A method for detecting concentration of multi-component gas based on composite signal

Abstract

The application discloses a kind of multi-component gas concentration detection methods based on composite signal, it is related to gas concentration detection technical field, comprising: using digital micromirror device according to preset wavelength sequence modulation incident to the light beam of gas chamber;The gas chamber is provided with mutually interfering absorption spectrum detection light path and photoacoustic spectrum detection light path, and real-time synchronous acquisition first signal generated by the absorption spectrum detection light path, and second signal generated by the photoacoustic spectrum detection light path;The first signal is processed to obtain first concentration value C1, and the second signal is processed to obtain second concentration value C2;Based on the first concentration value C1 and the second concentration value C2, according to preset rule determines target gas concentration value.The above scheme can take into account the synchronous detection of multi-component gas, the compactness of detection structure and the high sensitivity detection of low concentration gas.

Description

Technical Field

[0001] This invention relates to the field of gas concentration detection technology, and in particular to a method for detecting the concentration of multi-component gases based on composite signals. Background Technology

[0002] In semiconductor manufacturing processes, vapor deposition processes such as chemical vapor deposition (CVD), metal-organic chemical vapor deposition (MOCVD), and atomic layer deposition (ALD) all require precise monitoring of reactant and byproduct gases to ensure process stability and yield. Therefore, gas detection equipment is widely used in semiconductor process chambers, piping, and environmental monitoring to perform real-time detection and concentration monitoring of multi-component gases. In these applications, the detection equipment not only needs high sensitivity but also requires fast response speed and system stability to meet the high demands of complex process environments.

[0003] Currently, a wide variety of concentration detection devices are available on the market, employing different sensing technologies such as acoustic, optical, and mass spectrometry. Among existing gas detection technologies, non-dispersive infrared (NDIR) schemes are widely used due to their simple structure and mature technology. However, in multi-gas detection scenarios, traditional NDIR typically relies on fixed filters combined with mechanical filter wheels to distinguish the absorption bands of different gases. This approach has two significant problems: first, the filter wheel structure is bulky and requires a rotation drive mechanism, which not only increases the size and power consumption of the device but also leads to a significant reduction in response speed (switching time is typically on the order of hundreds of milliseconds); second, there is overlap between infrared absorption bands, for example... and Absorption in the infrared band is prone to cross-interference, which leads to a decrease in detection accuracy. Complex calibration algorithms are required for compensation, which further increases the system complexity and maintenance costs.

[0004] Therefore, while traditional NDIR schemes are suitable for routine detection of single gases, they are difficult to simultaneously detect multiple gases, have a compact structure, and achieve high sensitivity to low-concentration gases in real-time detection of multiple gases required for semiconductor processes, which has become a technical problem that urgently needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-component gas concentration detection method based on composite signals, in order to solve the problems in the prior art that it is difficult to simultaneously detect multiple gases, achieve a compact structure, and achieve high sensitivity detection of low-concentration gases.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A method for detecting the concentration of multi-component gases based on composite signals, comprising:

[0008] A light beam incident on the gas cell is modulated according to a preset wavelength sequence using a digital micromirror device;

[0009] The gas chamber is equipped with an absorption spectral detection optical path and a photoacoustic spectral detection optical path that do not interfere with each other, and the first signal generated by the absorption spectral detection optical path and the second signal generated by the photoacoustic spectral detection optical path are collected in real time and synchronously.

[0010] The first signal is processed to obtain a first concentration value C1, and the second signal is processed to obtain a second concentration value C2;

[0011] Based on the first concentration value C1 and the second concentration value C2, the target gas concentration value is determined according to a preset rule.

[0012] The preset rules include: if both C1 and C2 are within a first preset concentration range, then C1 is output as the target gas concentration value; if both C1 and C2 are within a second preset concentration range, then C2 is output as the target gas concentration value; if C1 and C2 are within different preset concentration ranges, then based on the numerical difference between C1 and C2, C1, C2, or the value obtained by data fusion of C1 and C2 is selected as the target gas concentration value.

[0013] By adopting the above scheme, a composite detection method based on photoacoustic spectroscopy and spectral absorption signals is used. A digital micromirror device modulates the light beam, enabling simultaneous acquisition of absorption and photoacoustic detection within the same gas chamber. This approach leverages the wide dynamic range of absorption spectroscopy and the high sensitivity and anti-scattering interference characteristics of photoacoustic spectroscopy to achieve complementary detection performance, avoiding the limitations of single detection methods and improving overall detection accuracy.

[0014] Furthermore, the first preset concentration range is from 12 ppm to 990 ppm.

[0015] By adopting the above scheme, the high signal-to-noise ratio and wide linear range of absorption spectroscopy under medium concentration conditions are fully utilized to ensure the accuracy and stability of the detection results and avoid the nonlinear distortion that may occur in photoacoustic detection at medium concentrations.

[0016] Furthermore, the second preset concentration range is higher than 1010 ppm or lower than 7 ppm.

[0017] By adopting the above scheme, the sensitivity can be improved by utilizing the trace detection capability of photoacoustic spectroscopy at low concentrations, and the distortion of the absorption spectrum due to transmission signal saturation can be avoided at high concentrations, thus ensuring the validity of the detection results.

[0018] Furthermore, the statement that C1 and C2 are respectively within different preset concentration ranges includes:

[0019] One of the values ​​of C1 and C2 is in the critical range of 7 ppm to 12 ppm or 990 ppm to 1010 ppm, while the other value is not in that critical range.

[0020] By adopting the above scheme, discrimination is performed at the boundary of concentration ranges to prevent abrupt errors caused by the instability of a single path and to achieve a smooth transition of results.

[0021] Furthermore, the step of selecting to output C1, C2, or the value obtained by data fusion of C1 and C2 based on the numerical difference between C1 and C2 includes:

[0022] Calculate the absolute value of the difference between C1 and C2, |C1-C2|;

[0023] If |C1-C2| is greater than a preset difference threshold, then the smaller of C1 and C2 is output as the target gas concentration value;

[0024] If |C1-C2| is less than or equal to the preset difference threshold, then a data fusion algorithm is used to fuse C1 and C2, and the fusion result is used as the target gas concentration value.

[0025] By adopting the above scheme, when the difference is greater than a threshold, a smaller value is selected for output; when the difference is within the threshold, data fusion is performed. This allows for the removal of abnormal signals through difference measurement, while the fusion strategy enhances robustness under boundary conditions, avoiding measurement bias caused by fluctuations in a single signal.

[0026] Furthermore, the preset difference threshold is 1 ppm.

[0027] Furthermore, the data fusion employs a recursive filtering algorithm based on uncertainty metrics, the steps of which include:

[0028] S1: State Prediction

[0029]

[0030]

[0031] in Let k be the predicted gas concentration at time k. This represents the optimal estimate of the gas concentration at time k-1. Let k be the measure of prediction uncertainty at time k. Let Q be the estimation uncertainty measure at time k-1, and let Q be the process noise covariance, which is a pre-set confidence parameter close to zero.

[0032] S2: Sequential Update:

[0033] S21: First, update using the first concentration value C1:

[0034]

[0035]

[0036]

[0037] in, As the first fusion weight, This is the preset reliability parameter corresponding to the first concentration value C1. This is an intermediate estimate after incorporating C1. This is the corresponding intermediate uncertainty measure;

[0038] S22: Subsequently, and As a new prior estimate, the second concentration value C2 is used for updating:

[0039]

[0040]

[0041] in, As the second fusion weight, This is the preset reliability parameter corresponding to the second concentration value C2. This refers to the target gas concentration value after data fusion at time k. The updated uncertainty metric is used for calculation in the next time step.

[0042] A recursive filtering algorithm based on uncertainty metric is employed to dynamically adjust the weight distribution of absorption and photoacoustic spectral signals during data fusion. By calibrating the two types of optical paths in a stable gas environment, their respective statistical variances are obtained as uncertainty metrics. This allows the system to automatically reduce the weight of the corresponding signal when signal fluctuations are large or noise levels are high, thereby improving the reliability of the final estimate. This method not only effectively suppresses the impact of abnormal fluctuations in a single detection signal on the results but also achieves a smooth transition within the concentration critical range, avoiding abrupt changes in measured values. Furthermore, the recursive filtering algorithm has real-time update capabilities, continuously correcting the prediction results during continuous measurements, making the final concentration estimate more stable and accurate. Therefore, this invention maintains high detection accuracy and robustness even under complex environmental conditions, significantly outperforming methods that rely solely on simple averaging or fixed-weight fusion.

[0043] Furthermore, the pre-set confidence parameters and The measurement data of the absorption spectroscopy detection optical path and the photoacoustic spectroscopy detection optical path were recorded over a period of time after calibration in a stable gas environment, and their statistical variances were calculated. and The value assigned is proportional to the statistical variance.

[0044] Furthermore, when the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is less than that of the photoacoustic spectroscopy detection optical path, i.e. < During the sequential update process, the first fusion weight Always greater than the second fusion weight ;

[0045] When the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is greater than that of the photoacoustic spectroscopy detection optical path, that is... < During the sequential update process, the first fusion weight Always less than the second fusion weight .

[0046] Furthermore, it also includes a data storage step: storing the real-time collected first concentration value C1, second concentration value C2, and the target gas concentration value after data fusion. and the corresponding measurement timestamps and uncertainty measurement parameters. Fusion weight parameters and The data is stored in a structured data format in non-volatile memory.

[0047] By adopting the above scheme, not only is subsequent source tracing analysis and long-term operational data comparison and calibration facilitated, but a reliable data foundation is also provided for algorithm optimization and model retraining. This storage method ensures the integrity and traceability of historical detection records, enabling the system to adaptively adjust to different application scenarios. Simultaneously, non-volatile storage ensures that data is not lost after power outages or system restarts, improving system reliability and engineering application value.

[0048] The beneficial effects of the multi-component gas concentration detection method based on composite signals provided by this invention are as follows: Based on a composite detection method using both photoacoustic spectroscopy and spectral absorption signals, the method achieves simultaneous acquisition of absorption and photoacoustic detection within the same gas chamber by modulating the light beam using a digital micromirror device. This leverages the wide dynamic range of absorption spectroscopy and the high sensitivity and anti-scattering interference characteristics of photoacoustic spectroscopy to achieve complementary detection performance, avoiding the limitations of single detection methods and improving overall detection accuracy. Furthermore, the use of a digital micromirror device to modulate the light beam avoids the problems of large size and slow response caused by traditional NDIR relying on filters and mechanical switching, resulting in a more compact structure, reduced size and power consumption, while ensuring fast response. Attached Figure Description

[0049] Figure 1 This is a flowchart of a multi-component gas concentration detection method based on composite signals according to the present invention;

[0050] Figure 2 This is a flowchart of gas concentration analysis according to an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the optical path of a multi-component gas concentration detection system according to an embodiment of the present invention;

[0052] Figure 4 This is a schematic diagram of the air chamber structure according to an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of the overall structure of a multi-component gas concentration detection system according to an embodiment of the present invention.

[0054] Reference numerals: 1. Light source; 2. Gas chamber; 21. Optical signal detection module; 22. Acoustic signal detection module; 221. Microphone; 23. First optical window; 24. Second optical window; 25. Third optical window; 3. First optical modulator; 4. Processing unit; 5. Second optical modulation component; 6. Beam splitter. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein shall have the ordinary meanings as understood by those of ordinary skill in the art to which the present invention pertains. The words such as "including" used herein mean that the elements or items appearing before this word cover the elements or items listed after this word and their equivalents, without excluding other elements or items.

[0056] In the prior art, gas concentration detection has important applications in fields such as environmental monitoring, industrial process control, and medical diagnosis. The current mainstream technologies mainly include non-dispersive infrared, photoacoustic spectroscopy, and tunable diode laser absorption spectroscopy. Among them, non-dispersive infrared is widely adopted due to its low cost and high stability. Photoacoustic spectroscopy has the advantage of high sensitivity in trace gas detection, while tunable diode laser absorption spectroscopy features strong selectivity and high resolution. However, with the increasing demand for synchronous detection of multi-component gases in application scenarios, it is difficult for the prior art to balance sensitivity, dynamic range, and cost, facing obvious limitations.

[0057] Taking non-dispersive infrared technology as an example, its principle is based on the Beer–Lambert law, and the gas concentration is deduced by measuring the transmission intensity of infrared light at a specific wavelength. The non-dispersive infrared device has a simple structure and low cost, and is suitable for long-term industrial monitoring. However, its detection sensitivity for trace gases (<ppm level) is insufficient, and multi-component detection often relies on fixed filters and mechanical switching filter wheels, which not only results in a large volume and slow response speed, but also and In scenarios where absorption peaks overlap, cross-interference is prone to occur, requiring complex compensation algorithms for correction. Tunable laser absorption spectroscopy utilizes narrow-linewidth lasers to scan gas absorption lines, achieving high sensitivity and selectivity, suitable for precision analysis. However, this approach typically requires multiple distributed feedback lasers or quantum cascade lasers, supplemented by precise temperature control and drive circuitry, resulting in a complex and costly system. Furthermore, the limited laser output power leads to signal saturation and insufficient dynamic range when detecting high-concentration gases, hindering its widespread adoption in low-cost or portable applications. Photoacoustic spectroscopy, based on the principle that gas molecules absorb light energy to generate sound waves, offers advantages such as high sensitivity and zero background noise, making it particularly suitable for complex optical environments containing dust. However, photoacoustic signals rely on acoustic resonators, which have limited linear response range, and the sound pressure signal is prone to saturation at high concentrations. Additionally, multi-gas detection requires wavelength modulation or multiple light sources, resulting in poor real-time performance and increased cost of the light sources. Meanwhile, the microphone is sensitive to external vibrations and environmental noise, requiring additional vibration isolation and shielding, increasing system size and cost.

[0058] The following is in conjunction with the appendix Figure 1 - Appendix Figure 5 The specific embodiments of the present invention will be further described in detail below.

[0059] Reference Figure 1 and Figure 2 In some embodiments of the present invention, a multi-component gas concentration detection method based on composite signals is applicable to gas concentration detection systems that use a single broadband light source 1 and combine photoacoustic spectroscopy and spectral absorption detection principles, and can achieve high sensitivity and wide dynamic range detection of multi-component gases in the same gas chamber 2.

[0060] In some specific embodiments of the present invention, the method includes the following steps: First, the detection light emitted from the broadband light source 1 is modulated using a digital micromirror device. Through programming control of the micromirror array in the digital micromirror device, the detection light is selectively reflected according to a preset wavelength sequence, thereby forming a controllable beam that is incident on the gas chamber 2. The gas chamber 2 is equipped with two independent optical paths: an absorption spectral detection optical path and a photoacoustic spectral detection optical path. These two paths share the light source 1 and the gas path within the same gas chamber 2, but do not interfere with each other. The absorption spectral detection optical path receives the transmitted light after penetrating the gas sample through a photodetector and outputs a first signal; the photoacoustic spectral detection optical path excites the photoacoustic effect of gas molecules by modulating the beam, and the acoustic signal is collected by a microphone 221 and outputs a second signal. The system achieves synchronous acquisition of the first and second signals.

[0061] Subsequently, the first signal is processed to obtain a first concentration value C1 calculated based on the principle of spectral absorption; the second signal is processed to obtain a second concentration value C2 calculated based on the principle of photoacoustic spectroscopy. Since the two types of signals have complementary characteristics, the system introduces a dynamic selection mechanism based on preset rules during concentration inversion. Specifically: If C1 and C2 are both within a first preset concentration range (in some embodiments of the present invention, the first preset concentration range is 12 ppm to 990 ppm), then C1 is output as the target gas concentration value to take advantage of the stability and good linearity of the absorption spectrum in the medium concentration range; if C1 and C2 are both within a second preset concentration range (in some embodiments of the present invention, the second preset concentration range is below 7 ppm or above 1010 ppm), then C2 is output to take advantage of the high sensitivity and anti-saturation characteristics of photoacoustic spectroscopy under trace and high concentration conditions; if C1 and C2 are within different preset ranges, then it is further determined whether they are within a critical range (in some embodiments of the present invention, the critical range is 7–12 ppm or 990–1010 ppm), and the output result is selected or a fusion algorithm is used for processing based on the difference between the two.

[0062] In the fusion process, this invention proposes a recursive filtering method based on uncertainty measurement. First, state prediction is performed, using the concentration estimate from the previous moment and the uncertainty measurement to obtain the predicted value for the current moment. Then, two types of signals are sequentially introduced for updating: first, C1 is incorporated into the predicted value, and intermediate estimates and uncertainties are calculated; then, this is used as a new prior estimate to introduce C2, completing the final update and obtaining the target concentration value. The weights of different signals are determined by their respective uncertainty parameters, which are obtained through calibration experiments in a stable gas environment and are proportional to the statistical variance of the signal. Thus, in the medium concentration range, if the variance of the absorbed signal is small, its weight in the fusion is greater; conversely, the photoacoustic signal has a greater weight, achieving signal reliability-driven adaptive fusion.

[0063] Specifically, in some other embodiments of the present invention, selecting to output C1, C2, or the value after data fusion of C1 and C2 based on the numerical difference between C1 and C2 includes:

[0064] Calculate the absolute value of the difference between C1 and C2, |C1-C2|; if |C1-C2| is greater than a preset difference threshold, output the smaller of C1 and C2 as the target gas concentration value; if |C1-C2| is less than or equal to the preset difference threshold, use a data fusion algorithm to fuse C1 and C2, and use the fusion result as the target gas concentration value.

[0065] In some other embodiments of the present invention, the preset difference threshold is 1 ppm.

[0066] In some other embodiments of the present invention, the data fusion employs a recursive filtering algorithm based on uncertainty metrics, specifically a Kalman filtering algorithm, the steps of which include:

[0067] S1: State Prediction

[0068]

[0069]

[0070] in Let k be the predicted gas concentration at time k. This represents the optimal estimate of the gas concentration at time k-1. Let be the measure of prediction uncertainty at time k, which is also the current best estimate covariance. Q is the estimation uncertainty measure at time k-1, where Q is the known process noise covariance and is a pre-set confidence parameter close to zero.

[0071] S2: Sequential Update:

[0072] S21: First, update using the first concentration value C1:

[0073]

[0074]

[0075]

[0076] in, The first fusion weight at time K and C1 concentration. This is the preset reliability parameter corresponding to the first concentration value C1. This is an intermediate estimate after incorporating C1. for , This is the corresponding intermediate uncertainty measure;

[0077] S22: Subsequently, and As a new prior estimate, the second concentration value C2 is used for updating:

[0078]

[0079]

[0080] in, At time K, and with C2 concentration as the second fusion weight, This is the preset reliability parameter corresponding to the second concentration value C2. This refers to the target gas concentration value after data fusion at time k. The updated uncertainty metric is used for calculation in the next time step.

[0081] In some other embodiments of the present invention, the preset confidence parameter and The measurement data of the absorption spectroscopy detection optical path and the photoacoustic spectroscopy detection optical path were recorded over a period of time after calibration in a stable gas environment, and their statistical variances were calculated. and The value assigned is proportional to the statistical variance.

[0082] In some other embodiments of the present invention, when the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is less than that of the photoacoustic spectroscopy detection optical path, i.e. < During the sequential update process, the first fusion weight Always greater than the second fusion weight ;

[0083] When the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is greater than that of the photoacoustic spectroscopy detection optical path, that is... < During the sequential update process, the first fusion weight Always less than the second fusion weight .

[0084] In some other embodiments of the present invention, the method further includes structured storage supporting the result data. The system can store the real-time acquired first concentration value C1, second concentration value C2, and fused target concentration value. and the corresponding measurement timestamps and uncertainty measurement parameters. Fusion weight parameters and The data is saved to non-volatile memory. This design not only provides a foundation for subsequent data tracing and trend analysis, but also provides experimental support for device self-learning and algorithm iteration.

[0085] Currently, semiconductor process reaction gases mainly consist of two categories: metal-organic sources (MO sources) and hydrides. Hydrides are typically present in high-pressure cylinders and must be diluted to a safe concentration with a high-purity carrier gas (such as H2, N2) or other high-purity reaction gases before use. Metal-organic sources are precursors of elements such as Ga, In, Al, Group II (such as Zn, Mg, Cd), or Group IV (such as Sn, Ge). They are usually compounds formed by combining the metal with organic groups (such as methyl, ethyl). These are precisely controlled and delivered to the reaction chamber, where they undergo thermal decomposition and chemical reactions on the heated substrate surface, depositing the desired semiconductor material. Therefore, it is necessary to monitor the concentration of these multi-component gases to ensure accurate delivery of the process reaction gases. Furthermore, gaseous byproducts are generated after the reaction. During the thermal decomposition and chemical reaction of the MO source and hydrides in the reaction chamber, a large amount of volatile byproduct gases are also produced while depositing the target semiconductor thin film on the substrate. The main byproducts are hydrocarbons and hydrogen, depending on the organic groups in the MO source used. Most of these byproducts are highly toxic and need to be detected, then rapidly removed by a vacuum system and processed. Therefore, both semiconductor process reaction gases and volatile byproduct gases require high-precision multi-component gas concentration detection.

[0086] Reference Figure 3 and Figure 5 In some specific embodiments of the present invention, a multi-component gas concentration detection system used in the present invention includes a light source 1, a beam splitter, a gas chamber 2, a first light modulator 3, an optical signal detection module 21, an acoustic signal detection module 22, and a processing unit 4.

[0087] In some embodiments of the present invention, the light source 1 is used to emit broadband detection light covering the absorption band of the target gas; in some preferred embodiments, the light source 1 is preferably a broadband infrared light source 1, such as a MEMS blackbody radiation light source 1 or a quantum cascade laser, capable of covering the characteristic absorption band of the target gas, that is, covering the key absorption range of –μm. A beam splitter is used to disperse the detection light and at least split it into a first beam and a second beam; the gas chamber 2 is used to contain the gas to be tested, and is internally equipped with an optical signal detection module 21 and an acoustic signal detection module 22, with multiple optical windows on the outer wall of the gas chamber 2; a first optical modulator 3 is used to perform wavelength selection and modulation of the second beam; in some specific embodiments of the present invention, the first optical modulator 3 preferably employs a digital micromirror device, which can select modulated light of a specific wavelength to be incident on the gas chamber 2 according to a control signal to excite a photoacoustic effect. The processing unit 4 is used to synchronously receive and process the light intensity signal output by the optical signal detection module 21 and the sound pressure signal output by the acoustic signal detection module 22. In this design, the first light beam enters the gas chamber 2 through an optical window and is received by the optical signal detection module 21 after passing through the gas chamber 2, forming an absorption spectral detection optical path based on Beer-Lambert's law. The modulated second light beam enters the gas chamber 2 through another optical window, is absorbed by the molecules of the gas to be measured, and excites an acoustic signal, which is received by the acoustic signal detection module 22, forming a photoacoustic spectral detection optical path based on the photoacoustic spectral effect. This design integrates photoacoustic spectroscopy and spectral absorption technology in the same system, fully utilizing the advantages of photoacoustic signals (high sensitivity, unaffected by optical scattering and detector drift) and spectral absorption technology (wide dynamic range, easy calibration), thereby achieving efficient and accurate multi-component gas detection.

[0088] In some specific embodiments of the present invention, the beam splitting component includes a second optical modulator and a beam splitter 6 arranged sequentially on the optical path. The second optical modulator preferably employs a digital micromirror device for wavelength dispersion and spatial distribution control of the broadband detection light. The second optical modulator can act as a grating to disperse the broadband light source 1. The beam splitter 6 is preferably a 50:50 beam splitter, used to divide the dispersed detection light into two paths of equal intensity, which serve as the inputs to the absorption detection path and the photoacoustic detection path, respectively. Through the high-speed encoding control of the digital micromirror device and the beam distribution of the beam splitter 6, the system can achieve simultaneous detection of multiple components of gas without the need for multiple independent light sources 1. Compared with traditional methods relying on multiple light sources 1 or laser tuning, this scheme has a more compact structure, reduces system cost, and avoids the problems of large size and slow response caused by mechanical filter wheels.

[0089] Furthermore, both the first optical modulator 3 and the second optical modulator include a micromirror array and corresponding control components. The micromirror array consists of multiple micromirror units whose deflection angles can be independently changed. Each micromirror unit includes a rotatable micromirror, a pivot connected to it, and a circuit unit for driving the micromirror to swing around the pivot. The control components change the outgoing light path by adjusting the tilt direction of each unit. With this structure, the second optical modulator can perform dynamic band selection and dispersion of the broadband light source 1, and the first optical modulator 3 can rapidly modulate the beam of the incident gas chamber 2, realizing frequency division multiplexing control. In this way, the absorption and photoacoustic optical paths not only share the same broadband light source 1, but also can achieve independent control of different target gas bands through programming, replacing traditional filter combinations or laser arrays and significantly improving the system detection efficiency.

[0090] In some specific embodiments of the present invention, the air chamber 2 includes a housing with an air inlet and an air outlet. A first optical window 23 is located on one side of the housing and is equipped with a collimating lens to ensure that the incident light enters collimatedly; a second optical window 24 is located on the other side of the housing and is equipped with a converging lens to focus the transmitted light onto the optical signal detection module 21; a third optical window 25 is located on the top of the housing and is also equipped with a converging lens to effectively guide the modulated light into the air chamber 2 and make it opposite to the acoustic signal detection module 22. An acoustic resonant cavity structure can be further formed inside the air chamber 2. The acoustic signal detection module 22 is preferably a high-sensitivity microphone 221 with a frequency response of 20Hz–20kHz, installed in the area with the strongest acoustic field. Through the above-mentioned shared design of optical and air paths, the system can simultaneously complete spectral absorption detection and photoacoustic detection in the same air chamber 2, avoiding optical alignment errors and air path switching losses caused by the multi-air chamber 2 design, and greatly improving detection efficiency.

[0091] The specific optical path connection is as follows: The detection light emitted by the light source 1 first enters the second optical modulator, which includes a micro-mirror array and its control components. Through the coding control of the micro-mirror units, a portion of the beam is selectively diffracted according to a preset wavelength and reflected to the beam splitter 6. After being split by the beam splitter 6, the detection light is divided into two paths. One part is transmitted to form an absorption spectrum detection optical path, and the other part is reflected to form a photoacoustic spectrum detection optical path. After the transmitted light enters the absorption spectrum detection optical path through the beam splitter 6, it sequentially enters the gas chamber 2 through the first optical window 23, penetrates the gas sample, and exits through the second optical window 24. Finally, it is received by the optical signal detection module 21 to realize absorption spectrum detection. The other part of the light reflected by the beam splitter 6 enters the first optical modulator 3. The first optical modulator 3 selectively reflects the beam of the target wavelength into the gas chamber 2 through the programming control of its micro-mirror array. The beam then enters the interior of the gas chamber 2 through the third optical window 25, where the photoacoustic effect is used to excite an acoustic signal, which is collected by the acoustic signal detection module 22, thus forming the photoacoustic spectrum detection optical path.

[0092] Among them, the upper limit of the detection of the concentration of multi-component gases mainly depends on the micro-mirror array. At the same time, it is necessary to consider whether the gases to be detected have spectral overlap and whether there is interference in the spectral bands of the light source.

[0093] It is important to note that absorption spectroscopy detection, based on the intensity changes of broadband transmitted light, can simultaneously acquire the absorption characteristics of multiple gas components in a single measurement, thereby retrieving the concentrations of various gases in parallel, resulting in high detection efficiency. In contrast, photoacoustic spectroscopy relies on modulating specific wavelengths of light to excite the acoustic response of corresponding gas molecules. Therefore, in multi-component detection, the target wavelength band needs to be switched sequentially via the first optical modulator 3 to perform time-division detection of different gases. Thus, absorption spectroscopy is suitable for rapid parallel measurement of multiple components, while photoacoustic spectroscopy provides high sensitivity compensation for low-concentration components; combining the two can balance detection efficiency and sensitivity.

[0094] In some specific embodiments of the present invention, the optical signal detection module 21 includes a photodetector and an optoelectronic signal conditioning circuit electrically connected thereto. The photodetector can be flexibly selected according to requirements, such as a multi-channel pyroelectric sensor, a multi-channel thermopile sensor, or a charge-coupled device spectrometer. The optoelectronic signal conditioning circuit amplifies and filters the received light intensity signal to ensure stable and reliable output data. Compared with traditional solutions relying on a single fixed filter, the optical path and detector interface design provided by the present invention expands the selection range of detection devices, enabling flexible configurations of high resolution, multi-channel, or low cost according to specific application requirements.

[0095] In some specific embodiments of the present invention, the acoustic signal detection module 22 includes a microphone 221 and an electrically connected acoustic signal conditioning circuit. The microphone 221 is used to detect the acoustic wave signal generated by the photoacoustic effect and convert it into an electrical signal; the acoustic signal conditioning circuit amplifies and filters the electrical signal. By combining lock-in amplification or digital demodulation methods, environmental noise can be effectively suppressed, and the detection sensitivity for trace gases can be improved. Compared with the spectral absorption channel, the photoacoustic channel does not depend on the absolute value of light intensity, but directly reflects the gas absorption characteristics through acoustic response, thus providing higher detection accuracy in low-concentration scenarios.

[0096] In some specific embodiments of the present invention, in terms of data processing, processing unit 4 includes a signal demodulation module, a concentration inversion module, and a data fusion module. The signal demodulation module is responsible for frequency division multiplexing demodulation of the acoustic signal to distinguish the acoustic responses of different gas components; the concentration inversion module uses machine learning algorithms to model and calculate the light intensity signal and the demodulated sound pressure signal, decouple cross-interference, and invert the concentration of each gas; the data fusion module fuses and calibrates the two types of detection results, photoacoustic and spectral absorption, preferably using a Kalman filter algorithm for dynamic weighting, enhancing the weight of the photoacoustic signal in the low concentration range and enhancing the weight of the absorption signal in the high concentration range, thereby achieving high-precision detection across the entire range. Through this signal collaborative processing method, the high sensitivity of photoacoustic and the wide dynamic range of absorption are complementary, further reducing error interference.

[0097] In some other embodiments of the present invention, the system further includes an environmental parameter sensing module for real-time acquisition of temperature, humidity, and pressure data within the air chamber 2, and transmission of this data to the processing unit 4 for compensation calculation. This design can correct for the impact of environmental changes on the optical absorption cross-section and sound velocity, improving detection reliability in complex application scenarios.

[0098] The multi-component gas concentration detection system proposed in this invention organically combines a single broadband light source 1, a beam splitter, a first optical modulator 3, and a second optical modulator, taking advantage of both photoacoustic spectroscopy and spectral absorption principles. Structurally, it not only avoids the problems of large size, high cost, and slow response associated with multiple light sources 1 and mechanical filter wheels, but also significantly reduces system cost and detection time, while improving detection efficiency and flexibility through deep integration of optical path, gas path, and signal processing.

[0099] While embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations fall within the scope and spirit of the present invention. Furthermore, the present invention described herein may have other embodiments and can be implemented or carried out in various ways.

Claims

1. A method for detecting the concentration of multi-component gases based on composite signals, characterized in that, include: A light beam incident on the gas cell is modulated according to a preset wavelength sequence using a digital micromirror device; The gas chamber is equipped with an absorption spectral detection optical path and a photoacoustic spectral detection optical path that do not interfere with each other, and the first signal generated by the absorption spectral detection optical path and the second signal generated by the photoacoustic spectral detection optical path are collected in real time and synchronously. The first signal is processed to obtain a first concentration value C1, and the second signal is processed to obtain a second concentration value C2; Based on the first concentration value C1 and the second concentration value C2, the target gas concentration value is determined according to a preset rule. The preset rules include: if both C1 and C2 are within a first preset concentration range, then C1 is output as the target gas concentration value; if both C1 and C2 are within a second preset concentration range, then C2 is output as the target gas concentration value; if C1 and C2 are within different preset concentration ranges, then based on the numerical difference between C1 and C2, C1, C2, or the value obtained by data fusion of C1 and C2 is selected as the target gas concentration value. The step of selecting to output C1, C2, or the value obtained by data fusion of C1 and C2 based on the numerical difference between C1 and C2 includes: Calculate the absolute value of the difference between C1 and C2, |C1-C2|; If |C1-C2| is greater than a preset difference threshold, then the smaller of C1 and C2 is output as the target gas concentration value; If |C1-C2| is less than or equal to the preset difference threshold, then a data fusion algorithm is used to fuse C1 and C2, and the fusion result is used as the target gas concentration value.

2. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that, The first preset concentration range is 12 ppm to 990 ppm.

3. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that, The second preset concentration range is above 1010 ppm or below 7 ppm.

4. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that, The condition that C1 and C2 are respectively in different preset concentration ranges includes: One of the values ​​of C1 and C2 is in the critical range of 7 ppm to 12 ppm or 990 ppm to 1010 ppm, while the other value is not in that critical range.

5. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that, The preset difference threshold is 1 ppm.

6. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 5, characterized in that, The data fusion employs a recursive filtering algorithm based on uncertainty metrics, and its steps include: S1: State Prediction in Let k be the predicted gas concentration at time k. This represents the optimal estimate of the gas concentration at time k-1. Let k be the measure of prediction uncertainty at time k. Let Q be the estimation uncertainty measure at time k-1, and let Q be the process noise covariance, which is a pre-set confidence parameter close to zero. S2: Sequential Update: S21: First, update using the first concentration value C1: in, As the first fusion weight, This is the preset reliability parameter corresponding to the first concentration value C1. This is an intermediate estimate after incorporating C1. This is the corresponding intermediate uncertainty measure; S22: Subsequently, and As a new prior estimate, the second concentration value C2 is used for updating: in, As the second fusion weight, This is the preset reliability parameter corresponding to the second concentration value C2. This refers to the target gas concentration value after data fusion at time k. The updated uncertainty metric is used for calculation in the next time step.

7. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 6, characterized in that, The preset confidence parameters and The measurement data of the absorption spectroscopy detection optical path and the photoacoustic spectroscopy detection optical path were recorded over a period of time after calibration in a stable gas environment, and their statistical variances were calculated. and The value assigned is proportional to the statistical variance.

8. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that: When the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is less than that of the photoacoustic spectroscopy detection optical path, that is... < During the sequential update process, the first fusion weight Always greater than the second fusion weight ; When the statistical variance of the absorption spectroscopy detection optical path in the intermediate concentration range is greater than that of the photoacoustic spectroscopy detection optical path, that is... < During the sequential update process, the first fusion weight Always less than the second fusion weight .

9. The method for detecting the concentration of a multi-component gas based on a composite signal according to claim 1, characterized in that, It also includes a data storage step: storing the real-time collected first concentration value C1, second concentration value C2, and the target gas concentration value after data fusion. and the corresponding measurement timestamps and uncertainty measurement parameters. Fusion weight parameters and The data is stored in a structured data format in non-volatile memory.

Images