated.
  • a signal having a complicated waveform is a signal with an increased number of features (parameters) included in the signal.
  • the difference signal of two or more photodetectors is a signal that is easy to extract the feature amount extracted by the signal processing according to the present embodiment, and as a result, the particle signal or noise based on the feature amount according to the present embodiment. It is thought that the determination accuracy is improved.
  • the laser light source 301 and the detector 302 include inexpensive devices although they have a relatively large noise component. In that case, there is a case where it is desired to use an inexpensive device even when the SN ratio is 1.5 to 2.0 or less. In addition, the S / N ratio tends to be lower in an organic liquid than water. As described above, even when the SN ratio is 1.5 to 2.0 or less, it is possible to accurately distinguish the particle signal and the noise from the detection signal by the signal processing based on the waveform modeling according to the present embodiment. it can.
  • waveform extraction and signal determination results by waveform modeling according to the present embodiment can be accumulated and machine learning can be performed.
  • the range of variation is defined in advance by a threshold, and if the feature amount is outside the range of the threshold, the signal is discarded without being subjected to machine learning.
  • the control device 200 extracts an optimum waveform in real time from the waveform of the particle signal accumulated in the storage unit based on the learned result, and the waveform according to the present embodiment. You may make it apply to the signal processing by modeling of.
  • the present invention when detecting particles such as particles in a liquid or gas inspected object flowing in tubes attached to various devices, it is possible to accurately determine whether or not the signal indicates the particles from the detection signal regardless of whether the SN ratio is high or low.
  • the present invention includes a capacitively coupled plasma (CCP) device, an inductively coupled plasma (ICP) processing device, a plasma processing device using a radial line slot antenna, a helicon wave excited plasma ( The present invention can also be applied to devices such as HWP: Helicon Wave) Plasma devices, electron cyclotron resonance plasma (ECR) devices, and surface wave plasma processing devices.
  • CCP capacitively coupled plasma
  • ICP inductively coupled plasma
  • ECR electron cyclotron resonance plasma
  • the signal processing method according to the present embodiment can be applied to various measuring apparatuses that detect fine particles such as particles in a fluid (liquid or gas) inspection object.
  • the measurement mechanism 300 in the cleaning apparatus 100 has been described with reference to FIGS. 1 and 2 as an example of the configuration.
  • a particle monitor 133 index measuring device using laser scattered light as shown in FIG.
  • the particle monitor 133 includes a laser oscillator 134 that oscillates laser light, a CCD camera 135 that observes scattered light in the chamber 112, and a pulse generator 136 connected to the laser oscillator 134 and the CCD camera 135. .
  • the laser oscillator 134 oscillates a laser beam toward the inside of the chamber 112 during processing such as etching through a slit window 137 provided in the chamber 112.
  • the particles p in the chamber 112 generate scattered light when irradiated with laser light.
  • the generated scattered light is observed by the CCD camera 135 through the slit window 138.
  • the particle monitor 133 measures the amount of particles p floating in the chamber 112 through the number and intensity of scattered light generation (index measurement step). Then, the measured amount of particles p is transmitted as a measurement result to the PC 132 via the Internet 131 or the like.
  • the pulse generator 136 transmits a synchronization signal to the laser oscillator 134 and the CCD camera 135, and thereby adjusts the timing of oscillation of the laser light and the timing of reception of the scattered light.
  • the particle monitor 133 includes a transmission device 139 that receives a device status signal indicating the operating status of the chamber 112 and the presence or absence of a failure from the chamber 112 and transmits the device status signal to the PC 132 via the Internet 131 or the like.
  • the PC 132 performs signal processing according to the present embodiment.
  • the measurement apparatus to which the signal processing method according to the present embodiment can be applied is not limited to the case where the laser is irradiated into the chamber 112 as described above to detect particles floating in the space on the substrate.
  • the signal processing method according to the present embodiment detects particles adhering to the substrate by irradiating the substrate placed in the chamber 112 with a laser beam from directly above and detecting reflected light from the substrate. It is also possible to apply to an apparatus.

Abstract

Provided is a signal processing method in which a computer executes the processes of: detecting a signal of light radiated onto a fluid object being inspected; searching for a waveform peak in the detected signal; fitting the width of the waveform of the searched peak; calculating a characteristic value of the fitted waveform; and determining, on the basis of whether the characteristic value of the calculated waveform falls within a pre-determined threshold range, whether the signal with the fitted waveform is a signal representing a fine particle in the object being inspected.

Description

信号処理方法及びプログラムSignal processing method and program
 本発明は、信号処理方法及びプログラムに関する。 The present invention relates to a signal processing method and a program.
 液体又は気体に照射される検査光が、液体又は気体に含まれる微粒子により遮光されることを利用して液体等の被検査体内の微粒子を検出することが提案されている(例えば、特許文献1を参照)。 It has been proposed to detect fine particles in a body to be inspected such as a liquid by using the inspection light irradiated to the liquid or gas being shielded by the fine particles contained in the liquid or gas (for example, Patent Document 1). See).
 また、試料溶液内における顕微鏡の光検出領域の位置を移動させながら、光検出領域から実質的に一定の背景光を含む光を検出して時系列の光強度データを生成し、そのデータにおいて発光しない単一粒子が光検出領域内へ侵入した際に生じる光強度の低下を単一粒子の各々の存在を表す信号として検出することが提案されている(例えば、特許文献2を参照)。 In addition, while moving the position of the light detection area of the microscope in the sample solution, light containing a substantially constant background light is detected from the light detection area to generate time-series light intensity data, and light emission is performed in that data. It has been proposed to detect a decrease in light intensity that occurs when a single particle that does not enter the light detection region as a signal indicating the presence of each single particle (see, for example, Patent Document 2).
 上記の特許文献では、検査対象の信号と予め特定された波形のテンプレートとのマッチングの結果に応じて検査対象の信号中の目的とする信号とノイズとを分別する。 In the above-described patent document, a target signal and noise in the signal to be inspected are classified according to the result of matching between the signal to be inspected and a waveform template specified in advance.
特開2009-14702号公報JP 2009-14702 A 国際公開第2013/031309号パンフレットInternational Publication No. 2013/031309 Pamphlet
 しかしながら、上記特許文献では、目的とする信号の周波数帯域とノイズの周波数帯域とは異なる。また、目的とする信号の強度(スペクトラム)が、ノイズの強度よりも大きい。検査対象の信号に以上の特徴がある場合、上記特許文献では、検査対象の信号中の目的とする信号とノイズとを区別することは容易である。 However, in the above patent document, the frequency band of the target signal is different from the frequency band of noise. Further, the intensity (spectrum) of the target signal is larger than the noise intensity. When the signal to be inspected has the above characteristics, in the above-mentioned patent document, it is easy to distinguish a target signal and noise in the signal to be inspected.
 これに対して、SN比(ノイズ対信号比)が例えば2以下のような低SN比の信号であって、目的とする信号の周波数帯域とノイズの周波数帯域が近い場合、目的とする信号とノイズとを分別することは困難である。 On the other hand, if the signal-to-noise ratio (noise-to-signal ratio) is a signal with a low signal-to-noise ratio such as 2 or less, and the frequency band of the target signal is close to the frequency band of noise, It is difficult to distinguish from noise.
 更に、上記特許文献が検査対象とする信号は、ガウシャン波形やローレンツ形状等で表せる比較的単純な波形の信号である。これに対して、複雑な波形な信号を検査対象とした場合、上記特許文献に開示された信号処理方法では、目的とする信号とノイズとを分別することは更に難しくなる。 Furthermore, the signal to be inspected by the above-mentioned patent document is a signal having a relatively simple waveform that can be represented by a Gaussian waveform, a Lorentz shape or the like. On the other hand, when a signal having a complicated waveform is to be inspected, the signal processing method disclosed in the above-mentioned patent document further makes it difficult to separate the target signal and noise.
 上記課題に対して、一側面では、本発明は、液体又は気体の中の微粒子の検出において、SN比の高低にかかわらず検出信号から目的とする微粒子の信号を精度よく判定することを目的とする。 With respect to the above-described problem, an object of one aspect of the present invention is to accurately determine a signal of a target particle from a detection signal regardless of the S / N ratio in detecting a particle in a liquid or a gas. To do.
 上記課題を解決するために、一の態様によれば、流体の被検査体に照射した光の信号を検出し、検出した前記信号の波形のピークをサーチし、サーチした前記ピークの波形の幅をフィッティングし、前記フィッティングした波形の特徴量を算出し、算出した前記波形の特徴量が予め定められた閾値の範囲内かに基づき、前記フィッティングした波形の信号が前記被検査体の中の微粒子を示す信号か否かを判定する、処理をコンピュータが実行する信号処理方法が提供される。 In order to solve the above-mentioned problem, according to one aspect, a signal of light irradiated on a fluid inspection object is detected, a peak of the waveform of the detected signal is searched, and a waveform width of the searched peak And calculating the feature quantity of the fitted waveform, and based on whether the calculated feature quantity of the waveform is within a predetermined threshold range, the fitted waveform signal is a fine particle in the inspection object. There is provided a signal processing method in which a computer executes a process for determining whether or not the signal indicates a signal.
 一の側面によれば、液体又は気体の中の微粒子の検出において、SN比の高低にかかわらず検出信号から目的とする微粒子の信号を精度よく判定することができる。 According to one aspect, in detecting fine particles in a liquid or gas, it is possible to accurately determine the signal of the target fine particles from the detection signal regardless of the level of the SN ratio.
一実施形態に係る洗浄装置の縦断面の一例を示す図。The figure which shows an example of the longitudinal cross-section of the washing | cleaning apparatus which concerns on one Embodiment. 一実施形態に係る測定機構の縦断面の一例を示す図。The figure which shows an example of the longitudinal cross-section of the measurement mechanism which concerns on one Embodiment. レーザ光源から出力された光の信号の一例を示す図。The figure which shows an example of the signal of the light output from the laser light source. 一実施形態に係る波形モデルの定義の一例を示す図。The figure which shows an example of the definition of the waveform model which concerns on one Embodiment. ガウシャン関数をベースとしたHermite-Gaussモード式を用いてパーティクル経路の検出される確率の高い領域を算出した結果例を示す図。The figure which shows the example of a result of having calculated the area | region where the probability that a particle path | route is detected is calculated using the Hermite-Gauss mode formula based on a Gaussian function. 一実施形態に係る波形のモデル化による信号処理の一例を示すフローチャート。The flowchart which shows an example of the signal processing by modeling of the waveform which concerns on one Embodiment. 一実施形態に係る波形のピークのサーチを説明するための図。The figure for demonstrating the search of the peak of the waveform which concerns on one Embodiment. 一実施形態に係る波形のフィッティングを説明するための図。The figure for demonstrating the fitting of the waveform which concerns on one Embodiment. 一実施形態に係る特徴量のルールの閾値の一例を示す図。The figure which shows an example of the threshold value of the rule of the feature-value which concerns on one Embodiment. 一実施形態に係る波形の抽出と残信号を説明するための図。The figure for demonstrating the extraction of the waveform which concerns on one Embodiment, and a residual signal. 一実施形態に係るルールに基づく信号の判定結果の一例を示す図。The figure which shows an example of the determination result of the signal based on the rule which concerns on one Embodiment. 一実施形態に係る検出信号及び波形の特徴量と判定結果の一例を示す図。The figure which shows an example of the feature-value and determination result of the detection signal and waveform which concern on one Embodiment. 一実施形態に係るSN比に応じた信号処理の一例を示すフローチャート。The flowchart which shows an example of the signal processing according to S / N ratio which concerns on one Embodiment. 波形のテンプレートとのマッチングによる波形抽出の一例を示す図。The figure which shows an example of the waveform extraction by matching with the template of a waveform. 一実施形態に係る波形抽出の一例を示す図。The figure which shows an example of the waveform extraction which concerns on one Embodiment. 一実施形態に係る制御装置のハードウェア構成の一例を示す図。The figure which shows an example of the hardware constitutions of the control apparatus which concerns on one Embodiment. 一実施形態に係る測定装置の構成の一例を示す図。The figure which shows an example of a structure of the measuring apparatus which concerns on one Embodiment.
 以下、本発明を実施するための形態について図面を参照して説明する。なお、本明細書及び図面において、実質的に同一の構成については、同一の符号を付することにより重複した説明を省く。 Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In addition, in this specification and drawing, about the substantially same structure, the duplicate description is abbreviate | omitted by attaching | subjecting the same code | symbol.
 [洗浄装置の全体構成]
 まず、本実施形態に係る洗浄装置100の一例について、図1を参照しながら説明する。洗浄装置100は、レジストパターン形成装置等に組み込まれ得る。洗浄装置100は、ウェハWを略水平に支持する基板支持部31と回転機構32とカップ41とを有する。
[Overall configuration of cleaning equipment]
First, an example of the cleaning apparatus 100 according to the present embodiment will be described with reference to FIG. The cleaning apparatus 100 can be incorporated in a resist pattern forming apparatus or the like. The cleaning apparatus 100 includes a substrate support portion 31 that supports the wafer W substantially horizontally, a rotation mechanism 32, and a cup 41.
 基板支持部31は,ウェハWの裏面中央の領域を下方から支持する円板である。回転機構32は、基板支持部31とともにウェハWを回転させる。また、回転機構32は、昇降可能になっている。 The substrate support portion 31 is a disc that supports the center of the back surface of the wafer W from below. The rotation mechanism 32 rotates the wafer W together with the substrate support unit 31. The rotating mechanism 32 can be moved up and down.
 カップ41の上面には、ウェハWより大口径の開口部41aが設けられ、開口部41aを介し搬送アームとの間でウェハWの受け渡しが行われる。 An opening 41a having a larger diameter than the wafer W is provided on the upper surface of the cup 41, and the wafer W is transferred to and from the transfer arm via the opening 41a.
 洗浄装置100は、ウェハWの表面を洗浄するための表面洗浄ノズル6を有する。表面洗浄ノズル6は、洗浄液ノズル61とガスノズル62とを有する。洗浄液ノズル61は、ウェハWの表面に向けて、当該表面に付着したパーティクルを洗い流すための洗浄液(例えばDIW:De-Ionized Water)を供給する。ガスノズル62は、ウェハWの表面に向けてウェハ表面の洗浄液の乾燥を促進するために、例えば窒素(N)等の気体を供給する。 The cleaning apparatus 100 includes a surface cleaning nozzle 6 for cleaning the surface of the wafer W. The surface cleaning nozzle 6 includes a cleaning liquid nozzle 61 and a gas nozzle 62. The cleaning liquid nozzle 61 supplies a cleaning liquid (for example, DIW: De-Ionized Water) for cleaning particles adhering to the surface of the wafer W toward the surface of the wafer W. The gas nozzle 62 supplies, for example, a gas such as nitrogen (N 2 ) toward the surface of the wafer W in order to promote drying of the cleaning liquid on the wafer surface.
 洗浄液ノズル61及びガスノズル62は、例えば共通の支持部63に支持されて、駆動機構によりウェハWの径方向に移動自在及び上下方向に昇降自在に構成されている。 The cleaning liquid nozzle 61 and the gas nozzle 62 are supported by, for example, a common support portion 63, and are configured to be movable in the radial direction of the wafer W and vertically movable up and down by a driving mechanism.
 表面洗浄ノズル6の洗浄液ノズル61及びガスノズル62は、それぞれ供給路61a,62aを介して洗浄液(DIW)源65及び窒素ガス源66に接続されている。 The cleaning liquid nozzle 61 and the gas nozzle 62 of the surface cleaning nozzle 6 are connected to a cleaning liquid (DIW) source 65 and a nitrogen gas source 66 through supply paths 61a and 62a, respectively.
 洗浄液源65及び窒素ガス源66等は、塗布、現像装置全体の動作を制御する制御装置200により制御される。制御装置200は、メモリを有するコンピュータからなる。メモリには塗布、現像装置全体の動作を制御するためのプログラムが格納されている。制御装置200は、プログラムに設定された手順に従い洗浄装置100の動作を制御する。なお、メモリは、例えばハードディスク、コンパクトディスク、マグネットオプティカルディスク、メモリーカード等の記憶手段により実現される。 The cleaning liquid source 65 and the nitrogen gas source 66 are controlled by a control device 200 that controls the operation of the entire coating and developing device. The control device 200 is composed of a computer having a memory. The memory stores a program for controlling the operation of the entire coating and developing apparatus. The control device 200 controls the operation of the cleaning device 100 according to the procedure set in the program. The memory is realized by storage means such as a hard disk, a compact disk, a magnetic optical disk, a memory card, and the like.
 以上に説明した洗浄装置100には、洗浄液源65から供給される洗浄液に含まれるパーティクルを測定するための測定機構300が設けられている。本実施形態では、測定機構300は、洗浄液ノズル61から上流へ10cm~20cm程度の離れた供給路61aに設けられている。測定機構300は、供給路61aに流れる洗浄液中のパーティクルの状態を測定する。 The cleaning apparatus 100 described above is provided with a measurement mechanism 300 for measuring particles contained in the cleaning liquid supplied from the cleaning liquid source 65. In the present embodiment, the measurement mechanism 300 is provided in the supply path 61a that is separated from the cleaning liquid nozzle 61 upstream by about 10 cm to 20 cm. The measurement mechanism 300 measures the state of particles in the cleaning liquid flowing through the supply path 61a.
 なお、本実施形態における被検査体は、ウェハを洗浄する薬液、有機液、水等の液体であってもよく、減圧した空気等の気体であってもよい。以下に測定機構300の詳細を説明する。 Note that the object to be inspected in the present embodiment may be a liquid such as a chemical liquid, an organic liquid, or water for cleaning the wafer, or may be a gas such as decompressed air. Details of the measurement mechanism 300 will be described below.
 [測定機構]
 測定機構300では、図2に示すように、供給路61aに流れる洗浄液中のパーティクルを検出するために洗浄液にレーザ光を照射する。具体的には、洗浄液ノズル61の吐出口付近の10cm~20cm程度上流の供給路61aに設けられた計測部305に向けてレーザ光源301からレーザ光が出射される。レーザ光は、計測部305の内部に流れる洗浄液に含まれるパーティクルによって散乱する。検出器302は、レーザ光を受光し、光電変換により電気信号(以下、「検出信号」という。)に変換し、検出信号を解析して洗浄液中のパーティクルの状態を検出する。これにより、ナノレベルのパーティクルの状態を検出することができる。本実施形態では、検出器302に第1検出器303及び第2検出器304が用いられる。
[Measuring mechanism]
In the measurement mechanism 300, as shown in FIG. 2, the cleaning liquid is irradiated with laser light in order to detect particles in the cleaning liquid flowing in the supply path 61a. Specifically, laser light is emitted from the laser light source 301 toward the measuring unit 305 provided in the supply path 61a upstream by about 10 cm to 20 cm near the discharge port of the cleaning liquid nozzle 61. The laser light is scattered by particles contained in the cleaning liquid flowing inside the measuring unit 305. The detector 302 receives laser light, converts it into an electrical signal (hereinafter referred to as “detection signal”) by photoelectric conversion, analyzes the detection signal, and detects the state of the particles in the cleaning liquid. Thereby, the state of nano-level particles can be detected. In the present embodiment, a first detector 303 and a second detector 304 are used as the detector 302.
 第1検出器303及び第2検出器304は、受光したレーザ光を電気に変換するフォトディテクタである。本実施形態では、第1検出器303及び第2検出器304のそれぞれが検出した検出信号の差分信号(以下、「検出信号」ともいう。)を使用して洗浄液中のパーティクルの状態を検出する。差分信号を使用することで、第1検出器303及び第2検出器304のそれぞれが検出した検出信号に含まれるノイズ成分が相殺され、ノイズ成分の少ない信号に基づき解析を行うことができる。 The first detector 303 and the second detector 304 are photodetectors that convert received laser light into electricity. In the present embodiment, the state of particles in the cleaning liquid is detected using a difference signal (hereinafter also referred to as “detection signal”) of detection signals detected by the first detector 303 and the second detector 304. . By using the difference signal, the noise components included in the detection signals detected by the first detector 303 and the second detector 304 are canceled out, and the analysis can be performed based on the signal having a small noise component.
 ただし、検出器は、第1検出器303又は第2検出器304のいずれかであってもよい。この場合、差分信号でなく、検出器が検出した検出信号に基づき解析が行われる。 However, the detector may be either the first detector 303 or the second detector 304. In this case, the analysis is performed based on the detection signal detected by the detector, not the difference signal.
 明視野光学系の液中及び気中のナノパーティクルの検出では、パーティクルを検出した信号中にレーザ光源301や検出器302で生じたノイズが含まれる。よって、第1検出器303及び第2検出器304のそれぞれが検出した検出信号の差分信号にもノイズが含まれる。差分信号や検出信号中のノイズ成分が信号と同一又は比較的近い周波数帯域である場合、信号(以下、「パーティクル信号」という。)とノイズを区別することは容易ではない。特に、本実施形態において扱う、粒径が20nm以下のナノレベルのパーティクルの検出では、低SN比の条件での測定を行う必要があるが、SN比が2よりも小さい場合、パーティクル信号とノイズとの区別を精度良く行うことが課題となる。 In the detection of nanoparticles in the liquid and in the air of the bright field optical system, noise generated by the laser light source 301 and the detector 302 is included in the signal for detecting the particles. Therefore, noise is also included in the differential signal of the detection signals detected by the first detector 303 and the second detector 304, respectively. When the noise component in the difference signal or detection signal is in the same or relatively close frequency band as the signal, it is not easy to distinguish the signal (hereinafter referred to as “particle signal”) and noise. In particular, in the detection of nano-level particles having a particle size of 20 nm or less, which is handled in the present embodiment, it is necessary to perform measurement under conditions of a low S / N ratio. It becomes a problem to make a distinction between and accurately.
 そこで、本実施形態では、波形のモデル化により信号波形の抽出を行い、ノイズの波形と区別することで名目上のSN比を高め、SN比が2よりも小さいレベルの信号においても99%程度の確率でパーティクル信号を検出できる信号処理方法を提供する。 Therefore, in the present embodiment, the signal waveform is extracted by waveform modeling and is distinguished from the noise waveform to increase the nominal S / N ratio, and about 99% even for a signal with a S / N level lower than 2. Provided is a signal processing method capable of detecting a particle signal with a probability of.
 [レーザ光源の信号例]
 レーザ光源301が出力するレーザ光の出力強度にはバラツキが生じる。レーザ光源301の信号の一例を図3に示す。図3の横軸は周波数を示し、縦軸は出力されたレーザ光の光電変換後の電圧を示す。
[Signal example of laser light source]
Variations occur in the output intensity of the laser light output from the laser light source 301. An example of the signal from the laser light source 301 is shown in FIG. The horizontal axis in FIG. 3 indicates the frequency, and the vertical axis indicates the voltage after photoelectric conversion of the output laser light.
 波形aは、レーザ光源301のパワーを0にしたときのレーザ光源301から出力される信号の一例である。波形bは、レーザ光源301のパワーを12(mW/ch)にしたときのレーザ光源301から出力される信号の一例である。波形cは、レーザ光源301のパワーを9(mW/ch)にしたときのレーザ光源301から出力される信号の一例である。図3に示す波形bにおいて、3×10(Hz)及び1×10(Hz)の高い周波数側で、破線の枠で囲ったレーザ光源301のノイズが、パーティクル信号と近い周波数帯域に存在する場合があることを示している。 A waveform a is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is set to zero. A waveform b is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is set to 12 (mW / ch). A waveform c is an example of a signal output from the laser light source 301 when the power of the laser light source 301 is 9 (mW / ch). In the waveform b shown in FIG. 3, the noise of the laser light source 301 surrounded by a broken line frame exists in a frequency band close to the particle signal on the high frequency side of 3 × 10 5 (Hz) and 1 × 10 6 (Hz). Indicates that there is a case.
 図3に示すレーザ光源301から出力される光は、供給路61aを流れる液体(被検査体の一例)に照射され、被検査体の中のパーティクル(微粒子)によって散乱する。したがって、第1検出器303及び第2検出器304が検出する検出信号は、被検査体の中のパーティクルにより生じた散乱光の信号である。検出信号は、レーザ光源301、第1検出器303及び第2検出器304にて主に発生するノイズを含んでいる。 The light output from the laser light source 301 shown in FIG. 3 is applied to the liquid (an example of the object to be inspected) flowing through the supply path 61a and scattered by particles (fine particles) in the object to be inspected. Therefore, the detection signals detected by the first detector 303 and the second detector 304 are scattered light signals generated by particles in the object to be inspected. The detection signal includes noise mainly generated by the laser light source 301, the first detector 303, and the second detector 304.
 検出信号におけるSN比(ノイズ対信号比)が例えば2以下のような低SN比(つまり、信号に対するノイズ成分が多い)の信号であって、目的とする信号(本実施形態ではパーティクル信号)とノイズとの周波数帯域が比較的近いところにある場合、信号とノイズとを分けることは難しい。ガウシャン波形等で表せる比較的単純な波形の信号と比べて、複雑な波形の検出信号の場合には、更にパーティクル信号とノイズとを分けることは困難になる。 A signal having a low S / N ratio (that is, a large noise component with respect to the signal) such as an SN ratio (noise-to-signal ratio) of 2 or less in the detection signal, and a target signal (in this embodiment, a particle signal) When the frequency band with noise is relatively close, it is difficult to separate the signal and noise. In the case of a detection signal having a complicated waveform, compared to a signal having a relatively simple waveform that can be represented by a Gaussian waveform or the like, it becomes difficult to further separate the particle signal and noise.
 そこで、本実施形態に係る波形のモデル化による信号処理では、検出信号におけるSN比の高低にかかわらず、検出信号からパーティクル信号とノイズとを精度よく区別する手法を以下に提案する。 Therefore, in the signal processing based on the waveform modeling according to the present embodiment, a method for accurately distinguishing the particle signal and the noise from the detection signal regardless of the level of the SN ratio in the detection signal is proposed below.
 (波形モデル定義)
 まず、本実施形態で行う波形モデルの定義について、図4を参照しながら説明する。図4の横軸は時間、縦軸は信号強度を示す。本実施形態では、単純なガウシャン波形でのモデル化を行う。検出信号に含まれる波形には、任意性があり、例えば、必ずしも正負が異なる波のピークが発生しない非対称な波形も含まれる。そこで、本実施形態では、波形の特徴に合わせて波形モデルを定義することが可能である。
(Waveform model definition)
First, the definition of the waveform model performed in this embodiment will be described with reference to FIG. In FIG. 4, the horizontal axis represents time, and the vertical axis represents signal intensity. In the present embodiment, modeling with a simple Gaussian waveform is performed. The waveform included in the detection signal is arbitrary, and includes, for example, an asymmetric waveform that does not necessarily generate wave peaks having different positive and negative. Therefore, in this embodiment, it is possible to define a waveform model according to the characteristics of the waveform.
 本実施形態に係る波形モデルは、交差したガウシャン波形の差分により定義される。図4に示す波形の例では、波形の正のピークの高さA2、負のピークの高さA3、それらのピーク間の時間Offsetにより、波形をモデル化する。その際、まず波形のピークの位置を合わせ、そのピークの波の幅を最小にすることで波形がモデル化される。波の幅は、ガウシャン関数を示す式(1)の「ω」で示される。より具体的には、検出信号の波形モデルを定義する際、正のピークの高さA2の波形の幅や、負のピークの高さA3の波形の幅は、ガウシャン関数のωで表され、ωを最小することで、これらの波形の幅のフィッティングが可能になる。 The waveform model according to this embodiment is defined by the difference between intersecting Gaussian waveforms. In the example of the waveform shown in FIG. 4, the waveform is modeled by the positive peak height A2 of the waveform, the negative peak height A3, and the time offset between those peaks. At that time, the waveform is first modeled by aligning the peak position of the waveform and minimizing the width of the peak wave. The width of the wave is indicated by “ω” in Expression (1) indicating the Gaussian function. More specifically, when defining the waveform model of the detection signal, the width of the waveform having the positive peak height A2 and the width of the waveform having the negative peak height A3 are represented by ω of the Gaussian function. By minimizing ω, the width of these waveforms can be fitted.
Figure JPOXMLDOC01-appb-M000001
 レーザビームスポットの強度分布を表すガウシャン関数を示す式(1)では、I(x、y)は、検出信号に対応する散乱光のx方向及びy方向のレーザビームスポットの強度プロファイルを示す。
Figure JPOXMLDOC01-appb-M000001
In Expression (1) indicating the Gaussian function representing the intensity distribution of the laser beam spot, I (x, y) indicates the intensity profile of the laser beam spot in the x direction and the y direction of the scattered light corresponding to the detection signal.
 式(1)は、単純なガウシャン関数ではなく物理学的には“Hermite-Gaussモードのレーザ光パターンであり、Hermite多項式の係数Hkを組み合わせた関数として表される。Hermite多項式の係数Hkを組み合わせた関数については、例えば、https://en.wikipedia.org/wiki/Transverse_mode#Laser_modesに説明されている。 Equation (1) is not a simple Gaussian function, but is physically a “Hermite-Gauss mode laser beam pattern, and is expressed as a function combining Hermitian coefficients Hk. Combining Hermitian coefficients Hk The functions are described in, for example, https://en.wikipedia.org/wiki/Transverse_mode#Laser_modes.
 また、この関数で表される物理量は、レーザの共振時に発生するレーザの進行方向に対して垂直方向に発生するビームが分割される現象を示す。本実施形態では、レーザビームを位相シフト板に透過させることで同様な2分割パターンを生成するが、この際に生成する2分割ビームスポットが、Hermite-Gaussモードの10のパターンで近似できるために、本実施形態ではこの関数を用いる。システム上は、実際には、図5の00モードのビームパターンを、図2に示す分割された第1検出器303及び第2検出器304をまたぐように光学系を設置することで図5の00モードと10モードでほとんどの差のない結果が得られる。00モードでは、本実施形態の信号波形処理と同じ関数が使える。また、10モードでは上記式(1)にて示される関数が使える。よって、本実施形態では両方のモードに適用することが可能である。 Also, the physical quantity represented by this function indicates a phenomenon in which a beam generated in a direction perpendicular to the traveling direction of the laser generated at the time of laser resonance is split. In the present embodiment, a similar two-divided pattern is generated by transmitting the laser beam through the phase shift plate. However, since the two-divided beam spot generated at this time can be approximated by 10 patterns in the Hermite-Gauss mode. In this embodiment, this function is used. In practice, the optical system is actually installed so that the beam pattern of the 00 mode in FIG. 5 crosses the divided first detector 303 and second detector 304 shown in FIG. The result is almost the same between 00 mode and 10 mode. In the 00 mode, the same function as the signal waveform processing of this embodiment can be used. In the 10 mode, the function represented by the above formula (1) can be used. Therefore, this embodiment can be applied to both modes.
 ガウシャン関数は、パーティクルの経路となり、図2の計測部305で表される部分のパーティクルの通過する際の検出できる確率を算出することができる。本実施形態では、パーティクルが通るレーザ光パターンが図5の10モードのものでは、パーティクルの経路の確率ではなく、その領域を通過するパーティクルの散乱光や影を誘起し、その光の変化が各光検出器にはいった場合の検出感度の強度を表す。 The Gaussian function is a particle path, and the probability of detection when the part of the particle represented by the measurement unit 305 in FIG. 2 passes can be calculated. In the present embodiment, when the laser light pattern through which the particle passes is in the 10 mode of FIG. 5, instead of the probability of the particle path, the scattered light or shadow of the particle passing through the region is induced, and the change in the light Indicates the intensity of detection sensitivity when entering the photodetector.
 図2及び図5に示すように、白で示した光の間に黒の部分があり、この黒い部分は、位相シフトによる干渉で光が相殺された部分である。この部分をパーティクルが通過することが、システム全体の感度を向上させることに貢献する。この領域を多くのパーティクルが通過するように、光学系、フローセル、光検出器の位置関係をシステムで設定する。設定の結果、この部分を通過するパーティクルの信号強度は、2つの光検出器(第1検出器303及び第2検出器304)の差分信号を強くすることになるため、感度が高くなる。 As shown in FIGS. 2 and 5, there is a black portion between the light shown in white, and this black portion is a portion where the light is canceled by interference due to the phase shift. The passage of particles through this part contributes to improving the sensitivity of the entire system. The system sets the positional relationship of the optical system, flow cell, and photodetector so that many particles pass through this region. As a result of the setting, the signal intensity of the particles passing through this portion increases the difference signal between the two photodetectors (the first detector 303 and the second detector 304), and thus the sensitivity is increased.
 この黒の部分がパーティクルが透過していることを示す。つまり、黒の部分は、パーティクルにより光が遮断された領域に対応する。 This black part indicates that the particles are transmitted. That is, the black portion corresponds to a region where light is blocked by particles.
 [波形のモデル化による信号処理]
 次に、本実施形態に係る波形のモデル化による信号処理について、図6を参照しながら説明する。図6は、本実施形態に係る波形のモデル化による信号処理の一例を示すフローチャートである。本実施形態に係る波形のモデル化による信号処理は、主に制御装置200により、波形のサーチ→波形のフィッティング→特徴量の抽出→ルールに基づく特徴量の評価→波形を差し引いた残信号の生成→信号判定の順で実行される。
[Signal processing by waveform modeling]
Next, signal processing by waveform modeling according to the present embodiment will be described with reference to FIG. FIG. 6 is a flowchart showing an example of signal processing by waveform modeling according to the present embodiment. The signal processing by waveform modeling according to the present embodiment is performed mainly by the control device 200 by searching for a waveform → fitting a waveform → extracting a feature value → evaluating a feature value based on a rule → generating a residual signal after subtracting the waveform. → Executed in the order of signal judgment.
 (波形サーチ)
 本処理が開始されると、制御装置200は、第1検出器303及び第2検出器304が検出した信号を取得し、その差分を検出信号として検出信号の波形のピークをサーチする(ステップS10)。
(Waveform search)
When this processing is started, the control device 200 acquires the signals detected by the first detector 303 and the second detector 304, and searches for the peak of the waveform of the detection signal using the difference as a detection signal (step S10). ).
 検出信号の波形のピークをサーチする場合、制御装置200は、図7の上図に示すように、予め、検出信号の波形(元波形)に対して、S-Golay Filterによりフィルタ処理を行い、フィルタ処理後の波形に対してピークのサーチを実行してもよい。図7の下図に示すように、ピークのサーチでは、例えば、1024個のデータ中の最大ピーク(Max)、最小ピーク(Min)及びその近傍のピークをサーチしてもよい。図7に示すように、検出信号の波形は、必ずしも上下対称な波形とはならず、非対称な形状の波を含む。 When searching for the peak of the waveform of the detection signal, as shown in the upper diagram of FIG. 7, the control device 200 filters the detection signal waveform (original waveform) in advance using an S-Golay Filter, A peak search may be performed on the waveform after filtering. As shown in the lower diagram of FIG. 7, in the peak search, for example, a maximum peak (Max), a minimum peak (Min), and a peak in the vicinity thereof in 1024 data may be searched. As shown in FIG. 7, the waveform of the detection signal is not necessarily a vertically symmetric waveform, but includes an asymmetrical waveform.
 (波形フィッティング)
 図6に戻り、次に、制御装置200は、サーチした結果、最大ピークの高さA2及び最小ピークの高さA3を設定し、ガウシャン関数を示す式(1)を用いて最大ピーク及び最小ピークの幅を最適化(最小化)することで、波形をフィッティングする(ステップS12)。
(Wave fitting)
Returning to FIG. 6, next, as a result of the search, the control device 200 sets the maximum peak height A <b> 2 and the minimum peak height A <b> 3, and uses the formula (1) indicating the Gaussian function to determine the maximum peak and minimum peak. The waveform is fitted by optimizing (minimizing) the width (step S12).
 例えば、制御装置200は、図8の上図に示すように、まず、サーチ結果から、最大ピークの高さA2及び最小ピークの高さA3を設定する。ただし、A2及びA3を設定する波のピークは、最大及び最小のピーク以外のピークでもよい。次に、制御装置200は、図8の下図の左側に示すように、ガウシャン関数により波形の幅をフィッティングする。ここでは、ガウシャン関数を示す式(1)のω値と相関のあるパラメータのみを抽出している。制御装置200は、ピークの位置と高さA2,A3を揃えて、各ピークの幅をフィッティングさせる。つまり、図8の下図の右側に示すように、最大ピークの高さA2の波形と最小ピークの高さA3の波形とを正規化し、それらの波形を示すガウシャン関数のωの差分の2乗和の最小値が、最大ピーク及び最小ピークの幅となる。 For example, as shown in the upper diagram of FIG. 8, the control device 200 first sets the maximum peak height A2 and the minimum peak height A3 from the search result. However, the peaks of the waves that set A2 and A3 may be peaks other than the maximum and minimum peaks. Next, as shown on the left side of the lower diagram of FIG. 8, the control device 200 fits the width of the waveform using a Gaussian function. Here, only the parameter having a correlation with the ω value of the equation (1) indicating the Gaussian function is extracted. The control device 200 aligns the peak positions and the heights A2 and A3, and fits the width of each peak. That is, as shown on the right side of the lower diagram of FIG. 8, the waveform of the maximum peak height A2 and the waveform of the minimum peak height A3 are normalized, and the sum of squares of the difference of ω of the Gaussian function indicating those waveforms is obtained. The minimum value is the maximum peak and the minimum peak width.
 (ルールに基づく特徴量の評価)
 図6に戻り、次に、制御装置200は、予め設定されているルールに基づき、フィッティングした波形の特徴量を算出する(ステップS14)。例えば、設定したルールとルールに示した特徴量の閾値の一例を図9に示す。図9の各閾値は、特徴量毎に設定された第1の閾値の一例である。
(Evaluation of feature values based on rules)
Returning to FIG. 6, next, the control device 200 calculates the feature amount of the fitted waveform based on a preset rule (step S14). For example, FIG. 9 shows an example of the set rule and the threshold value of the feature amount indicated in the rule. Each threshold in FIG. 9 is an example of a first threshold set for each feature amount.
 図9に示すルールでは、ルール1では、最大ピークの高さA2の波形と最小ピークの高さA3の波形とを正規化して、それらの正規化した波形の差分の2乗和の最小値Diffが、波形の特徴量の一つとして定められている。正規化した波形の差分の2乗和の最小値Diffの閾値は、0~0.08に設定されている。 In the rule shown in FIG. 9, in rule 1, the waveform having the maximum peak height A2 and the waveform having the minimum peak height A3 are normalized, and the minimum value Diff of the sum of squares of the difference between these normalized waveforms is normalized. Is defined as one of the feature quantities of the waveform. The threshold value of the minimum value Diff of the sum of squares of the normalized waveform difference is set to 0 to 0.08.
 ルール2では、最大ピークの高さA2の波形と最小ピークの高さA3の波形の偏りRが、波形の特徴量の一つとして定められている。波形の偏りRは、最大ピークの高さA2と最小ピークの高さA3の比(A2/A3)である。波形の偏りRの閾値は、1.1~2.0に設定されている。 In Rule 2, the waveform R having the maximum peak height A2 and the waveform having the minimum peak height A3 is defined as one of the feature quantities of the waveform. The waveform deviation R is the ratio (A2 / A3) of the maximum peak height A2 to the minimum peak height A3. The threshold value of the waveform deviation R is set to 1.1 to 2.0.
 ルール3では、ピークの幅Offsetが、波形の特徴量の一つとして定められている。ピークの幅Offsetの閾値は、0~60に設定されている。また、ルール3では、最大ピークの高さA2の波形の幅と最小ピークの高さA3の波形の幅ωが、波形の特徴量の一つとして定められている。波形の幅ωは、式(1)のガウシャン関数により求められる。波形の幅ωの閾値は、0~5.0に設定されている。制御装置200は、各ルールに基づき、フィッティングした波形の各特徴量を算出する。 In Rule 3, the peak width Offset is defined as one of the feature quantities of the waveform. The threshold value of the peak width Offset is set to 0-60. Also, in rule 3, the width of the waveform having the maximum peak height A2 and the width ω of the waveform having the minimum peak height A3 are determined as one of the feature quantities of the waveform. The width ω of the waveform is obtained by the Gaussian function of Equation (1). The threshold value of the waveform width ω is set to 0 to 5.0. The control device 200 calculates each feature amount of the fitted waveform based on each rule.
 (波形を差し引いた残信号の生成)
 図6に戻り、次に、制御装置200は、フィッティングした波形の成分を抽出し、その波形の成分を検出信号から差し引く(ステップS16)。例えば、図10の(a)に示すように、制御装置200は、フィッティングした波形S1の成分を抽出する。そして、図10の(b)に示すように、制御装置200は、検出信号から波形S1の成分を除く。検出信号から波形S1の成分を除いた、図10の(b)に示す残信号が、次の解析対象となる。
(Generation of residual signal with waveform subtracted)
Returning to FIG. 6, next, the control device 200 extracts the fitted waveform component and subtracts the waveform component from the detection signal (step S16). For example, as shown in FIG. 10A, the control device 200 extracts the component of the fitted waveform S1. Then, as illustrated in FIG. 10B, the control device 200 removes the component of the waveform S1 from the detection signal. The remaining signal shown in (b) of FIG. 10 obtained by removing the component of the waveform S1 from the detection signal becomes the next analysis target.
 図6に戻り、次に、制御装置200は、残信号の最大ピークが予め定められた閾値(第2の閾値)よりも小さいかを判定する(ステップS18)。制御装置200は、残信号の最大ピークが第2の閾値よりも大きいと判定した場合、ステップS10に戻り、ステップS18にて残信号の最大ピークが予め定められた第2の閾値よりも小さいと判定されるまで、ステップS10~S18の処理を繰り返す。 Returning to FIG. 6, next, the control device 200 determines whether the maximum peak of the remaining signal is smaller than a predetermined threshold (second threshold) (step S18). When determining that the maximum peak of the residual signal is larger than the second threshold value, the control device 200 returns to step S10, and when the maximum peak of the residual signal is smaller than the predetermined second threshold value in step S18. Until it is determined, the processes in steps S10 to S18 are repeated.
 これにより、ステップS10~S18の処理が繰り返される度に、検出信号からフィッティングした波形の特徴量を算出後、フィッティングした波形成分が検出信号から除かれる。図10の(c)には、特徴量を算出した後、フィッティングした波形S1~S3の信号成分が順に検出信号から差し引かれる一例が示されている。 Thus, each time the processing of steps S10 to S18 is repeated, after the feature amount of the fitted waveform is calculated from the detection signal, the fitted waveform component is removed from the detection signal. FIG. 10 (c) shows an example in which the signal components of the fitted waveforms S1 to S3 are sequentially subtracted from the detection signal after the feature amount is calculated.
 (信号判定)
 図6に戻り、ステップS18において、制御装置200は、残信号の最大ピークが第2の閾値よりも小さいと判定した場合、フィッティングした各波形の特徴量のそれぞれが、第1の閾値のうちの各特徴量に対応する閾値の範囲内であるかを判定する(ステップS20)。ここでは、例えばルール1~ルール3に示した4つの特徴量の全てがそれぞれの閾値の範囲内である場合、制御装置200は、特徴量のすべてが第1の閾値の範囲内であると判定し、判定対象の波形の信号は、パーティクル信号であると判定し(ステップS22)、本処理を終了する。
(Signal judgment)
Returning to FIG. 6, in step S <b> 18, when the control device 200 determines that the maximum peak of the residual signal is smaller than the second threshold value, each of the feature values of the fitted waveforms is included in the first threshold value. It is determined whether it is within a threshold range corresponding to each feature amount (step S20). Here, for example, when all of the four feature amounts shown in the rules 1 to 3 are within the respective threshold ranges, the control device 200 determines that all of the feature amounts are within the first threshold range. Then, it is determined that the signal of the waveform to be determined is a particle signal (step S22), and this process ends.
 一方、4つの特徴量のうちの一つでも第1の閾値の範囲外である場合、制御装置200は、特徴量に対応する閾値の範囲外であると判定し、判定対象の波形の信号は、ノイズであると判定し(ステップS24)、本処理を終了する。例えば、図11に、図9のルールに基づくフィッティングした波形の信号の判定結果の一例を示す。図11の(a)はノイズと判定され、図11の(b)及び図11の(c)はパーティクル信号と判定されている。 On the other hand, if any one of the four feature amounts is outside the first threshold range, the control device 200 determines that the feature amount is outside the threshold range, and the determination target waveform signal is , It is determined that the noise is present (step S24), and this process is terminated. For example, FIG. 11 shows an example of the determination result of the signal of the fitted waveform based on the rule of FIG. 11A is determined as noise, and FIG. 11B and FIG. 11C are determined as particle signals.
 以上に説明した、本実施形態に係る波形のモデル化による信号処理によれば、抽出した信号の波形の特徴量が、第1の閾値の範囲内の場合、フィッティングした波形の信号はパーティクル信号であると判定される。また、抽出した波形の特徴量が、第1の閾値の範囲外の場合、フィッティングした波形の信号はノイズであると判定される。 According to the signal processing by waveform modeling according to the present embodiment described above, when the waveform feature amount of the extracted signal is within the first threshold range, the fitted waveform signal is a particle signal. It is determined that there is. Further, when the extracted waveform feature amount is outside the first threshold range, the fitted waveform signal is determined to be noise.
 これにより、検出信号からパーティクル又はノイズと判定された波形を取り除くことができる。これにより、名目上のSN比を改善することができる。この結果、液体又は気体の中の微粒子の検出において、SN比の高低にかかわらず検出信号から目的とする微粒子の信号を精度よく判定することができる。 This makes it possible to remove the waveform determined as particle or noise from the detection signal. Thereby, a nominal S / N ratio can be improved. As a result, in the detection of fine particles in a liquid or gas, it is possible to accurately determine the signal of the target fine particles from the detection signal regardless of the SN ratio.
 また、従来の正規化相互相関法(XCOR:cross correlation)では、予め定めたテンプレートの波形と検出信号とを比較してマッチングスコアを生成し、波形がテンプレートと類似する、マッチングスコアの高い波形を信号と判定する。正規化相互相関法では、信号から除外されてノイズと判定された波形中に本来的な信号の成分があり、誤判定がなされる可能性がある。これに対して、本実施形態に係る信号処理では、テンプレートによるマッチングではなく、波形をモデル化して抽出し、抽出した波形の特徴量が第1の閾値の範囲内であるかを判定することで、その判定結果に基づき、抽出した波形の信号がパーティクル信号かノイズかが判定される。これにより、従来の正規化相互相関法では波形がテンプレートと類似しないためにノイズと判定された信号が、本実施形態ではパーティクル信号と判定され得る。これにより、本実施形態に係る波形のモデル化による信号処理によれば、任意性の高い波形の検出信号に対して、より精度の高いパーティクル信号の検出が可能になる。 In the conventional normalized cross correlation (XCOR) method, a matching waveform is generated by comparing a predetermined template waveform and a detection signal, and a waveform having a high matching score, which is similar to the template, is generated. Judge as a signal. In the normalized cross-correlation method, there is an original signal component in the waveform that is excluded from the signal and determined to be noise, and there is a possibility that an erroneous determination is made. On the other hand, in the signal processing according to the present embodiment, instead of matching by a template, a waveform is modeled and extracted, and it is determined whether or not the feature amount of the extracted waveform is within the first threshold range. Based on the determination result, it is determined whether the extracted waveform signal is a particle signal or noise. As a result, a signal determined to be noise because the waveform is not similar to the template in the conventional normalized cross-correlation method can be determined to be a particle signal in the present embodiment. Thereby, according to the signal processing by waveform modeling according to the present embodiment, it is possible to detect a particle signal with higher accuracy with respect to a detection signal having a highly arbitrary waveform.
 (判定結果の精度)
 例えば、図12には、本実施形態に係る検出信号、及び検出信号から抽出した波形の特徴量と抽出した波形の信号に対する判定結果の一例を示す。図12の上段は、レーザ光源301から出力されるレーザ光のスポットの直径が1.2μm、レーザ光のビームのパワーが20mW、検出器302が一つのときの検出信号の分布の一例を示す。図12の上段左のグラフで示す検出信号のSN比は1.2であり、中央のグラフで示す検出信号のSN比は1.5であり、右のグラフで示す検出信号のSN比は2.0である。
(Accuracy of judgment results)
For example, FIG. 12 illustrates an example of a detection signal according to the present embodiment, a waveform feature amount extracted from the detection signal, and a determination result for the extracted waveform signal. The upper part of FIG. 12 shows an example of the distribution of detection signals when the diameter of the laser light spot output from the laser light source 301 is 1.2 μm, the power of the laser light beam is 20 mW, and the number of detectors 302 is one. The SN ratio of the detection signal shown in the upper left graph of FIG. 12 is 1.2, the SN ratio of the detection signal shown in the center graph is 1.5, and the SN ratio of the detection signal shown in the right graph is 2. .0.
 図12の下段は、各検出信号からフィッティングにより抽出した波形の特徴量の分布と、抽出した波形が信号又はノイズのいずれかかを判定した結果及びミスカウントの確率を示す。 The lower part of FIG. 12 shows the distribution of the feature amount of the waveform extracted from each detection signal by fitting, the result of determining whether the extracted waveform is a signal or noise, and the probability of miscounting.
 図12の下段の3次元座標の3つの軸には、3つの特徴量である「A2/A3比」、「Offset」,「Diff」が示されている。図12の上段の検出信号から、図12の下段の特徴量を持つ波形が抽出されたことが示されている。本実施形態に係る信号処理によれば、図12の下段の「〇」が示す特徴量を有する波形成分はパーティクル信号と判定され、図12の下段の「×」が示す特徴量を有する波形成分はノイズと判定されている。 In the three axes of the three-dimensional coordinates in the lower part of FIG. 12, three feature amounts “A2 / A3 ratio”, “Offset”, and “Diff” are shown. It is shown that a waveform having the characteristic quantity in the lower part of FIG. 12 is extracted from the detection signal in the upper part of FIG. According to the signal processing according to the present embodiment, the waveform component having the feature value indicated by “◯” in the lower part of FIG. 12 is determined as a particle signal, and the waveform component having the feature value indicated by “X” in the lower part of FIG. Is determined to be noise.
 このようにして、ルールで設定した特徴量及び各特徴量に対する閾値(第1の閾値)条件で評価を行った結果、従来の信号処理では、信号波形の認識が困難であったSN比が1.2程度の信号においてもミスカウントが0.2%となった。また、SN比が1.5の信号ではミスカウントは0.05%となり、SN比が2.0の信号ではミスカウントは0%となった。この結果から、本実施形態に係る波形のモデル化の信号処理では、検出信号に含まれる信号又はノイズ成分の区別を99%以上の精度で行うことができることがわかる。 Thus, as a result of performing the evaluation under the feature amount set by the rule and the threshold value (first threshold value) condition for each feature amount, the signal-to-noise ratio, which is difficult to recognize the signal waveform in the conventional signal processing, is 1. The miscount was 0.2% even for signals of about .2. Further, the miscount was 0.05% for a signal with an SN ratio of 1.5, and the miscount was 0% for a signal with an SN ratio of 2.0. From this result, it can be seen that in the signal processing for waveform modeling according to the present embodiment, the signal or the noise component included in the detection signal can be distinguished with an accuracy of 99% or more.
 [SN比に応じた信号処理]
 次に、本実施形態に係るSN比に応じた信号処理の一例について、図13を参照しながら説明する。図13は、本実施形態に係るSN比に応じた信号処理の一例を示すフローチャートである。
[Signal processing according to SN ratio]
Next, an example of signal processing according to the SN ratio according to the present embodiment will be described with reference to FIG. FIG. 13 is a flowchart illustrating an example of signal processing according to the SN ratio according to the present embodiment.
 本処理が開始されると、制御装置200は、検出器302から検出信号を取得する(ステップS30)。検出信号は、一つの検出器が検出した信号であってもよいし、二つの検出器が検出した信号の差分であってもよい。 When this process is started, the control device 200 acquires a detection signal from the detector 302 (step S30). The detection signal may be a signal detected by one detector or a difference between signals detected by two detectors.
 次に、制御装置200は、検出信号を正規化相互相関法(XCOR)を用いてテンプレートの波形とマッチングし、マッチングスコア(検出信号の相互相関値)を求め、マッチングスコアからSN比を算出する(ステップS32)。 Next, the control device 200 matches the detection signal with the waveform of the template using the normalized cross correlation method (XCOR), obtains a matching score (cross correlation value of the detection signal), and calculates the SN ratio from the matching score. (Step S32).
 図14は、波形のテンプレートとのマッチングによる波形抽出の一例を示す。制御装置200は、正規化相互相関法(XCOR)では、特徴波形のテンプレートを検出信号内の特徴波形部分とマッチングさせる。制御装置200は、マッチングスコアが閾値以上の場合、その波形を検出信号(パーティクル信号)と判定する。 FIG. 14 shows an example of waveform extraction by matching with a waveform template. In the normalized cross correlation method (XCOR), the control device 200 matches the feature waveform template with the feature waveform portion in the detection signal. When the matching score is equal to or greater than the threshold, the control device 200 determines the waveform as a detection signal (particle signal).
 図13に戻り、次に、制御装置200は、SN比が1.5以上かを判定する(ステップS34)。制御装置200は、SN比が1.5以上であると判定した場合、図14に一例を示す正規化相互相関法(XCOR)によるマッチングスコアによって、検出信号がパーティクル信号か否かを判定する(ステップS36)。その後、制御装置200は、ステップS30に戻って次の検出信号を取得し、ステップS30以降の処理を繰り返す。 13, next, the control device 200 determines whether or not the SN ratio is 1.5 or more (step S34). When it is determined that the SN ratio is 1.5 or more, the control device 200 determines whether or not the detection signal is a particle signal based on a matching score by a normalized cross correlation method (XCOR) shown in FIG. Step S36). Thereafter, the control device 200 returns to step S30 to acquire the next detection signal, and repeats the processing after step S30.
 一方、ステップS34において、SN比が1.5未満であると判定された場合、制御装置200は、SN比が1.2以上であるかを判定する(ステップS38)。制御装置200は、SN比が1.2以上であると判定した場合、図6に示した本実施形態に係る波形のモデル化による信号処理を実行し(ステップS40)、ステップS30に戻る。一方、ステップS38において、制御装置200は、SN比が1.2未満であると判定した場合、直ちにステップS30に戻る。 On the other hand, when it is determined in step S34 that the SN ratio is less than 1.5, the control device 200 determines whether the SN ratio is 1.2 or more (step S38). When it is determined that the SN ratio is 1.2 or more, the control device 200 executes signal processing by waveform modeling according to the present embodiment shown in FIG. 6 (step S40), and returns to step S30. On the other hand, when determining in step S38 that the SN ratio is less than 1.2, the control device 200 immediately returns to step S30.
 これによれば、SN比が1.5以上であれば正規化相互相関法(XCOR)を用いた信号処理では所定以上の精度を得ることができる。よって、SN比が1.5以上の場合、例えば図14に示す正規化相互相関法(XCOR)を用いた信号処理により、パーティクル信号が検出される。 According to this, if the signal-to-noise ratio is 1.5 or more, the signal processing using the normalized cross-correlation method (XCOR) can obtain a predetermined accuracy or more. Therefore, when the SN ratio is 1.5 or more, for example, a particle signal is detected by signal processing using the normalized cross correlation method (XCOR) shown in FIG.
 一方、SNが1.5未満の場合、正規化相互相関法(XCOR)を用いた信号処理では所定以上の精度を得ることができない。よって、SN比が1.5未満の場合、本実施形態に係る波形のモデル化による信号処理(図6)が実行されることで、例えば図15に示すように、正規化相互相関法(XCOR)ではノイズに埋もれて検出できなかった特徴波形(パーティクル信号)を検出することができる。 On the other hand, when SN is less than 1.5, signal processing using the normalized cross-correlation method (XCOR) cannot obtain a predetermined accuracy or more. Therefore, when the signal-to-noise ratio is less than 1.5, signal processing (FIG. 6) by waveform modeling according to the present embodiment is executed, so that, for example, as shown in FIG. ) Can detect a characteristic waveform (particle signal) that is buried in noise and cannot be detected.
 このように、SN比が1.5~2未満になると、検出信号中のパーティクル信号とノイズとの区別がつき難い。SNRが1.5~2以上になると、検出信号中のパーティクル信号とノイズとの区別がつき易い。よって、SN比に応じて、検出信号の信号処理方法を変えることで、SN比の高低にかかわらず、検出信号からパーティクル信号かノイズかを精度よく判定することができる。また、SN比に応じて、検出信号の信号処理方法を変えることで、処理の付加を軽減することができる。 Thus, when the S / N ratio is less than 1.5-2, it is difficult to distinguish between the particle signal and noise in the detection signal. When the SNR is 1.5 to 2 or more, it is easy to distinguish between the particle signal in the detection signal and the noise. Therefore, by changing the signal processing method of the detection signal according to the SN ratio, it is possible to accurately determine whether the signal is a particle signal or noise regardless of the level of the SN ratio. Further, the processing addition can be reduced by changing the signal processing method of the detection signal in accordance with the SN ratio.
 なお、図13のステップS38では、SN比の下限値を1.2とした。しかしながら、SN比の下限値は1.2以外の数値(例えば、1.0)であってもよいし、SN比の下限値は設けなくてもよい。また、図13のステップS34では、検出信号の信号処理方法を変えるSN比の値を1.5とした。しかしながら、これに限らず、ステップS34では、SN比が2以上等、1.5~2の範囲のいずれかの値を検出信号の信号処理方法を変えるSN比の値としてもよい。 In step S38 in FIG. 13, the lower limit value of the SN ratio is set to 1.2. However, the lower limit value of the SN ratio may be a numerical value other than 1.2 (for example, 1.0), and the lower limit value of the SN ratio may not be provided. In step S34 in FIG. 13, the value of the S / N ratio that changes the signal processing method of the detection signal is set to 1.5. However, the present invention is not limited to this. In step S34, any value in the range of 1.5 to 2, such as an SN ratio of 2 or more, may be used as the SN ratio value that changes the signal processing method of the detection signal.
 (制御装置のハードウェア構成)
 最後に、制御装置200のハードウェア構成の一例について、図16を参照しながら簡単に説明する。制御装置200は、パーソナルコンピュータやタブレット型の端末等の情報処理装置である。制御装置200は、入力装置101、表示装置102、外部I/F103、RAM(Random Access Memory)104、ROM(Read Only Memory)105、CPU(Central Processing Unit)106、通信I/F107、及びHDD(Hard Disk Drive)108などを備え、それぞれがバスBで相互に接続されている。
(Hardware configuration of control device)
Finally, an example of the hardware configuration of the control device 200 will be briefly described with reference to FIG. The control device 200 is an information processing device such as a personal computer or a tablet-type terminal. The control device 200 includes an input device 101, a display device 102, an external I / F 103, a RAM (Random Access Memory) 104, a ROM (Read Only Memory) 105, a CPU (Central Processing Unit) 106, a communication I / F 107, and an HDD ( Hard Disk Drive) 108 and the like are connected to each other via a bus B.
 入力装置101は、キーボードやマウスなどを含み、制御装置200に各操作信号を入力するために用いられる。表示装置102は、ディスプレイなどを含み、各種の処理結果を表示する。通信I/F107は、制御装置200をネットワークに接続するインターフェースである。これにより、制御装置200は、通信I/F107を介して、他の機器(検出器302等)とデータ通信を行うことができる。これにより、制御装置200は、検出器302からレーザ光の検出信号を取得する。 The input device 101 includes a keyboard and a mouse, and is used for inputting each operation signal to the control device 200. The display device 102 includes a display and displays various processing results. The communication I / F 107 is an interface that connects the control device 200 to a network. Thereby, the control apparatus 200 can perform data communication with other apparatuses (detector 302 etc.) via communication I / F107. Thereby, the control apparatus 200 acquires the detection signal of a laser beam from the detector 302.
 HDD108は、プログラムやデータを格納している不揮発性の記憶装置である。格納されるプログラムやデータには、制御装置200の全体を制御する基本ソフトウェア及びアプリケーションソフトウェアがある。例えば、HDD108には、各種のデータベースやプログラム等が格納されてもよい。 The HDD 108 is a non-volatile storage device that stores programs and data. The stored programs and data include basic software and application software that control the entire control device 200. For example, the HDD 108 may store various databases and programs.
 外部I/F103は、外部装置とのインターフェースである。外部装置には、記録媒体103aなどがある。これにより、制御装置200は、外部I/F103を介して記録媒体103aの読み取り及び/又は書き込みを行うことができる。記録媒体103aには、CD(Compact Disk)、及びDVD(Digital Versatile Disk)、ならびに、SDメモリーカード(SD Memory card)やUSBメモリ(Universal Serial Bus memory)等がある。 External I / F 103 is an interface with an external device. The external device includes a recording medium 103a. Thereby, the control device 200 can read and / or write the recording medium 103a via the external I / F 103. The recording medium 103a includes a CD (Compact Disk), a DVD (Digital Versatile Disk), an SD memory card (SD Memory Card), a USB memory (Universal Serial Bus memory), and the like.
 ROM105は、電源を切っても内部データを保持することができる不揮発性の半導体メモリ(記憶装置)である。ROM105には、ネットワーク設定等のプログラム及びデータが格納されている。RAM104は、プログラムやデータを一時保持する揮発性の半導体メモリ(記憶装置)である。 The ROM 105 is a nonvolatile semiconductor memory (storage device) that can retain internal data even when the power is turned off. The ROM 105 stores programs and data such as network settings. The RAM 104 is a volatile semiconductor memory (storage device) that temporarily stores programs and data.
 CPU106は、上記記憶装置(例えば「HDD108」や「ROM105」など)から、プログラムやデータをRAM104上に読み出し、処理を実行することで、装置全体の制御や搭載機能を実現する演算装置である。例えば、記憶装置には、本実施形態に係る波形のモデル化による信号処理のプログラムが記憶されていて、CPU106は、上記記憶装置から前記プログラムを読み出し、プログラムに示される手順で処理を実行することで、検出信号に含まれるパーティクル信号及びノイズの判定が可能になる。 The CPU 106 is an arithmetic unit that realizes control of the entire apparatus and mounting functions by reading programs and data from the storage device (for example, “HDD 108”, “ROM 105”, etc.) onto the RAM 104 and executing processing. For example, the storage device stores a signal processing program by waveform modeling according to the present embodiment, and the CPU 106 reads the program from the storage device and executes the processing according to the procedure indicated by the program. Thus, the particle signal and noise included in the detection signal can be determined.
 以上、本実施形態に係る波形のモデル化による信号処理によれば、液体又は気体の中のパーティクルの検出において、SN比の高低にかかわらず検出信号から目的とするパーティクルの信号を精度よく判定することができる。 As described above, according to the signal processing based on the waveform modeling according to the present embodiment, in the detection of particles in a liquid or gas, the target particle signal is accurately determined from the detection signal regardless of the SN ratio. be able to.
 従来の信号処理では、比較的単純な(特徴部の少ない)ガウシャン関数による波形等で表せる信号が解析対象となっていたのに対して、本実施形態に係る信号処理の対象は、例えば、2以上のフォトディテクタ(例えば、図2の第1及び第2検出器)の差分信号であり、信号波形が複雑化している。つまり、複雑化した波形の信号は、信号に含まれる特徴部(パラメータ)が増えた信号である。以上から、2以上のフォトディテクタの差分信号は、本実施形態に係る信号処理が抽出する特徴量の抽出がし易い信号であり、その結果、本実施形態に係る特徴量に基づくパーティクル信号又はノイズの判定精度が向上しているとも考えられる。 In the conventional signal processing, a signal that can be expressed by a waveform or the like by a relatively simple Gaussian function (having few characteristic parts) is an analysis target. On the other hand, the signal processing target according to the present embodiment is, for example, 2 This is a differential signal of the above photodetector (for example, the first and second detectors in FIG. 2), and the signal waveform is complicated. In other words, a signal having a complicated waveform is a signal with an increased number of features (parameters) included in the signal. From the above, the difference signal of two or more photodetectors is a signal that is easy to extract the feature amount extracted by the signal processing according to the present embodiment, and as a result, the particle signal or noise based on the feature amount according to the present embodiment. It is thought that the determination accuracy is improved.
 また、レーザ光源301や検出器302には、比較的ノイズ成分が多いけれども安価な機器が存在する。その場合、SN比が1.5~2.0以下となる場合であっても安価な機器を使用したい場合がある。また、水等よりも有機液中ではさらにSN比は低くなる傾向がある。このように、SN比が1.5~2.0以下となる場合においても、本実施形態に係る波形のモデル化による信号処理により、検出信号からパーティクル信号とノイズとを精度よく区別することができる。 Further, the laser light source 301 and the detector 302 include inexpensive devices although they have a relatively large noise component. In that case, there is a case where it is desired to use an inexpensive device even when the SN ratio is 1.5 to 2.0 or less. In addition, the S / N ratio tends to be lower in an organic liquid than water. As described above, even when the SN ratio is 1.5 to 2.0 or less, it is possible to accurately distinguish the particle signal and the noise from the detection signal by the signal processing based on the waveform modeling according to the present embodiment. it can.
 なお、本実施形態に係る波形のモデル化による波形の抽出と信号判定結果を蓄積し、機械学習させることもできる。その際、予めバラツキの範囲を閾値により定義して、特徴量が閾値の範囲外にあれば、その信号は機械学習の対象とせずに廃棄する。このようにして、機械学習精度を高めることで、制御装置200は、学習した結果に基づき記憶部に蓄積したパーティクル信号の波形から、リアルタイムに最適な波形を抽出して、本実施形態に係る波形のモデル化による信号処理に適用するようにしてもよい。 It should be noted that waveform extraction and signal determination results by waveform modeling according to the present embodiment can be accumulated and machine learning can be performed. At this time, the range of variation is defined in advance by a threshold, and if the feature amount is outside the range of the threshold, the signal is discarded without being subjected to machine learning. In this way, by increasing the machine learning accuracy, the control device 200 extracts an optimum waveform in real time from the waveform of the particle signal accumulated in the storage unit based on the learned result, and the waveform according to the present embodiment. You may make it apply to the signal processing by modeling of.
 以上、信号処理方法及びプログラムを上記実施形態により説明したが、本発明にかかる信号処理方法及びプログラムは上記実施形態に限定されるものではなく、本発明の範囲内で種々の変形及び改良が可能である。上記複数の実施形態に記載された事項は、矛盾しない範囲で組み合わせることができる。 The signal processing method and program have been described in the above embodiment. However, the signal processing method and program according to the present invention are not limited to the above embodiment, and various modifications and improvements can be made within the scope of the present invention. It is. The matters described in the above embodiments can be combined within a consistent range.
 本発明は、各種装置に取付けられた管内を流れる液体又は気体の被検査体の中のパーティクル等の微粒子の検出において、SN比の高低にかかわらず検出信号から微粒子を示す信号か否かを精度よく判定する。例えば、本発明は、容量結合型プラズマ(CCP:Capacitively Coupled Plasma)装置、誘導結合型プラズマ(ICP:Inductively Coupled Plasma)処理装置、ラジアルラインスロットアンテナを用いたプラズマ処理装置、ヘリコン波励起型プラズマ(HWP:Helicon Wave Plasma)装置、電子サイクロトロン共鳴プラズマ(ECR:Electron Cyclotron Resonance Plasma)装置、表面波プラズマ処理装置等の装置にも適用できる。 In the present invention, when detecting particles such as particles in a liquid or gas inspected object flowing in tubes attached to various devices, it is possible to accurately determine whether or not the signal indicates the particles from the detection signal regardless of whether the SN ratio is high or low. Judge well. For example, the present invention includes a capacitively coupled plasma (CCP) device, an inductively coupled plasma (ICP) processing device, a plasma processing device using a radial line slot antenna, a helicon wave excited plasma ( The present invention can also be applied to devices such as HWP: Helicon Wave) Plasma devices, electron cyclotron resonance plasma (ECR) devices, and surface wave plasma processing devices.
 本実施形態にかかる信号処理方法は、流体(液体又は気体)の被検査体の中のパーティクル等の微粒子を検出する様々な測定装置において適用可能である。上記実施形態では、その構成の一例として図1及び図2を挙げて洗浄装置100内の測定機構300について説明した。他の構成としては、図17に示すようなレーザ散乱光を利用するパーティクルモニタ133(指標測定装置)が挙げられる。 The signal processing method according to the present embodiment can be applied to various measuring apparatuses that detect fine particles such as particles in a fluid (liquid or gas) inspection object. In the above embodiment, the measurement mechanism 300 in the cleaning apparatus 100 has been described with reference to FIGS. 1 and 2 as an example of the configuration. As another configuration, there is a particle monitor 133 (index measuring device) using laser scattered light as shown in FIG.
 図17において、パーティクルモニタ133は、レーザ光を発振するレーザ発振器134と、チャンバ112内の散乱光を観測するCCDカメラ135と、レーザ発振器134及びCCDカメラ135に接続されたパルスジェネレータ136とを備える。 17, the particle monitor 133 includes a laser oscillator 134 that oscillates laser light, a CCD camera 135 that observes scattered light in the chamber 112, and a pulse generator 136 connected to the laser oscillator 134 and the CCD camera 135. .
 レーザ発振器134は、チャンバ112に設けられたスリット窓137を介してエッチング等の処理実行中のチャンバ112内に向けてレーザ光を発振する。チャンバ112内のパーティクルpは、レーザ光によって照射されると散乱光を発生する。この発生した散乱光は、スリット窓138を介してCCDカメラ135によって観測される。このとき、散乱光の発生回数や強度がチャンバ112内に浮遊するパーティクルpの量に対応する。したがって、パーティクルモニタ133は散乱光の発生回数や強度を通じてチャンバ112内に浮遊するパーティクルpの量を計測する(指標計測ステップ)。そして、計測されたパーティクルpの量が計測結果としてインターネット131等によってPC132に送信される。 The laser oscillator 134 oscillates a laser beam toward the inside of the chamber 112 during processing such as etching through a slit window 137 provided in the chamber 112. The particles p in the chamber 112 generate scattered light when irradiated with laser light. The generated scattered light is observed by the CCD camera 135 through the slit window 138. At this time, the number of occurrences and the intensity of scattered light correspond to the amount of particles p floating in the chamber 112. Therefore, the particle monitor 133 measures the amount of particles p floating in the chamber 112 through the number and intensity of scattered light generation (index measurement step). Then, the measured amount of particles p is transmitted as a measurement result to the PC 132 via the Internet 131 or the like.
 なお、パルスジェネレータ136は、レーザ発振器134及びCCDカメラ135に同期信号を発信し、これにより、レーザ光の発振のタイミングと散乱光の受光のタイミングとを調整する。 Note that the pulse generator 136 transmits a synchronization signal to the laser oscillator 134 and the CCD camera 135, and thereby adjusts the timing of oscillation of the laser light and the timing of reception of the scattered light.
 また、パーティクルモニタ133は、チャンバ112から該チャンバ112の稼働状況や不具合の有無を表す装置ステータス信号を受信し、且つ該装置ステータス信号をインターネット131等を介してPC132に送信する送信装置139を備える。PC132は、本実施形態にかかる信号処理を行う。 In addition, the particle monitor 133 includes a transmission device 139 that receives a device status signal indicating the operating status of the chamber 112 and the presence or absence of a failure from the chamber 112 and transmits the device status signal to the PC 132 via the Internet 131 or the like. . The PC 132 performs signal processing according to the present embodiment.
 更に、本実施形態にかかる信号処理方法を適用可能な測定装置としては、上記のようにチャンバ112内にレーザを照射し、基板上の空間に浮遊するパーティクルを検出する場合に限らない。例えば、本実施形態にかかる信号処理方法は、チャンバ112内に載置された基板に真上からレーザを照射し、基板からの反射光を検出することで、基板上に付着するパーティクルを検出する装置に適用することも可能である。 Furthermore, the measurement apparatus to which the signal processing method according to the present embodiment can be applied is not limited to the case where the laser is irradiated into the chamber 112 as described above to detect particles floating in the space on the substrate. For example, the signal processing method according to the present embodiment detects particles adhering to the substrate by irradiating the substrate placed in the chamber 112 with a laser beam from directly above and detecting reflected light from the substrate. It is also possible to apply to an apparatus.
 本国際出願は、2016年12月8日に出願された日本国特許出願2016-238691号に基づく優先権を主張するものであり、その全内容を本国際出願に援用する。 This international application claims priority based on Japanese Patent Application No. 2016-238691 filed on December 8, 2016, the entire contents of which are incorporated herein by reference.
 6   表面洗浄ノズル
 31  基板支持部
 32  回転機構
 41  カップ
 41a 開口部
 61  洗浄液ノズル  
 61a 供給路
 62  ガスノズル
 62a 供給路
 63  支持部
 65  洗浄液(DIW)源
 66  窒素ガス源
 100 洗浄装置
 200 制御装置
 300 測定機構
 301 レーザ光源
 302 検出器
 303 第1検出器
 304 第2検出器
 305 計測部
6 Surface cleaning nozzle 31 Substrate support section 32 Rotating mechanism 41 Cup 41a Opening 61 Cleaning liquid nozzle
61a Supply path 62 Gas nozzle 62a Supply path 63 Support section 65 Cleaning liquid (DIW) source 66 Nitrogen gas source 100 Cleaning apparatus 200 Control apparatus 300 Measuring mechanism 301 Laser light source 302 Detector 303 First detector 304 Second detector 305 Measuring section

Claims (11)

  1.  流体の被検査体に照射した光の信号を検出し、
     検出した前記信号の波形のピークを抽出し、
     抽出した前記ピークの波形の幅をフィッティングした後の波形の特徴量を算出し、
     算出した前記波形の特徴量に基づき、前記フィッティングした波形の信号が前記被検査体の中の微粒子を示す信号であるかを判定する、
     処理をコンピュータが実行する信号処理方法。
    Detects the light signal irradiated to the fluid test object,
    Extract the peak of the detected waveform of the signal,
    Calculate the feature value of the waveform after fitting the width of the extracted waveform of the peak,
    Based on the calculated feature amount of the waveform, it is determined whether the signal of the fitted waveform is a signal indicating fine particles in the object to be inspected.
    A signal processing method in which processing is executed by a computer.
  2.  算出した前記波形の特徴量が予め定められた第1の閾値の範囲内である場合、前記フィッティングした波形の信号が前記被検査体の中の微粒子を示す信号であると判定する、
     処理をコンピュータが実行する信号処理方法。
    When the calculated feature amount of the waveform is within a predetermined first threshold range, it is determined that the signal of the fitted waveform is a signal indicating fine particles in the object to be inspected.
    A signal processing method in which processing is executed by a computer.
  3.  検出した前記信号から前記フィッティングした波形の信号成分を除き、
     除いた後の残信号から抽出したピークの波形の幅をフィッティングし、前記フィッティングした波形の特徴量に基づき、該波形の信号が前記被検査体の中の微粒子を示す信号か否かを判定する、
     請求項1に記載の信号処理方法。
    The signal component of the fitted waveform is removed from the detected signal,
    Fitting the width of the peak waveform extracted from the residual signal after the removal, and determining whether the signal of the waveform is a signal indicating the fine particles in the inspected object based on the feature amount of the fitted waveform ,
    The signal processing method according to claim 1.
  4.  前記残信号の最大ピークが、予め定められた第2の閾値よりも小さくなるまで、
     検出した前記信号から前記フィッティングした波形の信号成分を除き、
     除いた後の残信号から抽出したピークの波形の幅をフィッティングし、前記フィッティングした波形の特徴量に基づき、該波形の信号が前記被検査体の中の微粒子を示す信号か否かを判定する、処理を繰り返す、
     請求項3に記載の信号処理方法。
    Until the maximum peak of the residual signal is smaller than a predetermined second threshold,
    The signal component of the fitted waveform is removed from the detected signal,
    Fitting the width of the peak waveform extracted from the residual signal after the removal, and determining whether the signal of the waveform is a signal indicating the fine particles in the inspected object based on the feature amount of the fitted waveform , Repeat the process,
    The signal processing method according to claim 3.
  5.  算出した前記波形の特徴量が予め定められた第1の閾値の範囲外である場合、前記フィッティングした波形の信号はノイズであると判定する、
     請求項1に記載の信号処理方法。
    When the calculated feature amount of the waveform is outside a predetermined first threshold range, it is determined that the signal of the fitted waveform is noise.
    The signal processing method according to claim 1.
  6.  前記フィッティングした波形の特徴量として、フィッティングした前記ピークの波形の幅、フィッティングした前記ピークの波形の幅の差分の2乗和の最小値、フィッティングした正負が異なるピーク間のオフセット、及びフィッティングした前記正負が異なるピークの高さの比率を算出し、
     算出した前記波形の特徴量のすべてが、前記第1の閾値のうちの特徴量毎に設定された閾値の範囲内である場合、前記フィッティングした波形の信号が前記被検査体の中の微粒子を示す信号であると判定する、
     請求項1に記載の信号処理方法。
    As features of the fitted waveform, the width of the waveform of the fitted peak, the minimum value of the sum of squares of the difference in the width of the fitted waveform of the waveform, the offset between peaks with different positive and negative fitted, and the fitted Calculate the ratio of peak heights with different positive and negative,
    When all of the calculated feature values of the waveform are within the threshold value range set for each feature value of the first threshold value, the signal of the fitted waveform indicates the fine particles in the object to be inspected. It is determined that the signal is
    The signal processing method according to claim 1.
  7.  検出した前記信号のSN比が1.5よりも小さい場合、上記信号処理方法を用いて検出した前記信号を処理し、
     検出した前記信号のSN比が1.5以上の場合、正規化相互相関法を用いて検出した前記信号を処理し、前記被検査体の中の微粒子を示す信号であるか否かを判定する、
     請求項1に記載の信号処理方法。
    If the signal-to-noise ratio of the detected signal is less than 1.5, processing the detected signal using the signal processing method;
    When the signal-to-noise ratio of the detected signal is 1.5 or more, the detected signal is processed using a normalized cross-correlation method, and it is determined whether or not the signal indicates a particle in the inspected object. ,
    The signal processing method according to claim 1.
  8.  流体の被検査体に照射した光の信号を検出する処理は、
     洗浄装置に設けられた表面洗浄ノズルに接続される供給管の内部を流れる洗浄液の被検査体に照射した光の散乱光を検出する、
     請求項1に記載の信号処理方法。
    The process of detecting the signal of the light irradiated to the fluid test object is
    Detects scattered light of the light irradiated to the inspection object of the cleaning liquid flowing inside the supply pipe connected to the surface cleaning nozzle provided in the cleaning device,
    The signal processing method according to claim 1.
  9.  流体の被検査体に照射した光の信号を検出する処理は、
     チャンバ内に載置された基板上の空間にレーザを照射し、該基板上の空間に浮遊するパーティクルによる散乱光を検出する、
     請求項1に記載の信号処理方法。
    The process of detecting the signal of the light irradiated to the fluid test object is
    Irradiating the space on the substrate placed in the chamber with laser, and detecting scattered light caused by particles floating in the space on the substrate;
    The signal processing method according to claim 1.
  10.  チャンバ内に載置された基板に向けてレーザを照射し、該基板からの反射光を検出する、
     請求項1に記載の信号処理方法。
    Irradiating a laser toward the substrate placed in the chamber, and detecting reflected light from the substrate,
    The signal processing method according to claim 1.
  11.  流体の被検査体に照射した光の信号を検出し、
     検出した前記信号の波形のピークを抽出し、
     抽出した前記ピークの波形の幅をフィッティングした後の波形の特徴量を算出し、
     算出した前記波形の特徴量に基づき、前記フィッティングした波形の信号が前記被検査体の中の微粒子を示す信号であるかを判定する、
     処理をコンピュータに実行させるプログラム。
     
    Detects the light signal irradiated to the fluid test object,
    Extract the peak of the detected waveform of the signal,
    Calculate the feature value of the waveform after fitting the width of the extracted waveform of the peak,
    Based on the calculated feature amount of the waveform, it is determined whether the signal of the fitted waveform is a signal indicating fine particles in the object to be inspected.
    A program that causes a computer to execute processing.
PCT/JP2017/042854 2016-12-08 2017-11-29 Signal processing method and program WO2018105462A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2018554949A JP6742435B2 (en) 2016-12-08 2017-11-29 Signal processing method and program

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2016-238691 2016-12-08
JP2016238691 2016-12-08

Publications (1)

Publication Number Publication Date
WO2018105462A1 true WO2018105462A1 (en) 2018-06-14

Family

ID=62491019

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2017/042854 WO2018105462A1 (en) 2016-12-08 2017-11-29 Signal processing method and program

Country Status (2)

Country Link
JP (1) JP6742435B2 (en)
WO (1) WO2018105462A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20210000656A (en) * 2019-06-25 2021-01-05 오므론 가부시키가이샤 Appearance inspection management system, appearance inspection management device, appearance inspection management method, and appearance inspection management program
WO2022024389A1 (en) * 2020-07-31 2022-02-03 株式会社日立ハイテク Method for generating trained model, method for determining base sequence of biomolecule, and biomolecule measurement device

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05240770A (en) * 1992-02-29 1993-09-17 Horiba Ltd Particle counter
JP2003279484A (en) * 2002-03-26 2003-10-02 Sharp Corp Fine particle detection method, fine particle detection device, fine particle manufacturing device, fine particle manufacturing program, storage medium and solid-state component
JP2005502875A (en) * 2001-09-07 2005-01-27 インフィコン インコーポレイティッド Signal processing method for particle monitoring of in-situ scanning beam
JP2005537781A (en) * 2002-02-14 2005-12-15 イムニベスト・コーポレイション Method and algorithm for cell counting at low cost
JP2006153745A (en) * 2004-11-30 2006-06-15 Tokyo Electron Ltd Particle-detecting method and particle-detecting program
JP2009097959A (en) * 2007-10-16 2009-05-07 Tokyo Seimitsu Co Ltd Defect detecting device and defect detection method
WO2011108369A1 (en) * 2010-03-01 2011-09-09 オリンパス株式会社 Optical analysis device, optical analysis method, and computer program for optical analysis
WO2013031309A1 (en) * 2011-08-26 2013-03-07 オリンパス株式会社 Single-particle detector using optical analysis, single-particle detection method using same, and computer program for single-particle detection
WO2015064628A1 (en) * 2013-10-31 2015-05-07 栗田工業株式会社 Method and device for measuring number of particulates in ultrapure water

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05240770A (en) * 1992-02-29 1993-09-17 Horiba Ltd Particle counter
JP2005502875A (en) * 2001-09-07 2005-01-27 インフィコン インコーポレイティッド Signal processing method for particle monitoring of in-situ scanning beam
JP2005537781A (en) * 2002-02-14 2005-12-15 イムニベスト・コーポレイション Method and algorithm for cell counting at low cost
JP2003279484A (en) * 2002-03-26 2003-10-02 Sharp Corp Fine particle detection method, fine particle detection device, fine particle manufacturing device, fine particle manufacturing program, storage medium and solid-state component
JP2006153745A (en) * 2004-11-30 2006-06-15 Tokyo Electron Ltd Particle-detecting method and particle-detecting program
JP2009097959A (en) * 2007-10-16 2009-05-07 Tokyo Seimitsu Co Ltd Defect detecting device and defect detection method
WO2011108369A1 (en) * 2010-03-01 2011-09-09 オリンパス株式会社 Optical analysis device, optical analysis method, and computer program for optical analysis
WO2013031309A1 (en) * 2011-08-26 2013-03-07 オリンパス株式会社 Single-particle detector using optical analysis, single-particle detection method using same, and computer program for single-particle detection
WO2015064628A1 (en) * 2013-10-31 2015-05-07 栗田工業株式会社 Method and device for measuring number of particulates in ultrapure water

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20210000656A (en) * 2019-06-25 2021-01-05 오므론 가부시키가이샤 Appearance inspection management system, appearance inspection management device, appearance inspection management method, and appearance inspection management program
KR102284095B1 (en) * 2019-06-25 2021-07-30 오므론 가부시키가이샤 Appearance inspection management system, appearance inspection management device, appearance inspection management method, and appearance inspection management program
TWI767229B (en) * 2019-06-25 2022-06-11 日商歐姆龍股份有限公司 Appearance inspection management system, appearance inspection management device, appearance inspection management method, and program
WO2022024389A1 (en) * 2020-07-31 2022-02-03 株式会社日立ハイテク Method for generating trained model, method for determining base sequence of biomolecule, and biomolecule measurement device

Also Published As

Publication number Publication date
JP6742435B2 (en) 2020-08-19
JPWO2018105462A1 (en) 2019-10-24

Similar Documents

Publication Publication Date Title
TWI686718B (en) Determining coordinates for an area of interest on a specimen
KR102514134B1 (en) Correlation of SEM and Optical Images to Identify Wafer Noise Newsons
US7301645B2 (en) In-situ critical dimension measurement
JP7052024B2 (en) Detection of asymmetric structures and dimensional measurement
TWI673489B (en) System and method for generating inspection results for a specimen with an adaptive nuisance filter, and non-transitory computer-readable medium
TWI625806B (en) Metrology optimized inspection
KR102352702B1 (en) Detecting ic reliability defects
KR19990072263A (en) Method and apparatus for critical dimension and tool resolution determination using edge width
WO2018105462A1 (en) Signal processing method and program
US10332810B2 (en) Process modules integrated into a metrology and/or inspection tool
JP5593399B2 (en) Measuring device
US10254110B2 (en) Via characterization for BCD and depth metrology
US8014891B2 (en) Etching amount calculating method, storage medium, and etching amount calculating apparatus
CN101459049A (en) Apparatus and method for detecting etching terminal
CN113412485B (en) System, computer-readable medium, and method for selecting design files
TWI508161B (en) An etching amount calculating method, a memory medium, and an etching amount calculating means
KR20130096228A (en) Data perturbation for wafer inspection or metrology setup
KR20070002125A (en) Method of inspecting a defect and apparatus for inspecting a defect using the same
JP2010101894A (en) Deterministic component identification device, deterministic component identifying method, program, storage medium, test system, and electronic device
CN113677980B (en) Defect candidate generation for inspection
WO2018118805A1 (en) Method and system for weak pattern quantification
JPH0560540A (en) Measuring method for pattern size using charged beam
JP5291586B2 (en) Calculation apparatus, calculation method, program, storage medium, test system, and electronic device
JP4964400B2 (en) Anomalous photoresist line / spacing profile detection by measurement waveform signal process
TWI794982B (en) Plasma treatment device and plasma treatment method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17879242

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2018554949

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17879242

Country of ref document: EP

Kind code of ref document: A1